JPRS ID: 10265 USSR REPORT ELECTRONICS AND ELECTRICAL ENGINEERING

• Document Type: CREST
• Collection: JPRS
• Document Number (FOIA) /ESDN (CREST): CIA-RDP82-00850R000500020034-7
• Release Decision: RIF
• Original Classification: U
• Document Page Count: 364
• Document Creation Date: November 1, 2016
• Sequence Number: 34
• Content Type: REPORTS

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAI. USE ONLY
JPRS ~/10265
20 January 1982
USSR Report
ELECTRONlCS AND ELECTRICAL ENGINEERING
(FOUO 1 /82)
FBIS FOREIGN BROADCAST INFORN9ATION SERVICE
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
NOT'E
JPRS publ.ications contain information primarily from foreign
newspapers, periodicals and books, but also from news agency
transmissions and broadcasts. Materials from foreign-language
sources are translated; those from English-language sources
itre transcribed or reprinted, with the original phrasing and
o;her characteristics retained.
. Headlines, editorial reports, and, material enclosed in brackets
are supplied by JPRS. Processing indicators such as [Text]
or [Excerpt] in the first line of each item, cr following the
last line of a brief, in3icate how the original information was
Frocessed. Where no processing indicator is given, the infor-
_ mation was summarized or ex:racted.
i,nfamiliar names rendered phonetically ur transliterated are
enclosed i.n parentheses. Words or namas preceded by a ques-
tion mark and enclosed in parentheses aere not clear in the
orig-,'6nal but have been supplied as appropriate in context.
- Other anattributed parenthetical notes with in the body of an
item originate with the source. Times with in items are as
given by source.
The contenty of this publication in no way represent the poli-
cies, views or at.titudes of the U.S. Government.
COPYR?GHT LAWS AND REGULATIONS GOVERNING OWNERSHIP OF
MATERIALS REPRODUCED HERELN REQUIRE THAT DISSEMINATION
OF THIS PUBLICATION BE RESTRICTED FOR OFFICIAL USE UNLY.
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFiCIAL USE ONI.Y
JPRS L/i0265
20 Jaiiuary 1982
USSR REPOQT
ELECTRONICS AND ELECTRICAL ENGINEERING
(FOUO 1/82)
CONTENTS
ANIPLIFIERS
~
- Noise in Ferroprobes and Magnetic Amplifierrs... 1
Amplifier for KNNIP201UPI Autosnatic Switch 41
ANTENNAS
- Antennas and Feeder Lines 42
- Antenna Device With Electrical Beam Oscillatian in Two Planes..... 75
~ CER7'AIN ASPECTS OF COMPUTER HARD AND SOFT WARE: CONTROL, AUTOMATION,
" TELEMECHANICS, TELEMETERING, MACHINE DESIGNING AND PLANNING
Magnetic Bubble Data Stora,;e and Processing Levices 77
Device for Taking Median Value 120
, Microprocessor Psychodiagnostic Device Patent 123
CERTAIN ASPECTS 0F PHOTOGRAPHY, MOTION PICTURES AND TELEVISION
TV Contour Detector 124
Parameter Selection o� Optical Systems for Scanning Reproducing
_ Devices With Kinescopes 134
Ndultimicroprocessor System for Processing Audio Signal in
Frequency Region 141
~ Automatic Restoration of Movie Films by Computer 150
- a- [III - USSR - 21E S&T FOUO]
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAI. UfiF: ONI.Y
COMMUNICATIONS, COMMUNICATION EQUIPMENT, RECEIVERS AND TRANSMITTERS,
NETWQRKS, RALIO PHYSICS, DATA TRANSMISSION AND PROCESSING, INFURMATION
Z'HEORY
Microwave Delay Line.........e ...............................t.... 156
Methods for Findf.ng Breaks in Optical Cables 158
Control System, for Arc Discharger Switc:hing ilsing Field
Distortion Technique 166
Wideband Photoelectronic keceiver 170
Device for Controlling PIN-Diode Attenuator. 176
Silicon Radio-Frequency p-i-n Diodes for Coiamunications
Equipment 179
Principles of the Design of Nonlinear Pulae-Frequency SignaJ.
Selecti:ors 184
Device for Ad3ition of Pewers 204
Radio Receiver Patent 207
CONVERTERS, INVERTERS, TRANSDUCERS
Waveguide Reflecting Phase Inverter 208
Solid-State Phase Invet-ter Control Device......................... 212
INSTRZJMENTS, MEASURING DEVICES AND TESTERS, METHODS OF MEASURING,
GENERAL EXPERIMEPITAL TECHNIQUES
- Precision Digital Infrared Phase Meter With Frequency
Conversion of Laser Radiation 216
Panoramic Measuring Device for Complex Parameters of Microwave
Devtces 221
Wldeband Integrated Circuit Digital Power Meter for Laser
_ Radiation 225
Subnanoaecond Pyroelectric Infrared Radiation Detector......... 229
= Integrated Temperature Sensor Using Silicon Diode Matrices........ 233
~
- b -
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400540020034-7
FOR OFHECIAL USE ONLY
MICROELECTRONICS
Prospects of UsinR Media With Higher Die?.ectric Constant in
Microwave GIS Techiiology 236
Problems of Modeling Elements of Large-Scale Integrated
Circuits With Submicronic nimensions 2414
Selected Abstracts of Articles From Journal 'MICROELECTRONICS',
September-October 1981 269
OSCILLATORS, MODULATORS, GENERATORS
Technological AMTs 1550 Crystal Oscillator 273
PUBLICATIONS, INCLUDING COLLECTIONS OF ABSTRACTS
Abstracts From Journal 'METROLOG'Y', September 1981 275
~ Acoustical Distance Measuremer.t and Control Methods.....:......... 278
Cascade Generators 283
Classification of Thyristors 286
Control Systems for Thyristor Frequency Converters 289
Digital Encoding of Television Images 292
Electrical and Optical Propertfes of Semiconductors 29i:
Elec.eronic Microwave Devices 2.99
Industrial Telephone Communication, Electrical Signaling and
Installation of Electric Clocks 307
Measurement Elements in Power System Emergency Automatic
Equipment 310
- Mechanical Section Design for Overhead Superhigh Voltage Lines.... 313
Modeling and Testing of Radio Equipment 317
� Modern Methods and Devices for Data Dieplay........................ 321
Pulse Radar Circuits 328
- Problems of Optimal Detection and Localizar.;.on of Malfunctions
in Radioelectronic Equipment 332
- c -
FOR OFFICIAL IfSE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
F()R ()N'H'1('IAI. IltiE ()NI.Y
,
.
Methods of Optimization in Controlling Electrical Power
Systems 335
Prospects for Development tor Maritime Elecc-rical Engineering..... 338
- Radar Data Processing Against Interference Background 342
- Radio Measureanents 351
Superconducting Accelerating Microwave Structures 355
- d -
FOR OFFICIAL USE ONLY
*
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00854R000540020034-7
,
FOR OFFICIAL USE ONLY
AMPLIFIEkS
:
t3DC 620.179.143:621.375.3
NOISE IN FERROPROBES AND 'NIAGNETIC A*PLIFIERS
Moscow IZMERENIYA, KONTROL', AVTOMATIZATSIYt, in Russian No 4(38),. 1981 pp 37-53
[Article by Yu.V. Afanas'yev and V.N. Gorobey, engineers]
[Text] Introduction
Ferroprobes (FZ's) and magnetic amplifiers (MU's) have received extensive applica-
tion in measuring equipment and automation [1-4]e FZ'a and MU's are distinguished
by high reliability and long life, noise imtnunity and the ability to operate over
a w~.Lde temperature range,.low power requirement and low cost. The high sensitivity
characteristic of FZ's and MtJ's has been responsible for their use for detecting
and measuring very low values of a magnetic field or current.
~ The maximum capabilities of FZ's and MU's are 3.imited by the level of their in-
trinsic noise. This concept usually includes noise frAm induction from the power
circuit, the noise of cores, detectable duriYae cyclic reversal of magnetization,
and the thermal noise of windings. Noise in cores, which has begun to be called
magnetic noi.se, has the highest value.
The nature of magnetic noise has been studied already for more than 30 years,
- However, in spite of the successes achieved, many aspects of this phenomenon,
f, which is interesting from the viewpoint nf physics and important from that of
' practical application, have continued to remain unclear.
In 1971 N.N. Kolachevskiy's monograph was published, which summarized the results
of theoretical and experimental studies of the nature of magnetic noise during the
preceding period. New experimental data requiring analysis have been gained since ,
then f5].
In this article an attemPt is made to discuss these data, and also to turn atten-
tion to certain experimentally established facts making it possible to lower the
level of magnettc noise in ferroprobes and magnetic amplifiers by onP to two
orders of magnitude. Transformation Ratios of Ferroprobes and Magnetic Amplifiers
Ferroprobes and magnetic amplifters, are active inducttve transformers, i.e.,
additional energy is induced in them (by means of an excitation winding supplied
1
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
with alternating current), because of whi.ch transformation itself is rendered
possible.
The e.m.f.,in the output winding (called the measuring winding below) is determined
by the equation:
C'y cm - $ffJ ~M w BE (HE) r
where s is the total cross-sectional area of the cor;eos covered by the measuring
winding; w is the number of tuxns of the winding; i is the magnetic axis
i normal to the plane nf the winding"s turns); 1-Ts t~e design factor (2];
is the vector of total rgnetic induction; H= H1 + Hp is the vector of
tge total fi ld strength; H1 is the vector of tie strengtFi of the excitation
field; and ~0 is the vector of the strength of the measured field.
For magnetic amplifiers the relationship between the fi,eld strength and input .
current, ivkh ' is determined by the expression
. No = TOlex ~n't1 1 -
where w is the number of turns of the input winding and R is the length
of the cen~er line of a ring core. �
Experience has shown that components e(t) , multiples of the frequency of the
excitation field, because of the nonlinearity of the characteristics of ores of
ferroprobes and magnetic ampl~fiers, prove to be functions not only of but
also of % , and even when H0 = const . ,
Let us expand nonlinear dependence i E (I}iE ) into a Taylor series, being limited for
sufficiently low.values of HD to two terms of the series [6]:
A
Bi~ (Nr) = �n I �'Hl -F- �'Ho
where u* is the tensor of the relative magneti.c lermeability of the core,
u* is the normal relative permeability of the core and uo is the magnetic
constant.
(2)
Tensor A* takes into account the anisotropic properties of the core acquired by
it in superposition of field k , whereby the material of the core is assumed to
- be isotropic. Ad usted for the principal.axis (the axis of sy~mmetry of the core,
x, y and z; x tensor Q* becomes diagonal and is defined by three com-
ponents, whereby one of them, u* , is the same as the differential permeabflity,
u* = u* _(1/u )/(dB/dH) , and~he other two, u* and u* , as the normal per,
meabiligy, u* =~(1/u0 )/(B/H) . 1~1 all cases the'sterisk fndicates the necesaity
of taking into account the de-magnetization coefficient of the core (in a ferro-
probe) in the respective directions. . Siibstituting (2) in (1), we find
2
F'OR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
� -tr d ro's
�A . di
eE - - swE�o e -I-
r ~ �0 d~l C~ ~
�A -sin a sin ~
-I- .T: Ho ~
cos a coa ~
(3)
where e� is the so-called imbalance coefficient, taking into account the non-
ideal isolation of the measurinb circuit from the excitation circuit, and angles
a and 0equal: a= H~0' I~1 and s=~;~H1 .
_ The expression obtained makes it poesibte ta study transformation grocesses in
ferrogrobes and magnetic amplifiers with any relative orientation of vectors
H, H and t� , in particular, with a longitudinal (a = 0), transverse (a =Tr/2)
and rolating (am~._ wt) excitation field, as wetl as with various spectral composi-
tions of the excitation field.
Having become widespread is the operating mode of a ferroprobe and magnetic ampli-
fier utilizing the central syarmetry of the major hysteresie loop of the cores.
In this mode the cares are excited by means of a strong vgrfable field whose spec-
trum contains only the fundamental wave and odd harmonics. Then the spectrum of
permeability ua and u* will contain a constant component and even harmonics:
e
�n (f ) = RAo i cos (2 n (u t + +
~i
~
~t) = �o -F' ~ I'L Cos (2 R (u t+~~n) ,
~t
(4)
where * 2n and u*are the amplitudes of even harmonics or Fourier coeffi-
' cients computed in a certain manner from analytical dependences Ua[H1(t)] ,
'u*[H1(t)] 1, 2, 3, ; w is the angular f.requency of thc fundamental
wave of the er.citation field, Hl(t) ; and 0 Zn represents initial phases depending
on the value of the coercive force of the dynamic hysteresis loop.
Substituting (4) in (3) we find that the imbalance a.m.f., associated with the
first determinant of expression (3), will have an odd-harmonic spectrum, since the
net e.m.f., proportional to the value of H0 and associated with the second de-
terminant, will have an even-harmonic spectrum. Therefore ferroprobes and magnetic
amplifiers operating in this mode are often c.alled even-harmonic.
,
~
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ON1.Y
S1_nce any even harmonie in the spectrum of the output e.m.f.. (usually called the
second harmonic) can be easily filtered from odd harmonics of noise, then the
threshold sensitivity of ferroprobes and magnetic auipljfiers operating in the mode
indicated ts determined not hy the peak value of components of the imbalance e.m.f.,
but by the level of the magnetir.. noise in cores near the even harmonic used. It
is not difficult alsn to demonstrate thzt within the framework of the assumptions
made fluctuations in harmonics of the excitation field, H1(t) , in harmonics of
permeability u*(t) and ~:*(t) and the imbalance coefficient, e, can result in
broadening af tge lines of only odd noninformative harmonics of spectrum e(t) .
This is the main advantage of the ferroprobe and magnetic amplifier operating mode
discussed.
Taking into account expressions (3) and (4) it is not difficult to obtain expres-
_ sions for the transformation ratios of a ferroprobe and magnetic amplitier. For
_ example, for a ferroprobe and magnetic amplifier with longitudinal excitation with
a piecewise linear approximation of cur.ve B(H) oE the cores and sinusoidal,
triangular and square-pulse forms of the excitation field wave, we get the follow-
ing transformation ratios for the serond harmonic:
_ dE. 8 ~ H~ 7(H,
dB~ n 41 iP15 t FA mAxHm 1Y dEs / Hs `
G~� de - n u,s sin ~ n Hm.x
_ o
G (0 iQ1S F~ N'y, mexS1II (7t-t),
dE2 a4
~p dBp I Hmaa > HS 1L )
(5)
where BD = u0H is the induction of the measured field; E is the peak value
of the e.m.f. oP the second harmonic; H is the strength oi the saturation field
of the cores; H or H is thF peaksor maximum value of the excitation field;
and T is the duration o� t? Xa pulse.
The dependence of the transformation rattos af a ferroprobe and magnetic amplifier
on the degree of relative overexcitation of their cores is illustrated in fig 1.
Calculations performed according to equations (5) have demonRtrated that in prac-
tice ferroprobe and magnPtin amplifier tranaformation ratios of dozens of micro-
volts per nanotesla (uV/nT) are cmnpletely achievable. Since the noise level o�
elPCtronic amplifiers connected to the output of a ferroprobe and magnetic ampli-
fier is estimated in tenths and hundredths of a microvolt (at f.requencies above
103 Hz in the.l Hz band), thPn theae numerical values of the transformation ratios
of ferroprobea and magnetic amplifiers are totally aufficient in order to detect
~ and measure by their means magnetic fields at the level of single numbers of
picotestas. However, the�magnetic noise of ferroprobes and magnetic amplifiers
interf.eres with this.
4
11
FOR OFF[CiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00854R004500020034-7
FOR OFFICItiL USE ONY.Y
i
0,8
i 0,6
; 0,4
;
Figure 1. Dependence of Transformati.on Ratioa of a Perroprabe and Magnetic
Amplifier on Degree of Relative Overexcitation of Cores
Key:
1. Square-pulse
2. Triangular
3. Sine-rwave
Magnatic Noise Spectrum; RelAtionship Between Noiae Level and Threshold Sensitivity
of Ferroprobes and Magnetic Amplifiera
Cyclic magnetic reversal of the cores of ferroprobes and magnetic amplifiers is
accompaniad hy Barkhausen abrupt changes whose parameters--intensity, duration and
initial phase--fluctuate, varying �rom cycle to cycle, which reaults in the origin,
of magnetic noise.
It has been established experimentally that the spectrum of magnetic noise is
periodic. In it it is possible to isolate relatively uniform and sufficientl,y
extended sections s3tuated between diacrete lines--harmonies of the frequency of
magnetic reversal (excitation) of the core, as well as distinctly uniform and
considexably less extended sections directly abutting the discrere lines and charac-
terized by a rise tn spectral density.
~ The manifestation of noise in sections o� the firs*_ type can be described by means
- of the so-called "shot model" [S, 7, 8].
The appearance of noise in sections of the second type doeg not follow from any
kind of theoreti.caJ. structure, as the reault of which noise in these sections has
begun to be called excess. Excesa noise was detected experimcntally [5, 9].
_ Later it was established by many Authora that the rise in spectral density.of
excess noise obeys the rule 1/fY , whare f is the frequency of tuning out from
a line (for a demodulated fer.roprobe and magneti.c amplifier signal, thp excitation
frequencq) and Y ig an indicator which varies 4ver a range of 0.5 to 1.8. Ob-
servance of this rule has provided an occasion for making certain analogies, in
connection with whi.ch excess noise has begun to be called also f.licker noise [5].
5
- FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
0 1 2 J 4 Hmq,riN$

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004500020034-7
F()R nFb'1('IAI. iltiE t)N1.Y
It must be emphasized that in many practical cases (when measuring constant and
slowly varying magnetic fields and electric current) it is precisely excess noise
with its characteristic rise in keeping with the 1/f rule which determir.ES the
threshold sensitivity of ferroprobes and magnetic amplifiers. Therefore, studying
the nature of excess noi.se and finding methods of effectively suppressing it have
become of first-rare importance for practical applications.
Let us now dwell on the relationship between the noise level and the threshold
sensitivit;, ot ferroprnbes and magnetic amplifiers.
If the noise process is stationary and ergodic (this assumption agrees with the
experimental data), the root-mean-square value of the noise's e.m.f. equaZs
1/7'
Qe-where gf is the spectral rower density of the noise process in VZ/Hz and fl
and f2 are the lower and ugper limits of r_he frequency banci aelected.
Let us call the output noise level of a ferroprobe and magnetic amplifier the value ef Esh = KQE 'where K is a coefficient depending on the law of distri-
- bution of magnetic noise and the prescribed confidence level of the estima.te [4].
As illustrated in [10], the noise of a ferroprobe and aragnetic amplifier is dis-
tributed according to a normal law; therefore, with K= 3 the value of the esti-
mate of E equals the maximum (with a probabiiity.of 0.997) peak value of the
noise's e.~ f., which makes it possible to relate E to the frequently encoun-
tered estimates "for the width of the noise track" orh"from peak to peak," equal
_ to 2Esh '
As was demonstrated in the first section, the output e.m.f. of a ferroprobe and
magnetic amplifier depends on a whole series of factora; there�ore it is more
convenient to iise r_ot the value of Esh , but the value of Bsh , determined from
the equation

df
Ew r~
s= _ ~
m G G
' (6)
where G is the transformation ratio of a ferroprobe and magnetic amplifier
_ (when these devices operate at the Fsecond harmonic of the output e.m.f., G.= G2
We call the value of Bsh also the threshold sensitivity of a ferroprobe and
. maenetic amplifier [4]-
6
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
The noise level is of.ten also related to the vR1ue of bf _~/G (T/Hz1/2).
The value of bf can be called the spectra_1 denaity of peak v~.lues of magnetic.
noise. The relationship between bf and Beh ia determined by the e::presston
(h df .
r~
Let us note that the increase in values of B hand b when neaxing the discrete
= lines (excess noise) is considerab ly smoofihersfi-han in tie value of gf . This is
because of r.he fact that both quantities BSh and bf are Proportional to
f'
Let us make the following calculation. Let us represent the spectral der.sity
curve, gf(f) , of the demodulated signal of a ferroprobe and magnetic ataplifier
in the form of uniform (f > f2) and nonuniform (f < f2 ) sections, wher.e f2 is
the upper-limit frequency of excess noise, correapond3ng to the bend in the curve
(fig 2~. For the uniform section, gf(f) = gc = cunst ; for the nonuniform g�(f) _
= gc /f . Let Y= 1 and f2 = 1 Hz (with strong overexci.tation of the cores
of a ferroprobe and magnetic amplifier these values are close to real). Then
for the nonuniform section of the spectrum it is possible to write
ae = ~r ~df = ~8,~ -
ii
~
Taking into account the respective values of f and f and keeping in mind that
according, to (6) BSh tia E , we find that for ~1 = 10 32Hz , Bsh ti 2.6~ and
f or f 1= 10 7 Hz ,$Sh ti 4 gc '
Figure 2. Spectral Power Density of Magnetic Noise of a Demodulated Signal
of a Fprroprobe and Magnetic Amplifier
7
FOR OFFICiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE ONLY
Tt is obvious from this calculation that incr.easing the time for observing the
noise process from duzens of minutes to several mon*_hs does not result in any
substantial increase in the level of excess noise over the level of ordinary mag-
netic noise. For the same frequency band with a width of 1 Hz this excess is
2.6- to fourfold (*_heGe values correspond to the experimental data in [11]), wtliCh
makes it possible to estimate tYie threshold sensitivity of a ferroprob.e and mag-
t:etic amplifier from recordinga of noise obtained in relatively briPf rFalizations.
Let us also note that at frequencies below 10-3 Hz it is diff.icult to distinguish
~ the recorded noise of a ferroprobe and magnetic amplifier from zero fluctuations
and drif t whose origin can also not be related to processes taking place in cores.
This zero drift and fluctuations on one hand hamper obtainment of the spectral
distribution of excesG --.oise at ultralow frequencies and on the other explain the
- often encountered inctor_sistency between calculation and experimPntal data on the
noise level at these fiaquencies.
Prerequisites for the Origin of Excess Nolse Near the Second HarmonicAs was already mentioned, the appearance of excess noise does not follow from any
theoretical constructions. This relates especially to excesR noise near the second
harmonic of the magnetic reversal frequency. TYierefore, attempts have Leen made
from the very beginning and continue to be made to find some independent sources of
fluctuations causing the origin of excess noise.
The idea that the excess noi.se nf ferroprobes and magnetic amp:ifiers operating at
the secand harmonic is of a different physical. nature than ordinary magnetic noise
was first expressed by R. Yq. Berkamn [9], but he did not suggest any phyaical or
- formal mathematical model explaiaing the origin of excess noise.
- M. Vayner [12] attempted to explain the appearance of a false signal for the second
harmonic and the rise in noiGe nPar it on account of the magnetostr..iction pheno-
menon. He correctly establ.ished that magnetostrictive lengthening of the co.res of
a ferroprobe results in the movement of lines of force intersecting the turns of
the measuring winding twice per cycle; however, he did not study the change in
direction of these lines from half-cycle to half-cycle, as the result of which the
original equation presented by him proved to be invaltd. A critique of.Vayner'G
model is given in [13]. In spite of the erroneousness of Vayner's model, the fact
itself of a correlation between the level of excess noise and the magnetos tric t ion
coefficients of the ferromagnetic'materia].s used is interesting and has been es-
tablished by other ?uthors and, as will be demonstrated below, finds an explana-
tion within the framework of the dislocation model of noise.
N.N. Kolachevskiy and coworkers [5, 141 suggested that excess noise 3.n ferromag-
netics magnetirally reversed cyclically can be associated with fluctuations tn
magnetic permeabil.itv, similarly to how flicker noise in granulated resistors
observed during'the pzssing of electric current through them is related to fluctu-
ations in the value of resistanc.es. However, as already indicated abovP, fluctu-
- ations in the magnetic permeability of cores, on the assumption of central sym-
metry of the hysteresis loop and the absence of a constant component and even
harmonics in the excitation current, c:annot result in the or.igin of excess noise
near even induction or e.m.f. harmonics, i.e., this asaumptton r.aken by itself
8
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFiC1AL USE ONLY
cannot serve as the basis of the mechanism for the origin ot exceas noise near
even inductian and e.m. f. hartaont.cs.
ACtually, let us assume that the differential permeability of the cores, ud '
- fluctuates. Then instead of series (4) tfi is possihle to write
m
_ !~n ~t) = N~AO ~ Ft~n C6S7(2n w t + "n)
e �AO (r~ + A (t) cos (2n ~ t
+
i
nml . (7)
where yd0 and ua2n are the mean values of the parameters and Audd and
Aud2n are incremenrs in the parameters (r.heir fluctuation valves).
' By definition Au~0(t) = 0 and Aud2n(t) = 0. It is possible to assume algo
that the fluctuat on spectrum is concPntrated near zero frequency and even harmonic
lines. Substitutine (7) in (3) with H= 0 and H= H sin Wt , and assuming
that the fluctuation processec in diffe~ent cores are mutually not correlated,
we find the output e.m.f. of a ferroprobe and magnetic amp].ifier with longitudinal
excitation (a = S = 0):
~
e(t) _-co sw ~ �o N. cos uo t{ e[�no-1- cos (2n cu t-{-
~i
ym)] -i- k [A �nfl (0 +
00
+ I 0 cos (2n uo i -Fq)2.)]} ,
Rmi
(8)
where k is the number of cores. Ir is obvious from the expression obtained that
the output e.m.f. spectrum contains onl.y diffused lines of the fundamental frequen-
cy and odd harmonics.
Some investigators perceive the common physical natixre of both ordinary and excess
noi:se. The periodicity of the spectrum of g(w) with its chsracteristic rises
near even and odd harmonics of the magnetic reversa_1 frequency is explained by them by the correlation of Barkhausen abrupfi changes originating in different mag-
netic reversal cycles [15].
Let us note that the correlation approach was used earlier, too, for Pxplaining
the periodicity of the g(w) spectrum. Por example, F.V. Bunkin, based on ful-
fillment of the obligatory condition g(O) = 0, demonstrated that correlation o�
thp peak values nf Barlchausen abrupt changes separated by a whole number of mag-
netic reversal ha.lf-cyclES must result in rises in spectral density near odd and
in dtps near even harmonics. This conclusion was used for a long time in order to
_ prove rhe impossibility oP describing rises in noise near even harmonics by means
- of the correla*ion approach. However, as N.N. Kolachevskiy correctly observed,
9
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOid OFFICIAL USE ONLY
L-he result obrained by F.V. Bun.kin is only one conseq�ence of tha assumption of
preservation of thP strict symmetry of the hysteresis loop during cyclic magnetic
reversal.
P.Ye. Kotlyar [16], basPd on the opposite assumption (first expressed by K.A.
Goronina) demonstrated that the correlation approach is totally suitable also for
describing rises in noise near etien and odd harmonics. He obtained an equation
_ from which it followed that the level of noise neax even harmonics is higher the
more significant the asymmetry of the toon and the longer the correlation time for
Barkhausen abrupt changes.
The assumption of the asymmetry of the hysteresis loop as the major prerequisite
_ for the origin of excess noise near even harmonics was also used in studies by
the authors of the present article [4, 17]. We believe that this assumption does
not contradict the law regarding the mininum free energy in a ferromagnetic. This
_ law lies at the basis of the extreffie1y great synmetry of the major dynamic loops
of ferromagnetics but its statistical naturP is revealed with an increase in the
sensitivity and accuracy of ineasurements. The law begins to be fulfilled only
"on average," in a stretch of a great number of magnetic reversal cycles, which at
the same time indicates the fluctuation nature of the appearance of asynunetry of
the loop.
Formally, within the framework of the transformation mechanism whichwe discussed
in the first section, the appearance of excess noise near even induction anu e.m.f.
;iarmonics must be related to the presence of odd harmonics of the permeab3lity of
ferroprobe or magnetic amplifier cores. Actually, with violation of the symmetry
of the loop we must write, instead of (4),
~
�A (t) = �AO 57 cos (2n w t (p2,, + 0 (p) +
~i
~
-F v2�-1 (1) �A c2.-1> sin [(2n - 1) (J) 1+
nsi
-f- T2n-1 + A (p] I
(9)
where 0o is the phase increment caused'by asymmetry of the loop and
, v 2_1(t) represents random weighting functior.s with a mathematical expectation
o'~ zero (v2n-1(t) = 0).
Substituting (9) in (3) and limtting ourselves to a discussion of noise only near
even e.m.f. harmonics, we find by ana]ogy with (8) and under the same conditions
- that:
10
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
~
- e(t) w k-'I� scv t �o H,,, cos cu tv21-1 (t) x
nml
x �a (2fi-1) sin [(2n - 1) w t-}- (P2. + A(p]
(o /L-"' Sw t j Vgn_I (1) B2,, S1A (2ii W t + 4~,.) ,
n-i
(10)
where B represents coefficients of the series (the constant component and peak
vatueR oi neven induction harmonics); - ' qtn = F ((ps.-~ ~ Aq)) �
It is obvious i'rom expression (10) that the spectral distribution of excess noise
_ near the zero frequenc.y and even harmonics dppends on the spectrum of weighting
functions v2 1(t) . I'f the energy spectrum of functions v2n-l(t) varies
according to the 1/f rule, then the energy spectrum of excess noise will also
vary according to the same law.
In fig 3, in which a schematic representation is given of asymmetric hysteresis
loops corresponding to an individiial cycle or some sequence of msgnetic reversal
cycles, it is obvious that the asyntmetry of the loop can be related to a differ-
ence in positive and negative coercive forces (fig 3a), to a difference in residual
induction (fig 3b), to the different slope of loop back edges (fig 3c) or to their
bending (fig 3d), to different squarenPSS ratios of the loop or different valuea
of maximum induction (fig 3e). A constant component and even induction and e.m.f.
harmnnics are evidenced in all these cases.
M.V. Bukharov [18] directly related excess notse neax zexo fxequency and even
inducrion harmonics to fluctuations in coercive forces and residual induction of
the loop, considering these fluctuations statist3ca11y independent. He demon-
strated r.hat relative deviations in these values oi by 10 3 to 10 4 are completely
sufficient to cause excess noise at the level observed in practice and he estab=
lished a quantitative relationship between nuise levels at zero frequency and
in even harmonics. From the exnressions obtained by him it also follows that
the spectral digtribution of excess noise obeys the law of the apectral distribu-
tion of fluctuattonR in these parameters of,the loop.
Thug, fluctuations in parameters of the dynamic loop of a ferromagnetic, res.ulting
if only in short-dur.ation asymmetry of it, as well as the pr.esence of a correlation
- in Barkhausen abr�ipt changes, explaining rises in spectral density of the noise,
are basic, if not the major, prerequtsites �or the origin of excess noise near
- even induction and e.m.f. harmonics.
11
FOR OFF'ICIAL iJSE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000504020034-7
FOR OFFICIAL USE ONLY
Hc2 THO
872
By~
a) b) cY
AlTO4-2 Bma
J
~`Q7
Bmi d) . e)
Figure 3. Asymnetric Hysteresis Loops
Let us investigatp the physical easence of these prerenuisites.
"Disloration" Mode1 of Noise; Role ef Magnetostriction
Barkhausen abrupt changes originate primarily as the result of elastic interaction
- between moving domain boundartes and defects in the crystal lattice. Dislocations
are the largest defects in ferromagnetics [19, 20].
The interaction of a boundary with a dislocation is deacribed by the Peach-
Keller equation [19]: ..x " . .
. p� -Sdl (v b),
- (11)
where p is the force oi interaction; dt is an element of the length of the
dislocation line, R; 8 is the tensor of magnetostriction internal strgsses
caused by the change in the direction of magnetization at the boundarq; b is
- the Burgers vector. C. Rider made a calculation of internal stresses for the most important plane
boundaries encountered in iron, ntckel and cobalt. Iz was demonstrated that with
the proper choice of the system of coordinates (axis z is oriented along the
normal liRe to tlie boundary) it is possible to utilize only three components of
tensor 8. For a.straight-line dislocation sttuated para11e1 to the plane of
the boundary, from (11) it follows that
12
FOR UFFICIAL USF, ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
Pr = 6~ I yGz - QM lX6v Q~~
(12)
where ki and bi are components of vectors ~ and b.
Let us find the conditions under which force p is minimized and, consequently,
the intensity of Barkhausen abrupt changes and Ehe noise level associated with
them are diminished.
It is obvious from (12) that force p becomes equal to zero if vector ~ or b
- are perpendicu;ar to the plane of theZboundary. These orientations axe permissible
and possible for single nases of interaction betweeii boundaries and dislocations.
_ However, it is not known how to render so favorable orientations possible if,not
for a11 but if only for a sufficiently great number of theae typeR of inzeraction.
Therefore, for practical purposes another method of minimizing force MpZ i.j ob-
vAously more effective--reducing the absol9 te values of components Q11 , a22 and
a 12 of the'tensor of i.nternal stresses, 8.
M 'i M
As Rider demonstrated [19, 201, components 0,1 , J2 and o12 are.proportional
to magnetostriction coefficients a100 and lil oi the ferromagnetic material.
or alloy. In turn, coefficients X 100 and X 111 depend on the percentage content
of nickel in *_he alloy [21]. Both coefficien'A pass r.hrough zero wifih a nickel
content of 80 to 83 percent. Consequently,.a~ccording to expression (12), alloys
with this nickel content must possess minimum intensity of Barkhausen abrupt changes
and the lowest noise level.
This conclusion agrees totally with experimental estimates of noise levels found
earlier (for a wi'.de range of alloys) [2, 12, 16, 221.
' Elaboration of the conditions for minimizing force p for alloys with values of
caefficients a~~l`. and 1~11 closp to zero is of inEerest. The point is that in
- the nickel contenf range ina~cated (80 to 83 percent) coefficients .X 100 and X 111
become eq>>al _o zero not simultaneously (cf. table 1) [23]. Therefore, in mini-
mizins, 'Loree P it; is necessary to take into account both the absolute values
and signs of *hese coefficients, and a relationship with which both coefficients
(the positive, X 100' and the negative, alll ~wi11 differ from zero can prove to
be optimum. . .
Lxperiments and calculations [24, 25] which we have made have confirmed this assump-
tion.
Curves, plotted trom our experimental data, characterizing the xelationship between),,
the noise level and the saturation magnetostriction,. a, of alloys iised for vari- '
ous types of ferroprobes are gresented.in Pig 4. (Thesparameters of the ferro-
probes are presente3 below in the note to table 2.) It is obvious from fig 4 that
the noise minimum is shif.ted into the regivn of negative values of N , The
values found in the experiment, A=-(0.5 to 2.0)�10 6, are achievea only when
coef�icients X 100 and N 111 have different signs Pnd coefficient X 111 is neg-
ative.
13
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000504020034-7
FOR OF'FiCIAL USE ONLY
`Pab1e 1. Dependence of MagnPtostricCion Coefficients a100 and X 111 on Per-
centage Content of Nickel in Ni-Mo-Fe A11eys .
flponexsna coaepMaaee 1
A1oo , X 10 6
>�iii y 10-11
NI I Nb
. 80.36 4,26
2,11
0,22
81.0 3,50
1.12
-0 , 89
, 82,0 5,0
0,63
-0,45 �
84,0 2,69
-2,12
-2.34
Key:
1. Percentage content
. !
' 9D3N2
~r 9~3 N1
~ I
-6 -q `2 0 2 (X 10's)
Figure 4. R,elationship Between Noise Level, B , and Saturation Magneto-
striction, as , of Nickel-Iron Alloys for Various Types of Ferro-
probes
Key:
1. BSh, nT
The experimental data do not agree completely wi,th the calculated, since expression
(12) was obtained with certain assumptions and, in addi-i.:ion, was extended by us to
polycrystalline specimens of a11o}rs. However, the results obtained testify to
the advantage o� the "dislocation" model of notse develored here and indicate a
14
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR UFFICIAL USE ONLY
different, as comparPd with khat suggested eaxl.ier [12, 16], xole of magnetostric-
tion in the mechanism of the oxigin of noise.
Because of magnetos t ric tion interaction is established between elements of the
domain and dis'location structures of the ferromagnetic. Therefore, its role re-
duces to regulating these form3 oP interaction. Since the noise is determined
by the intensity of these forms of interaction, the pattern described above i.s
observed: With str.ong interaction the noise level corre]ates with the value of
Ix s I ; with slight interaction, when it is necessary to take into account the
signs of coefficients a100 and X 111 ~ the notse minimum shifts to the region of
negative values af X s .
The relationships found relate to an equal degree both to ordinary and to excess
noise.
Let us now explain or_ thP basis of the "dislocatiort" model the excess noise pheno-
menon itself.
Since action equals counteraction, then from (11) it follows that not only does
a dislocation act on a boundary with a.force of p, preventing its motion, but
- the boundary acts on the dicloration with the same force, but with the oppnsite
sign. But the reaction of a dislocation to this effect is considerably slower
than the opposite. This is due to kinetic restructuring of the dislocation's struc-
ture., obeytng the laws of plasticity [19]. A dislocation has low mobility. Rates
of their motion (under low stresses) equal 10 9 m/s and less. Domain boundaries,
on the other hand, are exceptionally mobile. Therefore, "repeatability" of events
of interaction between boundar.ies and dislocatinns for the duratinn of hundreds
and thousands of magnetic rev-ersal cycles is precisely the physical Prerequisite
for the establishment of r.he correlation of Barkhausen abrupt chan;es ands con-
sequently; the appearance of excess noise near discrete lines --harmonicsof the
magnetic reversal frequency.
id Let'us now discuss aspects of the interaction of magnetization with dislocations
resulting in asymmPtry of the hysteresis loop and fluctuations in it.
In [16] attention was turned Eo the presence in ferromagnetics of asymmetric micro-
hysteres3.s loops. It was noted that in spite of, considerable averaging, taking
place with the superposition of these loops in the bulk of the ferromagnetic,
the resultiag or macrohysteresis loop can remain asymmetric. Here two questions
remain unclear: What causes the asymmetry of microresistive loops and why averag-
ing does nat resu1t in strict symmetry of the resulting loop. The answer to the
first question is given in [20]. The authors discovered a"gate effect," con-
sisting in the fact that when a domain boundary passes through a dislocation in
a number of inatances a new domai.n originates, and the very fact of its origin
depends on the direction of motion of the baundary. If a doma.in originates in
forward motion of the b oundary, then its reverse motion "erases" the doma.in. This
fact also causes asymmetry of microresisr.ivP loops. The answer to the second qlies-
- tion inust be found in the kinetics of the reGtructuring of domain and dislocation
- srructures. In spite of the fact that a domain structure rather quickTy "adapts"
to a dislor_ation structure, tlie very fact of the restructuring of the latter
creates situations in which the overall gate effect in the ascending portion of
15
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
the loop can predomina`L.. aver thaC in the descending poxt3.on and vice vexsa.
This predomtnance is evidenced in a difterence in fihe coercive forces of the loop
(cf. fig 3).
5tre-3ses r.reated by dislocations influence not only displacement processes but
also processes of rotation and center formation. Tnducing anisotropy, these
stresses prevent syn.chronous (i.e., with the frequency of the excitation field)
_ magnetic reversal of "inflexible" regions and fix for a certain time centers for
the origin of regfions of opposifie magnetization. The characteristic delay times
here are also determined by the kinetics of restructuring of the dislocation struc-
ture, which results in if only a short-duration difference in the residual and
maximum induction of the loop, as well as in a difference in its squareness ratio
(cf. fig 3b and 3).
As mentioned in the preceding section, fluctuations resulting in asymmetry of the
loop are completely sufficient for the origin of excess noise near even harmonics
of the magnetic reversal frequency.
Thus, within the framework of the "dislocation" theory it is possible to find con-
ditions under which minimization of the magnetic noise of ferromagnetics subjected
to magnetic reversal cyclically is achieved, which is of great practical signifi-
cance. In addition, it is possible to explain the role of magnetostriction in the
mechanism of the origin of noise (which previously appeared puzzling), to under-
stand the nature of the correlation of Barkhausen abrupt changes separated by a
great_ number of cycles and the nature of fluctuations c.~f the parameters of the
resulting hysteresis loop and, at the same time, to exgiain the origin of excess
noise near discrete lines--harmonics of the magnetic revisrsal frequency--including
near even harmonics.
Factors Influencing the Noise Level in Ferroprobes and Magnetic Amplifiers
The difficulties in studying the basic mechanisms and in constructing an appropri-
ate theory of magnetic noise lie primarily in the multifactor nature of this
phenomenon. Naise depends not only on the.properties and-parameters of the ferro-
magnetic material or alloy used, but also on the� volume and form of cores, their
fabrication technology, heat treatment features, parameters of the excitation field,
. external mechanical stresses and the temperature gradient in the bulk of cores.
Let us discuss the influence of these factoxs on the noise level in ferroprobes and
magnetic amplifiers.
Composition of the Alloy
As was already indicated above, the cores-of low-threahold ferroprobes and magnetic
amplifiers must be made of aiioys containing 80 to 83 percent nickel. The fact of
a correlation between noise level and the percentage content of nickel in the
alloy was noted in [2, 22]. An experimental alloy containing 81.2 percent Ni,
2.5 percent Mo and 16.3 percent Fe was recommended in [16] as a 1ow-noise a11oy,
and in [26] an a11oy containing 81.5 percent Ni and 6 percent Mo.
16
FOR OFFI(IAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
In connection with the developmeni of new nickel-iron al].oys, we made studies of
the noise properties of new alloys and ones previously used with a percentage
content of nickel close to that indicated [17, 24].* The results of these studies
are given in table 2, in which the noise 7,evels measured axe campared not only,
with the values of saturation magnetostriction, AS , but also with other paramptpYg
of alloys.
- Table 2. Relationship Between Noise Level in Ferroprobes and Parameters of
Nickel-Iron Alloys

n\ BQ9MCT~N CflillWD
2 J
~ 3) ~~DO!l11D IIlyMa 031 BIi), H7JI (B 11NIOCt 48fTOT
IO.OOI...I.G r4) A71A 9lT6Ir!% pssnenxda o6paauoR 0Pi I-Ns 41
P,tapKa ennesa
i
Y. 10-6 I
Hc . A!M I
�a x1M
I
002. 7-1 ~
M 1 I
N12 I
No 3 I
Nr 4
79HIM
2.0
3,2
14
0,85
0,28
-
-
-
. '911,M
2,0
2,4
20
0,73
0,22
0,82
0,17
-
BQH.XC
-
2,4
22
0,63
0,14
-
-
-
83H N
0,5
1,6
58
0,60 ,
0,065
0,40
0,12
0,028
81H MA2
0,5
0,88
66
0,iO
0,060
0,34
0,065
0,015
82HMn
-1 ,3
1,35
64
0,67
0,050
0,27
0,04
0,012
84H -14 - -
- - 2,0 - , -
Key:
1. Type of a11oy
3. Noise level in ferroprobe, B
h '
2. Parameters of alloys
s
nT (in the 0.001 to 1.0 Hz fre-
'
quency band) for four different`
samples (Nos 1 to 01
4. T
1Parameters of ferroprobes: No 1--core
dimensions 130 X 2.8 X 0.1 mm, frequency
- f= 1 kHz ; No 2--core dimensions 10 X
1.2 X 0.05 mm, frequency f= 25 kHz ;
No 3--core dim?nsions, loop (20 X 6) -
(18 X 4) X 0.05 mm, frequency f= 25 kHz;
No 4--ring (diameter 13.2 mm, width 1.5
= 12~5 kHz .
mm, thickness 0.12 mm), f
v
2Annealing was performed at 800 �C, holding for 3 h, cooling at a rate of 200 �C/h
to 600 �C, then at a rate of 400 �C/h.
*The alloys were developed by the TsNITChermet [Central. Scientific Research In-
stitute of Ferrous Metalluxgy] imeni T.P. Baxdin Institute of Precision Alloys.
The parameters and characteristics of the new alloys (except a11oy 82NMP) are re-
gulated by GOST [A11-Union State Standard] 10160-75, "Alloys, Precision, Magneti-
cally Soft."
17
FOg OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFiCIAL USE ONLY
From table 2 it is obvious that rtew alloys 81NMA, 83NF and 82NMP have a lower
noise level than a11oy 79NM anti �Ci:KhS used in ferxoprobes and magnetic amplifiers
previously. Furthermore, alloys 81NMA and 83NF are characterized by a lower value
of saturation magnetostriceion, W = 0.5�10 6,' Alloy $2NMP (82NMP-VT), character-
ized by a negative value of S mS-(0.5 to 1.5)010 6, has the lowest noise level.*
Shifting of the noise minimum into the region of negative values of 71 is in
agreement with the F'dislocation" model of nciise and the theoretical postulates
discussed ih the preceding sectton.
Magnetic Properties of the A11oy
_ The magnetic properties of alloys and the parameters characterizing them are di-
- vided into structure-sensitive and structure-insensitive. Under the heading of
structure-sensitive parameters come the coercive fo,r.;:e, H, permeability, u,
and hysteresis losses, W; under structure-fnsensitive come baturation induction,
B, and other parameters. Within the framework of the "dislocation" theory of
noise developed by us it is to be expected that it is precisely the structure-
sensitive parameters of alloys which will correlate with the noise level observed.
- It is also obvious from table 2 that for the grnup of alloys with positive magneto-
striction (a > 0) the noise level measured actually correlates both withthe
- coercive force, H; and with a value which is the inverse of the initial magnetic
permeability, ua c These data do not contradict those obtained earlier in [22, 271.
If the main contribution to noise power is actually made by the halting of domain
boundaries in dislocations, then the spectral density of the noise's e.m.f., g,
must be proportional to the volume density of dislocations, p, and, consequently,
according to what was said in the second section, the following proportionality
must also be observed:
On the other hand, from applications of the theory of the ductility of materials
to ferromagnetics it follows that [19]:
Hc �Q ' P
Therefore, within the framework of the "dislocation" model a relationship between
the noise level, B h , and quantities H and u 1 is completely natural. The
curve reflecting t~iis relationship is shown in ftg 5[17].
In [14] on the basis of certain experimental data and assumptions regarding the
capacity for the fulfillment of an inverse proportion relationship between the
*The parameters and pxoperties of a1.1oy 82NM axe presented in the catalogue .
"Novyye pretsizionnyye splavy" [New Precision A11oya], Moscow, Chexinetinformatsiya.
1979.
18
FOR OFFICIwL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
coercive force, Hc , and ulAximum permeabi],f,ty, p ma , fihe conc],usion is drawn
that for the purpose of improving the signal-to--nolse ratio it is necessary to
_ strive to increaae the raLio V X/1% 5I . Tn thfs connection it is necessary to
make the following comments. rst, wtth an increase in u not only the le-
gitimate signal can increase, but also the noise, especisllyawhen the increase in
u is caused by formation of the alloy's texture. Therefore it is hardly pos-
sHYe to ac'.ii_eve considerable improvement of the signal-to-noise ratfo only by
increasing the value of u . Second, a distinct inverse proportion relationship
has been established up to now only for quantities H and u [17, 19], but not
for quantities H and u . Third, as was already explained in detail, it is
not possible to u,ilize theaabsolute value (ag~ , since the noise minimum is shifted
into the region of negative values.
Let us discuss the relationship between noise and hysteresis losses. Hysteresis
losses, W, are proportional to product Hc B ; therefore the spectral density of
the noise must be proportional to this quanttty. However, since the structure-
sensitive parameter of an alloy is all the same the coercive force, H� and not
the saturation induction, B, then the noise correlates more strongly with quan--
tity Hc than with Bs (c~. table 2).
~~fl, NTA
0,2
1 2 J 4 Hr, A/,w
~ -
(5'l04) :#a
p ~/Z
' Figure 5. Dependence of Noise Level, B ~1~~, nn Coercive Force, 0, and the
- Inverse Value of the Original-Iagnetic Permeability, X, with a
Coupling Constant of 5�104
Key:
1. Bsh , nT
19
FOR OFFICIAL �USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000504020034-7
FOR OFFICIAL USE ONLY
The squareness of the hysteresis J.oap also inf :luences the noise 1eve1. Already
with the data of [1] it was possible to conclude that the noise increases with
an increase in pr1 tareness of the 1oop. NEw data [14, 281 confirm this. It was
noted in [14] th in alloys with a square loop the noise increases on account of
the consolidatior. )i indivi:dual regions magnetically reversed simultaneously. It
is also noted that the consolidation of these reg:tons and, consequently, the re-
duction of their number, must result chiefly in an increase in excess noise. In
our opinion the increase in ooth excess ar.d ordinary noise in alloys with a square
loop is associated primarily with the presence of a texture preventing the magnetic
r,::versal of "inflexible" regions. This question wi'_1 be discussed below.
Volume ond Form of Cores; Their Fabrication Technology
_ It was demonstrated above that the spectral noise density, gf , is proportional
to the volume density of dislocations, p. From this the conclusion can be drawn
that with a given p the spectral noise density must be proportional also to the
volume of the core, i.e., gf ti p.
:iowever, since with a change in the volume of cores, v, not only quantity gf
changes, but also quantity G--the transformation ratio of ferroprobes and mag-
= netic amplifiers--and both these quantities enter into expression (6) for the
threshold sensitivity, B , then dependence B (v) will be more complex. Let
_ us discuss some particular cases. sh
Let there be a ferroprobe with relatively short cores. The maximum differential
:
permeability of these cores is determined by the exp.ression [1, 21
. ~2
�,a~xl~ - � a m - A--'
A max s
(15)
where u is the maximum differential permeability of the material; m is the
permeabiii~yXOf `the form; and A= A(1C, s) is a function depending slightly on
Q and s .
Substituting expression (15) in (5) and disregarding dependence A(R, s) , we get
G ti QZ , i.e., the transforniation ratio of a ferroprobe witli short cores does not
depend an the area, s, of their cross sectian. Therefore, with one and the samR
internal maximum strength of the excitation field, disregarding the inf luence of
the demagnetization factor on processes of the origin of Barkhausen abrupt changes
and their space correlation [5], we get from (6):
ti
BW ca Im . �A max;
i -conet ; 0
(16)
It is obvious from expxession (16) that in ferroprobes it is feasible
the volume oP cores only on account of incxeasing their length. When
of cores is increased on account of inereasing their cross-sectional
20
FOR OFFICIAL USE ONLY
to increase
the volume
area with
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00854R004500020034-7
FOR OFFICIAL USE ONLY
- an assigned length, as the experimenta7, data in [1, 191 dempnsxx'ate, the xeduction
in the noise 1eve1 B , is exceedingl.y~ slight. Moxeover, as wi11 be demonstrated
- below, with an assigne~ ~ength ferroprobe cores have an optimum cross-sectional
area (with regard to the noise minimum).
The situation is otherwise with magnetic amplifiers. In closed cores of magnetic
amplifiers a demagnetization factor for the circular �ield measured is lacking, as
the result of which the transPormation ratios prove to be proportional to the cross-
sectional area of the cores, i.e., G ti s. Therefore, with the same assumptions
we get from (6) :
~w.%'k Y I=con't r0~U .
- (17)
From this it is obvious that, unlike ferroprobes, in magnetic amplifiers it is
feasible to increase their volume precisely by increasing their.cross-sectional
area.
It must be emphasized that the transformation ratio of a ma$netic amplifier, be-
cause of the absence of a demagnetization factor in its cores, is always higher
than the transformation ratio of a ferroprobe. Therefore, in spite of the higher
exponent for v in expression (16) the noise level, B h, is higher in a ferro-
- probe than in a magnetic amplifier [1]. With a great ~ncrease in the length of
ferroprobe cores condition m� ud X is violated and expression (16) becomes
invalid. At the limit, when the inf~uence of the demagnetization factor in ferro-
probe cores can be disregarded, dependence Bsh(v) wi 11 be described by expression
(16).
Expression (17) agrees well wi.th the experimental data in [1, 101 (as noted in [2],
this agreement with the experimental data indicates the absence of substantial
transverse correlation of noise in cores) and reflects rather graphically the fact
that the spectral density of noise power is actually proportional to the volume
- of the cores, gf ti v. On the basis of this, in [10] it is suggested that the
- noise proper ties of materials and alloys be characterized by parameter CB
(let us use our symbols--cf. the second section):
l/, i' glu
- Cf; = efv = GntY ~
_ 2
' which does not depend on the volume of cores, since g�v ti v and GMU [magnetic
amplifier] v (with k = const
The cores of ferroprobes and magnetic ampliPiers are usually made ox a rolled
narrow strip 0.02 to 0.1 mm thick. Thinner strips are not used because of worsen-
i.ng of their magnetic properties (reduction of initial and maximum permeability ,
21
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400540020034-7
FOR OFFICIAL USE ONLY
and increase in coercive force [291), and thicker ones because of the spin
effect and losses associated with ifi [1, 291,
Cores are made by the method of die forging or winding onto a supporting frame
(a spool). In the latter instance a strip of fihe prescribed width is cut by
means of precision roller shears. After the stock 3s cut out and cut burr s are
- removed and their surface is chemically or electrically polished. These opera-
_ tions reduce the density of defects (dislocations) originating in large quantity
in the vicir.ity of the cutout line or cutoff line.
A further reduction in the density.of defects is achieved by heat treating the
cores.
Heat Treatment of Cores
The great influence of this factor is explained by the fact that in the heat treat-
ing process certain types of defects are partly eliminated and internal stresses
are reduced [19, 21].
The mode of heat treatment of cores is characterized by the maximum annealing tem-
perature, the medium in which annealing takes place and the rate of cooling of
cores.
- The dependence of the noise level on the maximum annealing temperature f or cores
- was studied in [24, 30]. It was established that the minimum noise level for cores
made of alloy 81NMA is achieved a.t annealing temperatures lower than those recom-
mended, and precisely at 800 �C. According to the data of [30], at the same an-
nealing temperature a noise minimum is observed also in cores made of alloys
79NM and 83NF, and for alloy 79NM this effect occurs if the alloy is produced by
the electron beam remelting method, but not by open induction melting. In [24]
the effect of a reduction in noise level with a reduction in the maximum anneal-
ing temperature for cores is related to an increase in the hardness of the alloy,
and in [30] to the formation of its fine-grained structure. These.explanations
are contradictory and the effect itsplf requires further study.
- The dependence of the noise level on the medium in which annealing of cor es takes
place was noted in [26, 30]. Annealing in a hydrogen atmosphere as compared with
annealing under vacuum results in an approximately 1.5-fold reduction in noise
level.
The dependence of the noise level on the cooling rate of cores has been d iscussed
in a number o� studies. In [22] the following general rule was established for a
considerable number of binary nickel-iron alloys: The noise is reduced with rapid
cooling of cores. Unfortunately, in this study data are not given on the procedure
and rate for cooling cores and an explanation is not given of the pattern observed
itself. It is necessary to judge regard3.ng this from the data of other studies.
F'or example, in [21, 31] it is noted that with the rapid cooling of cores the
alloy becomes less ordered and more magnetically soft than with slow coo ling.
Rapid cooling is conducive to a reduction of magnetic anisotropy [21] and, con-
sequently, to the probability of the existence of "inflexible" regions which
22
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00854R004500020034-7
FOR OFFICIAL USE ONLY
result in a.s.ymmetry of the loop arkd is one at the reasons fox the appearance of
noise near even harmonics of the output e.m..f.. of afexxoQxobe and magnetic am-
plifier.
The question regarding the magnetic texture of the material acquired in the process
of coaling of cores deserves attention. The permeability af the form of ferro-
probe cores varies over a range of (1 to 20)�103: Therefore, even in the magnetic
field of the earth induction in cores can have values on the order of 0.1 T and re-
sult in the formafiion of a magnetic texture. The appearance of a texture will also
be conducive to the securing of "inflexible" regions.
In [24, 30) a study was made of the influence of the rate of cooling of ferroprobe
cores at 100 and 1500 �C/h over a fiemperature range of 700 to 350 �C. A naticeable
difference was not observed in the minimum achievable noise levels corresponding
_ to these rates. However, it was found in [30] that two- to threefold lower values
of the excitation current correspond to a faster rate of cooling (1500 �C/h) for
cores made of alloy 79NM, whereby the same noise minimum is achieved. These studies
are of undisputed interest since they make it possible to obtain information on
- the mechanisms for the orinin of noise at the stage of rotation processes and
- therefore should be continued.
Excitation Field Parameters
_ These include peak value, degree of inhomogeneity, direction and frequency.
- A curve characterizing the dependence of the noise level, B , on the peak value
of the excitation field, H, for ferroprobes and magnetic amplifiers with longi-
tudinal excitation and outpnt at the second harmonic ia shown in fig 6.
It is possibe to explain this shape of the B(H ) curve if this curve is repre-
sented as the resultant of two other curves, Psamshown in fig 6. The first charac-
terizes the growth in volume density of the interaction of domain boundaries with
defects (dislocations) or, which is the same thing, the increase in the number of
Barkhausen abrupt changes; the second characterizes the drop or reduction in fluc-
tuations of the par.ameters of abrupt changes and consequently parameters of the
hyst"eresis loop as the maximum magnetic reversal cycle is approached.
With very low peak va]ues of the excitation field, interaction of domain boundaries
with dislocations is reversible (the R.ayleigh region), Barkhausen abrupt changes
still do not originate here and noise in the ordinary sense of this word (i.e.,
- as some random process) is absent. However, strong displacements (on the order of
10 8 to 10 6 T) are observed in magnetic amplifiers and ferroprobes. 'rhese dis-
placements are the result of a great number of "inflexible" regions not partaking
in the magnetic reversal process, and can be rEgarded as a 01frozen" fluctuation of
- a specific parameter of the hysteresis loop.
As the peak value of the excitation field increases, the interactton of boundaries
= with dislocations becomes irreversible, Baxkhausen abrupt changes origfnate and
iridividual "in�lexib],e" regions begin to be magnetically reversed (although not in
step). The "brittle springs" model [32] gives a clear idea of these processes.
23
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00854R004500020034-7
FOR OFFICIAL USE ONLY
The sudden rise in noise in this. section is directly related to the increase in
rhe volume density of these !'orms o,f interaction. Finally, with a specific peak
value of the excitation fie1d the growth in noise ceases. In spite of a continued
indrease in the number of Barkhausen abrupt changes the mechanism for the suppres-
sion of fluctuations of the garameters of abrupt changes and for stabilizatian of
the parameters of the hysteresis loop goes into effect. The balance of increases
in these two trends also determines the point of the ma.ximum on the Bsh(H ) curve.
The noise 1eve1 is reduced, first rather quickly and then more slowly, witW a
further increase in the geak value of ti-ie excitation field.
omN. ed
4
3
2
~
1 ~
1 2 Nm.rnW.ea
Key:
Figure 6. Dependence of Noise Level, B (Normal Component) .on Peak Value
of the Excitation Field, H s~H curve characterizing
intensity of Barkhausen abrupt TRIAges; curve characterizing
number of "inflexible" regions and level of fluctuations caused
by them
1. Bsh , relative units
A similar dependence for BSh(H ) is observed in testing various designs of ferro-
probes and magnetic amplifiers T1, 2, 9, 10, 16, 331. Furthermore, as emphasized
in [10], the drop in noise is associated precisely witYt the peak or maximum value
of the excitation field and not with the form of its wave. At the same time,
since the value of B is inversely proportional to the transformation ratio,
G,(cf.*expression (e~), which in many cases diminishes with an increase in the
excitation field, effective lowering of the noise level on account of an increase
in the maximum value of the field is made possible with a square-pulse wave form
~ (cf. fig 2). .A close to square-pulse wave form is schieved, for example, in the
ferroresonance mode of exciting magnetic amplifiers and ring ferroprobes [34].
24
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
Experience has clemonskxZked CliAC w.i.tli itil ittcxectse in the peak vA7.ue of the excltA-
tion Field both ordinary and excess noise are xeduced. Fuxthexmore, excess noise
is "retained aga,inst," as it were, discrete lines--ha,xnnonics of the magnetic re-
vet9al frequency [11]. TAis behavior of excess noise can be re7.ated to the ever
more compl.ete and sysichronous (i..e., with the frequency of the excitation fie1d)
magnetic reversal of "ittflexible" regions, as we11 as to symmetrization of the
hysteresis 1oop, which was talked about in the third and fourth sections.
~
~
, The fairly slow drop in noise in strong fields is associafied with fihe existence
of crystallographic and induced anisotropy. Tt is precisely anisotropy which pre-
vents the synchronous magnetic reversal of "inflexible" regions. Therefore, for
the purpose of the effective suppression of noise on account of an overexcited
operating mode for ferroprobes and magnetic amplifiers it is necessar}r to use ma-
terials possessing not onlp 1ow values of saturation magnetostriction, but also
low values of crystallographic anisotropy constants, as we11 as slightly textured
materials. Under the heading of these materials come the already discussed nickel-
= iron alloys, in which the effectiveness of noise suppression on account of overex-
citation increases with an increase in the rate of cooling of cores after they are
annealed [22, 30, 31]. The strong influence of the rate of cooling of an alloy on
the noise level leads to the conclusion that not only these alloys but also the
amorphous materials under intense development [35] can prove to be promising for
use in ferroprobes and magnetic amplifiers. The fabrication of amorphous materials
involves the exceedingly rapid cooling of a melt put into a rotating drum. The
_ cooling rate can reach 166 �C/s. As a result a magnetically soft material is pro-
duced with a homogeneous structure and lacking magnetocrystalline anisotropy.
Data on magnetic noise in these materials is presented in the section "Ring Ferro-
probe with Low Threshold Sensitivity."
It has also been demonstrated in an experiment that the effectiveness of the sup-
pression of noise by increasing the peak value of the excitation field depends on
the degree of inhomogeneity of this field. In the majority of studies [1, 2, 271
_ devoted to studying the nature of excess noise, this dependence is directly related
to the existence of "inflexible" regions and the hysteresis memory of cores, de-
termined from the level of the residual signal of a ferroprobe or magnetic ampli-
fier after the short-duration influence on its cores of a strong steady magnetic
field. It is known that the noise is reduced if, other conditions being equal,
arrow-shaped or ring cores are used, making possible high uniformity of induction
along their length. The external constant field (the measured or compensating)
musf also be homogeneous. It has been noted also that inhomogeneity of the external
field results in asymmetry of the hysteresis loop [36].
An attempt at a"model" interpretation of the phenomena discussed here was made
in [4]. Based on a model of "reversible ellipsoids of striction" various experi-
- menral facts w?re compared and again the role of crystallographic and induced an-
isotropy in suppressang noise by means of changing the parameters of the excita-
tion field was emphasized. It was demonstrated that the elongated (sheetlike)
shape of "inflexible" reg:tons in the direction of the excitation field prevents
their magnetic reversal and is one of the main reasons for the origin of excess
_ noise, hysteresis memory and asymmetry of the magnetic reversal loop of ferroprobe
and magnetic amplifier cores.
25
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USIE ONLY
'i'he model of "xeversiblc e7,7,ipsaids og stxicCion" mke.& 1t poasible to undexstand
also the dependence of noise, level nok on7.y 4n a change in peak val,ue, but also
on a change in dixection of the excitation fie1.d, Tn pAxticulax, the fact of a
reduction in noise in ferroprobes with transvexse exei,tatioa w3:th'the application
of an adde.d longitudi~nal variab.7,e fie7,d to cores [37], in a Schonstedt ferroprobL
with "skew magnetic reversal" of cores [38], and in a T,angvagen ferroprobe with
the excitation of cores by means of a rotating fie1d [39] becomes totally explain-
ab1e. From the same model, it is not difficult to draw-the conclusion that the
methods of excitation enumerated wi11 be effective only with the presence of aniso-
tropy. If the anisotropy is eliminated, the noise must depend slightly on the
change in direction of the excitation field, which has been observed experimentally.
With the excitation of ferroprobe and magnetic amplifier cores by means of a rotat-
ing field with a sufficiently high "peak value," displacement processes:and, con-
sequently, events of the interaction of domain boundaries with dislocations re-
sulting in Barkhausen abrupt changes are practically eliminated from the magnetic
reversal cycle. This means that, at least, the level of ordinary magnetic noise
in a ferroprobe and magnetic amplifier with a rotating excitation field must be
considerably lower than in ferroprobes and magnetic amplifiers with loi;gitudinal
excitation. A rotating field acts effectively on excess noise and hysteresis mem-
ory only with the presence of crystallographic and induced anisotropy. The spec-:
tral distribution of noise in ferroprobes and magnetic amplifiers with a rotating
excitation field requires additional study. The promise of these ferroprobes and
magnetic amplifiers for measuring the parameters of weak and ultrawaak magnetic
fields is undisputed.
Let us now discuss the dependence of the noise level on the frequency of the ex-
citation field. It is obvious that with an increase in the frequency of the ex-
cication field the number of interactions of domain boundaries with dislocations
per unit of time also increases. By analogy with what was said above we conclude
that the spectral density of noise, gf ,(cf. the 4:hird section) must be propor-
rional to the fr2quency, f - w/27r , of t?:c c::c:.t~tion fiel.d, i.e., b^. f
However, the noise level, ~Bsh , depends not only on gf , but also on the trans-
formation ratio of the ferroprobe and magnetic amplifier. Since according to (5)
G ti fb , for Bsh we have [1,2]:
B ti f-1/2
sh v
(18)
It is obvious from (18) that the noise is reduced with an increase in the fre-
quency of the excitation field. However, at least two factors pr.event an unlimited
increase in the frequency of the excitation field: an inereasF in magnetic rever-
sal losses and ever greater maniPestation of the spin effect; and approaching sub-
harmonics ar rhe fundamental frequency oP the magnetostrictive resonance of cores.
Both factors result in an increase in noise [1, 2, 16]. Let us dwe11 in greater
detail on the seccand factor. It was experimentally established in [16] that both
orclinary and excess noise increase abruptly as magnetostrictive resonance is ap-
proached. In [16] this rise in noise is associated with flucttiations in the para-
meters nf defects (dislocations) and with their possible migration. Within the.
framework of the model of "reversible ellipsoids of striction" the rise in noise
26
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
can be associated a1.so with induced (on account of mecha,ni.ca], vibxations of the
core) macroscopic and microscopic anisotxopy [4, 351. It is posaibl.e to avoid
these undesirable phenomena if the length o!' aferroprobe and magnetic amplifier
rvd core and the diameter of a ring core, with an assigned frequency, f, of the
excitation field, do not equal tlte forbidden values, dete~cmined by~ equations
(19): .
,
1senp k E
= -
4nj~ p
d�nP 2 n,n/, Y p~,-}- (1- k)2 ,
(19)
where n= 1, 2, 3, represents the order of perturbing vibrations (taking
Ynto account the parity of the magnetostriction effect); k= 1, 2, 3,... is the
order of inechanical vibrations of the cores during resonance; E is the elastic
modulus; and p is the density of the core material.
External Mechanical Stresses
As we know [21], external unidirectional mechanical stresses resulting in the
extension or compression of cores alter substantially the shape of magnetization
curves. Here the effects observed depend on tha sign of saturation magnetastric-
tion, a. If a> 0, then extension of a sample results in a sudden increase
in maximum magnetic permeability and a reduction in initial permeability. With
a< 0 the opposite picture is observed. Already from these facts, taking into
account what was said above, it would be possible to conclude that the noise should
increase with the extension of cores made of alloys with as > 0 and decrease
for those made of alloys with X S < 0.
The results of direct measurements are presented in [40]. The author relates to
the induc*_ion of magnetic anisotropy the rise in noise in samples with a> 0'
with the presence of a tensile load. Tension results in an increase in domains
the direction of magnetization in which is close to the direction of tension. This
direction also becomes an easy axis. If the external field is directed along this
axis then on average Barkhausen abrupt changes should increase. In samples with
a< 0 the easy axis is perpendicular to the direction of tension. Therefore,
wflen a field is applied in the direction qf tension the intensity of Barkhausen
abrupt changes wi11 be reduced. With this samples become less anisotropic than
before the beginninfi of tension. -
The model of "reversible ellipsoids of striction" [4] gives a clear idea of the
mechanisms described here.
It is hardly feasible to utilize these mechanisms for the purpose of suppressing
noise. However, they should be recalled in a11 cases when cores are subjected to
external stresses because of working conditions, e.g., when they are rigidly
fastened to a base and the transformers themselves--the ferroprobe and magnetic
amplifier--are designed for operation over a broad temperature rar.ge.
27
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
Temperature Changes
FOR OFFICIAL USE ONLY
Let us dwe11 first of all on the dependence of thP nqi,se 1.eve1, 8 , on the r..hange
in the mean temperature of cores. Direct measurements of the spectral density of
magnetic noise in various fexromagnetics aver a wide tempexature range--from
helium temperatures to Curie points--made by N.N. Kolachevskiy [5] have demonstrated
that the noise 1eve1 remains unctanged for nickel-iron alloys in the 100 to 300 �K
tpmperarure range.
Invariability of th? noise level in the range indicated for the� variation of inean
remperatures of the samples studied can be associated f.irst with the rel2tive
constancy both of saturation magnetostriction and of magnetic anisotropy [21] and
second with the relative constancy of magnetic characteristics, i.e., the coercive
force and initial.permeabilitg, which bear information regarding the volume density
of dislocaLions [19]. With an increase in temperature to values close to the Curie point, saturation
magnetostricticin and magnetie anisotropy begin to di.minish suddenly (more rapidly
than saturation induction), the coercive force is reduced and initi�1 permeability
is increased, which, taking into account relationRhips (11), (13) and (14), as
well as what was said above, should result in a reduction in Barkhausen abrupt
changes and noiRe level. The dependence of the noise level on the change in tem-
pPrar.ure as the Curie point is approached for alloy 72NrIDKh with a low Curie point
was anaTyzed in detaiT.in [41]. In fig 7 these dependences, Bs}, (T) , are compared
with dependences B(T) and u(T) for the same alloy, taken by us from [29].
It is obvious from the dependences presented that a maxtmum on the u curve
corresponds to the noise minimum, which agrees with equations (13) ana (14).
BuU. ommed.
10
S
es.A l*a(dnM)
1.)
BS
2)
BW
~
O,Z
~o
B~
~
I v11a~s,
'Y
Q 5!J lUU 4
40
40
Figure 7. Dependence of B , B and u on Temperature as the Curie Point
Is Approached forhAlloy 72NMDKha
Key:
1. BSh , relative unita
28
2. BS , Ts �a (relative)
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
- FOR OFFICIAL USE ONLY
Tn the region of low tempexatuxes (less than 100 aK) tqagn?tic anisotropy begins
Co increase [21], whiGh according to the model of "reversible ell,ipsoids of
str.iction" and wha.t wa.s said above should xesu9.t in an increase at least in excess
hc+i9e [4]. An increase in noise in a fer.xop.xobe at low tempexatures was described
in [42]. Tlie fexroprobe was p7.ace,d in a superconducting shiel,d and cooled from -
room temperature to the temperature of liquid helium. At room temperature (300 �K)
the noise lPve1 equals 0.3 nT, at *_he temperature of liquid nitrogen (77 �K) 0.9
nT and at the temperature of lfiqutd helium 3.0 nT, i.e., an order higher than at
room remperature.* Th.e ferroprobe was also heated to 423 �K. Within the range of
300 to 423 �K a noticeablP change in noise 1eve1 tn the fexroprobe was not detected
(the fPrroprobe cores were made of a11op 79NM with a Curie point of 0= 723 �K).
Let us now proceed *_o a discussion of the question of the inf luence of time and
spare changes in temperature on the noiRe level.
In [S] a brtef curvey of publications is given in which a study was made of thP
relationshin between the nonrecurrence of Barkhausen abrupt changes and temperature
changes. Referring to [43], in which astrong dependence of the noise level in
nir.kel-iron alloys on the rate of change in temperature was observed, the author
expressed the hypothesis that this dependence fs caused hy'the "settling" of
energy states and by a change in the magnetic structure of the ferromagnetic.
Another interpretation of this dependence is also possible. With a sudden change
~ in the temperature of the envi.ronment, because of the slower propagation of heat
in the core, described by the heat conduction eqtiation, a temperature gradient
originates in its depth. Both fluctuations in the parameters of Barkhausen abrupt
changes and fluctuations in thermoelectric current [44], which add their contribu-
; tion to excess noise, can be associated with the origin of a gradient and :tts sub-
sequent elimination on account of equalization of the temperature in the core.'
~
- A direct relationship between the noise level and the temperature gradient compon-
ent for the length of ferroprobe cores was established in [45]. Cores of uniform
. cross section were used, the magnetic induction in which, as we know [1, 21, is non-
' uniform even with a homogeneous external field: Induction i4 maximum at the center
of the core and diminishPs toward the ends according to a parabolic law. The
~ temperature was measured along the length of cores when the latter were retained
against bases made of AG-4S or DSV molding material with low heat conductivity
- and it was established that the shape of the T(x) curve, where x is the dis-
_ tance from the center of the core, in general features agrees with the shape of
the B(x) curve. Fiirtfiermore, the noise level in the ferroprobe was measured.
Then the same cores were placed on bases made of VT1-0 titanium, the thermal con-
ductivity of which is two orders of magnitude higher than that of bases made of
*This experiment was partly repeated at thd A1.7.^Union Scientific Research Institute
of Metrology imeni D.I. Mendeleyev ScienCific Production Association in 1978 by
V.I. SherPmet and V.N. Gorobey with a ferroprobe of a different design. At the
temperature of liqui.d nitrogen the noise 1eve1 proved to be three- to fourfold
higher than at room temperature.
29
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
rnolding materials. The measureme1lts demonatxated that wi.th the pxesence o� ti-
tanium bases the tempexatuxe al,ong th.e~ 7.engrh of the coxes ts equal.ized, and the
noi5e (chief7.y. its excess component) is reduced two- to 2,5-fold.
Later the authors oP t.his. axticl,e znade a similar exQexi,mp-nt (without plotting the
T(x) curve)-with arrow-shaped cores making possilile su�fieient hQmogeneit}r of
magnr:tic induction along their-length. Replacemenfi of the bases did not result in
such a consideralile cha_:&oe in the noise leve7. in.the fexxoprobe.
From these experi_ments it follows that, first, the noise level is actually associ-
ated with the temperature gradient in cores and, ser_ond, the temperature gradient
itself in many cases is caused by the inhomogeneity of magne*_ic induction in cores,
appearing either on account of their shape or on account of the inhomogeneity of
the eYcitation field. Both tY!ese facts must be taken into account in designing
a 1ow-threshold ferroprobe anrt magnetic amplifier.
In nur opinion studies of temperature influences on the noise level, espe,cially
in the area of explaining r.he physical pattern of these influences, should be
continued.
Other Factors
Let us dwell �briefly on some other factors the mechanism of the influence of which
however still requires its own interpretation.
It has been observed that the noise is reduced when ferroprobe and magnetic ampli-
- fier cores are placed in a viscous medium or fluid, e.g., glycerine. The core of
a rin.g ferrop;.obe was placed in GKZh-94 fluid,. and a three- to fourfo].d reduction
in noise was observed.* This effect is apparently associated'with a reduction in
- the intensity of magnetostriction vibrations of the core on account of sup.pression
of the reflected acoustic wave [46].
Adependence of the noise level on the diameter of the turns of an excitation coil
has been observed. It has been established that with the same excitation current
a threefold reduction in the diameter c+f the turns of an excitation winding results
in a two- to 2.5-fold reduction in noise level. Wa observe this effect both for
rod and for ring ferroprobes.
A dependence of the noise level of a ferroprobe on the diameter of the turns of
the measuring winding was obserned even earlier [33] and again confirmed by our
experiment_s. It was established that the greater the diameter of the turns the
lower the noise 1eve1. In our experiments a twofold increase in the diameter of
the measuring coil resulted in a 1.5- tc twofold reduction in noise (with a simul-
taneoiis and quite considerable reduction in imbalance e.m.f.). One
*The experiment was prepazed by L.Ya. Bushuyev and was conducted in a superconduct-
ing shield by V.I. Shere.met, Yu.N. Bohko-17 and V.N. Gorobey.
30
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00854R000540020034-7
rOk OMHICtnL tISw: Otvt.v
noncon.tradictoxy explanafiion of the noise reduction ei'tect is given in [33] and
other explanations can be given on the basis of talcing finto account featuxes of
the space correla,tion of the e.m,f. of the noise in the vi,cinity of fexxoprobe
and magnetic ampli#iex Gores� [16]. .
Ring Ferroprobe with T,ow Threshold Sensitivity
Taking into account tlie ideas developed in this arti,cle regardfing the nature of
magnetic noise and earlier established and newly discovered experimental factors
resulting in a reduction in noise, we developed a ring ferroprobe with low thresh-
o1d sensitivity. .
The advantage of a ring ferroprobe over a rod type is first of all the high homo-
geneity of the excitation field, which eliminates a number of sources of noise
_ indicated in the preceding section.
- Actually the ring ferroprobe was invented by Ashenbrenner and Goubou. Considerably
later improved designs of ring ferropr.obes were suggested hy Geiger and, indepen-
dently of him, R.Ya. Berkman [2]. Gordon [13, 26] substantially improved the para-
meters of ring ferroprobes on account of using a new nickel-iron alloy (close in
- chemical compositi.on to domestic alloy 82NMP) and using a metallic supporting loop
onto. which a strip of this alloy w2s wound and was annealed tog?ther with it.
As a material for the supporting loop Gordon used the alloy Inconel X-750, with
negligibly low magnetic susceptibility and a temperature coefficient of linear
expansion close to that for permalloy. Structurally the ferroprobe ia designed so
that the ring core, fastened to the supporting loop and surrounded by the excitation
cainding, can turn in relation to a fixed measuring coil rigidly fastened to the
ferroprobe'.s case. Because of rotation balancing of the ferroprobe was achieved,
making possible mi.nimum imbalance e.m.f.* This ferroprobe was used with success
in making nagnetic measurements in outer spac.e.
The ring ferroprobe described here imitates Gordon's design in ir.s general features,
hut domestic materials were used for the surporting loop and core and other di-
- mensions were selected for the core.
In keeping with the considerations discussed above, a strip 0.02 mm thick and 1.5
mm wide of a nickel-iron alloy--permalloy 82NMP--was used for the core. Ohmic-
resistance alloys, in particular, alloy NM23KhYu [29], were sPlected as materials
for the supporting loep. After fabricating and pickling the supporting loop, the
turns of the permalloy strip are laid into its grnove and the last turn of the
strip is fastened ro the one preceding it by spot welding. After this the core
together with the supporting loop is heat treated.
*Of course, the balancing of rod-xype ferxopxobes is accomplished by matching cores
in pairs.
31
FOR OFFICIAL USE bNLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
'Che dimensiong of the core wexe s.elecCed out of desi,gn considexa.tiona and fTOm
r.he assigned frequency of the excitatihn fi:eld--12.5 kHz. The diameter of the
core, d a 13.2 mm , was calculated by taking i:nto accounz equation (19). The
thicknegbtof the pexma,l,loy strip (Q r~ 0.02 maa) was selected fxom the relationship
in [1, 2]:
A < A,~.x = 4fe-'l� ,
As we know, a.ring core possesses the property of demagnetization in relatien to
the external measured fie1d. Therefore, as we11 as taking into account the com-
plex nature of the dependence of the noise 1eve1 in aferropxobe on the area of
the core's cross section, we experimenfially found the number of turns ot the strip
in the core with which the noise is minimal; this number proved to equal six.
The permeability of the shape of the ring core was calculated from the equation
in [4]: M _ 1.2end = 419,
46h(lna+h--1)
- where d, d and h are the diameter, widt_h and thickness of the core (8 = 1.5
- mm , h= 6 and A = 0.12 mm
On top of the loop with the core is wound a 5ingle-layer toroidal excitation
winding whose number of turns equal5 130. At the moment of the core's saturation
a current of 0.6 A flows through the winding, which creates a circular field
strength of Hmax - 1700 A/m (approximately equals_20 Oe). The loop with the core and excitation wir_ding is placed inside the measuring cotl,
designed as two sections and wound in order to make possibls identical capacitive
counling between its output terminals and the excitation circuit. The number of
cojl turns equals 500. The maximum achievable transformation ratio under no-load
conditions is calculated by taking into acco�nt equations (5) and (15) by the equa-
tion
0sfnsx
B..o; ~ 16fO hrv,t m .
im < �Ama=
With fb = 12.5 kHz , dh = 0.18 tran2 , w2 = 500 , ~ = 0.6 [2] and m= 419 ,
G = 4.5 t1V/nT . The computed value of G2~X agrees well with that determined
experimentally (approximatPly 5 UV/nT).
On top of the measuring coi7. on a. special loop are wound a compensation winding,
feedback winding and other windings.
32
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
NOR OFHI('IAI. USw: ONLY
The circuit of the meas+iring apparatus b}r means of which the noise level, in the
ferroprobe developed was estimated ts presenrecl in fig 8. It includes the ele-
ments of a standard mabnetameter, a recordex, a spectrum analyzer and a shield
(superconducting or fexromagnetic) inside of which the ferroprobe is placed in
thP measuring process.*
Figure 8. Diagram of Measuring Apparatus for Estimating *_he Noise Level
in a Ferroprobe: 7--ferromagnetic and/or superconducting shield;
. Il--electronics of magnetometer; III--testing and recording devices;
1--ferroprobe; 2--input circuit capacitor; 3--differential ampli-
fter; 4--driver; 5--selective amplifier; 6--synchronous detector;
7--d.c. amplifier; 8--spectrum analyzer; 9--recorder; 10--cali-
brator; 11--calibration coil
Voltage of triangular form with a fundamental frequency of 12.5 kHz produced by a
precision oscillator (a 100-kHz quartz sel.f.-excited oscil'lator., an 8:1-countdown
frequency divider and an integratnr) js supplied to the ferroprobe's excitation
winding. The level of even harmonics in the oscillatox's output voltage does nqt
exceed 0.01 percent, which in combination with effective balancing of the ferro-
probe (suppression at a frequency of 2f = 25 kHz is not less than 50) guaran-
tees minimum zero driPt. As already men~ioned, the current at the moment of satu-
ration of the ferroprobe's core equals 0.6 A.
The ferroprobe's measuring coil is tuned by means of a capacitor in xesonance
to a frequency of 2f = 25 kHz . With this the ferroprobe`s trans�orma.tion ratio
with the maximum exci~ation current indicated increases to 40 uV/nT. The loop thus
formed is connected to the input of a wideband differential amplifier. Then the
ferroprobe's signal sequentially enters a selective amplifier with a gyrator
filter, a synchr.onous detector and d.c. amplifier. The real band of the measuring
channel in terms of the demodulated signal is 0 to 100 Hz.
*The circuit diagram of the magnetometer was developed by Xu.N. Bobkov.
33
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00854R000540020034-7
FOR OFFICIAL USE ONLV
= The channel.'s ouzput is connected to a type, N-327 xeGOxdex and type 54-54 spectrum
- analyzer. - The measuring procedure reduces to the fo1l,owing~ The fexroprnbe bei,ng tested
is placed in the woxking space of the texromagnetic or s-uperconducting shield and
is fastened to a special turning table by means of which the ferroprobe is turned
- around its center 180 degrees. As the result of this turning, the f.erroprobe's
steady zero drift is revealed in the shield's residual field. Then the position
of the ferroprobe is fixed and the noise is recorded. The duration of the record-
ing (realization) is determined by the lower limit of the section of the spectrum
analyzed and by the conditsons of fihe experiment: Tn a ferromagnetic shield whose
residual field is subject to changes, short, e.g., 30 minute realizations, are re-
corded (which corresponds to a lower-limit frequency of 0.001 Hz); in a super-
conducting shield whose residual field is stable over time [42], extended, e.g.,
24-hour realizations are recorded. The noise recording scale (the transformation
ratio of the measuring channel) is tested periodically by supplying the required
current value to the ferroprobe's calibration winding or to a standard coil put
onto it.
A fragment of a recording of noise (10-minute realization) and of the excess com-
ponent of rLoise for eight hours of the operation of a ferroprobe in a supercon-
ducting shield is presented in fig 9.
gm,,vTn 1)
0, f
t
fQrru~
2)
a)
em,M n
0'2 -0 -0-0--o a-
41 o 0
~ Z J q s s 73)t,t
b)
Figure 9. Results of Tests of a Ring Ferroprobe: a--fragment of recording
o.f noise; b--excess component of noise
- Key:
- 1, gsh ' nT 3. Hours
- 2. Minutes
34
FOR OFFICI.AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
The processing of experimental data (in the infxa.7,ow, freq,uency x'egion of the
~ spectrum processing can be performed both manual7.y arid automatical1.y [471),
obtained for no less than five samples of the ferxopxobe develaped: showed that ,
with the excitation condifi.ions selected the spectra'1 di.stritautian of noise actually
has the form presented in �ig 2. In the 1 to,100 Hz section the noise is uniform
and, ad j usted for the 1 Hz liand, equals B ti 0.005 nT ; in the 0.001 to 1 Hz
section B ti 0.01 to 0.015 nT; and finally the zero drift after 8 h of observa-
tion does not exceed 0.1 nT. As men':ioned earlier, at the present time it is
still dif ficult to say whether the zero drif ts observed are a reflection of noise
processes in the ferroprobe's core or are the result of the existence of coherent
noise in the channel's electronics. We consider only the zero values measured in
the 0.001 to 100 Hz frequency band reliable.
The data presented testify to the fact that in the ring ferroprobe developed the
- noise level adjusted for a unit of volume of the core is an order of magnitude
lower than the adjusted noise level obtained earlier [9J, which was achieved mainly
on account of using new nickel-iron alloys with a low value of X and primarily
_ alloy 82NMP, as well as on account of taking into account other factors enumerate3
above. The noise level reached, of course, is not the limit. In our opinion in the
immediate future �erroprobes and magnetic amplifiers will be developed with a noise
- level on the order of 0.001 nT (1 pT). Information is already available to the
effect that on the basis of using new alloys the noise level in a magnetic ampli-
fier nas been brought to the level of the thermal noise of the input winding.*
In conclusion let us note that in addition to alloys 82NMP, 81NMA and 83NF (cf.
table 2) amorphous alloy 84KSP.-A was also tested in the ferroprobe developed.
- A strip 0.04 mm thick and 1.5 mm wide was used, which, just as the permalloy,
was wound and fastened to a supporting loop. The core was not annealed. Under the
ferroprobe excitation conditions and with the measuringt procedure indicated earlier
in the 0.001 to 1 Hz frequency band a noise level of B ti 0.1 nT was achieved.
And although this is an order lower than that achieved w~en using alloy 82NPfP,
the result all the same must be considered quite encouraging since amorphous ma-
terials are being studied and developed intensively. In addition, in a number of
cases the requirement of ensuring the maximum low threshold sensitivity is not
- imposed on ferroprobes and magnetic amplifiers. Therefore, amorphous naterials
can be used already now in ferroprobes and magnetic amplifiers. The advantages of
these materials are obvious. In particular, cores made of them do not have to be
heat treated and are not very sensitive to mechanical stresses.
Conclusion
- The picture of the origin of magnetic noise which we have discussed is obviously
not complete and lacking internal contradictions. It will of course be refined as
the result of furthex theoretical and experimental xeseaxch. This relates
*R.Ya. Berkman`s report at the third interdepartmental "Magnetic Noise in Ferro-
probes and Magnetic Amplifiers" conference, Moscow, MFTT [Moscow Physieotechnical
Institute], October 1979.
~
35
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FoR oFFIcInI. Iisw: orvi.v
_ especi.ally to precise],y defining the role of magnetic anisotropy. $ut what has
alre 0.2, is computed from (3.21); in this case, the constant
factor is chosen equal to 30. The graph of Figure 3.6 can b e used for this same
purpose. The more complex formula of [7] is used to calculate RE, of a medium
wave loaded antenna.
47
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
(b) 19 �

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
If the inequality Z9 < 0.35 (Z3 is the equivalent antenna length), then the radia-
tion resistance referenced to the cur-rent at the feed points is:
REO = RE,,/s1II2 kl,.
If Zo/a > 0.35, then the law governing the current distribution along a vertical
antenna is close to a hyperbolic sine and the radiation resistance referenced to
the current at the feed points can be determined fram the approximate formula [7]:
RZ
II
Rz0 (R,o/1~8)'-Fsinskl, .
where WB is the characteristic impedance.of tYle vertical section of the antenna.
Another way of increasing RE consists in using complex antennas of severa'1 vertical
coupled dipoles (antennas with several downleads). Such antennas are employed for
high power very long wave and long wave radio stations. A schematic of a complex
antenna is shown in Figure 13.3, which in essence consists of three T-antennas con-
tiguous with each other. The f eed frwm the transmitter is delivered to the central
downlead. All of the downleads are tuned by means of the variable inductances L
so that the currents in them prove to be in phase and approximately equal. The
radiation resistance of each vertical conductor is the sum of the inherent resis-
tance of the given conductor and the resistances induced in it by the other vertical
conductors. The induced resistances can be calculated using the induced e.m.f.
technique (see �4.2), and the internal radiation resistances of the downleads are
calculated using (5.8). Because of the fact that the currents in all of the dawn-
- leads are in phase and approximatelq equal in magnitude, and the spacings between
the downleads are small as compared to the wauelength (the spacings between the
downleads amount to a f ew hundred meters), the induced resistances prove to be close
in magnitude to the internal resistances. For this reason, the overall radiation
resistance of one downlead is RE1 = nREll, where RE11 is the internal radiation
resistance of the downlead; n is the number of downleads. The total radiation
resistance of the entire antenna is RE = RE1 + RE2 + REn = n2RE11�
Thus, the radiation resistance of an antenna with several (n) downleads increases
approximately as n2 as compared to the radiation resistance of one vertical antenna
of the same length.
Such a large rise in the radiation resistance is achieved at the cost of a consider-
able increase in the complexity and expense of the antenna. It should be kept in
mind that the radiation resistance of complex very long wave and long wave antennas,
consisting of several downleads, is nonetheless low and usuaYly amounts to units of
ohms. Since the radiation resistance of an antenna with several downleads increases
- as compared to a single antenna by a factor of n2 times, while the loss resistance
increases approximately n times, the efficiency of a complex antenna is increased.
Efforts are ma.de in the design of VLF, LF and MF antennas to red,.Ce the electro-
magnetic energy losses and increase antenna efficiency. The losses in the surface
48
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
~ Figure 13.3.
bJ
l
Eigure 13.4.
of the ground which is incorporated directly in the antenna current circuit, and
in the tuning elements (especially the loading coils) play a maj or role. The losses
in the coils are reduced by means of an efficient structural design and the use of
wires with a low impedance at high frequencies. To reduce losses in the ground,
a grounding system or a counterpoise is set up, the underlying concept of which
_ consists in reducing the currents flowing directly in the upper layer of the soil.
- A ground can be made in the form of a system of radial conductors buried a small
depth in the ground (down to abaut 0.5 m), whirh are joined together at the center
and connected to a terminal of the generator (see Figure 13.1b). Since the specif ic
resistance of the ground conductors is considerably less than th e soil resistance,
the majority of the currents shorted to ground flow to the generator terminal via
these conductors, because of which, losses in the ground are red uced.
A counterpoise is made in the form of a grid of wires suspended below the antenna
_ at a slight height above the ground. All of the conductors are joined together and
connected to on e of the generator terminals. In this case, th e majority of the
displacement currents branctiing off to ground fram the antenna is intercepted by
the counterpoise conductors and routed in the form af conductance currents to the
generator terminal. Thus, it is as if the counterpoise shields the antenna fram
the ground; approximately 60 to 70 percent of the antenna current is shorted throughit and 30 to 40 percent through the ground. Counterpoises are usu311y employed at
medium wavelengths in the case of soil which is poorly suited for grounding devices
(hard or poorly conducting soil), as well as at short wavelengths for mobile radio
stations with asymmetrical vertical antennas.
' A ground or counterpoise should encompass tne area on which the bulk of the near
field is concentrated. The number of radial ground conductors is usually 80 to 120.
Vertical antennas without a horizontal section or umbrella antennas have all of the
radial wires of the same length, exceeding 10 to 20 percent of the antenna height.
_ In the case of T or 1' -shaped antennas, the length of the ground (counterpoise)
conductors increases as the projection of the horizontal portion of the antenna
onto the ground is approached and should exceed the amount of this projection by
- approximately the mast height. Complex (sectionalized) grounds are sometimes used
in high power LF and VLF transmitters. Antenna efficiencies close to 90 percent
49
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00854R000540020034-7
FOR OFFICIAL USE ONLY
have been successfully achieved at VLF
by means of special measures to reduce
the losses in the ground and 'grounding
density distribution in th e near f ield
B.V. Braude [60] have made a substanti,
methods f4r grounds.
frequencies using such grounds, as well as
losses in tuning elements. To calculate
systans, it is necessary to know the current
of the antenna. S.I. Nadenenko [59] and
al contribution to the development of design
The permissible power in an antenna is governed by the value of the zlectrical
field intensity components, En, normal to the conductor, at which eZectrical break-
down of the insulators can occur or the ionization of air close to the antenna
begins (the "corona" phenomenon), as well as by the current level at which the
conductors begin to heat up. The cross-section of the antenna cunductors should be
chosen so that the current density in them doe s not exceed 5 amp/mm2 (the e,ffective
value) .
Porcelain antenna insulators sustain a voltage of approximately 1 kilovolt. In the
case of comparatively low power tr ansmitters (tens of kilowatts and less), insula-
tor strings sustain a voltage of up 50 KV are used for the antenna. In the case of
high power transmitters (hundreds of kilowatts), special insulators are used which
- sustain voltages above 100 KV.
The permissible field inten:;ity at the conductors should be less than the critical
value (see �12.2). The amplitude of Ecr at VLF and LF frequencies in the case of
moist air amounts to about 14 K".'/cm [60]. In the case of a vertical antenna with
a horizontal portion (Figure 13.~~.), the voltage at any section of the equivalent
vertical antenna, UZ = Uncos k(Z + b3 - z), where U, is the voltage amplitude at
the voltage antinode (point C).
The voltage between the end of th e vertical section of the real antenna (point B)
and ground is:
'UB = U,; Cos kb,...
(13.1)
It is known from the theory of long lines that the voltage and current in the
corresQonding antinodes are relat ed by the exp ression:
U,~ = I~Wv ' Uo=1n:.W~,~ ; (13.2)
where Wv is the characteristic impedance of th e vertical part of the antenna;
(13.3)
By replacing Uw in (13.1) with expression (13.2) and substituting expression
(13.3) in place of I., we obtain:
,U8 ~lo.WDcos'kb,lsin,.kl,;' (13.4)
, , .
_ 50
FOR OFFICiAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00854R004500020034-7
FOR OFFICIAL USE ONLY
The voltage in a real antenna is maximum at the end of the horizontal part (point A).
Since the voltage distribution is UX = UA maxcos k(b - x) along the horizontal
section AB, where x is the present coordinate read out from point B, then the
voltage at the end of the vertical portion of the antenna (point B), and the maxi-
mum voltage at point A are related by the expression: ,
~ UB = Uq mQi cos,kb. (13.5)
Equating (13.4) and (13.5), we find:
101ne cos kb, (13.6)
UA% cos k6 sin kl, �It can be seen from the given expression that to reduce the voltage in the antenna
- it is necessary that: for a specified input power, the current Ip be reduced (i.e.,
increase the antenna radiation resistance); the characteristic impedance of the
vertical pa t of the antenna be reduced, i.e., increase the downlead capacitance;
the equivalent length increase bz) be increased, for which the ratio Whor/Wv ghould
b e reduced, as can be seen from (5.9), i.e., increas e the capacitance of th e hori-
zontal portion of the antenna as compared to the vertical (Clhor > Clved�
We will note that expression (13.6) is also justified for T-shaped and tnnbrella
antennas. In all cases, b is to be understood as the length of one arm (b eam).
The voltage UA maX should notexceed the permissible level, i.e., UA maX < U,qon[Uperl�
The power delivered to the antenna is Pp = IgRin/2. Substituting its expression
from (14.6) here in place of Ip, we obtain the formula for calculating the permis-
sib le level of power delivered to the antenna:
Rsi A;,n cos kb sn kl, (13 . 7)
PO per PoAO"2(.UA
� Wacoskb,.
For an antenna without a horizontal section (b = 0), Pxpression (13.7) assumes,the
form:
�
- R2x [U sin kl ~
Po Aofl =
C ` e
Keeping expression (12.5) in mind and replacing WA in it with Wr [WhorJ, expression
(13.7) can be written in the form:
' n aon na lt~,. cos kb sin k/s (13.8)
po lton =.2 Re~~~ . 6( f Ws cos kb9
where a is the conductor radius; n is the number of conductors in the horizontal
part.
51
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
In those cases where l/a > 0.35, the current in the antenna should be considered as
having a hyp6rbolic sine distribution. In this case, if the antenna does not have
a horizontal part, the permissible power can be computed from (12.9), keeping in
mind the fact that WA = WB Wverl�
To calculate the permissible pawer, the permissible value of En as well as other
quantities, it is necessary to know tha characteristic impedances of the vertical
and horizontal parts of the antenna. The characteristic impedances are calculated
by any approximation technique, for example, Hou's method [6, 50].
The characteristic impedance of an antenna mast or antenna tower can be approxi-
mated with the formula WA = 60(ln[Z/pt1 - 1), where ps is the equivalent radius of
the tower. If the antenna mast or tawer has a triangular or square cross-section,
then p. represents the radius of a circle inscribed in,the cross-section.
_ In the case of a large number of conductors in an antenna curtain, B.V. Braude [601
- proposed the substitution of solid metal surfaces for the wire curtains in the cal-
culations, where these surfaces h ave the size and shape of the corresponding wire
sections of the antenna.
The passband of VLF, LF and MF antennas is governed by the dependence of the input
_ impedance of the antenna on frequency (the directional properties of the antenna
change little with a change in frequency). With small changes in frequency as.,com-
pared to the resonant frequency, the antenna behaves as a conventional resonant
circuit. The antenna bandwidth is defined on analogy, viith such a circuit as the
frequency passband, at the edges of daMich the current amplitude in the antenna with
a constant voltage at its input drops off by a factor no more than F2 as compared
to the current amplitude at resonance. The current phase within the bounds of the
passband should change linearily with a change in frequency and be symmetrical
relative to the resonant frequency.
We shall consider an antenna operating in a heavily loaded mode (ap � ap or
� Z3 � ap, where ap is the resonant wavelength of the antenna; ap is the working
wavelength), representing it in the form of a series resonant circuit, consisting
of an input reactance Xin =-iWv cot k13, having a capacitive nature in this case,
a loading coil inductance Ly (Figure 13.5a) and the resistances Rin = REO + Rloss.
The relative passband of such a circuit when operating close to resonance is
2pfpQA, where QA is the quality factor of the antenna circuit, as is well known,
equal to QA = Xin.r/Rin; fp is the resonant (working) frequency of the antenna
- circuit; Xin.r is the reactive component of the input impedance of the antenna when
' f = fp.
Since in this case:
Consequently, QA = Wv/kPZ3Rin and the relative passband is 2Af/fp = RinkP(Z + bo)/Wv.
Thus, to increase the antenna bandwidth, Rin must be increased (by virtue of
increasing the radiation resistance), the characteristic impedance of the antenna
must be reduced and its equivalent length increased.
The bandwidth of antennas operating in lightly loaded or shortened modes can be
calculated fram the formulas available in [6], 621.
52
~
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000504020034-7
FOR OFFICIAL USE ONLY
110,crtrui frc(l. Tlic, fee4 I'or wire zintenniiH is usually ciccompliahed by means of
, transmitting the electromagnetic energy from the output circuit of the generator
using special coupling elements to the antenna downlead. In this case, the antenna
downlead is brought directly to the radio station building. Antenna masts and
antenna towers are fed using coaxial feed lines. Sometimes, coaxial lines are made
in the form of a number of conductors, wound along the generatrix of a cylinder.
In the first case, the load on the generator is the input impedance of the antenna.
To transmit the ma.ximum power from the generator to the antenna, this load should
- -be purely resistive and have a magnitude determined by the operating conditions of
the generator. In the second case, for the same purpose, the feeder should b e
loaded into a resistance equal to its characteristic impedance W, i.e., it should
operate in a traveling wave mode. The generator in this case,is loaded into a
resistance equal to the characteristic impedance of the feeder.
_ . � . � , , , , -
A ,
. , Lc6 '
CyK�P
~ , , � , ~ 2

p . ~ - . ~ .
T,
r.
' . Q) (a) 6)4b).~ / i:/..;i.i /////ii/ . .
i � , . ~
Figure 13.5. Figure 13.6.
The task of matching the antenna to the generator or to the feed line can be broken
down into two parts: tuning the antenna to resonance by means of compendating for
the reactive component of its input impedance and transforming the input resistance
of the antenna to the reguisite value. Matching is accomplished by means of lumped
reactive tuning and coupling elements. To illustrate the design procedure for
coupling elements [50], we shall consider a capacitively coupled matching circuit
(Figure 13.6). If the antenna is tuned to resonance by means of tuning elements
Lg LLloadl and Cg [Cload), then the impedance at points 1-1 is:
� �
. � . ~
, 7 - Rax'I (il ~.Ca' Rez~~ � 1 (sYli^'g Rgz/ i
-1-1- R.x -f- l /i w C
'
ebuP~�
or, by eliminating the imaginary component in the denominator,
2 2 2 x/(1 Wa Cca Rea); .
1'+~~ Ccs ~ax~ C' H Re
_ ~ � , ,
So that the cable operate in a traveling wave mode, the impedance at points 2-2
should be purely resistive and equal to the characteristic impedance of the cable.
This will be the case when the following conditions are met:
. . , ,
Rax/~1-f" a L'c~ Rsx~I~ W'v~1'aRix~~l Ccs Rea) -=r 0,
53
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R004500020034-7
FOR OFEtC1AL USE ONLY
where LCB [Lcoupling] is the inductance inserted to compensate for the reactive
component of the impedance Z1-1�
The unknowns CCg and LCg can be determined from these equations:
:
' 1. RB:- - W 1~~RBZ (13.9) ; (13.10)
CcouP a: 1.-CCb ruR
. �
As can be seen from the formulas cited here, this circuit can operate only when
Rin/W > 1, i.e., it is a step-down configuration.
13.2. Medium Wave Broadcasting Antennas
Guide antenna masts are used primarily for radio broadcasting at medium wavelengths.
In antenna masts with an insulated base (see Figure 13.2), the insulator not only
provides electrical isolation of the antenna'.from ground, but is also the mast
support. This insulator should have good electrical and mechanical strength, since
the weight of the tower resting on the insulator can 100 to 200 tons.
Because of the nonuniform pressure distribution over the cross-section of the
insulator for other reasons, mechanical damage to support base insulators is pos-
sible, something which can lead to a serious accident. The use of support insu-
lators increases the antenna cost and reduces its operational reliability. For this
reason, antenna masts with a grounded base which do not require support insulators
are of considerable interest, where these towers are mounted on metal foot bearings,
fastened to a f irm reinforced concrete foundation.
Also interesting is the possibility of transforming already finished and existing
towers and masts at radio centers which were constructed for other purposes (tele-
vision towers, etc.) into a radiator. There are two types of antenna masts with a
grounded base: shunt fed and top fed antennas, proposed by G.Z. Ayzenberg [63].
In the case of the shunt fed antenna (Figure 13.7), the radio frequency voltage is
fed to a certain point a, located'at a distance 12 from the lower end of the tawer,
by means of an incl.ined conductor, which is the continuation of the center lead of
the coaxial line.
If the length (height) of the mast, Z1 + Z2 = Z, amounts to a/4, then the react-
ances of the segments Z1 and Z2 cancel each other out. The input impedance of the
antenna is purely resistive in this case. By selecting the feed connection point
a, it can be made equal to the value W of the feed line and matched so that the
antenna and the feed line have no additional matching devices. If the input imped-
ance of the antenna is complex, then its reactance is compensated by a variable
- reactance inserted in the sloping line. Since the sloping line, the ].ower part of
- the tower and ground form a loop antenna, the radiation of which is superimposed
on the radiation of the antenna tawer, the directional pattern of the latter is
slightly distorted.
54
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
To explain the operational principle of a top fed
' antenna tower, we shall represent it in the form of a
circular metal cylinder with a grounded base, inside
which the vertical conductor is fastened to insulators
, (Figure 13.8a). A metal disk which represents a capaci-
' tive load, is connected to the upper end of this con-
ductor, which extends out beyond the cylinder. In
stxuctural terms, this load is usuallq a-portion of the
upper level of guys, while the cylinder is a conven-
tional type tawer. The feed from the generator to the
antenna uses a coaxial line, the shield of which is
connected to the body of the tower, while the inner
conductor is connected to the inner conductor of the
tower. Thus, the cylinder (the tower) and its internal
77777- 7 - conductor represent a coaxial line with a large cross-
'section. Because of the fact that the inner conductor
Figure 13.7 extends out beyond the limits of the cylinder and has a
load at the top end, the current flowing alongi:the
- internal surface of the cy3.inder also continues to its
outer surface. The currents flowing through the outer surface of the tower (the
cylinder) radiates the electromagnetic field. For this reason, from a radiation
viewpoint, it is as if the generator is inserted between points a and b, i.e.,
is located at the top of the tower (Figure 13.8b). At the grounded base of the
tower (point 0) there is always a current antinode (using the method of mirror
images, it is easy to show that the point 0 is the short circuit point). 'The
antenna is tuned to resonance and matchdd to the feed line by means of a special
circuit. Top fed antenna towers are widespread among the radio broadcasting centers
of the Soviet Union.
To increase the reliable reception
area, -the antenna radiation at angles
greater than 45� to the horizontal
should be reduced, so that;-the angles
of arrival of the space waves in the
near fading field amount to 45 - 85�.
OJ- Antennas which have directional pat-

terns in the vertical plane which pro-
vide for maximum radiation along the
Figure 13.8. surface of the earth and little radia-
tion at comparatively high angles
(starting at 40 to 50�) are called
QYItZfCLdZYIg. It is conventionally
assumed that an antenna has satisfactor_yantifading properties if the field intens-
ity at elevation angles exceeding 50� amounts to no more than 10 to 15 percent of
- the field intensity in a horizontal direction. The best directional pattern, from
- the viewpoint of antifading properties, is had by an antenna with a relative length
of Z/a = 0.53 (kZ = 190�) (Figure 13.9a). It is well known that an antenna with a
relative length of Z/a = 0.63 has the maximum gain in the horizontal direction.
However, there is a sidelobe in the directional pattern. of such an antenna which
= produces marked radiation at angles of 40 to 60� to the horizontal.
55
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000504020034-7
_
(a) �I (b) a~
Figure 13.9.
FOR OFFiC[AL USE ONLY
I
t
0
e
N
II
N
N~
. ~
~
,
b
a
KncpeBamvurry b) (b)
--~o Xmtr
d) (a). . .
Figure 13.10.
The use of antifading antennas makes it possible to increase the reliable reception
area by a factor of 2 to 2.5 times as compared to a short antenna. It should be
kept in mind that the thinner antenna (the greater its characteristic impedance),
the better its antif ading properties will manifest. With an increase in antenna
thicknPqs, the current at the ncie increases (Figure 13.9b), the sidelobe widens
and ra-iation is found at angles wrich are the most dangerous from the viewpoint of
near field fading. A large characteristic impedance leads to a rise in the potent-
ials in the antenna (a reduction in the r,ermissible power) and to a.decrease in
the bandwidth. For this reason, the characteristic impedance is to be chosen
taking into account obtaining sufficiently good antifading properties, the requi-
site bandwidth and the specified power.
Both antenna masts with an insulated bhse and antenna masts with a grounded base
can be used as antifading antennas.
Wire antennas and antenna masts with grounded or insulated bases cannot operate
throughout the entire radio broadcasting band (200 to 2,000 m). The working band
on the long wave side is limited by the reduction in the radiation resistance and
the corresponding decrease in the efficiency; on the short wave side it is limited
by the sharp reduction in radiation along the surface of the earth, if Z/a > 0.7.
The broadband antenna with a variable current distribution (ARRT) proposed by
G.Z. Ayzenberg, can operate throughout the entire broadcast range at wavelengths
of 200 to 600 and 600 to 2,000 m. In this case, this antenna has antifading pro-
perties in the 200 to 600 m band [64].
" One of the variants of an ARRT attenna [65, 66] takes the form of an antenna mast
260 m high, insulated at the base (Figure 13.10a). The lower portion of the mast,
H, which arnounts to one-third of its height Z, is surrounded by a cylindrical
shield with a diameter of about 10 m, consisting of 9 to 12 wires. The lower ends
of these wires are connected to the shield of the wire coaxial line, coming from
th& generator. In a first approximation, the wires of the shiectd can be treated
as a solid cylinder and the entire lower portion of the antenna enveloped by this
cylinder can be treated as a coaxial line.
56
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OF'FiCIAL USE ONLY
- Tu reduce Lhe chararCc~rlstic impedance of the tintenna (down to 170 to 100 ohms),
widen the passband correspondingly and increase the maximum transmitter power (more
than 150 KW), the upper portion of the tower is surrounded by a wire cylinder, which is insulated from the lower cylinder. The lower and upper ends of the wire
of the upper cylinder are connected to the start of the upper portion of the tower
and to its top end respectively.
The points a and b(the upper ends of the lower portion of the tawer and the
lower cylinder) can be considered the generator connection points. The radiation
is produced by currents flowing through the wires of the lower and upper cylinders
(Figure 13.10b). We will note that the current exiting at point b to the exter-
ior surface of the lower cylinder coincides in phase with the current flowing
through the upper portion of the antenna. The current distribution can be adjusted
by a variable reactance Xg inserted between ground and the lower end of the wire
shield. It is made in the form of a short circuited loop, for which the external
shie]_d of the feed line is used. The quantity Xg is adjusted by moving the short
" circuiting device. The current antinode is located at the short circuit point and
is moved when the short circuiting device is moved. The antenna operates without
current distrihution adjustment in a range of 600 to 2,000 m. It has an elevated
radiation resistance in this mode.
An antenna was subsequently designed having two feed points 44, 66]. With the
appropriate alignment and a tower height of 300 m, this antenna has a greater gain
(up to four).
A specif ic feature of ARRT antennas is the extremely slight dependence of the
directional pattern on the characteristic impedance of the antenna. This is
explained by the fact that the power is delivered to the antenna at its base, but
at a height of H= 0.3Z to O.SZ, because of which, the nature of the current dis-
tribution depends slightly on WA an3 is close to sinusoidal. For this reason,
even for a low value of WA (100 to 170 ohms), the antifading properties remain good.
In a number of cases, it is desirable for medium wave radio broadcasting to use
antennas which are directional in the horizontal plane. This makes it possible to
distribute the radiated power in.specif ic directions, as well as to reduce cross-
talk interference of radio stations.
To provide broadcast coverage for a territory having the shape of a sector, an
antenna system has been designed which consists of four shunt fed antennas, located
at the corners of a square (67]. Two of them are fed from the transmitter, while
the other two play the part of a passive reflector. One can obtain four coverage
sectors by combining the dipoles wh ich operate as the antennas and reflectors in
various ways by means of an appropriate switching system.
To provide broadcast coverage fAr territories which are considerable distances
away, a medium wave antenna system has been developed which consists of eight shunt
fed antenna towers, arranged in two rows [6$]. Four towers, arranged in one row,
are fed from a transmitter; the other four play th e part of a tuned passive reflec-
tor. The spacing between the raws is 75 m. The antenn&has a controlled direc-
tional pattern in a sector of +30�. The gain of the antenna system in a range of
57
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
~ x nepeDamvany . �
Pxc. 13.11 To thQ. transmiEter
Figure 13.11.
185 to 575 m varies from 28 to 5. Shunt fed antenna towers with a reduced charac-
- teristic impedance (WA 2:~ Ohm) were chosen as the radiator. The reduction in
WA is achieved by means of suspending a wire cylinder with a diameter of 15 m
around the trunk of the tower, where the cylinder is made from 15 to 16 wires with
a diameter of 16 mm. The wires ofithe cylinder at the top of the tower and at a
height of 0.5 H are connected to the grounded trunk of the tower. The height of
the antenna tower is H= 110 m.
The antennas considered here require an extremely complex tuning and switching
- system. Moreover, when the wavelength changes, the directional pattern in the
- vertical and horizontal planes change substantially.
When the vertical antennas are replaced by four logperiodic antennas, suspended on
one common support (Figure 13.11), each antenna, with the appropriate choice of
the configuration and number of elements, makes it possible to obtain a directional
pattern with a width of 70 to 100� in the horizontal plane and a satisfactory,
directional pattern in the vertical plane (�rom the viewpoint of antifading pro-
perties) [69). Changing the irradiation sector is accomplished by switching the
transmitter to the corresponding antenna. The directional properties of the an-
tenna practically do not change in a wavelength range of 200 to 600 m. The travel-
ing wave ratio in the feeder in this range does not drop below 0.5.
The development of strong polymer film technology created the conditions for the
3esign of new antenna structures. G.Z. Ayzenberg and V.N. Uryadko in the USSR
developed a pneumatic antenna mast made of solid polymer material. It takes the
form of a truncated cone 60 meters high (Figure 13.12)[not reproduced] made of
high strength polymer, maintained under an excess air pressure (the pressure in the
balloon is somewhat higher than atmospheric pressure). Either metal guys intended
at the same time for also supporting the antenna in a vertical position or a system
of wires tightly fitting the cylinder are used as the radiators. The advantage
58
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
of such antennas consists in the set-up speed, the possibilities for adjusting the
height, etc.
Whett placing several antenna towers on a site of limited size, it is necessary to
~ take steps to reduce the electromagnetic coupling between them (to increase the
isolation), which leads to the distortion of the directional patterns of the an-
tennas, i.e., imprbve the electromagnetic compatib ility of the antennas. Work which
has b een dono- in this area indicates that it is essential to provide for special
tuning of the antennas by means of inserting tuning impedances in them in conjunc-
- tion with filter circuits, the value of which depends on the spacing between the
antennas and their height.
13.3. Long Wave and Medium Wave Receiving Antennas
The marjor types of receiving antennas. 4ZF, LF and MF receiving antennas differ
significantly from transmitting antennas both in terms of the structural design
and types. The problems related to delivering large powers to the antenna and the
appearance of large potentials in it are absent in receiving antennas. The effic-
iency of these antennas down to a certain minimum is likewise of no great import-
ance. In the wavelength ranges considered here, atmospheric and industrial inter-
ference, as well as interference.from other radio stations have a high level, and
to increase the signal/noise ratio it is necessary to use antennas which have a
more or less directional pattern. However, the use of such antennas, because of
their large dimensions, is possib le only in the case of professional radio recep-
tion, and even then not always.
The main types of receiving antennas are: loop antennas of various types, traveling ~
wave antennas, ominidirectional vertical unbalanced antennas. The latter differ
from transmitting antennas only in the small dimensions and structural design.
Loop antennas. A loop antenna takes the form of one or more series connected
turns of a conductor. In this case, the loops can have different shapes (circular,.
rectangular, triangular, etc.). The planes of the loops are arranged perpendicular
to the surface of the earth (a vertical loop). Loop antennas are used as the re-
ceiving antennas for radio communications, radio broadcasting and various special
purposes. The dimensions of loop antennas vary in a wide range: from very small
(tens of centimeters) multiturn loops with magnetic dielectric cores, housed inside
the receivers, to external loops installed in a loop antenna field, the linear
dimensiotis of which can exceed 100 m. The size of all loop antennas (even the
largest) used in the VLF, LF and MF bands are small as compared to the wavelength.
For this reason, these loops can be considered elementary (see �1.2).
If a plane vertically polarized wave impinges on a single turn rectangular loop
(Figure 13.13a), the sides of which, ab and ce are perpendicular to the horizontal
plane, where the direction of arrival of the wave makes an angle ~ with the plane
_ of the loop (Figure 13.13b), then the electrical field.intensity at the point a
ig -kdC07T , where E is the field intensity in the center of the loop.
E, - Eoe 2 ~
59
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00854R004500020034-7
FOR OFFICIAL USE ONLV
_ Corrospondingly, the field intensity at point c is EZ = EoeMd/2)cos~. Since
the loop is an elementary one, the e.m.f.'s induced in conductors 1 and 2 of the
loop are determined by the expressions:
+ ke
- cos W
31=Eohe. 2
a
f
I
hi
1
9
+ cos (P
9$�-Eo h e .
Ei
n - N
li npucnnaKy (a)
To the receiVer
Figure 13.13.
These e:m.f.'s are identical in value, but shifted in phase because of the dif-
ference in the path of the rays and act counter to each other. The resulting
e.m.f. across the terminal of the loop is equal to their difference:
3P = 3a -31= i 2Ea h sin (kd cos qp/2).
Since the loop ditiensions are small and d/a � 1, the sine can be replaced by its
argument. Then:
3p = i Eo kdh cos cp.
(13.11)
As follows from the given formula, th e directional pattern of a loop takes the form
of a regular nunber eight. This has already been established in 51.2 from the
analogy between an elementary loop and an elementary magnetic dipole. The phase of
the e.m.f. changes to the opposite sign when the direction of wave arrival changes
by 180�. If a wave arrives from a direction perpendicular to the plane of the loap
= 90�), then the e.m.f.'s 31 and '32 induced in conductors 1 and 2 have identical
phases (the difference in the travel of the rays is zero) and tne resulting e.m.f.
is equal to zero. In this case, there is no current in the loop and the voltage
across the terminals is zero: there is no reception. The reception tnaximum is
obtained in the case where the direction of wave arrival coincides with the pla:.e
of the loop = 0).
60
FOR OFF[CIAL USE ONLY
y}
I
a'/z rns ~p n
a
p ~ x
6)
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFiCIAL USE ONLY
By employing the concept of effective length to a loop anteZna, based on (6.12),
- we obtain:
IA-3p/Eo- 2n S,.
where S= hd is the area encompassed by the loop.
Since the effective length of a loop is extremely small, the radiation resistance
is also low (even for large outdoor loops, it amounts to only tenths of an ohm).
The radiation resistance of a loop antenna is significanly less than the loss resis-
tance and the antenna efficiency is very poor. For this reason, lopp antennas are
not usually employed as transmitting ones (radio beacons are an exception). To
increase the effective height of a loop antenna, multiple turns loops are employed,
i.e., loops made of a number of series connected turns.
The reactive component of the input impedance of a loop is usual':ty')of: an fnductive
nature (a short circuited line of short length), and a capacitor is inserted to
compensate for it (tune the loop to resonance). When tuning the resonant circuit
of the loop--capacitor, the voltage across the receiver terminal increases by a
factor of Q times, where Q is the quality factor of the loop circuit. The
_ effective length of a tuned loop also increases by a factor of Q. The quality
factor of a loop circuit can reach values of 200 to 300. To increase the signal/noise ratio, the
i
directional pattern of a loop must be
'
turned so that the direction of arrival of
the interference coinci3es with the perpen-
2
dicular to the plane of the loop. Of
L
primary importance for reception quality '
is the depth of the directional pattern
I
,
\
N ~N-
minima of the loop. Minima in the direc-
Q)
tional pattern are equal to zero only in
the case of a completely electrically
Figure 13.14.
symmetrical loop; this symmetry is disrupted
if conductors 1 and 2 of the loop have
different impedances for the current flow-
; ing in them (for example, if the
sides of the loop',have different capacitances�to
'
ground because of their unequal
positioning with respect to it or any other objects)
(Figure 13.14) as well as in the
case where the '_oop is connectdd to the unbalanced
, input of a receiver.
If the impedances of the vertical sides of the loop are not equal, then with a wave
incoming f.rom a direction perpendicular to the loop plane, the identical e.m.f.'s
induced in them generate unequal but constant currents over the length of each side
of the loop, 11 and 12. Because of this, a certain resultant current Ip= I1 - 12
is established in the loop, and a potential difference appears across the loop
terminals, i:e., the reception takes place. This phenomenon is called the antenna
effect of a loop. As a result of the antenna effect, the directional pattern is
distorted, the null (reception) directions disappear and instead of them there are
shallow minima. Where the antenna effect of a loop is present, the ability to tune
61
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFICIAL USE ONLY
Figure 13.16.
Figure 13:15.
nprreaHUK
Receiver
J
or
,
.
po
9so
y~O
O
900
0
Pso
yp
0
$00
'
?00
!SO
.
!00
.
/0
'
�
70.
SO
.
p J, /v '/J tv co pv vr
Figure 13.17.
Ic.
2a'
JO
out interference incoming from a particular direction is degraded. If the loop
is used for direction finding, the antenna effect leads to errors in determing the
location of the target.
To eliminate the antenna eff ect, the signs of the loop and the conductors of the
feed lines should be completely symmetrical relative to ground and surrounding
objects. Moreover, the input circuitry of the receiver f ront end should also be
symmetrical. However, such a circuit is used rather rarely, since in this case,
the design is complicated and the cost of the receiver is increased. Considerably
more widespread is the connection of the loop to the asymmetrical input of a re-
ceiver using any kind of balancing device.
To combat the antenna effect, it is most effective to use shielded loop antennas
(Figure 13.15), where the receiver is housed in a common shield with the loop. A
shielded loop antenna takes the form of a single turn or multiple turn loop 1,
enclosed in a metal tube (shield) 2. There is a cutout 3 at the top of the tube,
which is usually covered with a dielectric sleeve. The lawer part of the tube
makes a transition to metal housing 4, in which receiver 5 is housed. An incoming
wave induces an e.m.f. in the outer surface of the shield, because.of which, at
points a an.d b of the gap there proves to be an applied potential difference,
which generates:;a current in the inside surface of the shield. The outer surface
_ of the shield and the conductors of the loop represent a coaxial line. For this
_ reason, a current appears in the conductors of the loop which is equal in magnitude
but opposite in direction to the current flowing.in the inner surface of the shield,
and a voltage is generated across the terminals of the receiver. The most import-
ant fact is that the potential difference applied to the gap, which generates the
current in the loop, is governed exclusively by the currents in the outer surface
of the shield and does not depend on any imbalance in the arms of the loop, in the
input circuits of the receiver, etc. The sides of the shield (more preciselS�, their
exterior surfaces) are usually synnnetrical with resgect to ground, and when a wave
- arrives perpendicularly to the plane of the loop, the e.m.f.'s occurring in the
62
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
exterior surfaces of the two sides of the shield prove to be equal in magnitude and
_ phase, because of which the potential difference applied to the gap is zero and no
current appears in the conductors of the loop. Thus, a shielded loop proves to be
free of the antenna effect.
Magnetic dielectric cores are used to increase the effective length of a loop
(Figure 13.16). The e.m. f. induced in a loop with a magnetic core can be deter- ;
- mined from the formula [7]:
3p i 2:t � nSEo cos (p. (13.12)
It can be seen from a comparison of (13.12) and (13.11) that inserting a core in
the loop which is made of a material having magnetic properties increases the
e�fective length of the loop by a factur of u times.
Magnetic dielectrics are used the cores, i.e., materials havi.ng a high magnetic
permeability and a low conductivity (ferrites, and Alsifer, etc.).
_ The effective loop length and the e.m.f. increase by a factor only in the case af
a rod of infinite length. In actual cases, the increase in the effective loop
length amounts to ueff times, where ueff < u because of the end effect (the de-
magnetization of the ends of the core). Thus, 2n
- Zd - (27r/X)ueffnS. lk= ~ �900 nS.
The quantity ueff depends on the relative magnetic permeability of the core
material and its shape. The demagnetizing e�fect of the ends o; the rod, and con-
sequently, the rediiction in ueff as compared to u, is greater, the smaller the
ratio of the rod length Z is to its cross-section (Figure 13.17). The effective
length of a ioop antenna is proportional to the area of the loop, however, it is
not advantageous to increase the loop area, since in this case, ueff decreases.
The number of turns in the loop cannot be chosen too great: in this case, the
internal capacitance of the coil (the loop) rises and it becomes difficult to tune
the loop with a capacitor. Loops with magnetic dielectric cores (magnetlc antennas)
are widely used broadcast receiver antennas; the sma11 overall dimensions of such
antennas make it possible to place the.m directly inside the receiver. The use o�
magnetic antennas is not limited to the LF and MF bands. The range of applications
for these antennas extends in the direction of shorter wavelengths in step with
the improving parameters of magnetic dielectLics, and reaches the VHF band.
The ef.fective ].ength of a loop antenna is proportional to the loop area; ~therefore,
_ for professional radio broadcast reception (radio repeaters) or for other types of
professional reception, large stationary loop antennas are used, whibh are set up
in an antenna field using towers. ;he major advantage of loop antennas over
verical asymmetrical antennas is the presence of reception nulls in the horizontal
plane, which maice it possible to tLtne out interfering signals, while in t.he case
6-3
FOR OFFiCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00854R000540020034-7
FOR OFFICIAL USF ONLY
- of small loop antennas used for individual reception, the directional patterns are
controlled by simply rotating the loop or the entire receiver, complex and expens-
ive structures are required to rotate large loops.
. aI 2 ~ Bf ( c ) ` .
(b)
4 r - - , ,
Search coil.
Q = ' Naram~na
(a) a) To receiver "c. 13.18 . �
Figure 13.18.
The directional pattern of large stationary loop antennas can be controlled by
means of an extremely simple (goniometric) device. A goniometric antenna system
(Figure 13.18a) consists of three main parts: an antenna of two mutually perpen-
dicular loops 1 and 2, two balanced feed lines 3 and 4 and the goniometer G. The
- goniometer consists of thr.ee small coils: two stationary (stator) 1 and 2, posi-
tioned in mutually perpendicular planes, and one moving (rotor) so-called search
coil. The stator coil 1 is inserted in the circu it of the first loop, while stator
coil 2 is inserted in the circuit of the second loop. The rotor coil rotates about
an axis passing through the center of th e stator coils. It is connected to the
recE.ver. Tuning elements are inserted in the ci rcuit of the search coil and the
circuit formed in this fashion is tuned to resonance with the station being re- i
ceived. The goniometer is set up a considerable distance away from the antenna in
the receiving station room.
We shall considgr the operational principle of a goniometr{.c antenna. The normal-
ized directional characteristics of loops 1 and 2 are determined by the expressions:
Fl (q)) = cos ry; F, (q)) - sin (p,
where � is the angle between the direction to the station being received and the
plane of the first loop (Figure 13.18b).
With the action of the e.m.f.'s induced in the loops, currents appear in the sta-
tionary coils inserted in these loops. The current flowing through the coil con-
nected to loop 1 is I1 = Ipcos~, wh ere Io is the current in the coil when the
direction of the incoming wave coi ncides with the plane of the loop.
Thc current flowing through the coil inserted in loop 2 is 12 = Ipsin~. In both
formulas, th e current Io is th e same, since botfi loops and the coils inserted in
64
FOR OFFICIAL USE ONd.Y
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000500020034-7
FOR OFFICIAL USE ONLY
them are completely identical. The coefficient of mutual induction between the
stationary coil 1 and the search coil is M1 = MpcosO, where 0 is the angle between
coil 1 and the search coil (Figure 13.19c); Mp is the coefficient of mutual induc-
tion when the planes of the stationary and moving coils coincide. The coefficient
of mutual induction between coil 2 and the search coil is M2 = MpsinO.
The electromotive forces induced by the stationary coils 1 and 2 in the search coil
are equal to the following respectively:
31=10 (D NJocoscpcos(D ;3~=1oco Mosin(psin(D. .
The resulting e.m.f. in the search coil is equal to the sum of the e.m.f.'s
inducFd by the stationary coils: 9p = 711 + 32 = Ipm io (cosocosl) + sinosin(D) . This
formula can he transformed: 3 p= Io wMo cos(~ -0). Thus, the normalized direc-
tionai j~attern of the goniometric antenna is:
. 13.13
F (y) = cos lW-cD).
This directional pattern is no way different from the directional pattern of a
conventional loop antenna. As can be seen from (13.13), by changing the position
of the search coil (the angle I,), one can control the directional pattern of the
antenna. The maximum reception direction is obtained when 0 _~D; there i,s a
reception null when 0 -0 90�. Thus, the rotation of the search coil yields
the same effect as the rotation of the loop antenna.
A goniometric antenna makes it possible to tune out interfering stations by means
of setting the null of the directional pattern on the interfering station. This
antenna also allows for reducing the level of directional atmospheric interference.
Each outside loop of the goniometer is usually made ~z the form of an isosceles
triangle, for which the base is approximately four times the height. Only one
tower is needed to string such an antenna. The lower conductor is suspended at a
height of 2.5 m. At LF wavelengths, the height of the loop runs up to 70 to 100 m
and at MF wavelengths, up to 10 m and more. Antennas are being developed in which
small ferrite antennas are used instead of the large outside loops.
If a vertical unbalanced antenna is added to a loop or goniometric antenna,. posi-
"tioning the vertical antenna/, for example, in the center of the loop (Figure 13.19a),
then unidirectional recepti,)n (a cardioid directional pattern) can be obtained by
means of a special phasing device. This is easily illustrated if the loop antenna
is replaced by the magnetic dipole equivalent to it. Then the loop''antenna and the
electrical dipele can be treated as two mutually perpendicular dipoles: an electri-
cal and a magnetic one (Figure 13.19b). It is we11 known (see �1.4) that the direc-
tional pattern of such an antenna (see Figure 1.8) is defined by the formula
(F(~) = 1+ cos~. A unidirectional cardioid type pattern (Figure 13.19c) has ad-
_ vantages over the bidirectional pattern of a loop or goniometer, since it permits
a significant reduction in the level of an interfering signal incoming from
65
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047/02/09: CIA-RDP82-00850R000504020034-7
FOR OFFICIAL USE ONLY
directions in a range of angles of 90 to 270� without markedly reducing the useful
signal level. Moreover, such a directional pattern makes it possible to unambig-
uously determine the direction to the station bei.ng received. Detailed data on
- loop antennas can found in [701.
The singZe wire traveZing wave anterma.(OB). Radio broadcast reception quality
(H 7 [as publisfled] must be
fulfilled for bubble devices, since otherwise selfageneration of a bubble is
possible under the influence of field Hr . The ratio Q= Ha/IHrI = Ha/M is
called the figure of inerit of a bubble medium material. It is desirable tflat
q> 2. For orthoferrites Q� 1 and not infrequently is greater than 100, and
for ferrite garnet films usually Q= 2 to 5.
In the absence of an external m.sg:ietic field,strip domains, illustrated in �ig 2,
where the black and white strips correspond to opposite directions of magnetiza-
tion, form in a bubble medium :naterial under the influence of the internal de-
magnetizing field. . Here the mean magnetization of the wafer practically equals
zero. In view of the fact that the diameter, d, of a bubble depends on the
strength, H, of the external (homogeneous) magnetic field, a bubble medium ma-
terial is o?ten characterized by the width, bP , of strip domains with He = 0.
Each domain is bounded in tne body of the wafer by a cylindrical boundary (wall)
having a finite width of A ti 0.1 to 0.5 u. Within the limits of a domain bound-
ary a gradual transition is made f~rom one dixection of magnetization to another,
corresponding to the next domain. This transitip-i can be of various natures,
which substantially influences the dynamic properties of bubbles. So-called
Neel boundaries exist in which the direction of magnetization rotates around the
~ axis perpendicular to the plane of the wafer, and Bloch boundaries in which the
change in magnetization between neighboring domains takes place by means of grad-
ual rotation of the direction oE magnetization around.-en axis situated parallel to
the surface of the wa�er and perpendicular to the wa11s of the boundary. Some
combination of N6e1 and Bloch boundaries is often observed in bubbles.
80
FOR OFFICYAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFIC.IAL USE ONLY
Figure 2. Strip Domains in the Absence of an External Magnetic Field
- Key: 1. b
P
Assuming that W= 0 (magnetization of the bubble is parallel to the axis of
easy ma.gnetization (OLN)), for the free energy of an isolated bubble under the
influence of field H directed along the OLN counter to the magnetization of the
bubble, we get W= We + W + W , where W is the energy caused by the in-
ternal field of the butble,grW eis the energy of a domain boundary and W is
the energy of the bubble in exEernal fi.eld H. Energy W tends to expand the
bLbble and W and W to compress it. Inea certain range of values of H an
equilibrium sg te is established (8W/8d = 0), characterized by a certain value of
the diameter of the bubble, d= f(H e The value of d also depends on the
- thickness, h, of the bubble mediummaterial and on its physical properties [4, 5,
_ 6, 7, 8, 9], which can be characterized by parameter !Co , called the character-
~ istic length of the material:
R.Q = Qgr/u~MS ~
where cr is the specific energy of the domain boundary in J/m2 and u~ = 4II�
�10 7 H/~!r is the magnetic constant.
For bubble materials h is usually chosen to equal (5 to 10)!CO , whereby the
mean diameter of the bubble equals d = 0.5(dk -F d) ti lORO , where dk is the
minimum diameter of a stable bubble ur~~h H= H Pand d is the maximum dia-
meter of a stable bubble with H= H (8 H annihila-
tion of the bubbl.e takes place, and w~iR HPp< Hmin the bubbge ismtransformed
81
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00854R000540020034-7
FOR OFFICIAL USE ONLY
into a strip domAin (fig 3). Therefore, the condition fqx szable operation of a
bubbl,e device is H a H 5 H . Ratios d/d and H /H depend on the
ratio h/~,~ . With~~i/R~ e 5,~~ /dk~ = 3.5 aid ~ a/H~na~ ~:~n, The re-
quired value of d , determined Ey fi-he prescribed Tn~orma,tion storage densityy
is achieved by,selectin, materials with values of cF and M making it possible
to obtain R,0 ti O.ldsr , whereby~ vgr depends on thIrtype of aomain boundary.
JAN -
d
cp
2)
F~v
3)
Gi~iQC~,c~\
~ 5 )
Konnanca \
O NMNH HMAKC He
- Figure 3. Dependence of Diameter of a Bubble on Strength of the External
Magnetic Field
Key:
1. Strip domain region
2. Stable bubble region
3. Collapse reg;.on
4. H
5. Hmin
max
The following equation is true:
d = (dpHmax ' dlcHmin)/(Hmax - Hmin) - (dP - dk)He~(Hmax - Hmin) ' H < H < H . With H = 0.5 (H + H) = H ~ d= d
min e ~X e min ~X sr [mean] sr .
Let x equal the coordinate of the center of the bubble along a certain direction
in the plane of the bubble medium material. If the components of energy W of
the bubble axe functions of x, then the bubble is influenced by the force
F tlu', ci W~~ d~1'~p ~ ~ u W.
+ da. u'., + d, . d, ~
82
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
wiiich c::iuses clisp;lac:enlent a,t tlle bubble if iL exceeds a cextain [hreshol.d value
of Fo , corres.ponding to the coercive fozce, Hc , oi' the bubbl.e medium material
(it is assumed that zhe values of individual derivatives are constant within the
limits of the axea occupied bythe bubble). Then i'rom tr.is expxession it is
possible to find the velocity, v, of the bubble along direction x[10]:
V= E2� (al Ah - a_ JK, -4 ae A.I, - a4 X
X A6rP H. sign
n
where u r is the mobility of the domain boundary; Ah , DH ,0 M and Qo
represen~ changes in the respective quantities along a distance ofsone bubblir
diameter in direction x; and al , a2 , a3 and a4 are positive coefficients.
t1h = d `h d grail h(x) ,
dx
AHC- dgradH,,(a'), etc.
It is obvious from this expression thai motion of the bubble can be accomplished
by creating a gradient in the thickness of the wafer, external field, magnetization
or domain boundary energy in the required direction. By the creation of a speci-
fic distribution of the thickness, magnetization or external field gradient it is
possible to produce preferred motion "paths" or steady positions of the bubble,
which makes it possible, in particular, to reduce the influence of random fields
(interference). Other conditions being equal, the bubble will move in the direc-
tion of the thicker section of the TMP (wafer), of the section with the higher
value of M, in the direction of the lower value of H(aH < 0), etc. By
utilizing,jor example, the dependence of M on temperaturee it is possible to
control the motion of a bubble by means of aslaser beam which causes 1oca1 heating
of the wafer. In particular, if M increases with the temperature (AM > 0) and
it is possible to disregard the influence of a change in agr , then thesbubble
will follow the motion of the laser beam.
In the majority of bubble devices control o� the bubble's motion is accomp'lished
by means of local changes in H with h= c~~nst , M= const and a = const
In this case the absolute valueeo.f the velocity of thesbubble equals gr
AN > N(,,
where u is the mobility of the bubble, AH is the absolute value of the differ-
ence o� components, normal to the surface of the �i1m, of the strength of the
external (1oca1) magnetic field at the front and rear boundaries of the bubble,
determined in relation to the direction of its motion, and H0 is the threshold
field caused by the coercivity, H, of the bubble medium material. Equations
u= ugr/2 and H0 = 8Hc /7 are of an approximate nature.
83
FOR OFFICIAL USE UNLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/42/09: CIA-RDP82-40850R000500420034-7
FOR OFFIC'IA1. USw: ONI.Y
t, specific rel.ationship exis.ts between the dynamic and static paraqeters of
bubble inaterials [5, 6, 7, 8, 9]:
4y= A M,.
U'I'p a Q1,P r
where Y is the gyromagnetic ratio, equal to 2.22�105(M/K) for electron spins;
a is the Landau-Lifshits damping parameter, characterizing the influence of vis-
cous drag and having dimensionality of the frequency; and A= cr2 /8H M is a
constant characterizing the exchange energy of the material. Sr a s
_ Obviously,
ia Qrp 72 A )1/2
l -
E rp - ?7~N,t X ?n Q ~
i i.e., other conditions being equal, an increase in the figure of inerit, Q, of a
- bubble material results in a re3uction in u and, consequently, in U. An
_ effective method of increasing u(the speeg rof response of bubble devices) is
the creation of materials with heightened vaiues of y and/or -low values of a
With an increase in AH there is a reduction in u. In addition, the vzlocity of
a bubble is limited at the higher end by the maximum value of V . The value
of ~ also depends on the strL::ture of a bubble boundary. Gene?ally, the boundary
oi a bubble consists of sections with a Bloch boundary structure separated by
sections with a Mel structure and called vertical Bloch Iines (VBL's). The
number, n, of Mel sections or VBL's is always even and can be positive or nega-
tive. In the particular case when the structure of a bubble boundary conforms
everywhere to a B1och boundary, n= 0. The value of n is positive in those
cases when in moving along the direction of magnei:ization of Bloch sectio;as of the
boundaries of a bubble Nee1 sections are encountered in which the direction of
magnetization corresponds to the general direction of the transfer of magnetic f lux
from the bubble into the surrounding region. Otherwise n< 0. The state of a
bubble boundary is characterized by parameter S= 1+:t/2 .
The expression presented above for velocity V cflrresponds to S= 0. It is
important that with S= 0 the direction of motion of the bubble coincides with
the direction of the maximum value of grad H . With an increase in the abso-
lute value of S, u is reduced and angle $ (angle of drift) increases, com-
posed of the direction of motion of the bubble and the direction of grad He [9]:
(p = aresin a s V
y d tH,. The static properties of the bubble change simultaneously: H is increased and
dk is reduced. The absolute value of S increaser~ in the process of motion of
84
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAI, USF ONLY
the bubble and dynamic restructuring of the domain boundary takes p7,ace. 8ubbles
with high ab.sol,ute va],ues of S are called ha,rd or inflexib].e.
Amottg the methods eliminating the formation of hard bubbl,es., the most widespread
is the method oi' forming on the s:urface ot the tiubb7.e medium material a thin
(less tran one micron) layer Grlth anisotropy of the "easy plane" type. This
method is distinguished by the fact that any direction in the plane of this layer
is a direction of easy magnetization. As a resulfi, the boundary of the bubble can
receive only a definite structuxe corresponding to S= 0 or S= 1. Tn bubble
media fabricated in the form of thin f erxite garnet waf ers with a negative magneto-
striction coefficient, an "easy p lane" is most often foxmed by the method of ion
introduction (fon-implantation), e.g., by the implantation of hydrogen or neon
ions, which causes compression of the surface layer. .
With S= 1, the change in mobility of a bubble as compared with the mobility of
a bubble with S= 0 is inconsiderable, although the angle of drift, 0 , can be
quite noticeable. This phenomenon is utilized in certain bubble memories which
have been developed for increasing the density of the storage of information, which
is written in the form of a defini te structure for bubble boundaries, e.g., a
structure with S= 0 corresponds to a binary 1 and a structure with S= 1 to
binary 0 [9, 11].
Bubble Medium Materials
Orthoferrites, representing antiferromagnetics with a chemical composition of
RFe03, where R is any rare earth element or yttrium or a combination of two
elements (R = R' R" were first used as bubble media [4, 6]. Orthoferrites
have high values oi-Neir figure of inerit, Q, bubble mobility and the magneto-
optic Faraday coefficient. However, the mean diameter of bubbles in them is rela-
tively great and for materials wi th a single rare earth component it equals 80
to 100 V. Only for mixed orthofer rites is it possible to obtain d = 20 to 25
u, but with considerable worsening of the temperature stability ofsEhe material's
parameters. Therefore, orthoferri tes are used primarily in various'magnetooptic
devices where a small diameter of bubbles is not required.
' Of other materials in which stable mobile bubbles can exist, in bubble memories,
only ferrite garnet thin films are widely used, grown on a nonmagnetic garnet
substrate (usually of gadolinium gallium garnet, Gd Ga5015) by the liquid-phase
epitaxy method [.4, 9, 12, 13, 14, 15, 16, 171. At ?he present time many composi-
tions of epitaxial ferrite garnet films have found an application, having the
formula (R1R2...)3(FeM) 5O1 , whe re R is a rare earth elemPnt or yttrium,
M is a metal (usually aluminum o r gallium, sometimes Ge, Si or Mn, partly replac-
ing Fe). By changing the composit ion of the film it is possible to change over
a widc~ range the values of Ms , Ha , Q, R (d U and their dependence on
the temperature. Uniaxial anisotr opy with 2he OER perpendicular to the surface
~ of the film in ferrite garnet films is usually the result of great magnetocrystal-
line or Exchange anisotropy induc ed in the process of growing the film. In cer-
tain cases it can also be the result of mismatching of the lattice constants of
the substrate and �i1m, which caus es great elastic stresses in the film and re-
sults because of magnetostriction in the forznation of the required Ha field.
85
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFIC(AL USE ONLY
The key paxameters c: certain texXite garnet epitaxial fi_lms tqaking it possible
to obtain values ot d ot 0.5 to 10 v are given in table 2[13, 14, 15, 16,
171. The parameters jra number ot other filnts and the.ix temperatuxe character-
istics, which are of first-rate importance, are a1.so presented in [4, 9, 121.
Table 2. Key Parameters of Some Ferrite Garnet Thin Magnetic Fi1ms
vraKC.
1~ I t~~ ~cf~ I I I 3jA%cK I4lr~%A�c
COCTBB NIlmCN k. MNM Q Ha. A/e. I 5,M/C
Gd2
2 Luo
B Fe4.3 Mno.4 Ato.s 012
0.92
2,32
25
3370
28,5
1880
450
.
,
Eu2
7 Lua.3 Fe4.3 Alo,7 0t2
0,62
2,54
12
2380
66
MO
30()
,
Eu2.7 Luo
a Fe,.2 Mno., A10,7 0tz
0.32
2,66
22
4400
95
350
100
.
Eul.4s Yo.4s Cat.t Fe.3.9 Sio 6 Geo.s 012
0,6
4
7,5
1200
65
875
300
Y1,2 Lu0,5 Sm0,4 Ca0,9 Fe4,1 Ge0.9 O12
0,35
-
5,3
1390
-
325
-
Y1 Lu0,7 Sm0.5 Ca0,8 Fe4_ Ge0.8 012
0,19
-
3,5
1400
200
377
-
Lao.e Luz.1 Smo.s Fe4.i Gao.s 0tz
0,14
-
2,6
1040
175
630
-
Lao.: Lul.g Smo.e C80.6 Fe+.+ Geo.6 012
0,12
-
2,9
1800
320
380
-
Eu, Tm2 Fe4.3 Gap
7 012
0.11
1,66
10
3500
-
230
-
,
Euo.s Tm2.2 Pe4.s Geo,s P12
1,0
3
1690
-
1250
-
Eul.7 Yb1.3 Fe5 012
2,8
~
2,1
2440
-
500
I -
Key: 1. Composition of film 4. u, cm2/A�s
2. h, u 5. vmax' m/s
3. Hmax 9 A/cm
Studies are being conducted on the use in bubble devices of two-layer and even
three-layer films with layers differing in chemical composition and physical pro-
psrties [4, 91. Two-layer films can be used for eliminating the formation of
hard bubbles, for producing stable bubbles without an external field, H, or for
constructing bubble devices based on the use of interaction of the domains of
two different layers.
Basic Elements of Magnetic Bubble Integrated Circuits (IC's)
'Po the basic eZements of integrated circt,its employing magnetic bubbles belong
elements making possable the generation of bubbles, the propagation of bubbles
along shift registers, �ixing of the poGition of hubbles in the information storage
mode (vrithout propagagion), char.oing the direction oP motion of bubbles, the ex-
change of bubbles between two shift registers, the multiplication (splitting) of
bubbles, the readaut of information, the erasure of bubbles and logic operations.
The major portion of the area of integrated circuits employing magnetic bubbles
is occupied by propagation elements, torming shift registers. Since all functianal
elements must be Pabricated in a single technological process with a minimum number
of lithographing and registration operations, the choice of design principle and
86
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00854R000540020034-7
FOR OFFICIAL USE ONLY
type of propagation elements to a considerable extent determines the choice of
a11 remaining kunctional e7,ements o� integrate.d cixcuits utilizing magnetic
bubbles, which must match propagar ton elements in t;ie. technolog3:.ca,7, 7cespect and
wtth respect to regions of stable operation (Ot1R's).
Two main classes of bubble device shift registexs exist. The first includes in-
tegrated circuits in which 1oca1 changes of the magnetic fields causing the shift-
ing of bubbles are accomplished by an external magnetic �ield which is alternating
in terms of direction and is common to all integrated circuits; under the influ-
ence of this field magnetization of thin magnetically soft (permalloy) elem-zts
located on the surface of the bubble film or of structures implanted in the sur-
face layer of the film with anisotropy of the "easy plane" type takes place. A
char-=eristic feature of integrated circuits of this class, which we will call
bubble integrated circuits with common fLeld control, is the absence of conduc-
tion lines in the integrated circuits, used for propagating the bubbles. To the
second class belong magnetic bubble integrated circuits with local current control,
in which local changes of the propagating magnetic field are accomplished by means
of electric current flowing through conductors arranged 3.n a certain inanner on the
surface of the bubble f ilm. Bubble integrated circuits of the first class with
permalloy control elements have been preferred for use at the present time.
A circuit for controlling the motion of bubbles by means of magnetizing a permalloy
core is presented in fig 4. The bubble medium, 1, is covered with a dielectric,
2(usually Si02 with a thickness, hn , of 0.4 to 0.5 u), on top of which is placed
a rectangular core, 3, made of a permalloy thin magnetic fi-n (core thickness of
0.3 to 0.5 u)� An external magnetic field, H, aik-~s pogsible the existence of
stable bubbles (cf. fig 3). Possible initialepositioiis of bubbles are designated
by the numbers 4, 5 and 6. The external control fiE:Ld, H r, directed along the
longitudinal axis of the core, 3, from right to left, is ~~irned on the instant
the bubble is at position 4(fig 4b). The core is magnetized and at its left end
is formed a north pole, which attracts the bubble, causing it to move from posi-
tion 4 to position 7, corresponding to the minimum free energy of the bubble in
the total external field. If the bubble is at position 5, then it is repelled by
the right south pole of the core and also goes to position 7. A bubble at posi-
tion 6 is repelled by the core and moves to the rigtL*.
The force acting on the bubble and its stable position (UP) under the permalloy
core are determined by the distribution of component H, norma~l to the surface
of the film, uf the core's external field, averaged forZthe thickness, h, of the
bubble medium film. The distribution of this component along the longitudinal
axis (x) for a value of y= 0 is shown in fig 4c for the case when it is possible
to disregard the inflvence of the bubble on magnetization of the core. The bub-
ble's stable position (UP) corresponds to the region of minimum values of HZ ;
this region is usually called a magnetostatic well (MSXa). The problem of en-
abling a specific path of movement of the bubble reduces essentially to enabling
the correspondinf; law of displacement of the MSYa, which carries along the bubble
in it.
- With H = 0, the permalloy core can be magaetized, although to a lesser degrae,
- under t~i~rinfluence of a bubble located neax it. Hexe for a bubble at position 4
- or 7 with H = 0, position 7 is stable as previously, corresponding to the
minimum of frge energy. For a bubble at position 5 or 6, the nearest stable
87
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
- position is the right edge, 8, ot the core (the center of Che bubble is set
beneath this edge). Thus, the same pexmalloY e1,ements can be used not only for
controlling the motion oX bubbles, but a7,so for fixing the position 09 bubbles
when the power is cut off (HuPr = 0).
2
e~ 2 3 8 x
Hy~p
1 ?
y.' N 5 6 2)
a}
y J y
o~ X
v
b)
HZ
r

HYM
nGX
Key:
C)
Figure 4. Diagram for Controlling the Motion of a Bubble by Means of
Magnetizing a Permalloy Core
1. h
n
2' Hupr [control]
3. UP [stable position]
4. MSXa [magnetostatic well]
Field H, acting on a bubble and rhe magnetostatic we11, is directed counter
to H and results in an increase in the diameter of the bubble. Therefore, in
bubble devices with permalloy control elements H is selected tn be greater than
H (cf, fig 3) for the purpose of enabling a brosd region of stable
operI~tion~(OUR). As a consequence, the annihilation ok bubbles is sometimes ob-
served when H is cut off and there is a corresponding reduction in the depth
of the magnetosEatic well. This is oPten ~liminated by orienting the permanent
magnets creating field H so that a s.tall field component is formed in the
plane of the thin magnetic film of the bubble medium. This component, magnetizing
permalloy cores with H = 0, creates magnetostatic we11s whtch reliably retain
bubbles at specific pcsypions.
88
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFF[CIAL USE ONLY
I.et us discuas Che prlnciples of the design of bubble shift regi$tezs (SR's)
based on permalloy control elements having the form of a T and I(fig 5). Here
four directions are distin uished (1, 2, 3, 4), of the control field, H ,
clockwise, and on the periaalloy elements by these same numbers aze indicREed
positions at which with these directions of H positive poles are formed.
With the first position of H let a bubble~e at position 1 0f an element I.
When H is turned 90 degreis (to direction 2) a north pole forms in position
2 of thyErelement and the bubble goes into this position. One more turn of
H by 90 degrees moves the bubble to position 3 of element I. With the next
tu?n of H by 90 degrees (direction 4) a north pole forms at lower end 4 of
core A andu?he bubble goes into this position. When H is turned 360 degrees
- the bubble occupies position 1 under element II, i.e., '~E rshifted by a single
cycle by a li=:-3r extent of a. T-shaped elements I, II, III, V, VI and VII
and cores A and b located between them serve the purpose of shifting the bubble
linearly along the SR and elements 0 and IV of changing the direction of motion
of the bubble. With a change in the direction of rotation of H the direction
of motion of the bubble is changed to the opposite, which is theu?esult of the
_ symmetry of the shifting elements.
Y!! V/ ~ 2
O - 2 l 4 3 2/V
A 8 f 1 3
23 4 1 2 3 4 ! 2 3 4 Hy~P
i
~
d ~ .
1 �
_ Figure 5. Magnetic Bubble Shift Register Utilizing T-I Permalloy Elements
.
~ Key:
l. H
upr
in recent times preference has been given to shi�ting elements with a single gap
, with a shift cycle of N , which are less critical with regard to the width of
the gap, cS , between neighboring elements and make it possible to increase the
ratio d/d and to reduce the ratio A/d and the mutual influence of bubbles
_ moving along adjacent tracks (SD's) and r_osexpand the integrated circuit's OUR.
The use of these elements increases the information density of bubble integrated
circuits without increasing the resolution of the photolithography equipment used.
Some of the most widespread single-gap permalloy shifting elements, which have
many modifications, are presented in fig 6(a and b--asymmetric chevrons; c and d--
asymmetric subdisks; e--a "tapir"; f and g---symmetric chevrons; and h--a symmetric
Y element) [18, 19, 20, 21, 22, 231. In the first five types of elements bubbles
can be moved only from left to right with rotation of the control field, HuPr'
clockwise (fig 61) and in the last three in any direction depending on the
89
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
direction of roCation of H (fig 6j). The wt.der 4UR in asymmetric elements
(a to e) is caused by the faEE that the increased mass (area) of the permalloy
at the end of the element "receiving" the bubble facilitates the bubble's crossing
the gap between two neighboring elements. Symmetric elements g and h have a wider
OUR than simplest chevron f, which is used primarily in bubble expanders incltided
in magnetoresistive readout devices, passive magnetic bubble splitters and some
logic elements.
a) ~
b)
c)
d)
e)
f ) AAA^
g) ~~II Ull ZJ
h) V V ll Q~11
Rey:
r'igure 6. Single-Gap Permalloy Shifting'Elements
1. H
upr
Hyp
i)
Ny�M,
J)
Displacement field H is usually created by two flat permanent magnets (1 and 2
iri f.ig 7a) and Pield eH by current i and i flowing through two ortho-
gonally positioned coilsprX and Y, insYde of whYch there are one or more mag-
netic bubble integrated circuits. The coils with the integrated circuits and the
permanent magnets are placed in a single case, which serves simultaneously as a
magnetic shield protecting the bubble module from the influence of external magnet-
ic fields with strength up to 4000 ,A/m [24]. With identical permanent coils X
and Y current i= I sin wt and i= I cos wt flowing through them create
field H upr , whicK rotates in the p1anX of mthe bubble integrated circuit with an
90
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
angular velocity of w and a constant amplitude. However, the amplitude of
H can also not be constant. Therefore, control current i and i of
triangular form (fig 7b) are widely used, which are more simpl.e to form~ espe--
cially when a bubble memory operates in the staxt-stop mode [25]. In this case
the vector of H in a polar system of coordinates describes a rectang7.e with
corners situatedugn axes x and y(fig 7c), whexe axis x is in line with the
direction of motion of the bubble. The change from the bubble shift mode to the
storage mocle is carried out at moment wt = 2n7 (n = 1, 2, where i=
= 0, by changing cuzrent i from -I to zero (current i thereby musY re-
main equal to zero). MotionXOt the bubUle is restored by chaXging i from zaro
to its maximum negative value with i A Q and by subsequently changYng iX and
i y according to fig 7b. y
~
N
X
iy
Y
~x
L N 2
�O
~y ?l11P ~f
CTOR
!X 1 ~
X
I'igur.e 7. Diagram o.f Formation of Displacement and Control Fields �or a
Magnetic Bubble Modiile
[Key on following pa ge]
91
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
a)
!

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFF(CIAL USE ONLY
Key:
7.. SCop
2. Start
3. H
upr
Three main principles are known for designing magnetic bulibl.e generators (G's):
by means of a current loop, creating a local magnetic field directed counter to
H and sufficient for the origin of a bubble (fig 8a and b); by means of division
- by means of a current pulse, i, of a bubble se1f-originating beneath a massive
permallog element with a largegleakage field (fig 8c); by using a massive perualloy
element as the units generator (G1) and erasing by means of current iA the bubble
formed, if it is required to write a 0(fig 8d).
J.) 14 ~ t-2)
7 ~
a)
rn
r(
3p~
Nir, C)
+ ~r
rn
b) ~ys~p
u'"a u
vU
~
Ns"p e> /IA 4)
~n + t,,
r> ~H
pp
a
AA
N)W d) 5
Figure 8. Diagram of Functional Elements of Bubble Integrated Circuits
Key:
1. H 4. PA [passive annihilator]
2. TPPycurrent loopJ 5. AA [active annihilator]
3. Bubble
The erasure device, also called an annihilator, can be designed similarly to the
current generator (fig 8a), but with the reverse direction of current iA for the
purpose of creazing a total field greater than the collapse fie1d, H . In addi-
tion to such an "active" annihilator (AA), passive annihilators (PA'sjXare used,
designed similarly to passive "1" genexators (fig 8d). Bubbles approaching a
"massive" permalloy element are captured b}r it and are not transmitted further.
Current loops are formed by thin-fi7.m lines usually made of an Al-Cu alloy.
Bubble generators for other shifting elements are designed similarly (cf. fig 6).
32
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504020034-7
. FOR OFFICIAL USE ONLY 'r. _ .

_ Current pulses i and iA must be phased in a specific manner in re7,ation to
H , as is convgntionally illustrated in fig 8, for the purpose of enpbling
tHErproper operation of G's and AA's. In a generatox (fig 8c) the disfiance between
element 1 and G1 is such that with i= 0 element 1 cannot capture a domain
formed on the right face of Gl with tfle corr.esponding direcXion of R . Xn
. the circuit given (fig 8d) such capture does take place and with the rgfation of
H a domain is stre*_ched and then split, forming two iqagnetic bubbles, one of
w~iy~h continues utoving around generator G1, and the othex is directed toward the
A,A.
In essence generator G in fig 8c forms an active splitter or replicator (AR),
and in the circuit in fig 8d, a passive bubble splitter (PR). These bubble split-
ters (repiicators) are used in bubble integrated circuits for branching information
for the purpose of readout'without removing the bubble from the shift register
(SR), w'iiich serves the purpose of the prolonged storage of information. Many
variants of AR's exist, one of which is presented in fig 9[22, 23, 26]. Here the
AR serves the puxpose of trans�erring information from SR-1 to SR-2 while storing
the transferred information for thr� purpose of readout without removing the bubble.
_ SR-1 is designed in such a mannet that with a certain direction of H �'ne
bubble, traveling beneath it, is stretched along the entire element. u?! at this
moment current pulse i i.s supplied to the current loop (TP), then the bubble
is divided into two par.Es ;ahich are mutually repelled. Qne of them continues to
move througli SR-1, and the second goes to SR-2. If current pulse i is supplied
to the point o� stretchitig of the bubble along element l, the bubbleparriving at
this element is conveyed to SR-2 without splitting. Tn this case we get a one-
way switch (OP) for the 3irection of mdtion of tYe bubble (only from SR-1 to SR-2).
Two-way switches (DP's) exist, making it possible to transfer a bubble both from
SR-l.to SR-2 and from SR-2 to SR-1. The function performed is determined by the
- phase shift between the switching current, i, and the control field. With i=
= 0, bubbles are moved in step independentl~ of one another through SR-1 and p
, SR-2. In addition to these elements, according to the same principle are designed
switches (KO's) for swapping tiubbles between two shift registers and universal
- switches (KO-R's) which implement either the function of swapping bubbles or the
function of splitting bubbles depending on the instant of the supply of and phase
- of the control currertt [22, 23, 261.
In fig 10 is presented a diagram oi a passive bubble splitter (PR) designed with
permalloy elements of the symmetric chevron type. Moving from left to right along
element 1, the bubble is captured by elements 2 and 3, is stretehed and is di-
vided into two bubbles, which continue to move through channels I and II. If the
direction of rotation of H is changed, then the bubbles wiZl be moved through
channels I and II from rigOpEo left. In this case elements l, 2 and 3 form an
OR circuit implementing a logical OR function with the entry of a bubble through
one of the channels (the simultaneous entry of two bubbles can cause failure
because of their mutiial repulsion).
The readout of information, represented by the presence or absence of a bubble,
can be accomplished by any circuits reacting to the magnetic ,field of a bubble or
its alteration and based on utilization of the electromagnetic induction phenome-
- non, the magnetooptic effect, the magnetoresistive effect, the Ha11 effect, etc.
_ Only the magnetoresistive effect has received practical application in modern
bubble memories in view of the technological advantages: execution o� the
93
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL U5E ONLY
magnetoresistive bubble sensor (detector) from the same permalloy thin magnetic
film as the other iunctional elements of the integrated cixcuit in a single tech-
nological process. The level of the output signal of fihe sensor is U -
= I.ARd =IdkRd , where Id is the current of the sensor and ARd isSihe change
in ~he initial resistance of the sensor, Rd , under the effect of the magnetic
field of the bubble read. The value of Id is lfmited to a value on the order of
5 to 10 mA, and coefficient k is 1ow (0.02 to 0.03). Therefore, for the purpose
of raising U to the level required for reliable readout against a background
of interferenceh increasing Rd by increasing the length of the sensor is resorted
to, whereby it is necessary at the same time fio stretch the bubble read in order
for the sensor to be completely under the influence oP its field. The simplest
variont of a readout unit (SU) consisting of a chevron domain expander (RD) and
a sensor (NID) formed by a chain of series-connected permalloy elements is illus-
trated in fig 11. Behind the sensor usually comes a circuit which reduces the
domain ta normal dimensions, after which the bubble can be directed ta any SR or
annihilator (PA), if erasure is required after reading. Often a second compensat-
ing sensor is provided, positioned directly behind the first and connected together
with it in a bridge circuit .for the purpose of reducing noise caused by the field
of H Not infrequently in RD's as;~rmmetric chevrons of a special shape ara
used ygr che purpose of expanding its OUR, as well as other methods of connecting
MID elements for the purpose of raising the level of the output signal [20, 22, 231.
In bubble devices wfth a nominal domain diameter of 2 to 3 u the number of chevrons
in the IrID reaches 300 to 500, which makes it possible to produce a strip domain
1.5 mm long. This stretching of the bubble read is usually achieved in 15 to 30
cycles of the shifting field.
CP -2 1)
A n ~Z~
ip -
2)
\ .4P
4)
~ Li
J
1
O ^
I 1
1
CP-1
rn
3)
Figure 9.
Key:
1. SR-2
2. ir
94
Diagram of Active Bubble
Splitter
3. Current laop
4. Active splitter
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00854R000500020034-7
FOR OFFICIAL USE ONLY
J
. Figure 10. Diagram of Passive Bubble Splitter
- Kev:
1. H
upr
~ 3)
~
y
- -
AFA
4)
Figure 11. Diagram of Bubble Integrated Circuit Readout Unit
Key:
l. RD [domain expander] 3. Ia
2. I~ID [sensor] 4. PA [passive annihilator]
95
FOR OFFICIAL USE OI'rLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
u 2)
1) PQ, I9 + MR

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
Structure of a Magnetic Module (Integrated Cizcuit) for a Bubb].e Memory
A bubble memory magnetic module is a functional7,}* and structurally complete item
coneisting of one or more magnetic bubble integrated ctxcuits and elements creat-
ing magn,~tic displacement and control fields 'kcf. fi:g 7a) and placed in a shared
' shielding case wiCh the necessary electric terminals. The structure of a magnetic
~ module is determined by the structure of the bubb].e integrated circuit, which
~ depends on the purpose oP the memory and on the principle used for controlling
the magtietic bubble.
Z~ao main classes of bubble memories exist. In memories of the first class the
' sequential single-channel and/or multichannel readin and readout of binary in-
a formation are accoi-plished the same as when recording on moving magnetic tape.
To the second class belong memories in which information is written and read in
_ pages or blocks according to an address ass--'gned, for e.ach page. The writing or
. xeading of ir.formation (words) t?kes place sequentially within the limits of a
page. The addresses cf pages can be assigned in an arbitrary manner. Bubble
memories of the secon3 type make it possible to xeduce substantially the time for
accessing inform.ation, the minimiz3tion of which is one of the mai~Z objectives when
using bubble~memories as part of a computer.
A typical structural diagram of one memo-ry bubble integrated circuit of the first
class is presented in fig 12. The circuit works in the following manner. During
readin, current p-.1ses, iZ~ b~gadi~~ , phased with current creating a rotating
H field, enter the con r ed bble generator, G,iso that when a 1 is written
- a''Bubble appears in the output of the G.and does not when a 0 is written. The
bubbles gene-sted are moved under the influence of H in the direction of the
merger point, S, and enter the shift register, SR ."?I N is the capacity of the
SR and n S is tlie number of cycles between G and point S, then N+ nGg shift
periods (~otations of H ) after the beginnir.g of the entry the free 5R is
filled. Here the entry o~curs onl}- during the first N periods, and during the
- entry Period the current in the active annihilator, AA, equals iA = 0. The last
written bit will be at potpt S, and part of the written information in register
section SV. Often the entire capacity of the SR, N, is not used, but only the
capacity, N, of the section of the SR. from point S to the active splitter (re-
plicator), XR. In this case after the completion of an entry (after N+ N GS
periods) the first bit entered will be found opposite the splitter, AR,rand
can proceed to be read oiit immediately, by supplying current pulses, i, to the
- AR in step with the sbtfting current. The bubble in the'.AR is split: rOne part is
sent to the readout v.nit, Sch, and then to the passive annihilator, PA, and the
ser.ond *_o point S for further storage in the SR: A�ter kN periods, where k is
a whole number, *_he information written in the SR assumes its initial arrangement.
. If it is necessary ta erase or to replace information entered earlier, tinen when
the bubble arrives at the position of AA an erase pulse, i, is supplied. It is
_ posgib7,e to enter new information simultaneously, since the distance (number of
cycles) from G and from AA to the merger point, S, is made identical. ~
The readout (access) time for the ,�irst bit, t~ , and the average time for~readin5
out a random hit, tSr , after the entry of nr bits equals: ~
96
' FOR OFF[CIAL USE ONLY
,
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOP. OFFICIAL USE ONLY
tl _ np.ca ; tep _ 2Rp.cv+ NP ~
f 2t �
where nr, h is the number of cycles between the AR and detector, MD, of the
readout un~i(cf. fig 11) and f is the operating frequencl, of the rotating
f ield .
1) 4) ueax
5) Cy !lA 6)
2 y AA
C AP Gp B
3) ~ � 8) 10)
I CP
i 9)
~ I
I ~
F
Figure 12. Structural Diagram of Bubble Integrated Circuit with One Shift
Register
Key:
1.
iz
[readin]
6�
Passive annihilatdr
2.
p
G
7.
Acfive splitter
3.
S
8.
i
4.
Output
9.
SEift register
5.
Readout unit
10.
V
The number of electric terminals for a single integrated circuit equals !C =~8
when using a single magnetoresistive sensor in the readout unit. When twoelrID's
[magnetoresistive sensors] are used, t = 9. The total number of terminals of
a bubble module with a single integratea circuit, taking into account the terminals
of the coils creating the control field, equals L = t + 4. Up to eight in-
tegrated circuits are often placed in a single moduge for thP purpose of reducing
the cost of a memory. If recording and readout take place '_n them in psr,allel,
as in a multitrack tape recorder, then the lines controlling the operation of the
AA and R's can be connected in series, and one of the enus of a11 write and read
lines can be connected to a common point. Then the minimum number of module term-
inals equals I, = 2n I.+ 9, where n I is the number of integrated circuits
in the module. eB~utnif independent writingor reading are required, then all AA's
and AR's must have individual terminals, and then Z = 4n + 5. The value
of Le.min limits the maximum number of integrated cimrcuits in~a single module.
97
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFFiC1AL USF ONLY
There is another limitation in the fact that all integrated circuits of a given
module must have identical OUR's [regions of stable operation]. Otherwise the
OUR of the module wi11 be smaller than the OUR of an individual integrated cir--
cuit.
The key parameters of some magnetic modules designed on the basis oX single-regis-
rer 4ategrated circuits are presented in table 3[27, 28, 291.
- Table 3. Key Parameters of Some Bubble Memory Modules with Single-Register
Integrated Circuits
Type oF module, Module parameters
_ firm, country Mean domain Module Operating Mean Tempera- Maximum
y diameter, capacity, frequency, access ture range, Powex
dsrf lm nICNr, f, kHz time, �C require-
k6its tsr, s ment, P,
Operat- Infor- V�A
ing mation
storage
29A, Western
Electric, USA 3 4x68 48 0.71 0;+50 -40;+80 4
~ FBM31DA, Fujitsu, ,
- Japan 3 4X74 125 0.29 -20;+70 -40;+90 2
' POS-8, Rockwell
International,
USA 4 8X103 160 0.32 -10;+60 -40;+85 6.2
, In sing].e-register bubble integrated circuits usually N ti 100 kbits , since any
' defect in a sequential register puts the entire integrated circuit out of order
and increasing N increases the probability of the occurrence 6f defects. The
optimum value of rN increases in proportion to improvement of the technology for
producing bubbl2-infegrated circuits.
Typical examples of the application of Pirst-class bubble memories are airborne
high-reliability sequential information recording and readout systems, equipment
for the digital recording and repeated reproduction of voice messages, systems
for the numerical program control of machine tools and the like.
- Second-class bubble memories have been developed and used most widely. Numerous
, variants of the structure of bubble integrated circuits exist for these memories
(20-23, 26, 30-34]. One of the first structural diagrams which received practical
application is presented in fig 13. The integrated circuit contains one distri-
buting register, RR, also called an input/output register, and m storage regis-
ters, NR's, serving the purtase of stox'ing information. The distance between ad-
� jacent NR's equals two cyc.les. Ea,ch NR is connected to the RR via its own exchange
switch, K0. A11 KO's are ordinarily controlled in step by a single shared curr,ent,
i r , for the purpose of reducing the number of integrated circuit
t�ermiitayshaiee~ However, the use oP several independently controlled groups of
98
FORc OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09
CIA-RDP82-00850R400500020034-7
FOR OFFICZAL USE OPVLY
l:G's increasea cansiderably the flexibility of an fntegrated cixcuit and o:ten
makes it posaib7,e to apeed considerably the accessinp, of infox'mation.
6) ue~x
7) /lA
~8~
~2)
AA P 6p
pp 1 0)
io6M
/f0 m f-------131 IZ 111
NPm I I I 11 t HP i
Key:
` 5 HP
Figure 13. Structural Aiagram of Bubble Integratcd Circuit with Many
Storages and a Single Aistributing Rpgister
1.
2.
izp [readin]
G
7.
8.
Readout unit
FA [passive annihilator]
3.
Distributing register
9.
Active aplitter
4.
Exchange switch
10.
i
5.
Storage registers
11.
iobm [exchange]
6.
Output
With i = 0, the RR and NR's are not interconnected and bubbles move through
all reg~sters independently of one another. During a write oQeration, a block of
information, usually in the form of like bits of m worda forming.a.sing7.Q page,
is entered into the RFZ by means of the generator, G. When the first word (bit)
is opposite NR1 and the last opposite NRm, a current pulse, lobm , transfers the
eneire page from the RR to the NR's. Tf other ir.formation was previously entered
in the individual locations of the NR's, then by means of the I h pulse it is
first removed to the RR (the bubbles removed are placed between ~Tie bubbles of the
new page to ba entered), after which the new information is transferred Prom the
RR into the NR's. The page removed can be read out, just as in the circuit in
fig 12, and then erased by means of the AA and PA. Tt is also possible fio erase
without reading out, if i is set equal to zexo. k'or the purpose of reading out
without erasing, khe bubbl.es removed from the storage registers to the input/oufi�-
put register, MN - Mn + 7. periods after the removal, passing thP AR, AA and
point S, again a~xive at their KO's and are tr.ansferred by current I obm back to
the storage registers. Here MNR and MR are the capacities of the storage
99
FOR OFF''ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R440500020034-7
FOR OFFICIAL USE OPILY
and input/output registers, respective],y. They diffCer by one, since when in-
formation is xemoved from the NRFs to the RIZ this information is delayed by
a single cycle.
The circuit in fig 13 has two shortcomings: The speed of the input and output of
information is twofo7,d less than the speed of motion of bubbles thxough storage
regtsters, since there should be a distance ok two cycles between adjacent words
entered into the input/output register during a write operation, and the same dis-
tance is obtained in the removal of information from the storage register to the
input/output register fur readout; the time of a complete accessing cycle during
readout, made up of the access time and the regeneration time (input back into the
storage register), is relatively long. A number or methods have been suggested
for eliminating these shortcomings [23, 26, 30-33], the implementation of one of
which is illustrated in fig 14. In the circuit presented in this figure, at one
end of all sLOrage registers there is an input/output register, RR-1, connected
to the storage registers by means of exchange switches (KO's), and at the other
end an output register, RR-2, for the readout unit, Sch, connected to the storage
registers by means of active bubble splitters, AR's. The input and output of ir.-
formation with the speed of its movement through the storage registers ar.e made
possible by dividing the stezage registers into two groups: odd and even.
Bubble generator G1 operates at odd cycles of the shifting field and G2 at even,
i.e., tlie write operation is carried out in each cycle. When new information is
entered the old inforr.iation removed through the exchange switches is erased in the
passive annihilator. Information is always read out without removing it from
storage registers. The two information flaws removed from the odd and even storage
registers through the AR to the individual RR-2 registers merge at point S into a
single flow, which enters the readout unit and PA.
Cv /1A
1~ ~sa
CZ PP-2
~
~p APl AP ~
HP yBP
1 H 1 2 - ~
14p) Hll I t5)
If0 KO KO
~OJM
PP-1 PP- BP
1
2 I !lA ~o6M
/lA ry ' r2 107
Key:
Figure 14. Struczural Diagram of Bubb7.e Integrated Circuit with Reduced
Access Time and I:ncre$sed Data Txansfer Rate
1. S[merger point] 4. Even storage registers
2. Rg_1 5. VR [auxiliary storage register]
3. Odd storage registers [Cf. .figs 12 and 13 for remaini:ng symbols]
100
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
The mean time, t ,for producing a word (b~:t) of a random page in the circuit
in fig 14 duringereadout equals;
f ( ~ 2 -F~
ahere 2Sch ti m/2 is the number of cyclea between AR and Sch [readout unit].
v
The minimum value of t apptloximately equals 2N/i and is achieved with
~ ti 2m , where N=~ is the information capacity of the integrated circuit.
Tiie meatc time, tl , for accessing the first word of a random page equals:
tl ='f 1 M2 +
The mean time, tzam ' for replacing a random page after its preliminary readout
equals:
tolw = tCp -f" .P
2 �
1
With an assigned value of the capacity, N, of the integrated circuit, the choice
- of ratio MNR /m is determined not only on the condition of minimizing the access
time in a specific operating mode of the memory (readout, write, replacement of
infirmation after readout), riut also by the required size of pages (usually m=
= 2 , where k is a whole number) and by the permissible power dissipated by the
KO's and AR (the dissipated power increases in proportion to m), and by the
feasibility, for technologieal reasons, of using bubble media chips having a shape
close to square, which for the circuit 3.n fig 14 is consiatent with MNR ti 4m .
if in the circuit in fig 14 symmetric (reversible) ahifting elements (cf. fig 6)
zre used and a page to be read out ie moved to the terminals of the AR into RR-2
c:irough the shortest route, then with MNR ti 4m we get a minimum value of t er ti
-
f 1 -N .
"I-e speed of the input and output of information is also brought to the speed of
ira movement in storagp registers by increaaing twofold the length of shift ele-
' -~?nts in the input/output register as cumpared with shift elements in storage re-
E.'_;ters. In this case a bubble travels the distance between two neighboring
E;.orage regi.sters through the input/output register in only one cycle of the rota-
field and the necessity of dividing storage regigtexs iato odd and even dis-
p�:r,aara (cf. fig 14). BuDb1,e integrated cixcuits have been designed on this prin-
ci~:;e (fig 15), possessing the same properties as integxated circuits designed
accc~-ding to the circuit in fig 14, but simplex in the technological respect [23,
34].
101
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504020034-7
FOR OFFiC1AL USE ONY.Y
Figure 15. Variant of Structural Diagram of a Bubble Integrated Circuit
with Short Access Time
[CQ. figs 12, 13 anu' 14 for key]
Since storage registers (NR's) occupy the naJor portion of an integrated circuit,
defects in a chip or permalloy element mest oxten cause malfunctioning of only one
or several NR's of the class of inemories discussed (cf. figa 13, 14 and 15).
Therefore, an excess number, m , of etorage registere is usually provided in
them, which makes it possible ~o increase considerably the percentage of the yield
of suitable circuits under conditions of series production. The relative value
of mi b/m is determined by economic conaiderations and depends on thefproduction
techno~ogy level reached. The addresses of maliunetioning or unused NR s can be
stored either in an auxiliary NR (VR in the c4cuite in figs 14 and 154 which has
individually controlled exchange switches (KO 's) and replicators (AR 's) con-
necting this VR to RR-1 and RR-2, or in a read-only (programmable) storage (PPZU).
[dhen a VR is used it also serves the purpose of storing tags (a code) for the
starting position of all NR's, which is necessary for synchronizing the operation
of a bubble memory when it chanRes from the atorage mode to write or read modes.
Otherwiae information on the current position of information in the NR'a must be
- stored permanently in an external setaiconductor counter with a scaling factor of
MNR , including when the controlling magnetic field is switched off.
- In addition to redundant storage registers (idR's) which serve the purpoae of in-
- creaeing the yield of suitable integrated cixcuits, mk additional NR's are often
provided, employed gox using codes whi,ch detect and coxrect erxors. There are
four different kinds of erxors. xhey dxe "mil,d" errors oxiginating only during
readout (they cannot appear in repeated readout); "serious" errors originating as
the result of the dieappeaxa,nce or appearance of a super�luous bubble in the NR;
errors appearing in catastrophic f"ilure of an NR, KO ox of the AR connecting
NR's with the RR; and errors reaulting from failure of the generator, RR or reaaout
unit, i.e., complete failure of the integrated ci,rcuit. The probabilities, Pi ,
102
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R440500020034-7
FOR OFFICIAL USE ONLY
of the appearance of errors of each kind are intexrelated as fo].lows; P4 ucneWU~arNas
i"x 10) ~
oa e ~~p 4
A......,. L/JqQA OWNfM nnay
7) XoNmponmp 15)
Key:
Figure 18. Structural Diagram of Bubble Memory Contxoller
1.
Address and instruction line
10.
2.
Data line
11.
3.
Synchronization pulse genera-
12.
tor
13.
4.
Instruction register
14.
5.
A.ddress x'egister
15.
6.
Page length register
16.
7.
Data register
17.
8.
Tnstruction decoder
9.
Address decoder
Data oxdering circuit
Pulse counter.
Word counter
Error correction ci.rcuit
PPZU [progxammable ROM]
Controller
Control. instruction generator (FKZ1)
To memory units
Trends in Further Aeve7,opment of Bubble Memoxies
Major trends in fux'thex development o.f bubble memoxies xek7,ect the aspi.ration of
= 1oweri.ng the unit cost, of improving the speed of response and reli,ability and of
expanding the functinnal, characteristi,ca oP these memori,es. Reduction of the unit
cost of a bubbl,e memory (cost in terms of memory capacity) can be achieved by
increasing the informati.on storage densi,ty, j , or capacity of a single integrated
circuit upon the condir.ion that the capacity of the fntegrated cixcui:t tncreases
109
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
a
FOIt OFFICIAL USE ONLY
more rapidly than the cost of pxoduction. Hithexto an inc,xease in the intorma-
tion densit}r, j , of bubble integrated cixcuits haa been achi,eved by reductng the
mean (nominal) diameter, d , of the bubble and the spacing o# the shifting
striicture, adsr . Wit~irthis j ti N~2 ,
LIMQ
- ' 69QKU ncMpmu
) ^
, 3) lUuHa ad,oecoe (1 KOMQHg
NI ~
a,mpw�n. 2
_ 1) 2) 4) 5) 6)
yee M~ o3y ~n3y ~3y [~MQ
~
3y
I
- il
(~IUHQ dQHHb/X 8) yBe ILIni
Key:
Figure 19. Structural Diagram of Microcomputer Containing aBubble Memory
1. Input/output unit
2. Microprocessor
3. Addresa and instruction 11ne
4. RAM
5. Programmable ROM
6. ROM
7. Bubble memory contxollex
8. Data line
9. Bubble s.c;uory unit
T.he minimum value of a, limited by the mutual influence o,f neighboring bubbles
when usi.ng single-gap ahifting elements (cf. �is 6), equals 4. The minimum value
of d for bubble integrated circuits with the familiar types of permalloy
shifting elements (cf. fig 6) is limited by the resolution of the lithographing
equipment used and by substantial worsening of the effectiveness of the use of
permalloy elementa for controlling the motion of bubbles with a reductj.on ir. their
linear dimensions in keeping with d with a simultaneoua increase in the rp-
quired value of H f [40f~ At the preaent time integrated cirauits
are being produced"ff thCOan~ug b~~e diameter of d q 1.8 to 2 um (X R 7.5 to 8 u)�
It is assumed that the minimiim values o~ d Rnd A for integxated circuits
with permAlloy shifting elements equal respectively 1 and 4 um, which makes it
' possible to obtain an 1nformatiqn density, J , on the order of 6 Mbits/cm2.
110
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
Among bubble integrated circuits with shared field control the moat wideapread
have become integrated cixcuits with ion-implanted shifting structures [41, 42].
In these, tipon rotation of the external contxol fie].d, Hu x, bubbles move along
a chair. of abutting nonimplanted disks (ND's in fi:g 20a) gr rectangles (NP's in
fig 20b). Nonimplanted chains are formed by coating by the methor: of photolitho-
graphy individual regions of aferrite garnet fi1m with a prntective layer which
does not admit ions implanted in the surrounding region. Tn the implanted rPgions
(TO's) an "easy" surface is formed, magnetized 3n the direction of the influence
of field H . A,t the interface of implanted and nonimplanted regions magneti.c
holes form,uHich exert a correspondi.ng influence on the bubbles, which can be
moved simultaneously along both directions of the chain of abutting disks (rect-
angles), e.g., cloc:kwi.se. The chain of abutting nonimplanted disks thus forms a
closed bubble shift register. Elements have been developed which implement in
ior_-implanted biibble integrated circuits the functions required for constructing
a memory (bubble generation, switching, readout, etc.) [42].
2) - No 3~,yQ ~5) y~
~
NO 4) HP
a)
6) 10) r
PP-2
~P^+ 11)
?
PP-r
No NO
L..~~
~ Ko 8> xo
~y 9) b)
Figure 20. Ion-Implanted Bubble Shifting Structures
i Key:
I 1.
Nonimplanted disk
7.
Nonimplanted rectangle
! 2.
Implanted region
8.
Exchange switch
3.
Bubhle
9.
Readout unit
4.
Storage register
10.
Genexator
5.
H
l1.
Generator ct~rrent
6.
In~ut/output unit No 2
Ion-implanted ghifting structures have two important advantag?s as compared with
the permalloy discussed earlier: leas strict requirements for lithographic re-
soluti.on because of the absence o.f gaps which are critical in tezms o� their size
between neighboring elements (approximatel,y eightfold Lower resolution is required
111
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFEICIAL USE ONLY
for producing the same d and X in these structures), and the 7.ack of the
need to increase H in~`proportion to a reduction in d and X [40].
These structures, be~ause oE the absence of nontransparenfispexYnalloy elements
dn the surface of the ferrite garnet fi1m, can be used in various magnetooptic
domain devices (displags, changeable transparencies, etc.).
Another method of creating 1oca1 changes in the magnetic properties of ferrite
~ garnet films !'or controlling the motion of bubbJ.es is the selective annealing of
individual sections of a film by means of a focused laser beam. For example, by
this method are produced para11e1 annealed channels with enhanced magnetization,
M, along which bubbles are propagated under the influence of an appropriate
control field [43].
- The minimum distance, a, between neighboring bubbles in integrated circuits can
be reduced twofold (from 4d to 2d ) if the active section of the integrated
circuit (the storage) is compietely filled with bubbles. In this case a so-called
domain array is formed in which the interaction ot neighboring bubbles is of a
constant nature, which makes it.possib?e to reduce a and X = ad r. Here in-
formation (a 0 or 1) must be repreaented not in the form of the agsence or pre-
sence of a bubble, biit in the form of a certain state of its boundary, character-
ized by the value of S(cf. the section "Conditiona for the Existence and Pro-
perties of Mobile Bubbles"). With a given value of d the utilization of domain
arrays makes It possible to increase fourfold the surfa~e density of information,
j . The propagation of bubbles in the array, in particular, can be accomplished
by means of permalloy elements with large gaps of noncritical size [11].
A qiiite promising method of lowering the cost of a bubble module is the creation
of functional elements making it poesible to fabricate bubble integrated circuits
with permalloy or ion-3mplanted elements with the use of only one lithography
stage. In the first group of circuits for this it is necessary to form permalloy
conductors simultaneously with propagating elements [20, 44, 45, 461. A bubble
module can be simplif.ied by employing two-layer f.errite garnet films as bubble
media not requiring a permanent bias field [47].
A disadvantage of a bubble module wi-th a shared control field is the relatively
high, primarily rPactive, power which must be supplied to the coils creating the
rotating controlli.ng magnetic field, H . This power is proportional to the
space in which thp field is created anRpwhich is much greater than the volume of
the ferrite garnet film with control elements. The required power grows somewhat
- more rapidly than thP operating frequency, f, si.nce with an increase in frequency
it is necessary to i.ncrease H . Tt ie precisely the permissible value of the
power required for the cotls w~i~ch limits the maximum operating frequency of a'
bubble integrated cixcui.t with field control to values of approximately 300 kHz,
which is fax 6el.ow the maximum opexating frequency determined by the mobility or
maximum speed of buhbles. One way of eubstantially reducing the power requirement
and according.ly increasing the maximum operating frequPncy of bubble devices is
to change from shared field control to 1oca1 currenti control. In the latter in-
stance 1oca1 magnetic fieJ.ds accomplishing the pxopagation of buhbles are created
by the current flowing through conductors 1.ocated on the surface of the ferrite
garnet fi1m. At the same time there is a drasttc reduction in the space occupied
by the controlling magnetic field and the reactive power necessary for crearing it.
112
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
, A number of designs of bubble integrated cixcuits wfi.th laca,], cuxl;enX contxol
have been sugges-ted [45, 481, An opexa,t3,n$ fxequency pf 4t0 5 MHz haa been
obtained for some exper3;mental bubbl.e propagation cixcui.ts and thi,a is not the
11mit. Quite promisi:ng is the accomplishment of 1.oca], cuxxent contxol (without
- permalloy el,esuents) by means of two perforated conducti;ng 7.ayexs applted to the
surface of aferxi,te garnet i'i1m with an fntermediate ingulating layer [48].
_ The c.urrent flowing around a hole (perforation) creates a 1oca1 magnetic field
which influencea a bubble near the hole. By the appropriate distrtbut3,on of holes
in two condticting layers and by alternating the current f Iowing thxough these
layers it is possib].e to provide the required propagation of bubbles, as we11 as
to perform the functtons of genexating, sfiretching and changing the direction of
motion of bubbles. Tt has been predicted that on the basis of this 1oca1 current
- co.rLtroi bubble integrated circuits will be created with an information density of
j ti 107 bits/cm2 and with an operating frequency reaching 4 to 20 MHz.
One more impcrtant trend in the development of bubble integrated circuits is the,
creation of new tyPes df functional elements (coders, decoders, logic elements,
arithmetic units, etc.) compatible in the technological respect and in terms of
region of stable operation with propagation elements [49, 50, 51]. The creation
of these elements will make it possible to increase considerably the effective-
- ness of the application of bubble memories in computing systems (to reduce access
time and the amount and cost of external electronics and to improve the reliability
- and expand the functional capabilities of inemories).
Substantial improvement of the characteristics of bubble memories can be achieved
by structural and+software methods: by dynamic ordertng of information in the
memory; -oy organizing a memory with minimum access time, b}r accomplishing content
accessing, etc. [32, 51,,52, 53, 54]. Let us limit ourselves to discussing only
one possib7.e bubble integrated circuit structure making it possible to improve
considerably the speed of resgonse of a memory. We divide a bubble integrated
circuit into four groups of individually controlled registers: input register
RR-t, input register RR-2, a group of buffer registers (BR's) and a group of
storage xegisters (NR's) (f:ig 21). Tt is assumed that by means of control current
il , 1 2 , iBR and iNR it is possible to enable the oPeration of any group either
in the bubbie motion mode or in the stationary bubble storage mode, regardless of
the state of the other groups of regiaters. A bubble integrated circuit v;ith local
current control [48] affords this possibility. The BR's and NR's are intercon-
nected by means of exchange switches (KO's) and with RR-1 and RR-2 by means nf
universal switches (UK's) which implement, depending on the phase for S'upplying the
control current, either an exchange function or a switching function. The storage
in short buffer arrays of pages which are required before othera considerably re-
- duces access time in readout. Consequently, it is advisable to enter directly
into the BR by means of generator G2 pages which in the immediate future must be
- removed to the centrat processor or random-access memory. Othexraise, if all
_ buffer registers are fi11ed, it is possible to use G1., as in the circuits in figs
14 and 15. Tf the requixed page is in storage registers, then j.t can fixst be
- transferred to buffex regisr.ers and then to RR-2 for readout. Furthermore, the
mean access time for these pages is not longer than i,n bubble integrated circuits
with the ordinar.y structuxe (cf. fig 14 and 15), s:tnce the BR1s and NR's operate
in the independent start-stop mode, which makes it possible to coordinate their
operation. The addi.tion of an added readout unit for RR-2 makes it possible to
113
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
reduce the mean time for access from Ws. The structure of this cixcuit (fig 21)
possesses important advantages even without G2 and Sch2 [xeadouz unit No 21.
The ordering of information in NR''s and BR's is accomplished easily in it. The
page to be reordered is removed fram fihe NR to RR-1, where it is stoxed without
movement and the i;nfoxmation fin the NR is shifted the requi,red number of steps,
- after which the page pxeviously removed is put back into the NR, erc. Similarly,
by means of RR-2, information in BTt''s can be ordered, too. Tn connection with the
increase in the capacity of the bubble integrated circuit the implementation of
two memory levels (BR's and NR's) in a single integrated circutt can serve as a
quite effective means of reducing the access time wh3ch, of course, necessitates
complication of the controller.
4) 5) ueax6) cp
' 3~ PP-2 n~
- 10 Ko t
- i>
12>
~NP HP ~NP
13)
n - ~yr
PP-J
' Pf
6~ Ue /!A
Key:
Figure 21. Use of Buffer Registers in a Bubble Integrated Circuit for
Reducing Access Time
1.
Generator curr.ent
8.
Exchange switch
2.
Generator No 2
9.
Buffex registers
3.
Input unit No 2
10.
Buffer register cuxrent
4.
Passive annihilator
11.
Exchange curxpnt
5.
Readout unit No 1
12.
Storage registex current
6.
Output voltage
13.
Storage registers
7.
Universal switch
Conclusion
The above reflects the xesul.ts of on1.y the i'irat stage i,n the devel.opment of
bubble integrated circuits and memories based on them. Already now the feasibility
is clearly obvious of extensively using bubble memories in instrument making,
control system.g and computer equipme.nt, especiall}r taking into account the real
114
FOR nFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FbR OFFiCIAL USE ONLY
prospects for xeducing the unit cost, improving the speed o� xespvnse and reli-
ability and Pxpanding the Punctiona1 capabilities of bukrble memories.
- The highest technical and economtc result from the use of bubble memories can be
achteved only if the potential capabili:ties and disfiinctiv@ features of these
_ memories axe taken into account as early as the stage of designing a new tnstru-
ment, unit, comput2r or sygtem.
At the present time the most promising areas of appllcation of bubble memories are
the following: the bulk storage of mtcrocomputers and micxoprocessor systems;
a relatively high-speed computer buffer storage between fihe random-access memory
and main external storage; the bulk storage of dispersed computing, control and
data processing systems, e.g., a terminal storage; reliable and compact storage
of modified control progra:ns, e.g., in systems for the numerical program control
of machine tools; and measuring and testing instruments with mathematical process-
ing of ineasurement results.
In the next few years must be expected the appearance of bubble devices which
accomplish not only the storage but also the processing of information, including
for performing the functions indicated in the introduction.
Bibliography
1. Vonsovskiy, S.V. "Magnetizm" [Magnetism], Moscow, Nauka, 1971, p 1031.
- 2. Minnick, R.C. et al. "Magnetic Bubble Compufier Systems," AFIPS PROC., Vol
_ 41, pp 1279-1298.
3. Naden, R.A. "One-Chip Stored-Program Magnetic Bubble Processor" in "Proc.
IEEF National Aerospace and Electronic Conf.," 1974, p 42.
4. Balbashov, A.M. and Chervonenkis, A.Ya. "Magnitnyye materialy dlya mikro-
elektroniki" [Magnetic Materials for Microelectronics], Moscow, Energiya,
1979, p 216.
- 5. Boyarchenkov, M.A., Prokhorcv, N.L., Rayev, V.K. and Rozental', Yu.D.
"Magnitnyye domennyye logischekiye i zapominayushchiye ustraystva" [Magnetic
Domain Logic and Memory Units], Moscow, Energiya, 1974, p 176.
6. Bobek, E. and Della TorrP, E. "Tsilindricheskiye magnitnyye domeny" [Magnetic
Bubbles], Moacow, Energiya, 1977, p 192.
7. 0'Dell, T. "Magnitnyye domeny vysokoy podvizhnosti" [High-Mobility Magnetic
Domains], Moscow, Mir, 1978, p 200.
8. Bar'yakhtar, V.G. "Magnetic Bubbles," USPEKHT FTZICHESKZKH NAUK, Vo1 121,
No 4, 1977, pp 593-628.
9. Lisovskiy, F.V. "k'izika tstlindxicheskikh magnitnyye domenov" [Physics of
Magnetic Bubbles], Moscow, Sovetskoye Radio, 1979, p 192.
_ 115
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFIC[AL USE ONLY
10. Thiele, A.A. et al. "xhe Energy and General Tranalation Foxce of
Cylindrica7. Magnetic Domains," BSTJ, Vo7. 54, No 3, 1971, rp 711-724.
11. Brown, B.R. et al. "High-Density Devices Using Percualloy Propagation of
Wa11-Coded Bubbles," IEEE TR,ANS. ON MAGNETTCS, Vo1 15, No 6, 1.979, pp 1501-
1506.
12. Diidorov, V:N., Randoshkin, V.V. and Telesnin, R.V. "Synthesis and Phystcal
Properties of Single-Crystal F3.1ms of Rare Earth Ferrite Garnets," USPEKHT
FIZICHESKIKH NAUK, Vo1 122, No 8, 1977, pp 253-293.
13. Doyle, W.D. et al. "A Comparison of Frequency Limited Bubble Device Per-
formance and Material Characteristics," IEEE TRANS. ON MAGNETICS, Vol 14,
1978, pp 303-305.
14. Breed, D.J. et al. "New Bubble Materials with High Peak Velocity," IEEE
TRANS. ON MAGNETICS, Vol 13, No�5, 1977, pp 1087-1094.
15. Makino, H. and Hibya, T. "Comparison of Eight Mixed Garnet.Systems for
7 1im Period Bubble Devices," J. APPL. PHYS., Vol 50, No 11, 1979, pp 7815-
7817. .
16. Kestigian, M. et al. "Past, Present and Future Small Bubble Diameter
Magnetic Memory Materials," J. APPT.. PHYS., Vol 50, No 3, 1979, pp 2161-2166.
17. Blank, S.L. et al. "Design and Development of Single-Layer Ionimplantable
Smaii Buuble Materials for Magnetic Bubble Devices," J. APPL. PHYS., Vol 50,
No 3, 1979, pp 2155-2160.
18. Dimyan; M.Y. and Hubbell, W.C. "Cmmpari.son Between the Propagation Ma.rgins
of Asymmetrical Chevron and Half-Disk Circuits at 100 kHz," J. APPI,. PHYS.,
Vo1 49, No 3, 1978, pp 1900-1907.
19. I1'yashenko, E.I. et al. "Optoelectrical Study of Bubble Propagation in
Field Access Devices," IEEE TRANS. MAGNETTCS, Vol 15, No 3, 1978, pp 1120-
1123.
20. Lomov, L.S., Parinov, Ye.P. and Chirkin, G.K. "Magnetic Bubble Integrated
Circuite," ELEKTRONNAYA PROMYSHLENNOST', No 6(60), 1977, pp 65-75.
21. Yoahimi, K. et al. "Design and Characterization of 3 um Bubble Memory Chips
Using Patterns," J. APPL. PHYS., Vol 49, No 3, 1978, pp 1918-1923.
22. Orihar.a, S. et al. "An 8 um Period Bubble Memory Device wtth Relaxed
Function Designs," IEEE TRANS. MAGNETICS, Vo1 15, No 6, 1979, pp 1692-1696.
23. Bullock, D.C. et al. "Aesign and Fabrication of Large Capacity Bubble
Memory Devices," IEEE TRANS. MAGNETTCS, Vol 15, No 6, 1979, pp 1697-1702.
24. Hayes, D.J. "Single-Chip Bubble Memory Packaging," TEEE TRANS. MAGNETICS,
Vol 15, No 6, 1979, pp 1901-1903.
116
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500020034-7
- FOR OFFICIAL USE ONLY
25. Yamigishi, K. et al. "A New Proposal on Field Access Bubble Drives,"
IEEE TRANS. MAGNETTCS, Vo1 11, No 1, 1975, pp 16-20.
26. Bozyhard, P.I. "A Novel Major/Minor Loop Memorq Chip Organization for
High Performance Without Block Replir_atfon," J. APPL. PHYS., Vo1 50, No 3,
1979, pp 2213-2215.
27. Williams, J.E. "Magnetic Bubble Memory in Telephone Syatetas" in "Proc.
Conf. 'Electro'," 1977, pp 12-13.
28. Yamagishi, K. "The Pragress of Magnetic Bubble Development in Japan" ia
"Proc. 3rd USA-Japan Computer Conf.," 1978.
29. Becker, F.J. et al. "An 8 X 105 Bit Bubble Memory Cell for Spacecraft
Applications" in "Digests 1980 Zntermag. Conf., N.Y.," 1980, pp 12-2 [as
published]. 30. Ypma, J.E., Gergis, T.S. and Archer, J.L. "64K Fast Access Chip Design,"
PROC. AIP, Vo1 29, 1976, gp 51-53.
- 31. "Single-Chip Magnetic Bubble Memoriea with a 250-kBit Capacity," ELEKTRONIKA,
No 17, 1978, pp 3-5.
32. Dzhordzh, P.K. and Reyling, Dzh, "Magnetic, Bubble Memories--State of the
Art and Development Prospects," EI,EKTRONTKA, No 16, 1979, pp 23-26.
33. Meviti, U.K. "Use of Bubble Memories in a Bu1k Storage," ELEKTRONIKA, No 7,
1979, pp 36-43.
34. Koks, Dzh. "Series of Three Bubble Memories with High Packaging Density
and a Complete Set of Interface Integrated Circuits," ELEKTRONIKA, No 23,
1979, pp 41-49.
35. Brayson, D. et al. "Set of Large-Scale Integrated Circuits for Joint Use
with Bubble Memories," ELEKTRONIKA, No 9, 1979, pp 23-32.
36. "Fujitsu Subble Memory" (firm's catalogue), 1979.
37. Chester, M. "Magnetic Bubble Memory Update," COMPUTER DESIGN, No 5, 1980,
- pp 232-236.
38. Stermer, R.L. "Bubble Memories for Spacecraft Data Recordera" in "Digests
Intermag Conf.," Cal., 1977.
39. Kartsev, M.A. "Arkhitektura tsifrovykh vychislitel'nykh mashin" [Architecture
of Digital Computers], Moscow, Nauka, 1978, p 296,
40. Kryder, M.N. "Magnetic Bubble Device Scaling and Density Limits," IEEE
TRANS. MAGNETICS, Vol 15, No 3, 1979, pp 1009-1016.
117
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504020034-7
FOR OFFICIAL USE ONLY
41. Milchev, Yu.K. and Chirkin, G.K. "Ion-zmp].anted Structure for Propagation
_ of Cylindrical iriagnetic Domains," ZARUBEZHNAXA RAAIOELEKTRONIKA, No 12,
- 1978, pp 68-69.
42. Nelson, T.J. et a1. "Design of Bubb1e Device Elements Employing Ton-
Implanted Propagation Patterns," BELL SYSTEM TECHN. J., Vol 59, No 2, 1980,
PP 229-257.
- 43. Argyle, B.E. and De Luca, J.C. "Bubble Automotion Channels by Laser Anneal-
ing" in "Digests Intermag Conf., N.Y.," 1980, p 27-2.
44. Ahn, K.Y. et al. "Fabrication of SLM Conductor-First 2 um Bubble Devices,"
IEEE TRANS. MAGNETICS, Vol 15, No 5, 1979, pp 1333-1335.
45. Dekker, E.H.L.G. et al. "Conductor and Transfer-Gate Performance of Single-
Mask Field-Current Access Magnetic-Bubble Devices," J. APPL. PHYS., Vol 49,
No 3, 1978, pp 1927-1929.
46. Ahn, K.Y. and Kane, S.M. "Toward a Single Mask Processing of Ionimplanted
_ Bubble Devices," IFEE TRANS. MAGNETICS, Vol 15, No 6, 1979, pp 1648-1650.
47. Zebrowski, J.J. and Humpherey, F.B. "Dynamic Behavior of Domain 'Wa11s in
Double Layer Self-Biasing Garnet Fi1ms," TEEE TRANS. MA,GNETICS, VoZ 15, No 6,
1979, pp 1915-1921.
48. Bobeck, A.H. et al. "Current -Ac ces s Magnetic Bubble Circuits," BELL SXSTEM
TECH. J., Vo1 58, No 6, 1979, pp 1453-1540.
49. Boyarchenkov, M.A., Vasi1'yeva, N.P. and Rozental', Yu.D. "Logicheskiye
ustroystva na magnitnykh sredakh s upravlyayemym dvizheniyem domenov" [Logic
Devices Employing Magnetic Media with Controlled Motion of Domains], Moscow,
Energiya, 1978, p 160.
50. Teyerman, V.A. and Titov, G.V. "Circuits with Selective Accessing of In-
f ormation for Magnetic Bubble Memories," ZARUBEZHNAYA ELEKTRONNAYA TEKHNIKA,
No 16, 1979, pp 3-45.
51. Chang, H. "Major Activity in Magnetic`Bubble Technology," COMPUTER DESIGN,
No 11, 1979, pp 117-125.
52. Lee. S.Y. and Chang, H. "Associative-Search Bubble Devices for Content-
Addresaible Memory and Array Logic," IEEF TRANS. COMPUTERS, Vol 28, No 9,
1979, pp 627-637.
53. Bongiovanni,'G.C. and Luccio, F. "Permutation of Data Blocke in a Bubble
Memory," COMMITN. ACM, Vol 22, No 1, 1979, pp 21-25.
118
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
54. Lin, W.S. and Ji,no, M. "Intelligent Magneti,c Bubble Memori:es and Their
Applications in Data Base Management Systems," TEEE TRANS, COMPUTERS, Vo1 28,
No 12, 1979, pp 888-906.
COPYRIGHT: Tsentral'nyy nauchno--issledovatel'skiy institut informatsii i tekhniko-
ekonomichaskikh issledovaniy priborostroyeniya, aredstv avtoma.tizatsii i sistem
upravleniya (TsNIITETpriborostroyeniya), 1981
8831
CSO: 1860/59
I
119
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R440500020034-7
FOR OFFICIAL USE ONLY
11
UDC 681.355 (088.8)
DEVICE FOR TAKING MEDIAN VALUE
Moscow OPISANIYE IZOBRETENIYA- 702381 in Russian 7 Dec 79
[Description of USSR Patent No 702381 by F.I. Kerbnikov, N.E. Mengazetdinov
- and M.A. Rozenblat, filed 6 Jun 78, published 7 Dec 79, class G 06 G 7/12]
[Text] This invention is in the area of automation, computing and measurement
technology and can be used in high reliahility, high precision systems for meas-
uring and transmitting analog signals.
There exist devices which take the median value [1].which extract the median from
an ordered sequence of input signals. These devices contain an odd number of
- channels operating in parallel whose outputs are combined and connected to a com-
mon load.
The closzst known madian-extracting device to the proposed invention i:s that des-
cribed in [2], which has in each nonlinear conversion channel an amplifier which
is connected through a current limiter to the load and which has comaon negative
feedback.
In tihis device, the amplifier in each channel can be.in the linear region under
the condition !
l(uin;. Ubut)ki Ic Eo , .
(1)
where Uini voltage at input of ith channel;
uOL1t~NeAUlII; UoLlt)"",
- (U:tn; Uoub~lc 4,.:.~ (2)
CVin~, Uout)Kzn.4}'
- device output voltage;
Ki - gain of shorted amplifier of ith channel;
Eo - size of linear zone of current limiter;
2n+1 - number of channels operating in parallel.
When condition (1) is not met, the amplifiers in the nonlinear conversion r
channels are saturated. In practice, when iZ � 1, only one channel is
120
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
operating in the linear region, while the ampliftexs in the other channels are
saturated. The saturated condition of the amplifters has no-.effect when
dc signals are being converted, since there ts always at least one amplifier
in the linear state.
When processing time-varying input signals (eapecially if the s ignals are
close to one another), condition (1) is necessarily violated for the channel
in the linear condition due to changes fin frequency and level of input signal
and variation in K from amplifier to amplifier. This causes the cliannels to
be switched abruptly. When this happens, the amplifier which had been in the
linear region is� saturated, and an amplifier in another channel for which
condition (1) is� met must }ie swttched to the linear condition. However,
there is a delay involved in the latter amplifier exiting the s aturated con-
dition. The time difference between the entry of the amplifier into the
saturated condition and its extt from that condition cauaes excursions of
- the voltage at the output oF. the reproducing device, the presence of which
constricts significantly the frequency response of undistorted signal
transmission.
The purpose of the present inventlon is to increase the speed of the device
and expand the operating frequency band:
_ This goal is achieved by adding a smooth:tng capacitor in parallel with the
current limiter to the medtan-extractian device, which contains 2n:+ 1
channels, each of which consists of a series-connected tnput s caling resistor,
amplifier and current limiter, the output of which is connected to the output
of the device and also to the input of the amplifier through a scaling
feedback register.
- The .drawing shows the functional diagram of the device for extracting the
median value.
Each channel of the device contains input scaling resistor 1, amplifier 2,
current limiter 3, scaling feedback resister 4 and smoothing capacitor 5.
Amplifier 2 in each channel is connected to the common load through a
parallel network (capacitor 5- current limiter 3).
The device operates as follows.
When channels are switched, the change in the voltage at the output of the
amplifier which had been in the linear region. passes through smoothing
capacitor 5 and resister 4 to the input of that amplifter, thus causing
it to take longer for the amplifier to become saturated. When the times
required for the amplifier to reach the saturated state and to exit that
state are the same (which is easily achieved by selecting the value of
capacitor 5), the distortions at the output of the device are fully eliminatEd,
and the operating frequency band fis increased to the bandwidth of a single
amplifisr used in its charinels.
121
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
The use of the capacitor makes fit possible to expand the operating frequency
band maximally (up to the bandw3dth of the amplifiers used) at minimum cost
(one additional element per channel), while the operating frequency band of
the prototype is narrower than the bandwidths of the amplifiers used in it
by a factor of 2-4.
Patent Claims
A device for extracting a median value, consisting of 2n + 1 parallel channels, each of which consists of series-connected input scaling resister, amplifier
and current limiter, the output of which is connected to the output of the
device and also through feedback scaling resistor to the input of the amplifier,
distinguished by the fact that tn order to increase speed and expand operating
~ bandwidth, each channel contains a smoothing capacitor connected in parallel
to the current lintiter.
Information Sources
Used in Evaluation
1. Gil'bo, Ye.P., Chelpakov, I.B. "Obrabotka signalov na osnove uporyadochennogo
vybora" [Signal Processing Based on Ordered Selection]. Moscow, Sovetskoye radio,
1975. .
2, UK Patent No 1362378, class NKI G 4 G, published 1974 (.prototype).
Cnr
ut
uint
uiri2a.-~
.L
6900
CSO: 8144/0218-B
122
FOR OFFTCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
MICROPROCESSOR PSYCHODZAGNOSTIC DEVICE PATENT
Moscow OTKRYTIYA, IZOBF.ETENIYA, PROMYSHLENNYYE ORRAZTSY, TOVARNYYE ZNAKI in Russian No 29, 1981 (s.i.gned to press 10 Jul 81) p 296
[Text] Patent No 11032
- Claim Nc, 20122 ' .
" Class 24-02
Microprocessor Psychodiagnostic Device
Authors: F. I. Romanov, N. V..Danilin, N. N. Zubov, P. K. Mal,yutin, V. A. Merkulov,
A. V. Starshikov and M. A. Khadzhiakhmetov
~
~ Priority.as of 1 August 1979 �
C(1PYRIGHT: VNIIPI, 1981
CSO: 1860/80-?
123
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
CERTAIN ASPECTS OF PHOTOGRAPHY,
MOTION PICTURES AND TELEVISION
UDC 621.397.2
TV COIvTOUFt D'TiGTOR
Moscow TII{HNIKA KINO I TII,EVIDFNIYA in Russian No 9, Sep 81 pp 40-43
[Article by A. V. Anan'in. T. S. Gryaznova and M. N. Stinov, Institute of
Electrical Engineering of Communications. Novosibirsk]
[Text] Gne way of increasing the transit capacity of TV communications channels
is to use adaptive TV aystems capable, (depending upon the sub3ect of the trans-
mitted picture, dimensions of picture parts, the presence ar absence of movements
etc.), of controlling the TV channel parameters automatically. This makes it
possible to ' eliminate red.undancy in the transmitted data and multiplex the com-
municaticns channel with additionaY data. Contour and motion detectors may serve
as picture analyzers of auch adaptive syetems. The first are uaed to evaluate
the contour dimensions of the picture parts and their complexity (presence of
bends, sharp angles) and, therefore, the necessary width of the tranamission band
of the TV channel; the second are used to evalua.te the speed of movement of the
picture elements and, therefore, the posaibility of auch multiplexing in time at
which there will be a hardly noticeable or no stroboscopic effect in the repro-
duced picture.
'rhis paper considers one possible vaxiation of the T'ii contour detector, 1., which
the principle of electronic-optical filtration (EOF) is used. This method is the
simplest from the technical implementation viewpoint and makes possible its wide
use of standard TV equipment in the device.
As ia wel7.?rnown, the statistical distributlon of contour elements of real pic-
tures have maxima in the vertical and horizontal directionsi therefore, the prob-
lem of separating full contours can be considered solved to a great degree of
authenticity if the problem of separating the vertical and horizontal boundaxies
of the contours is solved. The proposed device uses the filtration properties
of the electron beam of an electron beam tube which has a certain aperture [1p2p3].
As shown in [w], the aimplest filter capable of solving the problem of sepaxating
contour boundaries is electron-optical shading (shading focus).
The spectral function of energy diatribution i.n the shading aperture of the beam
for a normal law may be described by expression [4]
124
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFFICIAL USE ONZX
~2 0)2
P(x, y) = 2. eXP - 4~ 4~2 ~ (1)
Khere ei and N are respectively horizontal and vertical dimensions of the shading.
As follows from (1), depending upon the orientation of the shading with respect to
- the scanning direction, it ma.y fulfill the funetion of the lower 0. as
well as the upper spatial frequencies 61 2)~ The EUF implementation process
is shown in Fig. 1.
The shading aperture of the electron beam of receivin electron-beam tube'can be
formed by a single or double quadrupole ma.gnetic lena~51, placed in the neck of '
the tube where there is no deflecting field.
The orientation of the stiading from frame to frame is determined by the direction
_ of the current in the coila of the quadrupole lens. If, in the first frame, the
shading is oriented normally to the scanning direction and the horizontal boundar-
ies of the contours are separated, then in the second frame, with a horizontal
orientation of the shading, the vertical boundaries of the contours are separated.
s
2
nrrFrirm
>
Fig. 1. lmplementation of electron-optical filtration in the TV receiving tube.
1-- initial picture; 2, 3-- for polarity inversion of ad3acent lines and
elements and undistorted beam aperturej 4, 5-- at EOF output of lower and upper
boundaries.
125
FOR QFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R440500020034-7
FOR OFFICIAL USE ONLX
t z
.1L. �
CC!! TU)
_F1_
xcu (b)
>s
u~ u2 vi u .
y f 6
US
~ U6 117
8
U1I 9 lo
r----OC/1 -,Beix.
1f f2 i 1~ >4 isudeo
.NL crH
ve u
J6 >7 18 /9
?0 f ---H 2 f
_ Fi~. 2. .~tructural arrangement of a device for separating contotu~s of optical
ob,jects. a. S5I d . OSP
b. K3I e. Yideo output
c. SGI
A two-gradation operating mode is used in the device in which the output signal
can assume only two valuest "1" (white level) and "0" (black level). Reducing
the signal to two grades (limiting the video signal output of the camera) im-
proves the signalboise ratio and eliminates the nonuniformity of the video aignal.
- Moreover, processing the video signal in the digital form is simplified consider-
ably. The block diagram of the device is shown in Fig. 2, and Fig. 3 shows the
signal time diagrams at corresponding points of Fig. 2.
Video monitoring device 13 and TV transmitting tube camera form an optically
coupled pair (CSP).
The vertical boundary signal is formed as follows.
The output signal from TV camera 1 is sent to video signal processing unit 2 in
which the level of black is established. F-rom the processing unit the signal ie
sent to amplifier-limiter 3 which reduces it to the two-grada.tion form. To im-
prove the tranaition boundaxies from black to white and back, the output signal
of the amplifier-limiter is sent to forming device 4. To make possible the transi-
tion from black to white and back, it is necessary to have direct, as well as in-
verse video signals for which purpose inverter 5 is introduced in the circuit.
Video signals are sent from the forming device to the strobing pulae former 6
that forms the pulses into decompoaition elements t3n long, with the atart of these
pulses coinciding with the boundaxies of changes in the brightnesa of the picture.
126
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-40850R040500024034-7
FOR OFFICIAL USE ONLX

c
a
0
0
0
0
1V( - OOP'02f1 O2QQMUYON(lN ~i~

t
UI ~

i
I -r
t
uJ
, 
t
U4
~
I
I
uS
tm (2)
~6 I
t i
~ ~
i
I
~
~
C
u7
e
e , i
r ~
- h"
,
~
,
~
h
i
r
Fig. 3. 'rlme diagrams of the separation procesa of vertical Up and horizontal UP
contour boundaries
1. Threshold boundaries 3. Tssc
2. . t.3J1 4� Tks i
127
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R400500020034-7
FOR OFFICIAL USE ONLX
To obtain element-by-elem.ent switching, an impact excitation oscillator 7SSI in The
troduced in the circuit ~rhich is synchronized bY theitimeaofsdecompositiCn element
A oscillation periad of the oscillator is equal to
t3!I . From the impact excitatlon oscillator the signal is sent tualrtogtwo de-d
we have symmetrical pulses at the trigger output with a perio eq
is Practically
composition elements 2tj~ , i.e�, ari element-by-element awitching
implemented.
then
If the initial picture contains fh8trobing pulse formerf6tthatrdisconnectsoimpact
a signal appears at the output o
excitation oscillator 7 for the duration of the decomposition element� In this
case, at the trigger 8 output, two ad~~obinetarri ese hlfithetinitial pice
the same polarity at the moment the s g pulse then there is no osc ture does not contain the vertical boundary of brightness drop, strclbing signal, the oscillation mode of Impact ofieach horizontal
change and this means that the si~als of all jacent
line have different polarities.
The forma.tion of the horizontal boundaxy signal occurs as follows. The video sig-
nal from procesaing unit 2 is sent to ampl8f~hroughiforming deviceh ?ge~d inverte
two-gradation form. Then the signal passe
er 18. The dlrect and inverse video signal2t prOf,ro~triggercl5~ninitiated by con
trolled by a symmei:rical pulse with per d Z
horizontal sync pulse (SSI). As a result, direct and inveri ict~urescontains
If the initia P
a apr
pear at the output of bthe ri electronic signal polarities of two adjacent lines
h~rizontal boundary ~
coincide where the drops occur. When there is no horizortal boundaxy of brig -
ness drop in the initial picture, the video signal polarities of adjacent lines
are different. .
- be obtained by sepa-
As noted preuiously, the information on the full contour maY
rating from the video signal one informa.ti~ie$ameFo �thishpurPoaecathe o tputes
19
and another one about the horizontal bound
signals f.rom element-by-element switchin~rlled b 8gymmetricallpulsesiwithiperiod
are sent to line electronic switch 9 con y reamplifier 10
2Tk from trigger 11. The awitch output si~al li~e ySGI are mixed into
and is then sent to nonlinear adder 12 where quenchin6 Pu mon it. The video signal formed in thia manieer~ igr ~~P~~viThe OSPiinrthi$dcase
t3 which is a component part of the co p 1ai
fulfills two basic functions: first, it provides for the output of information
on the centours in the form of electrical signals, and secondly, due to the inertia
of the transmittina tube of OSP camera l4,the efrom fra e tofframeiare t ans-
horizontal an d ve r t i c a l b o u n d a r i e s o f t h e contours
, formed into continouous information.
The current direction in the quadru ole ma~etic lens 21 coils, placed in the neck
of the VKU [Video monitoring devicelC)5P kinescope, must change from frame to
= frame. This is achieved by a devica for controlling quadruPole lens 20, done by
128
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504020034-7
FOR OEFICIAL USE ONLY
voltage-current converter [6] controlled by symmetrical pulses with a period of
2Tk produced by trigger 11.
A characteristic feature of the device is that, depending upon the spatial dis-
pogition of the initial picture elements, the contour boundaries observed on the
VKU OSP screen may be daxk or light with respect to background brightness.
_ The instantaneous brightness of the background BcP (averaging the brightness of
two ad,jacent lines or elements when scanned by the shading aperture) and of the
vertical BB and horizontal Br boundaries for double scanning of the phosphorus
elements with a high afterflow time 7J7 is evaluated by expressionss
8cb=0,b B. exp (-t(Tn), -Be-Br- Bmexp (-t/Tn),
where 'r,7 is the time constant'of the phosphorus glow of the VI(U kinescope= there-
fore, the brightness of light boundaries is found to be txice that of the back-
ground, while the brightness of the dark boundaries is equal to zero (BT=O), when
the modulation characteristic of the kinescope is fully utilized. In this case,
the output signal of camers CISP is found to have three gradations (m 9bix =3)�
c;Rne of the basic parameters of the device is the contrast sensitivity which may be
evaluated by the minimum signal/noise ratio that provides for a reliable separa-
tion of the optical object contours. For a small number of gradations, the fol-
_ lowing expression [71 is trueo
- V'ev: = mav: (m.u: -f- (2)
km
here km= coefficient that takes into account the reliability of detecting the
boundary signal.
Although (2) is true for linear systems in xhich there are no gradation distortions
=1), it ma.y be used to some approxima.tion at the given case because'Y~P _ 1.
Value km takes into account the presence of noises depending, as well as not de-
pending, on the signal; therefore [7]s
lk
v2 G km C 1/2 k,
where lower boundary km is set for systems in which there are no noises related
directly to the signal (only external noises are present), while upper boundary k
is for systems in which the noises are related directly to the signal. For
estima,ting calculations, it may be assumed that km=k.
129
FOR OFFTCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R440500020034-7
FOR OFFICIAL USE ONLY
Coefficient k determines the probability of detecting the signal of boundary pc
In agiven interval of the tSP output signal. When relating k and pc , determined
by the proba.bility interval, assuming pc =1-10-5 (error probability of 10-5 is
entirely acceptable for machine data processing of contours) we will find k N 4.42.
According to (2), In this case
The obtained value*uz must be fulfilled for the highest frequency components of
the boundaxy signal, in this case for the signal of vertical boundariea that have
frequencies of fB=b.S mega.Hz of the standard TV channel. If a vidicon transmit-
ting camera is used In the OSP, the basic factors influencing the +Vsu: will be the
noises of the output preamplifier stage (vidicon noises are insignificant and may
be neglected), as xell as aperture distortiona of the vidicon. The first deter-
mine the value of the equivalent noise current at the preamplifier input and the
second the reduction in the value of the signal current due to the aperture-
frequency characteristic of the vidicon.
By using a simple antinoise correction and a field effect transistor at the input
stage, the value af the equivalent noiae current at the preamplifier input can be
obtained by expression [81i
r I
m - 4R fWn -F' ewn R RQ2 df. (3)
l
L
_ where i~n- equivalent quadratic mean of noiae current of the field effect transis-
tor; 1 Wn equivalent quadratie means of noise emf of the field effect
transistor; R-- resistance of vidicon load; C-- capacitance at amplifier input
which includes the capacitance of the wiring, the input capacitance of the field
effect transistor and the capacitance of the vidicon target with respect to the
transistor source.
In accordance with (3), the aperture-frequency characteristic of the vidicon for
Gaussian diatribution of energy In the electron beam
$ (ai. = exp I - r nai~ l~l .
~ 2 (4)
where a1=2r118 relative transverse dimension of the electron beami rI its
ra(i lus; s=h/z scanning pitchi q =nd/z=f/fB relative scanning frequency.
Using (3) a,nd (4), the following expreasion can be used for the signal/noise ratioi
130
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLX
ic exp r nal~t ~2 _ _ . _
2
+ 4kT 2 e~. 4 � (5)
C R+' l mn R9 1f -I- 3 o[2 C2l~,f 3 ]AId
7
Fig. 4. Relationship between the signal/noise ratio of high freq,uency componenta
of the contowc boundary signal and transmission bandxidth of the device.
1. Vidicon LI415 3� f(megaHz)
2. Vidicon LI418 Fig. 4 ghowa the relationship between the signal/noise ratio of high frequency
comgonents of the contour boundaxy signal and the tranemission band of the device
- designed on the ba.sis of formula (5). In calculating curves in Fig. 4, the fol-
lowing paxameter values were useds 1c=0.2-microampa= C=20 picofaradsi R=1 megohmi
the field effect tranaistor was xP3o3v(iw,,= 0.5x10-16 p; emin = 5X10-9 B).
~z VTz
Tn the operating mode, taking into account the effect of reading the front ed.ge of
t.hp beam and for standard raater dimensions R,= 0.8 for L1415 and Qi'= 0.6 for
LI418. ,
It may be seen from Fig. 4 ttiat for fB=6.5 megaHz values ."y =7 and Y =14 do not
reach the neceasary value of ~=15. However, if the transmission band is
131
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
' 1 1 4 5 6

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000500020034-7
FOR OFFICIAL USE ONLX
reduced to 6.2 megaH2 (quality indicators of the device will be practically pre-
served), it is possible to reach the required values of owX for the OSP camera
usinG the LI418 vidicon. lt followa from what was stated above tha,t when using a
devtce for separating contours in adaptive TV communications channels, higher
resolution vidicons should be used in the OSP camera.
An experimental model of a device for sepaxalving contours made with an axrangement
as per Fig. 2 had the following chaxacteristics.
Swing of the output signal'boundaries with load P=150 ohms, not less tha,n 0.5
volts.
Signal/noise ratio for vertical boundary of contour not less than 18 (vidicon
L1418 used in the USP camers).
3win,;s of control signals not less than 1.5 volts.
'I'he investigations ma.de confirmed the possibility of using the methods of electron-
optlcal filtration in detectors of full contours of optical objecta, which make it
posslble to create them on the basis of adaptive TV communications channels.
In conclusion, it must be noted that the method described for separating contours
by the eiectron-optic filtration method may be implemented by two arrangements.
In the first, electron-optical filtration is used directly in the TV tranamittirig
tube as considered abnve. In the second case, it is necessary to have a apatial
coupling between the basic TV camera and the camera for contoux separation.
BIBLIOGRAPHY
1. Maxkovich, M. G.; 01'khovitskiy, L. A.; Tsukkerman, I. I." IIectron-(7ptical
Filtration contours." TEXHNIKA KINO I TELEVIDIIdIYA. 1965, No 7, pp 41-44.
2. Eaykin, I. A. "Spatial Filtration of Pictures by Simultaneaus 3canning with
Apertures of Different Dimensions and Shapes." Froblems of Radioelectronics,
series "Tekhnika televideniya." 1960, No 4.
3. Mezhov, F. D,; Noshchenko, V.'S.; Seredinakiy, A. V. "Two-Dimensional Filter
for Suppressing Interferences in TV Pictures." Problems on Radioelectronics,
series "Tekhnika televidentya." 96$, No 1.
4. Garelik, S. L.; Katz, B. M. "F,Lectron Beam Tubes in Data Proceasing Systems."
Moscow, "Ehergiya, " 1977�
5. Markovich, M. G. "Short Double, Four-, Six-, Eight-Pole Lenses as Aberration
Correctors When Deflecting IIectron Beams." ZhTF [Journal of Technical Physics],'
1972, No 1, PP 42-47.
6. Katayev, S. I.; Uvaxov, N. B. "Yoltage to Current Conversion for II,T [r3ec-
tron Beam Tube]. TII{HNtKA KINU I TII,EVIDFNIYA, 1974, No 3o Pp 55-57�
132
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500020034-7
FOR OEFICIAL USE ONLX
- 7. Curevich, S. B. "&ficiency and Sensitivity of TY Systems." Leningrad,
"aiergiya, " 1964.
8. Gershberg, A. Ye. "Transmitting TV Tubes wi�th Internal Photoeffect." Leningrad,
"a4iergiya, " 1973.
9. Ryftin, Ya. A. "TV System," Moacow, "Sov, radio," 1967.
COFYRIGHPi "Tekhnika kino i televideniya", 1981
- 2291
Cso1 1860/77
133
FOR OFFICIAL USE ONLX
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500020034-7
FOR OFFICIAL USE ONLY
- unc 621.397�13
PA.RA1i~.~TER SELECTIUN OF OPTICAL SYSTEMS FOR SCANNING REPRODUCING JEVICES WITH
KINFSCJPr.'
Moscow T&I-INIKA KINU I TII,EVIDN3NIYA in Russian No 9, Sep 81- pp 37-39
_ [Article by V. S. Babenko and A. I. Kravets, Institute of Civil Aviation
Engineers. Kiev3
[Text] 3canning systems (TVS) include systems with comparatively large scanning
anr;les along the horizontal (100� and greater). The scanning zone is increased
and the efficiency of spatlal perception is ratsed in acanning TVS which is es-
peclally important in sevoral TV applications. In developing scanning T V systems
it is necessary to solve a number of complicated problems related to the develop-
ment of the general principles of scanning TV and the crea+.!on of scanning
transmitting cameras, a communications channel and reproducing scanning devices
(PW).
_ In spite of the abundance of vaxious projects, the cloaest to realization axe
scanning TV systems of the multichannel type. Multichannel PVU are used in such
systems for the reproduction of pictures. The general visual field in such PVU
is. divi'ded optically in three parts a.nd is transmitted over three paxallel chan-
nels. The scanned picture is reproduced either on a cylindrical screen by means
of three TV projectors, or by several kinescopes. The basic shortcomings of pro-
- jection PVU are awkwardness, complexity and high cost. The shortcoming of PVU
using kinescopes is the relatively short observation distance and the presence of
visible seams in the picture at the 3unction zone of the screens. Multichannel
PVU with colltmation system kinescopes are free from such shortcomings. Such
systems, acting as ma.gnifiers, make it kossible to separate the pictures from the
kinescope screens and transfer them comparatively long distancea. In this case,
parameters of the system may be selected ao that there are no seams at the bound-
aries of the partial pictures. This improves the reproduction characteriatics of
- space sharply and increases the scanning effect of the PW.
- We will consider the basic relationshipa for the optical part of this type of PVU
and will determine the conditiona that its paxameters must satisfy. We consider
the dimensions and other.,parameters. of partial pictures in the kinescopes and the
collima,tion systams equal, while the observation angles of the partial pictures
are considered.comparatively small.
134
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLX
ThP basic arrangement nf the three-channel projection syatem (top viek) of the PW,
satisfying these conditions is shown in Fig. 1.
~ A
,r------= - -----y..
A~%~~� i / \ `~li
A`
, D~ i
Fig. 1. Basic projection system (top view) for a three-channel PW.
Partial pictures are formed on kinescope screens AV A2 and A3. Theae pictures
are viewed by the observer from point 0 through lenses L1, L2 and L3. The vi.rtual
images A1, A2 and A3 viewed by the observer are amplified and carried a compara-
tively large distance. With a correct choice of the parameters of the projection
system, the partial virtual pictureB are found to be joined.
The course of beams in the ohannel for forming one partial picture is shawn in
Fig. 2. In calculating the dimenaione of the projection system, it is necessaxy
to take into account the conditions for 3oining the pictures on the kinescope
s creens and the virtual pictures; the virtual partial pictures along the horizontal;
and the collima,tion lenses along the horizontal. The basic formulas relating to
the optical parameters of an n-channelPW axe shoxn in the Table.
Depending upon the value of bn If' the following cases are poseibler
1 nIf' ) 1, i.e., the eye of the observer ia at a diatance from the lens that
~ exceeds the focal distance, and b/bn C 1;
1n /f'=1, i.e., the eye of the observer is located at the focus of the lens and
b/bn =t;
~If'< 1, i.e., the eye of the observer is located a sma,ller than focal distance
from the lens and b/bn > 1.
135
FOR OFFICIAL USE OtJLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLX
Fio. 2. Course of beams and basic values that determine the projection arrange-
_ ment of the partial picture.
The analysis of the cited expressions indicates that the amallest optical distor-
tions and the greatest boundaries for possible location of the observer's eye with
respect to the lens corresponds to the case where the observer is removed from the
lens by a distance that exceeds the focal distance. However, in this case, the
angles of the viewed partial pictures are comparatively small (2Wrl =10 to 200);
therefore, a fairly large number of kineacopea must be used in the PW. The fact
that the film should be wider than the picture on the kinescope screen should also
be considered a drawback. Of interest is the case where the observer is located
in the focal plane of the lens since conditions for joini;ig the elemants of the
optical system in this vaxiation do not depend on amplification. But such a system
may be realized only when the lens parameters satisfy relationship
bn12'f'atg Wr.
In the third case (diatance between observer and lens is less than the focal dis-
tance), Zarge viewing angles 2Wr' may be obtained with a small number of kine-
sopes. But under auch conditiona, varioua kinds of optical picture distortions
are m4re prominent and the boundariee of posaible spa.tial location of the observ-
er's eye are minima,l.
In calculating the PW projection aystem, horizontal 2Wr and vertical 2WB angles
of the pictures viewed are initial. The remaining values ase partially given and
partially determined. The following variations of dimensional calcula,tion are
136
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLX
Table
Pa,: imet ers
Primary pa,rtial
p:tctures:
xidth
height
distance between
centers of adJacent
kinescope screena
Lenses
focal distance
Width
height
Projection system
distance from lens
- to kinescope screen
distance �rom lena
ta observer
system amplification
Virtual image
widt h
height
distance from lens
Observat ions
general horizontal
viexing angle
_ horizontal angle of
partial picture vtew-
- ing
- vertical viewing an-
gle
viewing dista.nce
Formulas*
6 - [b� 2/' (v-1) ta1D'r]/V
h- Jhn+ 2i'(V-1)tgW./v
d=2(!n-{--a)sinW;
_ tn IV (b/bn) - lI/(1- V)
bn=2lntgW'
hn=2lntgW,
n=f'(V -1)lV
1=1 p - a' ~ bn/2 tg W;
V= I h I� I b/ bn (ta Y)/(L�b - bnl~)
~ / \
b'=bV=2lptgW;
h'=hV=2lptgMr
a' = (b - 21 t8 Wr)1(2 tg W; - b/I')
2Wr = 2nW;
2W,. 2arctg (h12IP)
2WB = 2 arctg (h/2/p)
(,p= ln-f-a=b'/2tgW'
I
* Values in formulas explained in Fig. 2.
- typical for the considered problems. According to the given dimension of the pic-
ture on the kinescape screen and the dimensions of the resulting picture, it is
possible to determine the parameters of the lens and the mutual location of the
elements of the projection systQm. When the dimensions of the picture on the
kineacope screen and the pa.rameters o#' the lens are given it is possible to calcu-
late the values that determine the mutual position of the elements of the projec-
tion system. Another variation of the calculatian is.posaible if the dimensions
- of the resulttng picture and of the parameters of the le:is are given. in this
137
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R440500020034-7
FOR OFFICIAL USE ONLY
case, the position of the elements of the projection syatem and the dimensions
of the pictures an the kinescope screens axe determined
Fig. 3. Nomogram for calculating the basic PW parameters.
To illustrate the nature of the relationships and to evaluate the realizability
regions, a nomogram is shown in Fig. 3 that interconnecta the basic parameters of
the projection system. This nomogram eases the PW calculations greatly. When
calculating the dimensions of the PW projecLion systemp it is necessaxy to pro-
ceed from the minimum number of channels (the nutnber of kinescopes and lenses
respectively).
When using standard kinescopes and complste filling of their screen with pictures,
ratio b/h ia t.25 to 1.33 for horizontal and 0.75 to 0.8 for the vertical position
of the wldth of the kinescope acreen. The problem of eelecting the number of
kineacopes ma.y be simplified by using the nomogram in F'Lg. 4 that shoNe the rela-
tionship between viewing angles Wr /WB and the number of kine$copes for vaxious
138
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
6
~
:
I
I �,9
I ! IP.
s5 ~
W
Po ~
~07

I
1 2 3 y
R
S
Fig. 4. Nomogram interconnecting the rela,tionahip betxeen viewing angles Wi- IWB
and the number of kinescopes at various orientations of kinescope screenst
- horizontal orientation kinescope screensi
- - - - vertical orientation. orientationa (horizontal and vertical) of their screens. It ahould be taken into
account that the number n should be only diacrete. If ratio Wr /WB produces
- fractional n, then its values must be rounded out to the neaxeat xhole number and
a smaller value of.n will mean a less full utilization of the height of the sc;reen
while a greater value a lesa full utilize.tion of the Width of the screen. The
utilization of the vertical position of the kinescope acreens for a given Wr /WB
ratio ia related to an increase in the number of kinescopes and lenaea, but makes
it posaible to increase the dimensions of the resulting picture. Utilizing
kinescopes xith large screens makes it posaible ta obtain large picturea Hith
moderate amplifications, but is related to the neceasity of uaing larger diameter
lenses.
- Single lenses produce geometrical and chromatic d.istortions, especia.lly noticeable
at the edges of the field of the picture. These distortions are greater the
greater the system amplification and tha P^�.ater the deviation of the observer's
eyes from the central position. Practicaily, the amplification should not be
greater than 3 to 5, while tha spatial position of the observer's eyes when view-
ing the pictures should. not change essentially. Geometrical and chroratic dis-
tortions may be reduced by introducing correcting elements into the optical system
whose parameters are determined as a result of corresponding -calculations. .
The distortions may be reduced considerably in a two-lens system, although the
system complexity is increased while the coefficient of the optical transmisaion
is reduced.
139
FOR OFFICIAL USE ONLY
2W=
'11
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500024434-7
FOR OFFICIAL USE ONLX
The necessity of using lenses of fairly large dimensions causes serious difficul-
ties. The use of plastic Fresnel lenses may be considered a compromise solution.
However, the use of such lenses leads to noise caused by bea,ts betxeen the TV
picture and the lens rasters. PW of the mirror type axe free from many short-
comings of the lens systems, since their basic optical elements are spheric mirrors.
For spatial spacing between kinescopes and the observer, additional semitrans-
parent mirrors are used in mirror PW that complicate the system and, what is very
essential, reduce the optical transmission coefficient. The calculation of mirror
PW dimensions has, of course, specific features although the basic principles are
the same as for lens PW.
To obtain high quality pictures, apecial measures must be taken on the mechariical,
optical and electrical matching of components and projection system parameters of
partial pictures. Single type kinescopes with ma.tched characteristics (color,
brightness and contrast) must be used in the channels. Distortiona caused by
_ noncorrespondence in position, dimenaions, linearity and partial pictures (espec-
ially along the vertical) etc. should be reduced to a minimum.
In conclusion, we will note that the principles considered above may be used to
calculate the dimensions not only of scanning, but also of wide-angle W[Repro-
ducing devices] of the mosaic type whose picture field is composed of kinescope
screens set up on both the horizontal and vertical directiona.
CUPYRIGHTs "Tekhnika kino i televideniya", 1981
2291
cs Ot 1860/7?
140
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
UDC 681.84: 621.3.037.372
MULTIMICRQPROCESSOR 3YSTEM FfaR PROCb.53ING AUDIO SIGNAL IN FREQUENCY REGION
Moscow TIICHINIKA KINO I TELEI/IDIIdIYA in Russian No 9, Sep 81 pp 24-28
[Article by Ye. I. Grinberg and M. I. Peregud, All-Union Scientific Research Motion
Picture In,atitute]
[Textl The process for producing phonograms for cinema, films is related to a
large number of conversions implemented an the initial audio signala. In the
development of equipment and methods for the digital proceeaing of audio signals
it is possible to separate a number of atagee, each of which is typical in its
_ level of treatment of computer facilities.
In the first stage, at the beginning of the seventiee, investigations of basic
types of signal conversions on universal computers began. The most important re-
-j sult of the work at this stage was the understanding of the necessity to create
' specialized devices to process signals in real time and uae Fourier transforma-
tions far the implementation of almoat all kinda of filtration of the audio signal
C37 �
~ In the second stage, work was done on creating apecialized digital devices kith
low and medium degreea of integration. Due to the high cost of implementing com-
; puter operations on the given component base, digital deilay linea, reverberators
~ and other devices using the delay line principle became the first to be used
widely in audio recording equipment [1, 4]. .
, At present, considerable experience ha,s accumulated in our country and abroad on
simulating the digital aonversion procesae$ of audio signals, and creating spe-
cialized devices to implement individual functiona.
' The realization of the ma3ority of electroacouatic conversiona (adaptive noise
= reduction, adaptive interference compensation, frequency correction) ia related
- to the calculation of a signal pa.cket with a pulse chaxacteristic of a converter.
Tnasmuch as the most efficient method is calculating packets in the frequency re-
gion, a multimicroprocessor system for processing audio aignals xas created as a
basis, using equipment to transform initial Fourier signals.
141
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109
CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
The parameters of the system were selected on the basis of work carried out in the
electroacouetic laboratory of the N DCFI [All-Union Scientific Reseaxch Motion
Picture Institute] [3]�
According to the results of the indicated worko the system should be able to do the
followingi
1. Implement audio signal conversion in th~ frequency anddrestoreZtheon
frequencies of up to 40 kHz in real time, p oces the aPectrum
signal in the time domain.
2. Implement audio signal converaion into a 16-bit binary code with a total error
not exceeding 3 to 4 low-order bits (the number of spectral components up to 1024
co mplex numbers). '
3. 1'rovide the possibility of beinp increased to provide for proceasing in real
time signals with higher discretization frequency.
4. Provide the posaibility of simultaneous conversion of several audio signals
with a 3oint processing of the spectral characteriatics.
5. Be inexpensive, small and light, highly efficient and have a single type of
universal syatem to carry out inveatigations, as well ae to produce specialized
devices on ita basis.
6. Have a design anticipating the connection of new data sources and usere with-
out changing the equipment paxt of the syatem center.
_ 7. Ha.ve a lok level of internal audio noiaes, required for operation in apecial-
ized places.
The purpose of this paper is to describe the design of the system and its compo-
nent parts for procesaing signals with the greatest diacretization frequency.
In approprfate comments, posaibilities will be indicated for reducing the system
when the discretization frequency and the number of initial signals are lowered.
The basic problem in creating the system is the implementation of direct and in-
v erse Fourier transformations of the.high speed initial signal.
The realization of the Fourier procesaor using lox and medium degrees of integra-
tion circuits leads to very substantial expenditurea for microcircuits and cur-
rent consumption and, as a result, to a high cost of eq,uipment, large size, low
reliability etc.
In recent years, the appearance of domastic microprocesaors made it poasible to
design economical Fourier processora. The aeries K589 microprocesaor set was
selected as the basic ona.
The given set has the highest speed of opera.tion among the microprocessors being
manufactured.
142
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLX
The microprogram organiza,tion makes it possible to control, by expanded axithmetic,
external circuita simultaneously with the operation implementation in the procea-
sor. The indicated advantages determine practically a one-to-one choice of
series K589 series microproceasors for the Fourier analysis of the audio signals.
The general structure of the system ie shown in Fig. 1. Four basic parts may be
separated in the system3 apecialized processora with an expanded arithmetic on
the K589 microprocessor series,analog-digital and digital-analog converters for
communications with sources and users of audio data, a control device with the
- 580 series of microprocessors and service equipment.
The initial audio signals are converted by analog-digital converters from the
analog to the digital form and are stored in a buffer memory. As it becomes freed,
the first specialized processor readit the stored data. Naving implemented a given
number of algorithm "layers" on the obtained data, the specialized processor, in
the course of implementing its last own algorithm "layer," transfers the inter-
mediate results to the next specialized processor and introduces the next in turn
input readings. All apecialized processors included in the system operate the
same. Thus, all specialized processors operate in parallel which ma,kea it possi-
ble to provide for processing signals in vaxioue ranges of frequency discretiza,-
tion by changing the number of specialized processors. When increasing the
discretiza,tion frequency, the number of layers in the algorithm being implemented
in one processor is reduced, while the number of processors is increased, insuring
the required transit capacity of the system.
For compaxatively low discretization frequencies (in the case of proceasing apeech
~ signals, in the digital adaptive interference compensator etc.) the number of
algorithm "la,yers" implemented in one processor is increased, xhile the total
' number of processors in the system is reduced. This general principle for design-
; ing systems for processing rapidly flowing signals, whieh provide_ �nr the im-
plementation of the above-cited basic requirements, was laid at the basis of
creating a multimicroprocessor system to process audio signals in the frequency
i region.
~
~ The last specialized processor proceeses the obtained spectram (for example, mul-
, tiplication by tlie weighting function, smoothing etc.) and transfers the processed
- values to the first processor for implementating the 2nverse Fourier transforma-
' A tion.
After completing the algorithm of the inverse Fourier transformation, the data
; from the last specialized processor are entered into the buffer memory (or into
the required external device), which implements the equalization of the period
of output readings ahead of the digital-analog converter.
The calculation of the weighting functione for processing the spectrum, the re-
quired constanta, the loading of data into the last specialized processor and
the syatem control are lmplemented by the control unit made of aeries K580 micro-
processors. The use of the microprocessors is determinea by the requirements
of the relative simplicity of pragraming and the comparatively low speed in re-
ceiving the data. The syatem controller is uaed basically only in the investiga-
ting complex, since the required data is already calculated and stored in the
permanent memory of the last specialized, processor.
143
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504020034-7
FOR OFFICIAL USE ONLY
3y
6eca9ou
- NfIN
2
u
acxOd uu ~U
3DyKOQca ~xodNOe
~neuuona -
JuposaNNair
neuuonu- 6~xo woe
'upnQaNNao 6~�1Y0L
jy
a
~
e
llpu~od
TB /8~
1
cuzNan 6ymrpNOe
npaaeteoD
cc
npou
. n uqn
p
noHamopa
7Y
!
- Cu:Nan (9
Fig. 1. Structure of a multiprocessor system for processing audio aignals in the
frequency regions.
1.
,Lnitial audio aignal
6.
Output bixffer memory and analog-
digital convertar
2.
Analog-digital converter [ATsP]
and input buffer memory
7.
Cantrol unit
3.
Specialized proceasor 1
8.
Drive of TV monitor
4.
Buffer memory of the xeighted function
9�
Output analog audio signal
5.
Specialised processor n
Service devices make it possible to represent the spectral characteristics obtained
on the TV monitor and assign the required amplitude-frequency characteristics by
means of a "light pen."
The specialized processor consists of the folloxingt the 16-bit processor proper,
a parallel 16x8 bit multiplier, a permanent memory (PZU) of trigonometric functions
and output registers.
It is Well knoKn [1, 21 that conversion done on each reeAing in each "layer" of the
rapid Fourier transformation (BPF) algorithm consists of four multiplier operationa
and several adding operations xhere one of the multipliers is a sina or cosine of
the given argument. Since the multiplication operation and, especially, the calcu-
lation of the trigonometric functions mdde in the procesaor require large expendi-
tures of time, a multiplier and a permanent memory of trigonometric constants are
introduced into the apecialized procaseor as external devicae.
144
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLX
T--
(uuNa duA2l
( i,
U/twa
pnpoEni,vu,v
npcyeccooa
i
~o
;
~3)
,
~
~
a
yrrNOrru_
mena
i
2
~
b
u
my .
3kc

[(eNmponsNe~u
npoyeccop .
E4eccopa
(7)
mpuyeciruxj( $
[!/uNa esoda dcNNax /
~~a
~IIuNQ dQ BQMNb/X Y
. 10)
Fig. 2. Structural arrangement of the apecialised procesaor
1. To external devices
2. Processor data bus
3. Processor control bus
4. Central processor
5. Multiplier
6. Procesaor input bu8
7. Output regiaters
8. Permanent memory for trigonometric
funetiona
9. Da,ta input bus 1
10. Data input bua 2
The processor itself (Fig. 3) consists of eight microcircu.ite of the K5891K02 cen-
tral process unit [ToPE], a K5891x03 accelerated transfer circuit, a K5891K01
microprogram control unit (BMU), a microprogramed permanent memory mada of K556RT5
components and a main memory made,of K541RU1 microcircuits.
In the reprogramable permanent memory (PPZU) there are recorded microprograms for
implementing thnsa concrete "layers" of the algorithm done by the given processor,
set in a concrete position in the system's structure.
145
' FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLX
Mmwponpoz - RoNAe~-
pannMOt epwelu
0/1Jy peautip � 03y
enoK 5
ynpaQneMUa NoccuQ ueMmpane-
Heix npouictop-
Neix .7neneNm05
~
_
(7) l
u~
ll/tINQ
ynpaEneNUn
IY4N0
QQpdQ
U/f[HQ CS,
QQNMbIX
Fig. 3. 5tructure of the central proceasor
1.
Reprogramed permanent memory (PPZU)
5�
Data file of the central processor
2.
Conveyar register
components
3,
Control unit
6.
Control bus
- 4.
Main memory
7�
Input bus .
8.
Data bus
Besides the microprogram of the given part of the total algorithm of the system,
service microprograma for cleaning.out the main memory (OZU), the initial setting
of all processor registers and expanded arithmetic, as well as monitoring the ef-
_ ficiency of the processor are recorded in the PPZU.
Lach niicroinstruction along with standaxd control fields BMU and TsPm, includes
nonstandard control fields of record.ing and readout procesaes of all external cir-
cuits connected to the procesaor� arid the OZU. Additional regiaters are intraduced
into the proceasor circuit to increaae the operating speed of tha proceasor and
span the operations conveyor regieters in time xhich are installed at the outputs
of microprogram PPZU and make it possible to reduce'the synchronous signal period
of the proceasor. Input registera of data and.OZU addresses ma,ke it poasible,for
the central processor unita to do the converaions-without xaiting for the end of
the access cycle to the OZU.
The C'LU has a.capacity of 4K 16-bit Words. Ha.lf the memory is uaed for storing
the intermediate results of direct conversion, while the second half for atoring
the intermediate results of the inverse Fourier transformations (the reault of the
calculations up to 1024 absolute values of the spectrum and up to 1024 phase
pointa).
146
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
(1)
(2) ,
(3
U/u Me+ daNNei.q
hsacmp
napaAn�a&Nerv
nMA,,,,U ,
ynNOnramene
noto
>6X e
Om pJ9
QroOrau
mpata+one-
nynemu�
m0uwecxux
nnPKCOp
~
~
noNtmoNm
-
tUuma yn/QdpeNUn
. `
(5) -1
/aRaD- UktMs
Nou ~~~a
ylNb~-
nnawc ,
Fig. 4. Structural arrangement of the multiplier
1. Data bus
2. Multiplicand register
3. Parallel 16x8 bit mult iplier
4. OutpUt multiplexer register
The basiG proceesor functions arei
5. Input bus
6. From permanent memory of tri.gono-
metric functions constants
7. Control bus
8. Input multiplexer
1. Form addresses, data and control signals for microprogram PPZtJ, OZU and PZU
trigonometric conatants.
2. Do axithmetic operationa of adding and subtracting taking into account the
signs of the values.
3. Control loading and reading of multiplier data.
4. Control data recording into output regiaters and data reading from ATsP or.
previous specialized processors. Algo can:rol the weighting funetion and required
cYLU funetions.
In choosing the structural arrangement of the multiplier, it is neceseaxy to es-
tablish a compromise betxeen expenditures for equipment and speed of operation
taking into account the passibilities of the remaining parts of the apecialized
processor. The stz-uctural arrangement of the multiplier is ahown in Fig. 4. Its
operational part is a parallel 16x9-bit multiplier.
The multiplicand value is atored in a 16-bit regi$ter. Data to the multiplicand
regfster inputs is entered from the data buses of the central processor. The
147
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
control of recording in the multiplicand register is implemented by a separate
microinstruction field.
An 8-bit number is entered into the multiplier inputs from the multiplexer outputs.
In accordance with the address sent to the multiplexer input from the separate
microinstruction field, the multiplier may use the high-order byte in the PZi1 of
trigonometric constants, the low--order byte of the same PZU, the high-order byte
on the data proc&ssor buses or the low-order byte on the data processor buaes.
The output multiplier multiplexer sends eitner 16 high-order bits of the result or
8 high-order bits of the results to the input processor buses.
To reduce the aize and power eonsumption of the multiplicand reg'_ster, the input
and output multiplexers are made xith K555 8eries microcircuits. Such a connection
of the multiplier to the procegsor makea posaible rapid implementation of operationa
on the number in the procese of calculating the direct and inverse Fourier tra.~s-
formations, as well as in procesaing the obtained spectral characteristics.
= Besides the multiplier, two data sources that have a controlled triatable output
- (for example,, the ATsP and the prececling specialized output) may be connected to
the buses of the inputfoutput processor.
Permission for the passage of data from the ATsP or preceding specialized processor
is given by separate microinstruction fields. The specialized processors axe
coupled by means of an output regiater ma,de of K5551R16 microcircuits.
- The absolute value and phase are calculated in the last apecialized processor in
accordance with the obtained 1024 complex rsadings of the spectrum. Fach point
of the absolute value may be multiplied by the correspondin value of the weight-
Ing function (or processed. according to the given algorithm~ after which the re-
sults obtained are sent to the firat specialized procesaor for the start of the
inverse process of the Fourier transformation. Reading the weighting funetion
values and the required constants in processing the spectrum is implemented over
the second external.bus (instead of over the ATsP output).
To provide for the poasibility of a flexible change in the system of the number of
specialized processors, a complex of microprograms is being developed that takes
_ into account the distribution of the number of "layere" in the isPF algorithm in
each specialized proceasor.
Concluaiona
The analysis which was carried out indicated that in using three specialized proc-
essors (each mounted on three Kr1MAK plates), it is posaible to convert, process and
restore audio signals with diacretization frequenciea of 40 kHz in real time.
Connecting 10 specialized procesaors makes it poasible to expand the frequency
rarige of the processed signals to discretization speeds of 100-125 kHz.
If, in the investigation process, it is required to obtain apectra without restora-
tion of signals in the time do:aain, then for the same number of special processors
the discretization frequency may be doubled.
148
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00854R400504020034-7
FOR OFFICIAL USE ONLX
At present, only schematic diagrams have been developed. The deaign documentation
for the syatem ia being developed.
_ BIBLIOGRAPHY
1. Rabiner, P.; Gould, B. "The:ory ancl Application of Digital Processing of Sig-
nals." Moscow, "Mir," 1978�
2. Gold, B.; Reyder, Ch. "Digital Proeessing of Signals." Moscow, "Sov. Radio,"
- 1973.
3. Belkin, B. G. "Development of a Digital System for Recording-Reproducing
Sound in Motion Pictures. "Tekhnika kino i televideniya," 1976, No il, pp 3-12
4. Belkin, B. G.i Baryshnenkov, Yu. N.; Gordon, M. G.; Neverovskiy, K. V.;
Raver, L. Yu.; Tsygankov, M. V. "Digital Devices for Processing Audio Signals
in Real Time." Teletekhnika-80" exhibition. Scientific-Technical reports.
1980.
_ COPYRIGHTi "Tekhnika kino i televideniya", 1981
2291
CSO: 1860/77
149
FOR OFFICIAL USE ONLX
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500024034-7
FOR OFFICIAL USE ONLX
UDC 791.44.025: 681.322
AUTOMATIC RESTORATION OF MOdIE FILMS BY COMP(JTFR
Moscow Tfl{HINIKA KINO I TII,EVIDIIdYA in Russian No 9, Sep 81 pp 32-36
[Prticle by B. L. Borilir and V. V. Pospelov, USSR Scientific-Research Center
of Technical Documentation]
- [Dccerpts] In recent years, automated methods for processing pictures based on
using the computer are being used widely [7-16]. These methods axe used to solve
problems of various natures; automatic analysis of pictures, recognition of visual
Inages, improvement ef picture quality, eaeing of their visual interpretation etc.
The processing of pictures with computers consists of three basic sta,gesi
conversion of the picture into digital form and its introduction into the computer;
mathematical processing of the digital data in the computer;
withdrawal of the procesaed digital data from the computer and its reverse conver-
sirn into a picture.
Picture processing in digital form tnay be done in two modess automa.tically in
accordance with given programs, and in a d.ialogue mode with the computer, a proc-
ess in which the operator watching the picture on a TV screen of a special display
can intervene in the processing, selecting necessary programs and their paxameters.
'Phe automated methods used to solve the,problem of eliminating defects and ralsing
the quality of the picture make it posaible to do the folloxing:
eliminate a large numbAr of pfcture defects (correction of various distortions and
filtration of variaus kinds of interferences)i process the entire picture, as well as its individual parts up to its smallest
elements with dimensions of 10'2 to IO-3 mm;
carry out the proceasing xith high flexibility due to the dialogue selection of
programs zi,nd parameters;
obtain high reproducibility of procesaing results due to the ability of the com-
puter to repeat many times and convert the picture to a given accuracy;
150
FOR.OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
preserve the original picture fully since only its "digital image" is subjected to
the conversionsi
reduce the number of manual operations to a min'mum.
Due to its advantages, methods for processing pictures by computers axe used widely
for improving the quality of space, astronomic, medical and other pictures which
is confirmed by many publicatians [12-14, I7-20]. This article describes briefly
the results of the theoretical and experimental ir.vestigation of the automa,ted
method of restoring photographic documents, and discusses the possibilities and
prospects for automated restoration of films at the NITsTD SS5R.
Any black and white photagraph and photographic document can be repreaented mathe-
matically by a function of two variables U(ic, y) that describes the optical denaity
distribution in the photographic material plane. The defective photographic doc-
unent may be described by another funetion V(x, y) obtained from U(x,y) by several
conversions
v'A(U) ~1)
The difficulty in the problem of restoring the initial picture U(xly) is that, as
a rule, the exact form of operator A is unknown. Therefore, hypotheses are proposed with respect to the kind of conversion of A,
on the basis of the analysis of the distorted document, after which the pa.rameters
of this converaion are chosen in the procesa of solving equation (1) on the com-
puter. The correctness criterion of the proposed hypothesis and of the parameters
selected is an expert evaluation of the quality of the restored picture t(x,y).
Aa a result of investigating defects of archival photographic documents [21], it-
was propoaed to divide them into seven groupst frequency-contrast, gradation, cracks,
structural, spots, scratches, and other defects (not included in the basic groups).
Mathematical models of the most typical defects of photographic documents of the
- six basic groups are now available and theoretical a.nd experimental investigations
_ of inethods for their elimination by means of a computer were carried out [22]. It
was establiahed that the most efficient are dialogue (interactive) methods for the
restoration processing of photographic documents which allow the operator flexi-
bility in controlling the restoration process by watching the picture and selecting
the necessaxy programs and parameters to process it.
The investigationa done made it possible to formulate basic requirements for the
automated system for processing pictures (ASOIZ) intended for the restoration of
photographic dacuments.
The ASOIZ was created,on the basis of the LSI-2 minicomputer, apparatus for input-
output of photograr231. hs of the American ff.rm "Optronics" and the Pericolor" (France)
display processor At present, system software is being developed and pro-
duction work is being done (so far on a small scale) on digital restoration and
recording in digital form of unique and especially valuable photographic documents
of the USSR State Archives Hl.ind.
151
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504020034-7
FOR OFFICIAL USE ONLY
We will consider the possibilities and prospects of the automa.ted restoration of
movie films. The black-and-white movie film is described mathematically by a func-
tion of three vaxiables U(x, ,t) ~rhere t is time; a color movie film is a vector
function JR(x,y,t), G(x,y,t~, B(s,y,t)} where R, G, B are respectively the red,
grser_ and flue components of the optical density of the color movie film. If
each frame of the black-and-white movie film in the digital form represents a two-
dimensional da,ta file of.1.5 million eight-bit binaxy digits, then the volume.of
� digital data containing the information about the entire movie film will be tens
of billions of bytes. Proceseing auch large data files, even with modern powerful
computers, is impossible at present. However, the rapid development of computer
equipment ma,kes it possible to hope that digital restoration of movie films may be
done in the near future. In our opinion, promising ways to solve this problem
arec
usin; matrix processors that make it possible to process in parallel thousands
of picture elements (by using such a processor, a rapid Fourier tranaformation of
a data file of a thousand numbera may be made in 0.01 aeconds [251);
creating program4 controlled digital devices that do vaxious operations on procesa-
.ing plctures in real time (with the speed of the video signal pa,ssing through
them).
In this case, the ASOIZ could play the role of a system that would develop alg o-
rithm,5 for the elimination of typical defects in movie films. Then spect.alized
devices, created on the basis of these algorithms, would make it possible to
process hundreds of movie film frames at a high speed.
As already mentioned before, from the mathematical standpoint, a movie film ma,y be
considered as a function of three variables. An important fea,ture of this func-
tion is the presence of two groups of variables essentially different in natureo
these are spatial variables x, y and time variabie t. For auch functions a very
promising interference filtration method is an expansion of this funetion into a
bilinear series [24.11
u(X,y,t)=u1(X.y) T1(t) U2(X,Y) t (2)
Cibviously, in restoring movie filma, it is advisable to subdivide them into parts
(plans) containing uniform frames averaging 2 to 3 seconda in length. The general principles of deslgning a system of digital procesaing of movie films
were formulated by L. F. Artyuahin C261. Automa,ted systems for restoring movie
ftlms must coneist of a specialized ASOIZ containing three basic subsystemst con-
verting movie films into digital forms and backi processing and storing data files
of dlgital data= visua,l representation of digitalized movie films.
Requirements of these three subsystems are greatdr than the correaponding ASOIZ
subsystems tntended for restoration of movie filme. The aubsystem for converting
movie films into dig.ital form and back may be created on the basis of high epeed
TV or laser scanning d.evices (for example, electronic copying equipment for
152
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLX
elemont-by-element printing of awvie film pictures [271). The subaystem for pro-
cessing and storing digital da.ta files must contain a high speed computer operating
in the time shaxing mode (or a complex of several computers), a matrix processor,
specialized program control digital devices for processing pictures and high CapaC-
ity (about 108 bytes) magnetic diac memories.
The subsystem for visual representation of the digitalized movie films must bb an
interactive display proceasor (one or several), equipped with a main memory, a
high speed processor and flexible dialogue facilities for processing the pictures
(light pen, courser, trackball etc.) .
The processes for the input-output of movie film pictures and the t,ransfer of
digital data files between subaystems must be cantrolled by micro or minicomputers
without the loss of time by the central aystem processor. .
The development of an automated system for reetoring movie film pictures is a
complicated but realistic problen, substantiated by an already created electronic
multiplication system and by obtaining combined movie film pictures by computers
[28, 29]. In the future, when automated systems for processing movie film pic-
tures are introduced in practice, it will become possible to use digital methoda
to create movie films. Their realization will open up great prospects for the
restoration of movie filme, as well as for the development and improvement of
technical and creative cinematography processes.
BIBLIOGRAPHY
1. "Manual for Recognizing and Elitninating Defects in Movie Films and Microfilms."
All-Union Scientific Research Institute of Document Management and Archives.
Moscow, 1973�
2. Fridman, I. M.; Petrov, A. I. "Storing Lenin's Movie Films." TFKHNIKA KIldO
I TII,EVIDMIYA, 1970, No 4, PP 19-26i 1980, No 4, pP 3-7�
3. Volkman, H. "Thesen zur Dauerkonservterung und Restaurierung Audiovisueller
Informationstrager." Berlin, 1975�
4: "Storage and Restorational-Conservational Proceasing of Movie Films and 'Micro-
films." Interrepublic Technical Conditions (for pocument Standardization)
55-4-69. Moscow, 1969.
5. "Movie Filma. Additional (Restoration-Preventive) Proceasing of Initial Film
Materials at Cinema Studios and Movie Film Co;/ping F'actories. "RTM [Guiding
Technical Materials] 19-6-71. Moscow, 1971.
6. "Restoration of Faded Photographs. Review of Domestic and Foreign I,iterature
on Motion yicture ~4uipment." Moscow, 1973, No 1(5).
7. Rozenfeld', A. "Recognizing and Processing Pictures by Digital Computers."
Moscow, "Mir," 1972.
153
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
8. irishin, M. P.; Kurbanov, Sh. Mi Markelov, V. P. "Automa.tic Input and
Yrocessing Photographa by Computers." Moscow, "Energiya," 1976.
9. Khant, B. R. "Digital Processing of Pictures." in book "The Use of Digital
Processing of Signals." Noscow, "Mir," 1980 pp 192-267.
10. "Proceasing of Pictures and Digital Filtration." T. Khuang, editor. rloscow,
"Mir," 1979.
il. &idryus, G. "Use of Digital Computers for Processing Pictures." Moscow,
"&ergiya," 1977�
U. "Iconics. 3patial Filtra,ion of Pictures. Photographic Systems." Symposium
of articles, D. S. Lebedev, editor. Moscow, "Nauka," 1970.
13. "Iconica. Digital Holography. Processing of Pictures." Symposium of arti-
cles, D. S. Lebedev, editor. Moscow "Nauka," 1975.
14. "Iconics. Digital Processing and Filtration of Pictures" D. S. Lebedev,
editor. "Problems of Cybernetics." No 38, Moscow, ViNITI [All-Union Institute
of Technical Inf'ormationl, 1978�
- 15. "Processin$ Pictures by Digital Computers." Symposium of articles,
G. r3idryus and L. Inlo, editors. Moscow, "Mir;" 1973.
1 16. Yaroslavskiy, L. P. "Introduction to Digital Processing of Pictures." Moscow,
"Sov. Radio," 1979.
17. Belikova, T. P.; Kronrodt M. A.i Chochia, P. A.; Yaroslavskiy, L. P.
- "Digital Processing of Pictures of the Mars Surface transmitted by "Mars-3"
and "Ma.rs-4" AMS [Automatic Maxs Station]. "Space .inveatiga,tions," 1975,
13, No 6, PP 898-906.
18. "Space .Investigations of the khhrt.h. Mettiods of Processing Video Data by
Computers." Moscow, "Nauka," 1978.
19. Klimenko, V. L. "Digital Processing of Pictures." "Radioelectronics Abroad"
1975. No 4 pp 13-32.
70. Digital Processing of Signals and its Application." Symposium of articles,
L. P. Yaroslavskiy, editor. Moscow, "Nauka," 1981.
21. Borilin, B. L. "Qualimetric Evalua.tion of Archive Document Defects.
"SOV1~.~'SKIYF ARKHIVY, 1981, No 3 pp 37-40 �
22. Borilin, B. L.; Chochia, P. A. "Restoration of Photographs by Computers."
SUV~~'SKIYE ARKHIVY, 1980, No 3 pP 45-48.
23. Borilin, B. L.; Pospelov, V. V. "Reatora�tion and Conservation of Photographs
by Computera." &cpress information. VNII Documentation and Archive Manage-
ment, Series "Norms and Specifications for Storing Documents." 1980, No 2(11),
pp 1-7�
154
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFFICIAL USE OYLX
24. Pospelov, V. V. "Approximation of Funetiona of Several Variables by Products
of Flinctions of (?ne Vaxiable." Preprint IFM CInstitute of Applied. Ma,thematica]
1978, xo 32.
25. Integral Array Processor for r7clipse 5/250. Prospectus of Data Gerieral Co.$
United States.
- 26. Artyushin, L. F. "Prospects of Using IIectronic and Computer hquipment in
Technological Procesaes of Cinematography." TII{EIIVIXA KINO I TII,EVIDIIdIYA,
1979 No ii, pp 7-17.
27. Artyushin, L. F.; Toshin, 0, I.; Ovilko, 0. G.; Muchiyev, S. G. "Kinotele-
vision System of Printing and Recording Color Picturea by a Laser Beam on a
Movie Film from Film Ma,terials and Video Tapes. "TBXHNIKA KINO I Tffi+EVIDENIYA,"
1977, No 7, PP 3-11.
28, Tel'nov, N. I. "Proceasing Pictures by Computers and Electronic Nniltiplica-
tion." TFKHNIKA KINO I TII,EVIDENIYA," 1978, No 7, pp 70-81.
29� Holt, W. The New IIectronic Composite Photography and Image Modification
system. American Cinematographer, 1975, 5, No 4, pp 421-425, 438, 474.
COPYRtGHri "Tekhnika kino i televideniya", 1981
2291
cs0, 1860/77
155
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500024034-7
� FOR OFFICIAL USE ONLY
COMMUNICATIONS, COMMUNICATION EQUIPMENT, RECEIVERS
AND TRANSMITTERS, NETWORKS, RADIO PHYSICS, DATA
- TRANSMISSION AND PROCESSING, INFORMATION THEORY
UDC 621.372 (088.8)
MICROWAVE DELAY LINE
- Moscow OPISANIYE IZOBRETENIYA 637900 in Russian 15 Dec 78
[Description of USSR Paten t No 637900 by A.V. Vashkovskiy, V.I. Zubkov, V.N.
Kil'dishev and B.A. Murmuzhev, filed 9 Dec 76, published 15 Dec 78, class
H 01 1 P 9/00]
[Text] This invention is in the area of radio engineering and can be used in
- antenna feeder systems to delay microwave signals.
A microwave delay line exists which containe a rectangular magnetized ferrite plate
ar.d input and output coaxial antennas installed at opposite ends of the ferrite
plate [1].
However, this device has a small delay; furtherniore, parasitic resonances are
excited in it. The purpose of the present invention is to increase the delay time and eliminate
parasitic resonances.
In order to accomplish this, the ferrite plate in the microwave delay line contains
an additional flat metal spiral with w3dth and spacing which are multiples of half
the length of the surface magnetostatic wave; furthermoer, ttie beginning and end
of the metal spiral are lo cated directly at the coaxial antennas, which are in-
stalled at opposite corners of the plate; the portion of the ends of the ferrite
plate on which the metal spiral is wrapped is rounded.
The drawing presents a di agram of the proposed mir_rowave delay line.
The microwave delay line consists of rectangular magnetized ferrite plate 1 and
input and output coaxial antennas 2 and 3; ferrite plate 1 is furthermore magne-
tized by permanent magnet 4, and metal spiral 6 ie wrapped around its rounded
ends 5; in addition, input and output coaxial antennas 2 and 3 are fastened to
opposite ends 7 of ferrit e plate 1.
The microwave delay line operates as follows.
Input coaxial antenna 2 in the vicinity of face 7 of ferrite plate 1 excites a
surface magnetostatic wave signal which propagates along the waveguide region
156
FOIt OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
formed by flat metal spiral 6 and the adjacent upper suxface of ferrite plate l.
The surface magnetostatic wave signal follows rounded phase 5 to the bottom of
ferrite plate 1 and propagates to the poposite side along the second turn of flat
metal spiral 6.
Thus, the surface magnetostatic wave signal moves from the top of ferrite plate 1 to
its bottom n times (where n is the number of turns of flat metal spiral 6), each
time propagating over a different turn of flat metal spiral 6 to output coaxial
antenna 3, where the surface magnetostati c wave signal is converted to a delayed
electromagnetic signal.
25
Patent Claims
1. A microwave delay line containing a rectangular magnetized ferrite plate and
input and output coaxial antennas fastened to opposite faces of a ferrite plate,
distinguished by the fact that in order to increase delay time and eliminate
parasitic resonances, an additional flat metal spiral is wound on the ferrite
plate with width and spacing multiples of half the length of the surface magne-
tostatic wave; furthermore, the metal spiral begins and ends directly at the
coaxial antennas, which are fastened to opposite diagona.ls of the p].ate.
2. A microwave delay line as described in paragraph 1, distinguished by the
fact that the portion of the faces of the ferrite plate around which the metal
spiral is wound.
Information Sources
Used in Evaluation
1. L.H. Brundte, N.J. Freedmann 1, "ElecCron. Lett" 1968, 114, 7, p. 132.
6900
CS O: 8144/0218-B
157
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R440500020034-7
� � FOR OFFICIAL USE ONLY
UDC 621.372.8:535
METHODS FOR FINDING BREAKS IN OPTICAL CABLES
Moscow ELEKTROSVYAZ' in Russian.No 9, Sep 81 (manuscript received 4 Aug 80)
PP 25-28
[Article by S.M. Vernik and A.M. Kuznetsov]
[Text] Successful operation of communication links is deter-
mined to a significant extent by the capability for finding
breaks with sufficient accuracy. The solution to this
problem usually involves major difficulties both in developing
measurement methodology and in creating measurement instrumen-
tation. For optical cable communication links (OCCL), espec-
- ially long links with attenuation on the order of 30-40 dB,
' solving this problem involves additional difficulties arising
from the complexity of the propagation and ref lection of
light signals in optical fibers (OF).
We shall examine possible methods for finding breaks in OF and
OC as a whole, primarily methods suitable for use on long links.
It is known that not all actual breaks in an optical fiber (OF) result in complete
cessation of signal propagation. For example, a break in an OF with vertical
spalling and 2-5 um separation of the OF in the cable causes only an insignificant
increase in signal attenuation. Therefore, we ahall consider an OF break to be
damage which causes the propagation of light energy to crease completely. There
are three basic methods for locating OF breaks.
The method of ineasurin� the light energv emitted into the environment is used for
quality control of initial semifinished items, or "bars", from which optical fibers
are drawn, and for finding OF breaks and large (of the order of 2-3Y) imperfec-
tions. Quality control of bars is done by irradiating them with a visible-spectrum
gas laser (X = 0.62 um). When this is done, the spots st which the OF imperfec-
tions are located glow. The quality of the bar can be 3udged by the intensity of
the glowing and the number of glowing spots. By processing the visual observations
statistically, a correlation can be established between bar quality and OF imper-
fections.
The spots in which large OF imperfectiona are locatPd emit part of the light eignal
energy j.nto the environment. The measurement eignal input to the OF consists of
- powerful light transmissions (on the order of 1-3 W) transmitted at a frequency
158
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-04850R000500020034-7
FOR OFFICIAL USE ONLY
of 1-5 KHz. The light signals emitted into the environment are captured by the
photc,detector in the receiver and converted to current (voltage) pulses, which
are then amplified in a high-gain narrowband amplifier which ia tuned to the repeti-
tian frequency of the light pulses.
During the measurement process, the OF is moved with respect to the photodetector
and the radiation intensity is recorded, with the power distribution of the
emitted signals along the'OF characterizing the distribution of the imperfections.
The location of a break in the OF usually corresponds to high amplitude of the
emitted signal. Preliminary findings indicate that this method can be implemented
if the dynamic range of the receiver is between,110 and 140 dB.
The method of ineasuring the intensity of Rayleigh backscatter is widely used in
measuring the distribution of imperfections along OCCL and the internal attenuation
of OF and OC.
Figure 1 shows a functional diagram of a measuring device which implements this
method. The measured OCCL 5 is pulsed with short (2-5 ns) high-power light pulses
(1-2 W), which are input to the optical cable through differential device 4. The
reflected backward scattered flux reaches receiver 6; the ecreen of CRT 7(or other
display) displays the intensity of the backward scattered flux as a function of
the length of the line (or signal propagation time), as well as the location of
imperfections in the op.tical fiber. A break in the OF is characterized by a
sharp drop in scattered power.
The basic shortcoming oi this method i!s the low level of the backward scattered
flux, which prevents this method from being used to find breaks in long OCCL
because of the insufficient sensitivity of the receiver. In additiion, when there
are relatively large imperfections (which occurs at the existing level of OF and
OC fabrication technology) the power of the oncoming light pulse flux may be
significantly higher than the power of the Rayleigh scattering flux. This cir-
cumstance results in distortion of the measurement findings.
CD
Figure 1
The location method for finding breaks in optical fibers is identical to the im-
pulse method of ineasuring imperfection distribution on cable communication links.
The functional diagram of the instrument used in the location method is analogous
159
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE ONLY
to figure 1. Generator 1, which outputs light pulses 5-10 ns long is triggered
by electronic master oscillator 2, which sisnultaneously triggers CRT sweep
generator 3 with a slight delay ('[Z). The outgoing pulse is passed through
differential system 4 into the measured optical fib.er 5. Part of the energy
reflected from imperfections in the OF creates a backward light flux which
reaches photodetector 6 and is converted to voltage pulses. After amplification
in amplifier 7, these are applied to the verttcal plates of the CRT, causing the
electron beam to be deflected up or down from the middle of the y axis. The
arrival time of the reflected pulsee is registered in the scale of the sweep time
s(t), which is easily converted to the distance in kilometers to the break
(f igurz 2):
x = 0.5 s(t) v (1)
where v is the propagation speed of light energy.
F:igure 2
The location of the break is determined by the diatance x, beyond which no
reflected puJ.ses reach the receiver.
The main difference between the rarameters of the scheme in figure 1, which is
used to measure the location of breaks and distribution of imperfections using
the location method, and the parameters of analogous schemes used to measure
the intensity of the backward scattered flux, is that the receiver used in the
location method has a narrower operating bandwidth. This makes it possible
to implement a receiver amplifier with a dynamic range of 50-60 dB. Further-
- more, as was indicated above, the signals rEflected from imperfections and
' breaks usuaZly exceed the intensity of the backward scattered flux by several
orders of magnitude. This method thus makes it possible to make measurements
on long OCCL.
It is apparent from examining the functional diagram in figure 1 and the
description of its operation that it is analogous to the well known circuits
used in type R5 or UIP devicea [3], which are intended for investigating the
distribution of imperfections and finding breaka in cable conductors as a large
imperfection in which the coefficient of reflection is equal or close to unity.
This circumstance is of great significance, since iti makea it fairly simple
~ to implement a device to determine a location of optical fi6er breaks in the
160
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
NUR OFFIC'IAL USE ONLY
form of an attachment to the type RS devices which are in quantity production.
However, when measuring optical cables it ia necessary to allow for the sin-
gularities of optical signal propagat3on and reflection of optical flux in
_ optical fibers, which lead to the appearance of additional errors.
Error of location method. During transmission over opticsl fibers, optical
- signals are attenuated due to losses. tn the OF,. whieh causes the energy of
the outgoing light pulse Wo at the end of an optical calile with length L to
_ drop to
WL _ Woe_o .2 saL (Z)
where a is the OF attenuation factor, dB.
The energy of a pulse fully reflected from the end of the line as measured
at the beginning of the line
~ wref � Woe-0' 46aL (3)
Knowing the amount of attenuation of an OCCL repeater section arz , we can
- determine the mtnimum dynamic range of the instrument a~n = 2arep� In
order to determine the dynamic range of the instrument, we must ad4 to amin
the attenuation airef which characterizes the energy loss during reflection
at a point located at a distance of Li from the beginning of the OCCL.
These losses are random, and comprise the function ~i of the set of random
quantities made up of the absolute value of the coefftcient of reflection
at the ith point Pi, the distribution of the angles at which light waves of
various modes are reflected from the surface of the break in the optical
- fiber ~iu, where u is the mode number of the light wave. Furthermore,
, airef is a function of the OF aperture angle 6 and the propagation conditions
for individual modes in the OF. Thus,
airef (Pi. (4)
is random as well.
Figure 3 shows some possible profiles of optical fiber breaks and the oscillo-
grams of the ref.lected signals which correspond to them. The smallest losses
occur if the surface of the break is perpendicular to the axis of the fiber
(figure 3a), whj.le the greatest losses are obsenred if the angle between the
plane of the break and the axis of the fiber is 450 (figure 3b) ; when the
- surface of the break is complex (figure 3c), the amount of loss is random.
In addition to losses due to reflection, losses. in the differential section
adif of the measuring instrument must al.sa be alloeved for. Consequently,
161
- FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2047102109: CIA-RDP82-00850R400504020034-7
FOR OFFIC'IAL USF ONI,Y
ain = 2arep + axef + adef (5)
If we limtt ourselves to measuring the location of an optical fiber break
from both sides of a repeater sector,
ain min = areg + aref 't' adef (6)
The maximum value of ain which can b e realized in pract3ce is 80-90 dB with
outgoing pulse power of the order of the 1-2 W. Losses in the differential
device are usually 8-10 dB [1], while the attenuation of the repeater
sector is 30-35 dB. Tt follows from tfiis that aref caruiot be more than 0-10 dB
when measuring from one end of a repeater sector; the value is 25-35 dB when
measuring from both ends, which makes it possible to locate the OF break quite
accurately.
a~ . ~ .
- a,
break
I
I
b) " . . I
I
n' ~
_ I
---n1-----.
. I
I
I
I
c) I
I
----nt----- ~
Figure 3
THe duration of Che ulsin si nal also has an efPect on the error tn determining
the location oP an OF break [2 . In determining this error, allowance must be
made for pulse stretching in the 13.ne. Propagating through an optical fiber,
a light pulse with duration to at the beginning of the fiber stretches dne to
the dispersion of the OF inaterial, the relationship between the group delay time
and the mode number of the light waves, and the randcm dispersion of the group
162
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE: ONLY
delay time values in the optical fib.er.
These distortions caus.e the outgoi.ng pulsea to Uecome longer in proportfion to their
distance from the light source. As a result, the resolution of the method is uneven:
it is greatest at the beginning of the ltne and least at the end. The amount by
which a pulse is stretched depends upon the type of f3ber and the characteristics
of the radiation source.
Fo�r examnle, when, an incoherent light pulse with duration to is transmitted
through a multimode fiber, it stretches to a length of [1]
' (7)
f1 = (ro + eOa L'+ e t2 c2)
by the end of a repeater sector with length L, where AT is the dispersion of
the group delay time per unit of length in the multimode fiber caused by the
presence of different modes; ATd is the pulse expanaion per unit of length
diie to dispersion in the material and the dispersion of the intrinsic waves.
When the intrinsic modes are,mixed in fiber imperfections, the pulse at
the end of the fiber expands to a value of [1]
1 (8)
r~ = (t~ +e ~a' c=+e z� L, L) 2.
where Lc is the average length of the optical fiber within which there is a com-.
plete exchange of energy between modes.
In practice L> Lc; A'r � A'rd and the amount of pulse widening At is proportional
to `L. For example, for n/n = 0.01 the dispersion Ot of the group delay time
is 48.4 ns/km, while it is 24.2 ns/km for a coherent light source. For a grad-
ient fibe r, we have 4.84 ns/km and 2.42 ns/km, respectively. Thus, with an
outgoing pulse of to = 10 ns, :ahen the light source is incoherent the pulse
stretches to 34.2 ns by the end of the first kilometer of the line. These
data indicate that rhere is no point in making the outgoing pulses any shorter
than 5-10 ns.
The maximum measurement error cf the location of a break corresponds to the
maximum value of aref, Where no re,flected pulse is observed on the screen of
the instrument at the point of the break (cf. figure 3b). In this case, the
operator usually assumes that the bre,qk. is located at the fiber imperfection
which is closest to the pulse generator (figure 4a), which has the appearance
shown in figure 4b after the break. The measurement error here wi11 be deter-
mined by the highly regular sector lreg of the optical fiber where no reflected
pulses are observed.
As the quality of optical fibers improves, lreg will increase to the length of
the opti cal cable lines, since junction reflectj.ons which are recorded by
pulsed devices generally arise at the points at which cables are joined.
163
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAI. USF ONLY
With the help of the location method, these reflections can be used to determine
the structural length of the optical cable within which the break has occurred.
a) , .
~ . i ' .
x
L I
. ~ , j. � ,
~ i : I � . � . . , . .
b) ~ i .
1 I
I I '
~ I
~ X
I t~ ugN.~
k'igure 4
Conclusions. Analysis of the location method for finding optical �iber breaks
shows that possible measurement errors are a s.trong function of a number of
random factors which can lead to intolerably large error. In order to limit
this error, it is seen as advisable during OCCL construction to measure the
distribution of imperfections along the cable and include these characteristics
in the OCCL documentation. Changes in these characteristics, as measured from
both ends of a repeater sector and suppletnented with factual data concerning the
length of optical cable structural segments, the distribution of imperfections,
the amounts of reflections at structural interfaces and the actua.l optical
cable attenuation can be used to judge the steadiness of optical cable parameters
by periodically measuring the d3stribution of itaperfections, and to find
optical fiber breaks more precisely. The characteristic of the distribution
af imperfections obtained during these latter measurements can be used as a
baseline for finding optical fiber breaks. Measuring the distribution of imper-
fections from both ends of a repeater section to the point of a break and
beyond will promote a further improvement in the accuracy of locating fiber
breaks.
Bibliography
1. Unger, G.G. "Opticheskaya svyaz [Optical communication]. Moscow, Izdatel'stvo
svyaz 1979.
164
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
2. Grodnev, I.'_'., et. al., "Koaksial'nye kabeli avyazi" [Coaxial CoIDmunication
Cables]. Moscow, lzdatel'stvo"svyaz'", 1970.
_ 3. Kuchikyan, L.I. "Fiztcheskaya optika volokonnykh svetovodov" [Physical
Optics of Fiber Light Guides]. Moscow, Izdatel'stvo "Energiya", 1979.
4. Muradyan, A.G., Ginzburg, S.A., "Sistemy peredachi informatsii po
opticheskomu kabelyu" [Systems foz Transmitting Information Over Optical Cable].
Moscow, Izdatel'stvo "svyaz'", 1980.
5. Teumin, I.I., "Volnovody opticheskoy svyazi" [Opti:cal Communication
Waveguide]. Moscow, Izdatel'stvo "svyaz',", 1978.
6. Makhlin, R.Ye., Titov, I.V., "Measurewent of Reduced Coefficier,t' of
Reflection From End of Fiber During Break in Cable of Ftber Optic Communication
Link," Pis'ma�v ZhTF", Vol. 1, No. 11, 1975.
7. Makhlin, R.Kh, Kuznetsov, A.A., Titov, I.V., "Observation of Optical
Pulses Reflected From Distant End of Fiber", Radiotekhnika i elektronika,
Vol. 20, No, 6, 1975.
8. Guttman, J., Krumpholz, 0.. "Iacation of Imperfections in Ontical Glass-Fiber
Waveguides," Electronics Letters, 1975, Vol. 11, No. 10
9. Veno Y. Motoh Shimizu. Optical Fiber Fault I,ocation Method. Applied Optics.
1976, Vol, 15, No. 6.
10. Personik, C.D., Photon Probe - an Optical-Fiber Time Domain Reflectometer,
PSTJ, 1977, Vol. 56, No. 3.
11. "Detection and Localization of Fiber Breaks in Optical Communication Cables",
NTZ, Vol. 90, No. 11, 1977.
COPYRIGHT: IZDATEL'STVO "RADIO I SVYAZ "',"ELIICTROSVYAZ 1'1, 1981
6900
CS0:1860/61
165
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR 4FICIAL USE ONLY
UDC 621.373.43
CONTROL SYSTEM FOR ARC DISCHARGER SWITCHING USING FIELD DISTORTION TECHNIQUE
Moscow PRIBORY I TEKHNIKA EKSPERIMENTA in Ruasian No 4, Jul-Aug 81 (manuacript
received 30 Nov 79) pp 133-135
[Paper by A.I. Gerasimov and Ye.G. Dubinov]
[Text] A discharger-enhancer filled with a gas up to 1 MPa
is inserted in breaks in 26 cable sections, charged up to an
identical potential difference of 15 to 30 KV, each with its
internal conductor connected to the control electrode of its
own arc discharger. When all of the cables are shorted
(impedance of 2 ohms) hy a common switcher, the rise time of
their discharge current wave, T1 = 40 to 70 nsec, is reduced
by the enhancer down to 3.8 nsec. The scatter in the instants
of discharge enhancer is < 0.1 T1. The discharge ia of a'
multichannel nature when the discharge enhancer is filled with
nitrogen and has a gap of 1 to 2 mn between the electrodes.
In dischargers where the electrical field distribution is distorted, it is necessary
to rapidly change the polarity of the control zlectrode voltage [1]. When synchron-
ously switching a large number of diachargera, sectiona of coaxial cablea are used,
the inner conductors of which connect the control electrodea of the dischargera and
the switcher which is common to all the cables [2-4]. The preciaion of the mutual
actuation of the dischargera, as well se the uniformity of the development of the
current channels in them and the current diatribution over the channela depend
substantially on the rise time of the cable discharge current [2, 3]. If the
cables which are connected in parallel to the ahorting switcher number several tens
of cables, while their charging voltage Up is > 20 KV, then the rise time of the
current through the switcher is usually > 10 nsec.
A system of 26 sections of RK-50-11-13 cablea (overall impedance of p= 2 ohms)
connected in parallel, which makes it posaible to generate a cable discharge cur-
rent wave with a riae time of < 5 naec when Up > 15 KV, is described in this paper.
A schematic of the control syatem for the 26 dischargers being actuated, P1--P26,
and the switcher for the capacitive store C or the individual atorage devices
driving the load, V, is shown in Figure 1. The innner conductor of a RK-50-11-13
166
FOR OFF[CIAL USE ONLY �
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
cable section 10 m long is connected to each contirol electrode 1 of the dischargera.
The high voltage electrodea 2 and the storage device C have a potential of U
applied, while electrodes 1 and the center conductora of the cablea have a potential
Up of the same polarity applied; Up usually is 0.3U - O.SU. A comnon gas filled
discharge enhancer, R0, aimilar to that described in [5], is inserted in breaks in
the inner conductors of the cables. The equalization of the potentials shared by
the electrodes of the enhancer of the cables aections of the shaping line (FL) and
peaking line (OL) is accomplished through inductance coil L and resistor R1; the
peaking line section can be connected to the charging source just as the shaping
line section. The switch K serves to ahort the cable. The parameters of the dis-
charge current pulses of the shaping and peaking line cable sections are meaeured
across shunt resistors R2 and R3,'consisting of 20 TVO-0.25 resistors connected in
parallel which are positioned in breaks in the outer conductors, while the voltage
pulses at electrode 3 and at control electrodes 1 are measured by capacitive divid-
ers D1 (C1, C2,) and D2 (C3, C4) reapectively.
(A) K ocy. . t R~ 1K (A) K ocy. "
A
K acu.
SfOn
1OOK ' ~Mx !00
o
Za,z3 ~n on R3 ~ .
P
Ci OL , G'u Bn 2
S10n '
' R ' tOn p . f00K C Z
- K vcu. ~i U
� SJO �
aZ 0,23 ~n po O~ R 0, 23 ~ Z
L F'I.
OL p2e
- Figure 1. Schematic of the control system for the actuation of 26 dis-
chargers, which switch the capacitive storage device C into
the load. C1 and C3 are 10 capacitors of 51 pFd each.
Key: A. To the oscilloscope;
FL = shaping line; OL = enhancement line.
Following the actuation of switch K, a discharge current wave is produced in the
shaping line cablea; it is reflected from the enhancer diacharger as from a closed
end with a change in the polarity of the electrode 3 voltage. When Z1 � p, where
Z1 is the impedance of L and R1 in aeriea, the maximum pulse potential differ-
ence between electrodes 3 and 4 is 2Up. By selecting the spacing S between the
electrodes, the pressure p and the kind of gas in the enhancer discharger, with
a high overvoltage across the gap, one can obtain a multichannel discharge in;.the
peaking discharger, which enhances the rise time of the discharge current pulse of
the cables in the enhancement lines. The current wave regching electrodes 1 has a
leading edge shorter than the leading edge in the shaping lines, which stabilizes
the actuation of dischargers P1--P26� 167
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
The system was atudied for the case of Up = 15 to 30 KV, an enhancer discharger RO
filied with N2, an electron gas (SF6) and mixtures of them [6], where an air dis-
charger or a discharger with mechanically destroyed film insulation was used as the
switcher K. The studies were conducted in sequences consisting of one-time actua-
tions of the system with the subsequent polishing of the electrodes prior to each
successive start. The inspection of the electrode surfaces of the peaking discharger
showed that the number of erosion spota on the discharger electrodes with a diameter
of 135 mm, filled with N2, was greater than when filled with a mixture of SF6 + N2
and pure SF6 and amounts to > 3; as a rule, a single discharge channel is produced
in SF6. The recording of the current pulaes in two diaunetrically arranged enhancer
line cables coiifirmed the multichannel of the discharge in N2; the diff.erence in the
points in time wh.en the wave reached resistors R3 did not exceed the travel time for
the speed of ligtit around one-sixth of the peripheral length of the ring electrodes
of the peaking discharger. Uniform spacing between electrodes 3 and 4 over the
entire length of the periphery (a scatter of < 0.05 imm) is important for normal
operation of the discharger. The calculated electromagnetic wave propagation time
over half ofthe peripheral length of the electrodes amounts to about 0.7 nsec; in
mixtures of gases and in SF6, the indicated time ditference was observed to run up
to 0.6 nsec. The time scatter AT of the enhancer breakdown in N2 was also less with
the same gaps between the electrodes of the enhancer diacharger. Taking these
circumstances into account, reaults of ineasurements are given below for an enhancer
discharger filled with N2.
Typical oscilloscope traces of the discharge current pulse of the shaping line
cables and the voltiage changes across elect;odes 1 of the dischargers corresponding
to it when solid insulation of lavsan film [synthetic fiber similar to dacron] was
used in the switch K, and the cables were charged at voltage of Up = 23 KV, the
gap d was 2 mm and the pressure p was 0.85 MPa N2,are shown in Fig. 2[not shown]. It
is seen that with a switching current rise time of T1 = 40 nsec,.the voltage change
time is T2 = 5 naec. When d is reduced down to 1 mm, the time T2 decreasea down to
= 3.8 nsec. For six values of Up and d, incregsing the pressure p shifts the
moment of enhancer discharge breakdown to the peak of the current pulse, however,
TZ changes little in this case. The ecatter AT of the moments of enhancer dis-
charger breakdown relative to the start of the current pulse in the shaping line.
measured from the superimposition of 10 and more aignals (Figure 2b), was < 4 nsec
in a series of 100 actuations of the switch K. Under these same conditions, the
acatter in SF6 is greater by approximately a factor of two. Shortening the leading
edge of the current pulae T1 in the shaping line reduces the scatter in the time AT;
when an sir gap diacharger is used, T1 increasea up to 70 naec and AT correspond-
ingly increasea up to 7 nsec. Splitting each cable in the enhancer line near the
dischargers P1--P26 into conductora up to 2 m long which were matched to the charac-
teristic impedance (100 ohma) did not degrade the apeed of voltage polarity re-
versal at their ends, which made it poasible to control the operation of 52 dis-
chargers.
The enhancer discharger reliably operated throughout the entire range of variation
in Up, so that with static and pulsed charging after = 100 microseconda, no poten-
tial difference was produced between electrodes 3 and 4 of this diacharger.
168
FOR OFF[GIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFFICIAL USE ONLY
A similar system can be designed with a large number of cables. A further reducr
tion in T2 can be achieved by the sequential breakdown of several enhancement gaps
[7].
BIBLIOGRAPHY
1. Barnes P.M., Gruber J.E., James T.E., J. SCIENT. INSTRUM., 1967, Vol 44, No 8,
p 599.
2. Bosamykin V.S., Gerasimov A.I., Zenkov D.I. et al., "Gazorazryadnyye pribory
(Trudy konferentsii po elektronnoy tekhnike" ["Gas Discharge Devices (Proceedings
of the Conference on Electronics Engineering)"], Moscow, "Elektronika" All-Union ,,.Scientific Research Institute, 1970, No. 2(18), p 94, 95.
_ 3. Pavlovskiy A.I., Gerasimov A.I., Tananakin V.A., et al., PTE [EXPERIMENTAL
ENGINEERING AND INSTRUMENTS], 1970, No 2, p 122.
4. Pavlovskiy A.I., Gerasimov A.I., Zenkov D.I., et al., ATOMNAYA ENERGIYA, 1970,
Vol 28, No 5, p 432.
5. Pavlovskiy A.I. Kuleshov G.D., Gerasimov A.I., et al., PTE, 1976, No 6, p 134.
6. Gerasimov A.I., Saltykov V.B., PTE, 1979, No 4, p 265.
7. Vorob'yev P.A., Potalitsyn Yu.F., Collected Papers, "Elektrofizicheskaya
apparatura i elektricheskaya izolyatsii" ["Electrophysical Equipment and
Electrical Insulation"], Moscow, Energiya Publishers, 1970, p 140.
COPYRIGHT: Izdatel'stvo "Nauka", "Pribory i tekhnika eksperimenta", 1981
8225
CSO: 8144/0178
169
- FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
UDC 621.383.98
WIDEBAND PHOTOELECTRONIC RECEIVER
Moscow PRIBORY I TEKHNIKA EKSPERIMENTA in Russian No 4, Jul-Aug 81 (manuscript
received 3 Mar 80) pp 199-202
[Paper by R.R. Agiahev, G.I. I1'in and A.N. Pikulev, Kazan' Aviation Institute
imeni A.N. Tupolev]
[Text] The circuitry and operational principle of a photodector
with a photomultiplier are described, which hae a stepped amplitude
response and makea it poasible to receive optical signals in three
modes: with manual setting of the gain, with timewise gain control
and autrnnatic gain control. The dynamic input range is 80 dB, the
passband is 25 MHz and the maximum time for switching fr-om channel
to channel is 100 nsec.
The dynamic range of variation in optical signals fed to the input of a receiver
can reach 80 to 100 dB, and without taking special steps to compress the dynamic
range, this unavoidably leads to the overloading of the optical receiver and inform-
ation losses. The most well-known methods of dyanamic range compresssion are the
introduction of automatic and timewise gain control as well as the use of func.;- o
tional amplifiers; however, the preciaion of ineasurements of the output data in
such optical receivers is poor.
_ One of the promising methods of dynamic range compression with high measurement
precision and a large dynamic range of the signals being measured is the technique
of atepped gain control. However, there are no eufficiently universal photodetec-
- tors at the present time which employ the given principle of analog signal com-
pression. The circuit propoaed in [1]' for stepwiae automatic gain control (a.g.c.)
is convenient when working with suEficiently "smooth" signals, but cannot be
satisfactory when recording rapidly changing signals. The uae of delay lines
[1, 2] to prevent information losses during awitching leada to a narrowing of the
passband of the entire device. Moreover, delay linea are inefficient in the case
of repeated switching. A photodetector with timek�ise automatic gain control
(t.a.g.c.) yieldsi:Ghe best results [3], in which it is posaible to select tr.e
re-quisite portion of the signal by setting a definite program. it is necessary in
a number of other cases to analyze the entire signal at a fixed gain. Then it
becomes expedient to manually control the photodetector operation: the signal is
picked off;'of the selected dynode.
170
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFFICiAL USE ONLY
(
~
:(1y)
.4,+A4
D1=D4
BZ
(2)_ 470 e�v a~NOa ,u�a a~iioa
Anode
aN.a
(16)
Ba:oe
Output
Figure 1. Functional block diagram of the wide3and photoreceiver.
M1 - M3 = 1551A3; M[F, M5 m 1551A6; D1 - D4 = KD503.
Key: 1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
lE.
19.
20.
21.
Trigger;
Automatic gain control;
Time controlled automatic gain control;
Manual control.;
8th dynode;
lOth dynode;
EP1 = emitter follower 1;
EP2 = emitter follower 2;
Inv 2 = inverter 2;
EPO = emitter follower and limiter;
EP7 = emitter follower 7;
EP6;
Ogrl = limiter 1;
US1 � supplemental amplifier 1;
K4 = awitch 4;
SSV = output matching circuit;
Inv 1 = inverter 1;
GG = "Comb" generator;
RSP = voltage shift regiater;
AD1"m amplifier discriminator 1;
AD3 = amplifier discriminator 3;
171
FOR (1FFICIAL IJSE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFFICIAL USE ONLY
[Key to Figure 1, continued:]
22. SSL1 = logic level matchin g circuit 1;
23, LSVK = channel selection logic;
24. SU1 = control circuit 1.
A photodetector is described in this paper which can operate in any of three
modes: manual control, and with a.g.c. and t.a.g.c. The optical signal receiver
is designed in accordance with the circuit described in [4]. The output signals
are picked off of load resistora inaerted in the dynode circuits of the photo-
multipliers. As atudies have shown, to obtain a dynamic range of 104, the output
signals can be picked off of the 8th and lOth dynodes and the anode of the photo-
multiplier, and a aupplemental amplifier with a gain of 10 can also be used for
the signal fram the anode. In thia case, the relative error is < 5 percent [5].
A functional block diagram of the receiver is ahown in Figure 1. The aignals from
the 8th and lOth dynodes are fed to two identical channels. By way of example, we
shall consider the passage of the signal from the Sth dynode. The signal is simul-
taneously fed to inverter Inv 1 and amplitude discriminator AD1 through emitter
follower EP1. The negative polarity signal fram Inv 1 is fed through emitter fol-
lower EP3 to switch K1. The signal picked off of the photo multiplier anode is
simultaneously fed to emitter follower and limiter EPO and emitter follower EP6.
The voltage from the emitter follower and limiter is fed to discriminator AD3 and
emitter follower EPS, and then to switch K3. The signal from the output of EP6 is
fed to the additional amplifier US1 (to avoid overloading the amplifier by large
signals, there ia a limiter, Ogr 1 in the circuit) and goes through emitter fol-
lower EP7 to switch 4. At each fixed point in time, only awitch Ki is unblocked,
and the signal is fed from it through the SSV output matching circuit to the output
of the device.
We shall treat the operation of the circuit in varioua modes. When the switch
B1 is in the "manual" position, a logic 1 voltage is fed to the sliding contact af
the dynode switch B6. The switch B6 can feed the 1 voltage to logic gates M1-3,
M2_2, M3_1 and M3_4. The voltage which appears in this case at the output of one
of the four integrated circuit elementa M4 and M5 unblocka the correaponding switch
Ki through the control circuita SUi, and the aignal is fed at the selected gain
level to the output of the photodetector.
When operating in the time controlled a.g.c. mode, the switch B1 unblocks gates
M1_1, M1-4, M2_3 and M3_2 with a logic 1 voltage. A synchronizing pulse triggers
the "comb" generator GG, which generatea a series of four pulses. A voltage shift
register (rsp) generates a train of query gating pulses from the "comb", which are
fed through isolaCing diodes D1 - D4 to awitches B2 - B5, which accomplish the
requisite setting of the strobes. The interval between the comb generator pulses
can be varied by means of potentiometera R1 - Rl+, and the channel interrogation
time will change in this case.� Thus, the requiaite amplitude-time characteristic
of the t.a.g.c. is generated. The strobe pulaes in the aelected sequence unblock
the switches Ki in interrogatiag the channels, through the unblocked t.a.g.c. ;gates, integrated circuits M4 and MS and the SUi control circuita.
172
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500024434-7
FOR OFF'TCiAL USE ONLY
When operating in the a.g.c. mode, the logic 1 1eve1 incoming from awitch B1
unblocka gates M1_2, M2_1, M2_4 and M3_3. In thie case, the control of the
ewitchea ia tranafered to the channel selection logic (LSVK). As has already been
indicated, the input signals from emitter follower 1, emitter follower 2 and the
emitter foYlower-limiter are fed to amplitude discriminators ADi,"clesigned around
tunnel diodes. Then the logic 1 and 0 levels.are fed to the channel selection
logic through the logic level matching circuitry (SSL). Depending on whether the
discrimination threshold was exceeded by the signals fram the corresponding diodes,
enabl.e or inhibit signals are fed to the control inputs of the switches Ki through
a.g.c. gates MtF and M5, as well as through unita SUi, where each time the enable
pulse is fed to only one switch.
In the initial atate, the logic links between the channels are designed so that
only one switch K4 is open in this case. If the amplitude of the signal at the
photomultiplier input is amall, then the states of the discriminatora ADi do not
change, and a signal is fed through switch K4 at maximum gain to the device output.
If the signal rises so much that the discrimination threshold of AD3 ia exceeded,
then a 1 signal is fed from the SSL3 circuit, which provides for an inhibit logic
function to awitch K4 and turns on K3. The circuit operates in a similar fashion
with other changes in the input optical signal.
~
~ 3/71 47R
n y2~ 7
1V7n JO~--,
~V7n= 30~- _ =JO
I * ` 3J ~ o
~ o J~ I a d
4
` 1�u ~ 47n ~ T uM. Tt y' 4 h T, JL I MI
~ju ~ F77+-------------- -J
i JlFr NN02 Jny I
' 1S)(6) (7~
I ~
~
I
lunad I I
cay ;
Awnd � � - ~ I
47ii
T JU.T.,7/Z 117n
~
n N
n T
47n 1
~
o ' JD
f~T
T _17 T T
p ~ r~ IU ~ 47n
~f � U `�7
I~ 47 .
0 41, }
c N N A1
f rom
Clly ' t-a m
=47n ~ C Ml.y 5-7
70- 3.7 -!1 r---~y F~
30 --J CM~�~
F ~ ^ 47n r---Cy,
47n rw T$ ~V-t
+ c2f I .
^ o N Jo o ACOK M2-~ Mroa 0utput
ci N ~O q ~ 1 So /
& MaMr, To
M2_1
I � Q9 . 21D ~
NQ Mn .
~ Tli ~61f Tis ^ I K.Mr�y
~ a' As Q6 j & t na dfJ-z t o
IeZn
cy- I , M3-2
ZO a7n JJ r=~T i-----------~
~ ~JJ 12i'~ * � J0
J i
JO - LJO
+12 = ~JO -l1I ~ZZ 1 V'R ~1 1 1
! O JJ ~ -

~
N
T
~ il
30 .
To
47n
N
~ h

tp
N
a
19n y
AJ(r I", CC91 1
IS L---.J L---~
I AJ1; H CCA3 1-----
L-__J
(13) (14)
Figure 2. Basic schematic of the unit containing the analog processing
and switching stages as well as the threshold gates and the
channel selection logic.
173
FOR OFFICIAL USE ONLY
TJ I
~
(12)
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFF7CIAL USE ONLY
[Caption and key to Figure 2, continued:]
M1 = 190KT1; M2 = 155LR1; M3 = 155LA3; T1, T6, Tlp and T13 = KT315A;
T2 = KT316A; T3, T4, Tg and Tq = KT326A;.T5, T7 and T11 = KT361A;
T15 and T16 = KT339A; T12 and Tiq = GT311B; D1, D3, D4 and Dg = KA503A;
D2 = GI304A; DS - Dg = D814D.
Key: 1. EP1 = emitter follower 1;
2. Inv 1= inverter 1;
3. EP3 = emitter follower 3;
4. 8th dynode;
5. Emitter follower 2;
6. Inverter 2;
7. Emitter follower 4;
8. lOth dynode;
9. SUq = control circuit 4;
- 10. LSVK = channel selection logic; 11. SU1 = control circuit 1;
12. SSL1 = logic level matching circuit 1;
13. AD3 = amplitude discriminator 3;
14. SSL3 = logic level matching circuit 3.
- -!2 ~ R, 47x +s ' � p1 ~K+5 R~~+S
~E~ + 1
Tr~gger h ~~Nr- ---M~ t =iz ~ao i
lunycK J:Jx T pl 6 ~ 7Sn rill
YN � k
~
~ ~ n
.
Ry 47n ~ Tr ~ j : l 1. , t
~~~c L------------- n- f.f
r------, .
,~JN +s MJ C, RF cinpo0e~ Strobe
T r. I i Pulses
~
i ; Tv rs T. T?~ ~ I ~ J?
I i mn Mt.i N ~t y 4
~ ~ tS ov
� i /00 +S
� ~cepor '
Reset �
Figure 3. Basic shematic of the GG "comb" generator and the RSP
voltage shift regiater. MY and M2 = 155LA3; M3 = 155IR1;
T1 and T3 - T7 s KT361A; T2 and Tg = KT315A; D1 = KD503A.
A basic schematic of the unit in which the analog processor stagES, the switching
_ stages as well as the threshold gates and LSVK channel. selection logic are
174
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
- included is shown in Figure 2. A basic schematic of the GG "comb" generator and
the RSP voltage shift register is shown in Figure 3.
- /
U66lx ,8 Uout' volts
Figure 4. The amplitude-light response
of the photo receiver when
operating in the a.g.c. mode.
Phrel. is the normalized
light flux at the photo-
multiplier input.
0,1
0,01,
t
During testing of the photoreceiver in the a. g. c. mode, its amplitude-light re-
sponse was recorded, which is shown in Figure 4. The characteristics were re-
corded with an FEU-84 [photomultiplier] in a monopulse mode. The light signal was
generated in the form of a pulse 20 nsec to 1 usec wide by a high power LED, to
which a voli.age was fed from a G5-56 generator. The output signal was monitored
with a S8-12 oscilloscope.
The all-purpose photoreceiver makes it possible to register signals with a dynamic
range of 80 dB and t;ie passband of the devices is 25 MHz. The switching time from
channel to cha.nnel is < 100 nsec. When operati=ig in the t.a.g.c. mode, the device
makes it possible to produce an arbitrary amplitude-time characteristic with an
operating time for each of the four channels of 0.5 to 5 usec.
BIBLIOGRAPHY
- 1. Aver'yanov G.A., Safronov Yu.N., Savichev B.M., PTE, 1976, No. 4, p 162.
- 2. Brunov M.V.., Golovkov V.P., Corbatyuk V.V., et al., PTE, 1975, No 6, p 137.
3. I1'in G.I., Pikulev A.N., Pol'skiy Yu.Ye., PTE, 1980, No 5, p 199.
4. I1'in G.I., Pikulev A.N., Pol'skiy Yu.Ye., Collected Papers, "IV Vsesoyuz.
simposium po rasprostraneniyu lazernogo izlucheniya v atmosfere" ["Fourth
All-Union Symposium on Laser Radiation Propagation in the Atmosphere"], Tomsk,
Institute of Atmospheric Optics of tre USSR Academy of Sciences Siberian
Department, 1977, p 55.
~ 5. I1'in G.I., Collected Papers, "Vsesoyuz. simpozium po lazernomu i akusticheskomy
zondirovaniyu atmosfery" ["All-Union Symposium on Laser and Acoustic Sensing
of .*.he Atmosphere"], Tomsk, Institute of Atmospheric Optics of the USSR Academy
of Sciences, Siberian ilepartment, 1978, Part 4, p 138.
COPYRIGHT: Izdatel'stvo "Nauka", "Pribory i tekhnika eksperimenta", 1981
8225
- CSO: 8144/0178
175
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
" 10 !00 f000 10000 4omN

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500024034-7
b'UR OFFICtAL USE ONLY
UDC 621.376.5 (088.8)
DEVICE FOR CONTROLLING PIN-DIODE ATTENUATOR
Moscow OPISANIYE IZOBRETENIYA 746784 in Russian 17 Jul 80
[Description of USSR Patent No 746784 by A.V. Andriyanov, A.V. Goryachev and
G.B. Dzekhtser, filed S Jan 78, published 17 Jul 80, ciass H 01 P 1/22
x 01 ? 1/401
[Text] This invention is in the area of radio engineering and can be used to
construct high-speed pin-diode attenuators in the microwave band.
There exists a device for controlling a pin-diode attenuator which contains series-
connected master oscillator, delay element and first auxiliary oscillator, the
output of which is connected to the control input of the pin-diode attenuator; be-
- tween the other output of the first auxiliary oscillator and the control input of
the pin-diode attenuator is connected an element with two st3ble states, and a
second auxiliary oscillator is connected to the control input of the pin-diode
, attenuator [lj.
However, this device is not fast enough.
The purpose of this.invention is to reduce the duration of the leading edge of the
attenuator output pulse.
In order to accomplish this, the device for controlling a pin-diode attenuator,
consisting of series-connected master oscillator, delay element and first auxil-
iary oscillator with output connected to the control input of the pin-diode
attenuator, has an element with two stable staCes connected between the other
output of the first auxiliary oscillator and the control output of the pin-diode
attenuator; a second auxiliary oscillator is connected to the control input of the
oin-diode attenuator, the output of the mariter oscillator is connected to the other
input of the 6lement with two stable states, and an additional delay element is
connected between the output of the master oscillator and the input oi the
second auxiliary oscillator.
Thd drawing shows the functional diagram of the proposed device.
The pin-diode attenuator controller contains pin-diode attenuator 1, master
_ oscillator 2, element 3 with two stable states, f irst and second auxiliary os-
cillators 4 and 5, delay element 6 and additional delay element 7.
176
- FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102/09: CIA-RDP82-00854R000500020034-7
FOR OFFICIAL USE ONi,Y
The device operates as follows.
.
When a pulse is output by mastQr oscillator 2, element 3 with two stable states is
in a state which allows the working current to pass through the pin-diodes, as a
result of which there is practically no microwave signal at the output of diode
attenuator 1. The leading edge of the pulse coming from master oscillator 2 trig-
gers additional delay element 7, simultaneously switching element 3 with two stable
states to a state which cuts off the working current to the control input of pin-
diode attenuator 1(a slight cut-off voltage is usually applied to the diodes in
this case), as a result of which the charge begins to be dispersed from the base of
the pin-diodes. Sometime later, additional delay element 7 triggers second auxil-
iary oscillator 5, which outputs a powerful enough.pulse to the control input of
pin-diode attenuator 1 to accelerate the dispersal of the accumulated charge from
the base of the pin-diodes, as a result of which the output microwave power level
- rises rapidly to its maximum value. Additional delay element 7 is adjusted so that
when the pulse is applied a significant portion of the accumulated charge is dis-
persed from the base of the pin-diodes, but the microwave power level at the output
of pin-diode attenuator 1 does not exceed the value from which the counting of
the length of the leading edge of the pulse begins (for example, 0.1 of the ampli-
tude). As a result of this, the leading edge of the microwave power pulse is
short, and the speed increases. The cutoff of the pulse from master oscillator 2
activates delay element 6, which, in turn, activates first auxiliary oscillator
4 somewhat later. A pulse is applied to the control input of pin-diode attenuator
1, providing rapid charge accumulation and, consequently, short cut off of the
microwave power pulse.
Simultaneously, first auxiliary oscillator 4 acts upon element 3 with two stable
states, resetting it to its initial state. Delay element 6 can thus also act as
a pulse width regulating element.
The proposed device differs from the other existing device in that it is signift-
cantly faster, and has the capability of providing shorter microwave power pulses
and, consequently, a wider range of modulating frequencies during pulse modulation.
Furthermore, it also provides maximum suppression of the microwave signal during
pauses.
Patent Claims
A device for controlling a pin-diode attenuator containing series-connected master
oscillator, delay element and first auxiliary oscillator, the output oF which i.s
connected to the control input of.the pin-diode attenuator, zaith an element with
two stable states connected between the other output of the first auxiliary os-
- cillator and the control input of the pin-diode attenuator, and a second auxiliary
oscillator connected to the control input of the pin-diode attenuator, distin-
guished by the fact that in order to decrease the duration of the leading edge of
the attenuator output pulse the output of the master oscillator is connected to
the other input of the element with two stable states, and an additional delay
element is connected between the output of the master uscillator and the input of
the secon3 auxiliary oscillator.
177
FOR UFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFIC'IAL USE ONLY
Lt
Information Sources
Used in Evaluation
1. Dzekhtser, G.B., et. al. "P-i-n-diody v shirokopolosnykh ustroystvakh SVCh"
[Pin-diodes in Wideband Microwave Devices] , MoscoFr, Sovetskoye radio, 1970,
p 160, figure 4.19 (prototype).
6900
CSO: 8144/0218-B
J
178
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/42/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
unc 62t.389
SILICON RADI(3-FRDQUIIJCY P-I-N DIODES FOR COMMUNICATION5 IQUIPHOT
Moscow II,MR OSVYAZ' in Russian No 8, Aug 81 (manuscript received 15 Oct 79)
PP 51-52
[Article by Yu. R. Nosov]
[Text] Silicon pin diodes are used widely in radio and xire communications. They
are used in the SVCh [3uper-high frequency] range as switches, modulators, attenu-
ators and power limiters. Lately pin diodes axe being widely used in the VCh
(High-frequency) range in vaxious ARU CAutomatic gain control] circuits, variable
rsctifiers etc. In controlled attenuators, pin diodes are capable of providing
ARU, both flat and sloped of up to 60-70 db xith very high linearity and with a
minimum of introduced noises. However, the development of such domestic devices
was retarded by the lack of corresponding components. The proposed diode one
of the first such types of domestie diodes has good electrical characteristics.
!n recent year, the use of ailicon sxitching and limiter diodes with a pin structure
find wider and wider application. It was found that in the VCh range, such devices
may be very useful in ARU and power limiter circuits etc.
Fig. 1 shows a KD413A, B diode: a-- exterior viewi b-- the pin structure cir-
cuit; c-- the distribution of chaxge carriers in the switched-off atate (solid
line) and at various forward currents (broken lines); d-- equivalent circuit of
the pin diode (1 Cdif' z Ctri 3-' Ck are diffusion, junction and device
capacitances respectivelyt Lk inductance of the device), 4-- rdif is the
differential resistance of the diode, 5-- If is forward current. The pin diode operates as follows: At forrtaxd current flowing through the diode,
_ an intensive "pumping" of chaxge carriers into the base low conductivity region
(i region) and their accumulation in it are implemented due to the effect of a
double injection from the strongly alloyed n+ and p+ regions. First, at fairly
high frequencies, the rectifying action of p-i and n-i junctions ceases (diffuston
capacitance Cdif in Fig. ic is large) and the high frequency aignal can pass through
the diode without significant distortiona. Secondly, the reaistance of the base
region becomes small and ita value becomes dapendent on forward current. Thus,
- in some frequency range, the pin diode is a current controlled resistive component.
- 179
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFF(CIAL USE ONLY
~a~ �
(b) 61 1 n+
p.n (5)
~C~ D} Inp
. ~
CAM0 (4) X LK
(d) rA"O
t)
Puc. 1 r ne 2) CK (3)
Fi6. 1
Table 1
Electrical paxameters of KD413A. B devices
Diode type Constant forwaxd voltage
at If 20 ma, Uf, volts,
not over
KD413A
tcn413s
Diode
Capacitance
at Urev-0'
volta,
Cd, pico-
farads, not
over
0.7
0.7
Differential Switching
resistance at charge at
If 2ma and !f 2ma,
f=50MHz,
-10
U
rev
rdif'
ohms
volts,
not
not
Q
, nC,
less
tnore
$
not less
than
than
than
30

60
2
40
80
2
1
i
iso
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500020034-7
FUR OFFIC(AL USE ONLY
The operating mechanism cited is achieved in type KD413A,B devices whose electrical
parameters and maximum allowable operating modes are shown in Tables 1 and 2. The
lower bounda.ry of the working frequency range is determined fairly strictly by
switching char.ge Q s . Its value for the pin structure is related to the effective
lifetime of chaxgecarriers in the base T'eff by a simple relationship
Q s ~ jf ~ eff
Table 2
(?perating modes of KD413A,B devices in the -60 to + 100OC temperature range
Name of mode
itilaximum allowable constant or
average forward current
Nlaximum allowable constant
reverse voltage
Niaximum allowable scattered power*
Conventional designaticn
I f~7iax
Urevma'x
P max
Allowable
Value
20 ma
24 volts
20 inw
*The power liberated at the diode is calculated frunapproximate formula
P=IB..tf + I2x rdif , where If forward bias current, 1-- effective value of the
high frequency current through the diode, rdif - differential diode resistance at
the selected value of If. Condition P,< Pmax must be fulfilled.
The results of experimental investigations of the devices described are shown in
Fig. 2.
X!
B
6
4
2
qn,NIG(4) 2 1 '3
/
0 5 10 15 I4im
Fig. 2. 4-- Qg, nC; 5-- If, ma,
(5)
181
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/49: CIA-RDP82-00850R440500020034-7
FOR OFFICIAL USE ONLY
" In a broad range oF forward current changes, 'Ceff may be considered constant and
lying in the interval of 1 to 2 nicroseconds to a precision of up to 10-20%. The
_ cohdition for determining the loW value of the xorking frequency has the formt
Wmin Z eff n;>7
where the value of n dependa on the allowable level of signal distortion (incomplete
disappearance of the rectifying action of the pin structure) and, in the majority
of practical applications, may be assumed to be n=30...100. Thus, for the KD413A,
B diodes wP have
fmin nIf N 3...10 megaHz.
21T Qs
The upper boundary of the working range is determined by parasitic reactances
(inductance, capacitance) of the design and of the pin structure itself and lies
withln 300...500 megaHz. A typical value of the full capacitance af the diode at
fairly high reverse biases ( N 10 volts) is cloBe to 0.1...0.2 picofarad.s; the
parasitic inductance of the structure is 2 nanohenries.
T'he basic parameter of the pin diode as a resistive component is its differential
resistance at a given direct current measured at high frequenciea (f=50 megaHz)
and the form of relationehip rdif gId'
To simpl.ify the selection of the operati,,g mode of the device xith respect to the
forward :surrent, a two-sided limitation on the rdif paxameter is introduced (Table
1); its actual spread is still smaller.
o
97
60
40
20
70 rA',0'M` 5l
Fig. 3. 3-- K1413A; 4-- KD413Bi 5-- rdif' ohms=
182
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
,30 40 JO 60

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R400500020034-7
FQR OFFICIAI. USE ONLY
Fig. 3 shows the diode distribution function with respect to rdif valuea. 1, 2
are bounda,ries of a 95 percent spread. In the region of value change of If b,y
3...4 orders of magnitude (from 5...10 microamperes to 10...20 milliamperes), the
relationship between the differential resistance and the constant forward current
is described to a hign degree of accuracy by
rdif AjIf
where A=70...110 millivolts for group A and A790...140 millivolts for group B.
3tructurally, the KD413A,B diodes are made in the form of miniature all-glass
1.2x2.8 mm cylindez�s with axial lead-outs and are suitable for xiring in hybrid
microcircuits.
= COPYRIGHT: Izdatel'stvo "Radio i svyaz l", "Plektroavyaz l", 1981
2291
CSO: 1860/78
183
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
UDC 621.391:621.317.75
PRTNCIPLES Ok' THE DESIGN 017 NONLINEAR PUI,SE-FR,EQUENCX SIGNAL SEI,ECTORS
Moscow IZMERENIYA, KONTROL', AVTOMATTZATSIYA in Russian No 4(38), 1981 pp 54-61
[Article by V.I. Korchun, engineer]
[Text] Devices possessing the properties of frequency selection are most often
designed on the basis of classical analog structures (line filters) which, because
of their simplicity and universality, are widely used in various fields of engineer-
ing. However, in recent times the attention of inveatigators and developers has
been attracted by methods of designing nonlinear frequency-selective devices based
on digital logic circuits [1]. In these devices the methods and principles of
pulsed and digit. , equipment are used primarily; therefore, only digital circuits
and components required for implementing them. Interest in these methods can
be explained by following reasons:
The modern integrated technology for fabricating the components of electronic cir-
cuits is more compatible with engineering solutions basedon employing the methods
and principles of digital equipment, and not analog.
Frequency selection devices designed on the basis of digital logic elements make it
possible to obtain specific values of a number of parameters and characteristics
(speed of response, selectivity, stability of characteristics, simplicity of re-
tuning, etc.) which are difficult or impossible to obtain by means of analog line
circuits [2]..
_ A great number of terms are required for designating nonlinear frequency-selective
devices based on digital logic circuits, such as "nonlinear digital filter,"
"numerical filter," "discrete filter," "repetition rate pulse selector," "frequency
detector," etc. Taking into account the "act that in designing these devices digi-
tal principles of isolating the frequency trait of pulped signals are employed,
below we will use the teria "pulse frequency signal selectur" (SChIS), limiting our-
selves to the class of nonlinear devices.
SChIS's are used i.n various automation, computing and measuxing equipment devices.
They sxe uzied in pax'ticular for the precise fixing of the moment of time when
a parameter, varying over time, of a proceas characterized by a frequency tratt
reaches a prQdetexmined value. The basic functions of SChIS's are to analyze
electrical signals belonging to a specific frequency region and fixing the
moment of ^_rossing specific chreshold values.
184
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
Let us noze that in spite of the common natuxe of pxobl=a sal.ved by means of
- SChIS's and ].ine filtex'ing devi,ces it ts impos&ibl.e to dxaw a paxa7,iel between
these two classes of devices. Finding a sqathematica7, squiva7.eat of an SCUS in
the categorias customaxy fox the tlteox}r of anal,og ft1.ters is quite a comp].ex pxob--
lem which sti11 cannot be considered so3.ved. Nevertheless, fox the purpose of
characterizing an SChTS it is possible to select paxatneters (cutoff frequency,
transmission band, mean transmission fxequency, equivalent figure of inerit, etc.)
which in the physical sense are sufficiently close to the corresponding parameters
of line filters. This makes it possible to compare these devices to some extent.
A characteristic feature of nonlinear SChIS's is the ..:bsence of a distinctly pro-
nounced transition region between the disappearancs and appearance of the device's
output signal. Therefore, by the cutoff frequency, f , is understood the boundary
value of the input signal's frequency with which the existence of a reaction in the
SChIS's output is possible. The transmission band, 2Af , is defined as the differ-
ence between the high-Prequency, f2 , and low-frequency, f, limits of the region
of existence of the SChIS's output reaction, and the mean iransmission frequency,
fo , of band selectors is found from the equation:
fo= fi+i~ .
2
The sel.ective properties of an SChTS can be characterized by the percentage ratio
of the transmission band to the mean frequency, or by the equivalent figure of
i merit, Qekv ' determined by the expression
Q.p a f0 s fl + f! .
2e/ 2(f.-/1)
(1)
The physical meaning of this parameter differs from the concept of the figure of
merit of resonance systems; therefore, the term "equivalent figure of inerit" is
used.
The operating principle of an SChIS is based on comparing the time interval of
the input pulse signal with the time parameter fixed. For the purpose of revealing
I the general properties and features of various SChTS structures, in this study
; the classification of selectors is based on the principles of the �ormation of the
~ fixed time parameter and methoda of implementing the comparison operation. It is
pogaible to distinguish between the following groups of SChIS's according to this
~ classification: SChIS's with a signal delay for a fixed time [3]; SChTS's which
discriminate a time interval of specific duration [4]; SChIS's with direct compari-
son of the time interval o,f the input signal with the dvxation of a reference pulse
[5]; and SChIS's which compare the measured and referer-Lce frequencies [6].
Depending on the requirQd kind of output signal, the implEtieutati.on of each of the
design principles Por SChIS's is accompliahed by several methods, whereby special
185
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-40850R040500024034-7
FOR OFFICIAL USE ONLY
attention is paid to imp.raving the accuxAcy* of J'ixi,ng specific thXeshold va7.ues of
the frequency [7] and also to impxoving the noise immunity of the device [8].
SCbIS's wifih De1,ay ot the Signal for aFixed Time
A structural diagram and time di$gra;ms of the opexation of a selector making it
possible to discriminate pu7.ses reQeatipg at a frequency of f are shown in fig
la and b. The controlled pul,se train entexs the input of a deQay 1ine, LZ, with
a frequency of f(pexiod T- 1/P ) and one of the inputs of an AND gate. An
input signal delayed for a time of T= T enters the other input of the AND gate.
If the repetition frequency, f, of input pulses equals f0 , where f~ = 1/T =
= 1/T , then the delayed pulses are synchronized with the following input puQses,
U kh Z Furthermore, a pulse signal, Uvyh, with a#requency of f0 is formed in
t~ie output of the AND gate. It is easy ~o verify that the SChTS admits without
hindrance also those pulses of multiple frequencies (f - IcfO ; k= 2, 3, 4, )
the distance between which equals T0 .
-i
1~ Yex 2)Ue61x
a- N
H
3 ) /1.7 Ur
4) VJ'To a)
lir
b)
Key:
Figure 1. Circuit (a) and Time Diagrama of the Operatinn (b) of an SChIS
with a Delay Line
2. Uvkh [input]
vykh [output]
3. LZ [delay line]
4. TZ
The transmission band of the sel,ector at the mean fxequency of f 0 depends on the
duration, Ti , of input pulses:
1
2 = 1 I 12zR
~
-
T~ - zr T. sr 77.~
The equival,ent figure of inerit, Qekv ' accoxding to equation (1), also depends on
the duration of tnput pulaes:
186
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE ONLY
Qa~s ~ T0 .
2 Sn
Because of the tnfluence of transient pxoces-ses in the ANA gate and delay line,
LZ, as we11 as because of the effect of destabilizing factors causing instability
of the delay time, 'Cz , the duration, Ti , of input pulses must not be less than
a certain value. On the other hand, i:ncreasing the durati:on of input pulses re-
sults in a reduction in selectivity (selection accuracy). Therefore, it is neces-
sary to select duration Ti by taking into account these facts [9].
Since the SChIS (cf. fig 1) has an excepfiionall}* high speed of response (the time
for responding to the appearance of frequency f m f0 does not exceed the period
of the input signal), then the appearance of even a single noise pulse removed
from the information pulse by a time of T< t< T + T will result in the
appearance of a spurious output pulse. Fo~ the purpose of improving the noise im-
munity of SChIS's of this type it is possible to incrnase the number of inputs of
the AND gate to 3 to 10. With this signals with an tncreasing delay, =~T~~p ,
~ TZ = 2T0 ,'rZ3 = 3T0 , etc., are supplied to each successive input of l ~Fie AND
gaie, for which purpose a sectional delay line is most often used [10]. Combina-
tion logic circuits (KLS's) containing AND gates, OR gates and flip-flops are
often used instead of an AND gate [11].
With a high level of pulse noise and the presence of omissions of information
pulses in pulse trains being discriminated, instead of a delay line controlled
delay generators and gate generators synchronized with one another are used [12].
Selection is accomplished by gating the input signals with pulses of a set dura-
tion which follow at intervals of Tp By repeated triggering of the delay gene-
rator and gate generator it is possible to check the first signal picked up re-
peatedly for truth. At the same time gate pulses can be generated also with in-
dividual omissions of information pulses (up to a certain predetermined number of
information pulses omitted in succession), because of which tbe signal pickup time
with fading of the pulse train--the SChIS's readiness time-��is reduced to a mini-
mum (on the order of the period of the input signal).
A general shortcoming of SChIS's with delay of the signal for a fixed time is the
existence of a reaction at frequencies which are a multiple of the assigned. This
shortcoming can be eliminated by means of the method of n-fold comparison per
period [13], whose essence consists in the following. To a multi-ingut AND gate
-is fed an n-multiplet of additional pulses with intervals of T whose formation
is possible only upon condition of the arrival in succession of n+ 1 pulses
- of the controlled train with a period of TO . The method is implemented by means
of a so-called deluyed xegeneration loop executed on the basis of a sectional
delay line, registers and combination logic circuits. The time for the appearance
of a pu].se reaction in the output of an SChIS of this sort from the moment the
_ input signal is supplied equals nT0 .
187
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE ONLY
Thc ].owex limit o f the firequancy band o.E these SChZS, a ia determined by the maxi-
mum achievable delay CiuAs, , and the upper l.imit by the speed of xesponse of
logic elements. and it can xeach dozeas of mega,hextz, which taakea it possible to
use the selectors in vaxtous radio engineering devices..
SChTS's with Discriminatipn Qf aPertod of Specific Duration
Selectors of this Cype are designed on the basis of the principle of digital mea-
surement of the period of the input signal. Passage of the input controlled pulse
train to the device's output is permitted or inhibited depending on the result nf
- the measurement. A structural diagram and operating diagrams of a an SChIS for
narrowband selection are shown in fig 2a and b.

6) YM U~ nh ~ c ' 11)
1) M I QC N
a7
~
~M
~y~_~~~ 1 1 1 1
~nlllllllllll(llJlllllllllfllll _
r I111) 1111I II11 11111 111
rc 1 I I
_ - n
v'ee~x ~
b,)
Figure 2. Circuit (a) and Time Diagrams of the Operation (b) of a Narrowband
Digital SChYS
Key:
1.
2.
3.
4.
5.
jIIhgjtePutl
Countex
Decoder
Uvykh [output]
6. Bas.e pulses
7. Control signaJ.s
8. Refexence vol,tage
9. Timer
10. Clqck pu].ses
11. Gattng signal
12. D3:fferentiation
circui:t
188
FOR OFFICIAL USE QNLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFiCIAL US1E ONLY
From voltage U of the control].ed train, by means ot a one-shot multivibrator,
M, base pulses;'~ , are formed, with a duration of T. In turn, the base
pulses, U, are dTfferentiated hy means of dtfferentiahon cixcuit DS , in
- whose output are formed control signals, U , of short duration, which coincide
in time with the fiexmination of the ba,s.e Ases. Havi,ng received the next con-
trol signal, U , the pulse counter, ST , is reset to the zero state and auxili-
ary flip-flop ~r by means of the same signal is set to the "one" state at the
S input, thereby permitfiing the entry o,f clock pulses, U ,from the output of
the referPnce generator, GTT (with a frequency of f )~~irough the AND gate into
the counting input of the counter, ST . The decoder�,PDSh , analyzes the state
of the n outputs of fihe counter and f ixes the moment of the accumulation of
the set number, N, of pulses (in fihis case N= 5), thereby forming the gating
signal, U. With the arrival of signal U flip-flop T is set to the zero
state andscloses the AND gate, at the same hme inhibiting the entry of clock
_ pulses, Ut , into the input of the counter, ST . When signals U and U are
synchronized (for the purpose of clarity these signals are presenLd on the same
diagram) the output AND gate forms a pulse reaction, U k}, The counting process
_ begins from the start with the arrival of the next con~oI signal, Uupr '
From the time diagrams (cf. fig 2b) it follows that with an assigned duration,
T, of base pulses, U, a fixed value of the frequency, f , of the timer,
- G~I , and a set counter state decoding number, N, the syncgFonization of signals
US and Um and, accordingly, the appearance of signal U kh~~ in the SChIS's
ou_put are possible only in a limited region of input signa'~,--Uvkh ' frequencies
[4].
The mean transmission frequency, the effective transmission band and the equi-
valent figure of inerit of the SChIS are expressed, resper*_ively, in the following
manner:
f c, n 2 N- 1-}- Tn f cn .
f0 2 N (N - 1 -f- Tn /oo) ~ (2)
zn jon -1
- 2 Af _ f~ N (N - ! sn f on) ~ (3)
Q. = 2 N- 1-- sn /on (4)
2 zn / �
( on-1)
This circuit for an SChIS, as is obvious from the equations given, with the
appropriate choice of the values of N, T and f , can make possible,
_ generally speaking, as high indicators of Belection�~arameters in the narrowband
region as desired. As an example of the practical application of an SChIS of
this sort can be cited the typical problem of producing frequency markers for a
- narrowband swept-signal source. In the 215 to 217 kHz band the selector makes it
possible to produce: markers with an error net greater than 2 Hz [2].
An analysis of equations (2) to (4) demonstrates that for the normal operation of
the SChIS (fig 2) it is necessary to fu1�i11 the co-.ditions f�f and N�
� T f > 1. Since the value of the upper limit of Prequency fOP is aetermined
by tReoRpeed of response of the digital eJ.ement base and equals (fooP the most
189
FOR OFFIC'AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE ONLY
widespread series of digital e1,emenCs) 10 tn 50 Mfiz, then it i.ss, feasih7,e to i.m-
plement according to zhis circuit an SChTS in the 100 to SOQ kHz fxequency band,
This limit can he incxeased by one to two or-dexs af znagnifiude in the immadiate
future [14].
In the transmission band the output signals of this $ChlS can difi'er from the input
in terms oP the shap.e of pulses. Tn the majority of cases the following units of
- equipment are constructed on the basis of the direct use of tYie pulse reaction and
additional conversion is not required. Fiowever, sometimes the problem is posed of
discriminating pulse trains with a speciPied pulse repetitian rate without their
distortion j15].
In measuring equipment and in pattern recognition systems in signal processing the
need often arises for the wideband selection of pulses, whereby the upper cutoff
frequency is one to two orders of magnitude higher than the lower frequency. This
- type of selection can be accomplished by means of the SChTS whose structural dia-
gram and operaCing diagrams are shown in fig 3a and b. Unlike the preceding cir-
cuit, here two different states af the pulse counter are decoded, characterized by
the numbers K and N, whereby K� N. Number K is selected from the con-
dition f= f /K , where f is the reference �requency and f 2 is the upper
cutoff frequen�ci and number RP is selected from the condition fl = fop/N ,
where fl is the lower cutoff frequency.
The selector operates in the following manner. The circuft is set to the initial
state by the trailing edge of the base pulse, U(cf. fig 3b), after which the
counter, ST , begins to be filled by pulses, U m, from the reference generator,
GTI . During the time corresponding to fillingpof counter K with pulses, decoder
DShl by means of its output signal, U1 , resets flip-flop T1 to the "one" state.
If frequency f of the input signal sat-.isfies condition f< f< f , then by
the moment of the arrival of the next pulse of the input signal, U kh one"
levels will be present in the direct output of flip-flop T1 anu ~n the inverse
output of flip-flop T2 . The passage of input pulse Uvh through the AND gate
to the output of the SChIS is permitted by "one" levels o~ signa],s Utl and U
If f> f , then by the arrival of the next U kh pulse counter ST is not'atie.
to number2 K pulses and a"zero" level is main'ained in the output of flip-flop
T1 , inhibiting passage of pulse Uvkh through the AND gate.
If f< f , then counter ST is able to number N pulses. By means of signal
Ut2 fromlthe butput of decoder DSh2 flip-flop T2 is reset to the "one" state,
and signal U 2 (zero level) from the inverse output of the flip-flop inhibits
the passage o~ the next input pulse, Uvkh ' through the AND gate.
Thus, input pulses are transmitted to the output of the SChIS only upon the condi-
tion that fl < f< f2 .The selector makes possible the wideband selection of
pulse signals with easily retunable valuea of threshold frequencies. Improvemerct
- of accuracy at the J,imits of the transmission band is achieved by increasing the
' reference fzequency, f oP , and x'educing the duration, TP , of the base pulse, Um
[16].
The time for the appearance of a pulse reaction in the output o!: an SChZS designed
according to the principle of discriminating a pexiod of specified duration is
190
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500020034-7
FOR OFFIC'(A1, l1SE ONI.Y
not greater than the va],ue of TR 1/R , where f~ Rf f2 . Her.e sp~.itting of
_ the selected tiiae interva~. between adjacent pu7.ses, caused bythe supexposition
of pulse noise, reault5 in disabling of fihe SChTS and not on7.y pu9,-e noise but
alao the Zegitimate signa7, do not travel, to its output. Fox the purpvse of elimi-
nating this shoxtcoming special measures fqr bZ.ocking spuxiaus s.i.gnals are used
which complicate considexably the selector''s czrcuit [17].
a)
V�=rt
n n
_
rr"' ~

zrynP
-
z'' I t
-
t7i
~r~
r"X n n
~
b)
Figure 3. Diagram (a) aiid Time Diagrams oi: the Operation (b) of a Wideband
Digital SChIS
[Cf. fig 2 for key]
SChIS with Direct Comparison of the Period of the Input Signal with the Duration
of a Reference Pu1se
The operation of these SChIS's is based on utilization ok the phase relationships
between pulses of the input signal, and reference gulses [18]. The basic diagram
of the selector and the time diagrams explaining its operation are presented in
fig 4.
191
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
Vk 2 y UM D 5,V e.,x (f >A)
1) 3) T
~ u
eWx
~ a)
Ua . ~rfQ I a single pul.se is
~ formed fox each paix of input pu7,ses Jin the ci,rcuit"s output. With ffa7
f f , then a carr3~ pulse appears earlier in the output of divider DCh2 ,
settin~ flip-flop T into the "one" state. By the same signal divider DChl is
zeroed and, in addition, the presetting unit, BU , is triggered. The initial num-
ber, Nl, is entered ir~ counter-divider DCh2 by means of unit BU . Then flip-
flop T is in the "one" state until the controlled frequency becomes-less than
f (N - Nl)/N . Thus, the seleczor has hysteresis, which makes it possible to
- eQiminate erroneous operation at the cutoff limit from phasa noise. The value of
. the hysteresis is determined by initial number N1 and can be variea over a wide
range [25].
6e~xoa
- i) /'TN ,{on 441 R T
2) 6)
~ B,roa(f)
4) QV2
5)
I16yl
Figure 11. Diagram of SChIS with Frequency Dividers
Key:
1. Clock 4. Input
2. Reference frequency 5. Pxesetting unit
3. Frequency divider No 1 6. Output
The selector executed according to the circuit presented in fig 11 has a potential
"one" output with f?f and a."zero" with f 4!' . When it is necessarq to
produce an undistorted F or f pulse txain in t~iR output, SChTS's are supple-
menteci with switching cixcuits co�Rtxolled by the outputs of the f1ip-flop [26].
- The noise tmmunity of the se7,ector tncxeases with an inccease in scaling factor
N. However, it must be kept in mind tha,t the speed of response of the device is
198
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500020034-7
FOR OFFICIAL USE ONLY
reduced since the Cime fox the appearance of a x'a4ction in the autput of the
circuit fxom the moment og the suppl}* of the input si.gnal equals N(f .
This prfncipl,e of designing SChIS's makes i:t possib7,e to implement a1.so devices
for the band selectiort of pulses, which are designed on tlie basi.S of genexal-
purpose binary counters aperatiag in the krequency di;vision mode 1271. The
controlled pulse train with afrequency of P(cf. fig 12) 3s supplied to the
counting inputs of counter ST1 with a scaling factor of M and of counter ST2
with a scaling factor of N, whereby N> M. Reference frequenc}r pulses, fop ,
differentiated b}r means of the DS talock, are supplied from the output of
auxiliary generator GTT to the setCtng inpuvs of counters ST1 and ST2 .
After the next setting to the zero state counters ST1 and ST2 begin to be
filled with input signal pulses. with a frequency of f(f � f After period
of time M/f , a carry pulse appears fn the output of counter ??1 , which enters
the counting input of counter ST3 (resulta:nt pulse counter), whose reset input
is connected to the output of counter ST2 � The condition for formation of a
carry pulse in the output of counter ST1 , M/f < 1/f ; and the condition for the
absence of resetCing of counter ST3 to zero by the �cRrry pulse of counter ST2 ,
N/f > 1/f , determine the condition Por the accumulation of pulses in the re-
sultant c�ognter, MfoP < f< NfoP .
1)
Lfl M'"'i A'1
N
3)yon 4)
2 rr FN ,QC
Key:
Ba1xoa
6)
Figure 12. niagram of Band SChTS Employing General-Purpose Binary Counters
1. Input
2. Clock
3. foP
4. Differentiation circuit
5. Counter No 1
6. Output
The parameters of the SChIS are determined in the fol.lowing manner:
Mean transmission .�requency
to = fon M+N
2 r
199
FOR OFFICIAL USE ONLY
(5)
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONL'Y
Transmission band
2Af =fon(N-11),
(6)
Equivalent figurv of mexit
1V! N
Q�cs = 2 (N -M)
(7)
As follows from expressions (5) to (7) the required parameters of the selector
are made possible by selecting the appropriate values of M, N and f .
Thus, with M= 255 and N= 256 (two series-connected standard count�efs, e.g.,
series K155) the relative value of the transmission band equals less than 0.4 per-
cent (an equivalent figure of ineriz of about 255). The disadvantage of this cir-
cuit is its 1ow speed of response: The time for the appearance of a pulse reaction
in the circutt's output equals = K(M/f) , where K is the scaling factor of
counter ST3 . The SChIS has a wYae frequency band, determined by the choice of
the value of the reference frequency and the speed of response of the counters [14].
Use of SChIS's for Processing Harmonic Signals
The design principles for an SChTS discussed call for the input signals being in
the form of periodic pulse trains. For operation with signals whose shape differs
from rectangular (e.g., harmonic signals of frequency-response communications and
remote control systems), it is necessary first to convert the inpl?t signal, using
special shapers (an amplifier-limiter, a Schmitt circuit, comparator and the like),
which make possible the required curvature of the leading edges of the signal. In many cases better results can be obtained when using digital methods of select-
ing frequency signals than when using line filters. For example, a nonlinear
SChIS (fig 13) designed on the basis of two frequency discriminators is economical
and has high selectivity and speed of response. Tt is based on the circuit dis-
cussed earlier in fig 6. A multivibrator assembled from comparators K1 and K2
(series K140 or K157 microcircuits) and fiip-flops T1 and T2 makes possible
the formation of two intercorre3.ated reference pulses of stable duration. Flip-
- flops T3 and T5 form a lower cutoff frequency discriminator and flip-flops
T4 and T6 an upper cutoff frequency discriminator (a11 flip-flops are of series
K176). The mean transmisaion frequency can be varied over a wide range from 200
_ Hz to 20 kHz by changing the values of R and C of the timing circuit. The
circuit's se1,ecr.ivity is detertnined by xatio R1/RZ and does not depend on the
mean transmission fxequency. The typical value o~ Q is on the order of 100.
The speed af response of the selectox is one to two ox~~xs of magni,tude greater
than that of xesonance devices With s3,mtl.ar oelecti,vity. xt is �easibl.e to employ
additional input signal proceasing devices fox the purpose of impxoving the noise
. immunity [28].
200
FOYt OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
Af
Rl
R1
Figure 13. Diagram of Narrowband SChTS
Key:
1. Input 3. AND gate
2. Flip-flop No l. 4. Output
Bibliography
, 1. Mashbits, L.M. "Tsifrovaya obrabotka signalov v radiotelegrafnoy svyazi"
[Digital, Processing of Signals in Radio Telegraph Co=unications], Moscow,
Svyaz1974, pp 56-80.
2. Mashbit-?; L.M., Ya:ishina, G.F. and Chernomortseva, V.P. "Production of
Frequency Markers in Swept-Signal Sources for Observing Frequency Response
of Receivera," VESTNIK SVYAZI.' :io 3, 1968, pp 13-15.
3. Timakhov, O.N. and Lyubclienko, V.K. "Selektory impul`sov" [Pulse Selectors],
Mnacow, SovPtskoyP radi o, 1.966, p 33.
4. Mashbits. L.M. "Analytical ?telationships in Discriminating a Period of
Predetermined Duration in Voltage Varying in Terms of Frequency," RADIOTEKHNIKA,
Vol 24, No 6, 1969, pp 70-75.
5. McKinley, R.I. "Versatile Digital Cir.cuit Piltera Highs, Lows or Bands,"
ELECTRONICS, Vol 44, No 13, 1971, p 66.
6. USSR Patent No 387493. "Device for Comparing Frequencies," I.T. Zimin,
published in B.I. [BXULLETEN'-IZOBRETENIX�], No 27; 1973, p 168.
7. USSR Patent No 721909. "Frequency-Type Pulse Selector," V.I. Korchun and
V.V. Lebed', publiahed in B.I., No 10, 1980, p 210.
8. USSR Patent No 738137. "Pulse Selector," V.T. Korchun and V.V. Lebed', .
publiahed in B.I., No 20, 1980, p 311.
201
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
9. Itskhoki, Ya.S. and Qvchinnikov, N.I. "Impul'snyye i tsiirovyye ustroystva"
[Pulse-Driven and Digital Device&], Moscow, Sovetskoye radio, 1973, pp :;52-
553.
10. Maksimov, M.V. "Zashchita ot radiopomehh" [Protection from Radio Interferenee],
Moscow, 5ovetskoye radio, 1976, p 341.
11. USA Patent No 3732563. "Pulse Train llecoder-Degarbler," G.P. Nelson,
published 8 Ma}r 1973.
12. USSR Fatent No 660223. "Repetition Rate Pulse Selector," V.N. Zaytsev,
- L.M. Zhavoronkov et al., published in B.I., No 16, 1979, pp 276-277.
13. USSR Patent No 660222. "Repetition Rate Pulse Selector," S.V. Popov,
published in B.I., No 16, 1979, p 276.
14. Dubitskiy, L.A., Tafel', V.M. and Shvetskiy, B.I. "New Structures for
High-Speed Decimal Counters," IZMERENIYA, KONTROL', AVTOMATIZATSIYA, Moscow,
TsNIITEIpriborostroyeniya, No 3(11), 1977, pp 48-52.
- 15. USSR Patent No 699665. "Pulae Selector," A.I. Balandin, Ye.P. Larichev and
Ye.N. Okeanov, publiahed in B.I., No 43, 1979, p 236.
16. USSR Patent No 681549. "Pulse Train Selector," Yu.N. Yerofeyev and V.V.
Zaverin, published in B.I., No 31, 1979, p 210.
17. USSR Patent No 681550. "Frequency-Type Pulse Selector," E.Ye. Nesterov,
. published in B.I., No 31, 1979, p 210.
18. Alfke, P. "Frequency Discriminator Uses One-Shot and Flip-Flop," ELECTRONICS,
Vol 46, No 18, 1973, p 92.
19. Yen, T.T. "CMOS Flip-Flop Can Do More than Logic Tasks," ELECTRONICS, Vol 48,
No 6, 1975, pp 123-126.
20. Pearson, E.E. "One-Shot/Flip-Flop Pairs Detp-ct Frequency Bands," ELECTRONICS,
Vol 45, No 9, 1972, p 104.
21. Volk, A.M. "T~ao-IC Digital Filter Varies Pasaband Easily," ELECTRONICS,
- Voi 46, No 4, 1973, p 106.
22. Shah, M.I. "Programmable Monostable Is Immune to Supply Drift," ELECTRONICS,
Vol 46, No 3, 1973, pp 98-99.
_ 23. USSR Patent No 677087. "Device for Comparing Frequencies of Two Pulse
, Trains," L.N. Mel'nikov, A.V. Margelov et al., published in B.I., No 28,
1979, p 202. '
24. USA Patent No 3500069. "Pulse Repetition Frequency Discriminator," R.R.
Iohns, published 10 Mar 1970.
202
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
~ FOR OFFICIAL US1E ONLY
25. USSR Patent No 661761. "Frequency Threshold Device," A.A. samus', published
- in B.I., No 17, 1979, p 263.
26. Andreyev, M.I. "Digital Frequency Discriminator Employing Microcircuits,}'
PRIBORY I TEK.iNIKA EK3PERIMENTA, No l, 1979, pp 1116-118 .
- 27. Neww.in, S. "Tone Detector Sharpens* Digital Fi1tur's Response," ELECTRONTCS,
Vol 52, No 24, 1979, pp 118, 119, 121.
28. Glinchenko, A.S. and Chmykh, M.K. "Digital Device for Eliminating Erroneous
Crossing of a Signal Through Zero," PRTBORY T TEHIiNIKA EKSPERIMENTA, No 2,
1979, pp 112-116. - COPYRIGHT: Tsen*_ral'nyy nauchno-issledovatel'skiy institut informatsii i tekhniko-
ekonomicheskiTch issledo,;*aniy priborostroyeniya, sredstv avtomatizatsii i si,;tem
upravleniya (;'sNIITEIpriborostroyeniya), 1981
8831
CSO: 1860/59
- 203
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
UDC 621.872 (088.8)
DEVICE FOR ADDITION OF POWERS
Moscow OPISANIYE OZOBRETENIYA 733081 in Russian 7 May 80
[Desc.ription of USSR Patent No 733081 by I.V. Lebedev, V.G. A.lybin, D.Ya.
Ashkenazi and V.P. Belyayev, filed 8 Aug 77, published 7 May 80, class
H 03 B 7/14 H 01 P 1/151
[Text] This invention is in the area of radio engineering, and can be used
in variocs microwave devicQS.
There exists a device for addition of powers which contains a.waveguide segment
in the cross-aectional plane of which is located an exciter array in the form -
of inetal rods running parallel to the wide sides of the waveguide segment; diodes
placed on the narrow walls of the waveguide segment are connected to the ends of
the metal rods; furthermore, the middle of the rods are joined by posts which are
perpendicular to them and are connected to the voltage feed elements [1].
However, this device has low operating power and a narrow operating bandwidth.
This is because the diodes in the exciter array are not identical.
The purpose of the present invention is to increase the power and expand the
operating bandwidth. ' .
In order to achieve this, the device for addition of powerg, which contains a wave-
guide segment with an exciter array mounted in its cross section consisting of
- metal rods running paxallel to the wide sides of the waveguide segment with
diodes fastened to the narrow walls of the waveguide segment connected to the ends
of the rods; furthermore, the middle of the rods are joined by posts which are
- perpendicular to them and which are connected to the voltage feed elements, and
each post and rod of the exciter array separated by an additional dielectric layer
which is located in the planes of the longitudinal sections of the waveguide
segment passing through the axis of symmetry of the power feed elements.
The drawing shows a cross-sectional view of the construction of the proposed device.
An excited array is contained fn the croas-sectional plane of waveguide segment 1
comprised of inetal rods 2 which are parallel to its wide walls. Connected to the
ends of the rods are diodes 3, which are fastened to the narrow sides of the wave-
guide segment; furthermore, the middle of the rods are joined by posts 4 which
204
- FOR OFFICIAL USE OiVLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAI. USE ONLY
are perpendicular to the rods and which are connected to voltage feed elenents 5.
Each post 4 and rod 2 are separated by dielectric layer 6, which is loc.3,-ed in
the plane of the longitudinal section of waveguide segment 1, passing through the
axis of symmetry of voltage feed elements 5 to terminals 7.
The device opErates as follows.
When voltage is applied to terminal 7, direct current passes through diodes 3
and their impedance drops to an extremely low value, providing a condition in
the resonant array of total reflection of the microwave power passing through the
waveguide segment. Wfen voltage is absent from terminal 7, diodes 3 represent
a large capacitive impedance, and conditions are met in the resonant array for
unimpeded transmission of incident microwave power through it.
Dielectric layer 6 runs in the same direction as the microwave currents; con-
sequently, it introduces no additional losses in microwave energy, regardless
of the high-frequency properties of the dielectric.
The presence of dielectric layer 6 makes it possible to apply voltage separately
to each diode 3. This increases reliability and improves the parameters of the
devices which are dependent upon the degree of symmetry uf the array, for
example, the maximum generated or switched power.
Patent Claims
A device for addition of powers containing a waveguide segment in the cross-sec-
tional plane of which is an exciter array comprised of inetal rods running para-
llel to the wide walls of the waveguide segment with diodes located on the narrow
walls of the waveguide segment connected to the ends of the rods; furthermore,
the middle of the rods are joined by perpendicular posts which are also connected
to the voltage feed elements, distinguished by the fact that in order to increase
power and expand the operating frequency range, each post and excitar array rod
are separated by an additional dielectric layer running in the planes of the
longitudinal sections of the waqeguide segment passing through the axes of
symmetry of the voltage feed elements.
Infortnation Sources
Used in Evaluation
1. USSR Authorts certificate No. 566207, class H 03 B 7/14, 1975 (prototype).
205
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
� FOR OFFICIAL USE ONLY
733081
1
.
6900
CSO: 8144/0218-B
206
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
~ ~

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500020034-7
- RADIO RECEIVER PATENT
FOR OFFICIAL USE ONLY
Moscow OTKrYTIYA, IZOBRETENIYA, PROMYSHLENNYYE OBRAZTSY, TOVARNYYE ZNAKI
in Russian No 29, 1981 (signed to press 10 Jul 81) p 296
[Text] Patent No 11031
Claim No 20095
Class 14-03
Radio Receiver
Authors: I. I. Dezhurnyy, V. M. Kuz'min, V. S. Tsymbalyuk, 0, D. Fomin,
V. I. Zubkov, Ye. A. Lugovoy and V. I. Gvozdenko
Priority as of 26 July 1979
~
COPYRIGHT: VNIIPI, 1981
CSO: 1860/80-P
207
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
CONVERTERS, INVERTERS, TRANSDUCERS
UDC 621.372.852.27 (088.8)
WAVEGUIDE REFLECTING PHASE INVERTER
Moscow OPISANIYE IZOBRETENIYA 743084 in Russian 25 Jun 80
[Description of USSR Patent No 743084 by V.G. Vinenko and I.A. Ovechkin,
filed 13 Jun 75, published 25 Jun 80, class H 01 P 1/181
[Text] This invention is in the area of radio engineering and can be used in
microwave devices where discrete phase change is required.
There exist waveguide reflector phase inverters which use pin-diodes.
One of the existing phase inverters contains a waveguide transmission line contain-
ing a flat diaphragm with two slits, each of which contains a diode structure. A
- control signal causes the phase irYVerter to assume four phase states in sequence [1].
Design complexity is a shortcoming of the existing phase inverter, since it requires
that the diaphragm be fabricated with great precision.
The technical treatment which is closest to the present invention is a waveguide
reflecting phase inverter which contains a waveguide transmission line which is
closed at one end, two waveguide stubs placed in series with the waveguide line,
- and pin-diodes located at the end of the two stubs [2].
Relatively large size is a shortcoming of this phase inverter. The stubs are
located at a distance of ag/4 from one another (ag - wavelength in waveguide),
which provides the needed phase ahift, and the distance from the short-circuiting
switch of the main transmission line to the closest stub is also ag/4. The
- total length of the phase inverter is thus over Xg/2. The lateral dimensions
are relatively large because of the extending stubs which are placed at a right
angle to the main transmission line. The approximate length of the stubs is X/4.
The purpose of the invention is to reduce the size of the device.
This is achieved by placing the stubs, which are aa extension of the waveguide,
one on top of the other and connecting them to the wide wall, and fastening pin-
diodes to the input of each stub.
- Figure 1 shows the construction of the waveguide reflecting phase inverter;
figure 2 shows view A of figure 1.
- 208
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
ihe waveguide reflecting phase invertPr consists of waveguide segment 1, wave-
guide stubs 2, 3, pin-diodes 4, 5, which are fabricated as semiconducting struc-
tures, and taps 6 used to apply control current to them.
7'he conductivity of diodes 4 and 5 which are connected to the beginning of stubs
- 2 and 3 can take on two values under the influence of control signals. For
diode 5, which is connected to stub 2, these values will be JB'1 and JB"1.
The conductivity in the plane in which diode 5 is connected is the algebraic
sum of the conductivity of diode 5 and the conductiv3ty of shorted stub 2, (ym),
which can be written with the help of a familiar relationship as
11C
~~ysjZ1COfi AeeV
where Z1 - characteristic impedance of stub;
11 - stub length.
(1)
Thus, the total conductivity of stub 2 with diode 5 connected to its beginning can
_ be written as z r (2)
i:~a~ ~z~ot cD~+' .
where B1 - conductivity of diode, which can take on the two values B'1 and B"1
under the influence of the control signal.
Analogously, the conductivity of stub 3 in plane T with diode 4 connected to
its beginning can be written as
14z�jDz�j7.1C0t 79 eZ, (3)
where B2 - eondu�tivity of diode, which can take on the two values B'2 and B"2
under the influence of the control signal;
Z2 - characteristic impedance of stub 3;
12 - length of stub 3.
- The reflecting impedance in the plane T can be written using the obvi6us
relationship zT` vz `v z ' (4)
~ v 1 0
Thus, in the proposed construction the value o� the reflecting impedance, which
determines the phase of the reflected signal, depends upon the impedance of the
diodes and the impedance of the loops; consequently, in order to obtain the
needed reflecting impedance values, it is poesible to vary both the impedance
of the diodes, which is inconventent, as well gs the i~mpedance of the stubs,
which depends upon their length and characteristic impedances.
' 209
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
NUR OFFICIAL USE ONLY
The following formula, which aeaociates the phase shift of the reflected wave
(~p) with the value oP the reflecting impedance ia well known:
v ~
tan T (5)
~ .
Combining equations (2),(3), (4), (5) and keeping in mind the possibility that
the control stgnal may cause the conductivitq of the diodes to take on values
of B'1, B"1 and B'2, B"2, a system of equations of the following form can tie .
- obtained: - (D~ Zrotxt~X~2�Ztcotx~zlzo s t
q 2
b'lt 92t ZI COttQigZ,F0
CK~1 .
t4r' otKt~)c91.z2~cotKcllzo (6)
''---e,,e2 z,cotx~~'ZU:ot~1
.
:tan~YO ~
' (~~aZ~CptKt~)(D'z~2ZCOtK~1)ZO .t~y-~~'~1i
_ OtKe~ Z.jC 0 tKe2 .
(e;+z,cotKC,)cBzu � zzcof Keztz, _
d~ ~z,z~cotKt,�Z 2cot xt
z~0+470�` (61
tan % 2
where K = 27r/Xb
~a - initial phase.
Given the initial phase ~p, choice of conducti:viti:es B'1i B"1, B'2, B"2, of
characteristic stub impedance Z1 and Z2, and of stub lengths 11 and 12, the
equations in system (6) can be sattsfied.
The values of conductivities B'1, B"1, B12, B"2 are a function of the parameters
of the diodes used and the method by which they are connected to the transmission
line. .In the construction shown in the drawing, diodes 4 and 5 are soldered
between the inductive protrusion in the waveguide.
Patent Clai.ms
A waveguide reflecting phase inverter containing a waveguide segment and pin diodes
located in shorted stubs, distiriguished by the fact that in order to reduce size
the stubs, which are an extension of the waveguide, are located above one another
and connected to the wfide wall,�and the pfn diodea.are connected to the input of
each atub.
210
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500020034-7
FOR OFFICIAL USE ONLY
xnfox'uation Sources
Used in Evaluation
1. US Patent No 3478284, class 333-98, 1969.
2, UK Patent No 1325381, class H 1 W, 1973.
A ,
i
Fig. 2
6900
- CSO: 8144/0218-B
211
+
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
~
: ~ f

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
~
UDC 621.396.677: 621.382 (088.8)
SOLID-STATE PHASE INVERTER CONTROL DEVICE
Moscow OPISANIYE IZOBRETENIYA 750586 in Russian 25 Jul 80
[Description of USSR Patent No 750586 by V.A. Vinogradov and M.L. Slutskiy,
filed 26 Dec 77, published 25 Jul 80, class H 01 G 3/261
[Text] This invention is in the area of radio communication and can be used in
antenna array control devices.
There exists a device for controlling solid-state phase inverters which contains for-
ward and backward bias sources, in which the collector of the reverse bias source
switching transistor is connected through a resietance whfch limits current in
emergency conditions to the reverse voltage source in oxder to protect
the elements and the reverse bias source switciiing circuits from short circuiting
of the pin-diode or phase inverter coupling 1ine; a powerful current pulse used to
dissipate the body charge of the pin-diode at the moment of switching is created
by connecting an inductance coil to the collector circuits of this transistor [1].
A shortcoming of this device is the loss of power and impossibility of micro-
miniaturization.
The technical treatment which is closest to the present invention is a device
which contains a forward bias source switching transistor, reverse bias source
switching transistor with collector connected directly or through a current-limit-
ing resietor to the reverse bias source, and emitter or collector of forward bias
source swit!.hing transistor connected directly or also through a current-limiting
resistor, :o the forward bias source [2].
The lac'.. of protection of elements in the reverse bias source switching circuit
from line or pin-diode sho.rt circuits reduces the reliability of this device.
The purpose of the invention is to improve reliability. The solid-state phase
inverter control device, containing forurard and reverse bias sources, forward and
reverse bias transistors connected to an inverting amplifier, to the forward bias
source and to the solid-state phase inverters, achievea this goal by using series-
connected pulse shaper and awitching transistor, the collector of which is con-
nected through a resistor through the collector of the reverse bias switching
transistor, while the emitter is connected to the reverse bias voltage source and,
also, through a limiting resistor, to an inverting amplifier and to the base of
212
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICiAL USE ONLY
the reverse bias transistor.
- The drawing showa the circuit of this device for controlling solid-state pr.ase
inverters.
The device contains forward bias source switching transistor 1, reverse bias source
switching transistor 2, inverting amglifier transistor 3, current-limiting resis-
tor 4, switching transistor 5 with base connected to switching pulse shaper 6,
matching resl.stors 7 and 8, and pin-dicdes 9 and 10 (solid-state phase inverters).
The so lid-state phase inverter controlling device is part of a device for control-
ling a phased antenna array. The control signal applied to one of the control
inputs sets the forward or reverse bi$s mode of pin-diode pair 9 and 10 of one bit
of the phase inverter. The number of control inputs is equal to the product of the
numb er of phase inverters and the word length of one phase inverter.
The d evice operates as follows.
- Contr ol signals for pin-diodes 9 and 10 in the form of "high" and "low" voltage
levels are applied to the base of Eorward bias source switching transistor 1 and
inver ting amplifier transistor 3 from the phase distribution memory register of the
~ antenna array (not shown in drawing). The forward bias condition of pin-diodes _
9 and 10 corresponds to the "low" voltage level (diodes 9 and 10 are open, and
current from forward bias source +Efb flows through them). The reverse bias
condi tion corrPsponds to the "high" voltage level at the input (pin-diodes 9 and
10 ar e closed by voltage from reverse bias source Erb).
In th e forward bias condition, forward bias source switching transistor 1 is sat-
urated, and reverse bias source switching transistor 2 is cut off; the forward bias
sourc e is connected to pin-diodes 9 and 10 through the collector-emitter junction
of f o rward bias source switching transistor r and matching resistors 7 and 8.
Regardless of the forward or reverse bias condition of pin-diodes 9 and 10, the
collector of reverse bias source switching transistor 2 is connected periodically
to the reverse bias source through the collector-emitter junction of switching
trans istor 5 while switching pulses are applied to its base.
In the reverse bias mode of pin-diodes 9 and 10, forward bias source switching
transistor 1 and inverting amplifier transistor 3 are closed; negative potential
is established on the base of..reverse bias source switching transistor 2 with
respect to the emitter and the base-emitter junction is open, and cutoff voltage
tfrom the reverse bias source is applied to pin-diodes 9 and 10.
In o rder to accelerate dispersion of the body charge of pin-diodes 9 and 10 when
switching from the forward to the reverse bias condition, the collector of
reve rse bias source switching transistor 2 is connected to the reverse bias
sour ce for a time equal to the length of the awitching pulses.
213
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE UNLY
The switching pulaes are fornted by s.witching pulse ahaper 6.
If pin-diodes 9 or 10 or tfie connecting line are shoxt circuited, current with
a value close to tlie utaxfmum dtssfipattng current flows througfi current-limiting
resistor 4, reverse hias source switching tranststor 2 and matching resistors 7 and
8 for an amount of ttme equal to the lengtli of tTie switching pulse; when there is
no switc.hing puls-e current flows. from the reverse bias source through the collecto"'r
load of inverting anplifier trans�istor 3, the Fiase-emitter junction of reverse
bias source switching transistor 2, and matching transistors 7 and 8. These
currents are small and present no Iiazard for the circufit components.
This invention improves the reliabiltty of the dev3ce.
Patent Claims
A device for controlling s.oli.d-s.tate phase invertexs contafning foxward and
reverse bias sources, forward and reverse bias txansistors connecCed to an
inverting amplifier, to the forward li:ias source and to the solid-state inverters,
distinguished by the fact that in order to improve relfatii.lity additional series-
connected pulse shaper and switching trans-istor are used, with the collector of
the switching transistor connected through a resistor to the collector of the
reverse bias transistor, the emitter connected to the reverse bias voltage source
and to the inverting amplifier and the base of the reverse bias transistor
through a lirdting resistor.
Information Sources
Used tn Evaluation
1. US Patent No 3708697, class 343-854, 1973.
2. Georgopulos, Kh., "Proyektirovaniye modulyatorov dlya upravleniya pin-diodami"
[Designing Modulators for Controlling Pin-Diodes], translated from English, 1972,
(P-62874), figure 2 (proCotype).
214
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
ro
6900
CS O: 8144 /0218-B
215
FOR OFFICfAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFF'ICIAL USE ONLY
INSTRUMENTS, MEASURING DEVICES
AND TESTERS, METIiODS OF MEASURING,
GENERAL EXPERIlMENTAL TECHN14UES
UDC 535.854(088.8)
PRt:CISION DIGITAL INFRARED PHASE METER WITH FREQUENCY CONVERSION OF LASER
RAD~ATION
Mosc-)w PRIBORY I TEKHNIKA EKSPERIMENTA in Russian No 4, Jul-Aug 81 (manuscript
receiva3 13 Jan 78, after revision 4 Feb 80) pp 203-205
[Paper by V.A. Fedoseyev, V.V. Kalendin, V.I. Ktxkhtevich and i1.Ya. Sup`yan, All-
Union Scientif ic Research Institute of Phyaical Optics Measurements, Moscow]
[Text] The OTsF-2 precision digital infrared 9 to 11 um phase
meter, designed using a variant of a Mach-Zender interferometer
conf iguration using a homodyne technique with double frequency
conversion of laser radiation is described. The phase meter
serves for the measurement of constant and slowly changing phase
shifts in a range of 0 to 360� with an error of = 0.06� and a
measurement time of 0.72 sec.
High phase shift measurement precision is achieved in the visible and IR bands
when homodyne (with frequency conversion of the optical carrier) or modulation-I
- compensation phase measurement methods are employed [1, 2]. An IR band (9 to
11 um) compensation phase meter was described in paper [3] for the measurement of
constant and slowly changing phase shifts with an error of < 1�. The long measure-
- ment time (about one~minute) and the use of an insufficiently stable mechanical
= phase compensator are to be numbered among the drawbacks of this phase meter.
The OTsF-2 precision aptical digital phase meter, which is intended for the measure-
ment of constant and slowly changing phase shifts in the IR band, is described in
this paper. Homodyne conversion of the laser radiation frequency is'used in the
- instrument.
A block diagram of the phase meter ia shown in Figure 1. Double frequency con-
version of the laser radiation directly in the optical portion of the instrument
is used in the phase meter [4]. The optical scheme of the instrument takes the
form of a dual beam modified Mach-Zender interferometer, where there are acousti-�
cal-optical single band shift modulators, 8 and 13, in the reference and signal
- channels of the interferometer. The phase object under investigation 3(for example,
an'electromagnetic acoustical-optical modulator, optical delay line, eLC.) is located
in the signal channel.' The output of a C02 laser is fed through the input
window 1 of the optical unit to the light splitting plate 2 and routed into the
reference and signal channels of the interferometer. The portion of the radiation
216
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFF(CIAL USE ONLY
(the reference channel) reFlected from plate 2 impinges on mirror 6
and is focused by lens 7 onto the active Ge element of acoustic shift
modulator 8 with a LiNb03 converter. The IR radiation impinges
on the acoustical-optical modulator at the Bragg angle 4.5�) [5]. The acousti-
cal-optical modulator shifts the frequency of the optical carrier by 80 MHz. The
high frequency component of this frequency is fed to the modulator from a stabil-
ized crystal oscillator G1 with an output power of about 6 W. The C02 laser radi-
ation which is shif ted in frequency 80 MHz and deflected by twice the Bragg angle
- 90�) is shaped by lens 9 into a parallel beam and then goes to the mixer plate
5. The refracted portion of the radiation passes through the phase object being
studied 3(the signal channel, and in being reflected from mirror 4, is directed
to the second ac oust ical-opt ical single band shift modulator 13 with the focusing
14 and shaping 15 lenses and is then fed to mixer plate 5 where it is combined with
the reference beam. Generator G2 with an output power of about 4 watts powe;.s
modulator 13 at a frequency of 80,050 MHz. The combined signal and ref erence
beams are directed by reflector 10 and lens 11 to photodetector 12 with a detecting
element based on the ternary compound CdXHgl_X)Te, cooled with liqu id nitrogen.
Photodetector 15 converts the spatia11y combined optical signals at frequencies
of wp + 80 MHz and wp + 80,050 MHz (where wp is the op.tical carrier frequency) to
an electrical difference frequency of St = 50 KHz, which is fed to the signal input
of a digital display unit (BTsI). A reference signal at the same frequency,
50 KHz, is fed to the second input of the digital display unit, where this fre-
quency is generated in the internal mixer, SM, ff om the two high frequency voltages
- at 80,000 and 80,050 MHz incoming fram generators G1 and G2. Both generators are
stabilized by frequency-phase regulation using an automatic frequency and phase
control system (ChFAPCh) [PLL frequency and phase control], which provides for a
constant frequency and phase of the intermediate frequency signals at 50 KHz fed
to the input of the digital display unit.
The digital display unit, which takes the �orm of a digital radio frequency phase
meter, serves for the measurement of phase difference in a range of 0 to 360�
between sinusoidal signals at a frequency of 50 KHz and levels of 300 uV to to
0.4 volts with an error of < 0.05�. The phase difference is d isplayed on the
digital paneT of the frequency meter. The digital display unit consists of the
indicator unit (BI), the quantization pulse generator (GIK), the c lock oscillator
(VZG) and a Ch3-35 digital frequency meter. The display unit BI generates a train
of rectangular pulses, the width of which is proportional to the pha.se difference
in the 50 KHz signals fed to its input. These square wave pulses are f illed in
with pulses at a frequency of 50 MHz, generated by the quantization pulse generator,
and are fed to input A of the Ch3-35 frequency meter. The clock pulses are fed
to input B of the frequency meter fram the clock oscillator, the VZG. The
measurement time is specified by a frequency divider which div ides the 5 MHz by
360, which is housed in the clock oscillator unit. The frequency meter operates
by measuring the ratio of the frequencies FA/Fg, and the measurement results is
displayed in angular degrees on the digital display of the Ch3-35 frequency meter
for measurement time of 72 usec to 0.72 sec with a resolution of +0.01�.
The analog output of the instrument driving an autorecorder can b e used when meas-
uring slowly changing phase shifts. The square wave pulses, the width of which is
217
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00854R000500020034-7
FOR OFFICIAL USE ONLY
r 6~ ~A~ i i 6A114 (g) ;
~
- i  , ' r-F _ � r i
DI i i
rNK t
~ r y~any
- ~--,--C>-- Z--~ ~------E1 i B9r (G) i '
8 'i .
A yd-35
6 _ ID 60
(K) i
~ (g)
15 ' tJ i (Z) i
I 13
i ynr (J~ i
Laser I 14 ~ ----buN - ~
L---- _J
na3ep ~ � / 12 .

3
Figure 1. Block diagram of the phase meter.
Key : A. Generator unit;
B. Automatic frequency control unit;
C. Generatore 1 and 2; D. Internal mixer; E. Automati.c frequency and phase control system;
- F. Quantization pulse generator;
G. Clock oscillator;
- H. Display unit;
I. Direct current amplifier;
J. Digital display unit;
K. Ch3-35 frequency meter. proportional to the phase shift, are fed through integrating circuits to a DC
amplifier (UPT). The DC voltage at the input to the DC amplifier is propartional
to the phase shift between the signals in the optical channels of the interf er-
ometer.
To improve the seneitivity and; stability of the phase shifts being measured, the
optical elements of the instrument-:are rigidly fastened to a massive plate, placed
inside a thermastatically controlled multilayer housing. The housing serves to
_ reduce the impact'of acoustic interference and to maintain the temperature stabi-
lity of the phase meter. The absolute error in the radio frequency phase meter
measurements was determined by meana of comparisan with-standard phase shift
meters at a frequency of 50 KHz [6] and was 0.05� over a measurement time of 0.72
sec.
The OTsF-2 phase meter makes it possible to measure phase ahifts in a range of 0
to 360�. The region of instrument inaensitivity (the so-called dead zone) wh2n
218
FOR OFFlCIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE ONLY
measuring shifts close to 0� (or 360�) does not exceed +2�. There is an input for
a supplmental phase shift of 180� --o bring the instrument out of the "dead zone".
There is an output in the instrument for a digital printer using a binary-decimal
1-2-4-8 code to record the phase values.
To estimate the noise immunity and sensitivity of the OTsF-2 phase meter when
measuring slowly changing phase shifts, multiple recordings were Trade of the
drift and fluctuations in the phase difference of C02 laser radiation (Figure 2).
The internal "zero" phase drift of the OTsF-2 phase meter is about 0.01�/sec.
The phase shift measurement error is shown below as a function of the me.asurement
t ime .
Measurement time, sec 0.72 0.072 0.0072 0.00072
- Mean square error of ten
measurements, degrees 0.02 0.03 0.08 0.2
Maan square error of a
single measurement,
degrees 0.06 0.13 0.32 0.8
a
~
Figure 2. The typical signal drif t
at the output of the OTsF-2
phase meter.
The use of frequency conversion by means of two acoustical optical modulators
- [4] in the optical phase meter made it possible to measure phase shifts at the
- low intermed iate frequency 50 KHz, eliminate the impact of high frequency coherent
interference and parasitic coupling between the instrument channels as well as
reduce phase mismatching in the arms of the interferometer and achieve a high pre-
cision in the phase measurements.
The OTsF-2 phase meter measures phase shifts when coherent radiation is fed to
its input at a wavelength of 9 to 11 Um having a power of 10'2 - 10 W, a diverg-
ence of 10-2 - 10-4 radians::and a radiation beam diameter of 1 to 10 mm. It must
be underscored that the high precision of the phase shift measurements imposes
definite requirements on the parameters on the lasers employed. The radiation
source should be single mode, single frequency and frequency stabilized to no
worse than 10'7 and stabilize with respect to power at < 1-3 percent over a
measurement time of one second.
219
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
The OTsF-2 instrument consists of three units: the optical, electronic-phase
metering and the Ch3-35 frequency meter. The dimensions are: optical unit,
500 x 586 x 338 mm3; the electronic phase metering unit, 490 x 170 x 488 mm3; the
weight of the instrument is = 40 kg.
BIBLIOGRAPHY
1. l:armashov A.I., Etsin N.Sh,USPEKHI FIZ. NAUK [PROGRESS IN THE PHYSICAL SCIENCES],
1972, Vol 106, No 4, p 689.
2. Kalendin V.V., Kukhtevich V.I., Mukhmarov R.I., "Tr. VNII fiziko-tekhn. i radio-
tekhn. izmereniy. Metrologicheskoye obespecheniye izmereniy optiko-fizicheskikh
- parametrov izlucheniya OKG" ["Proceedings of the Al1-Union Scientific Research
Institute for Applied Physics and Radio Engineering Measurements. Metrological
Support for Measurements of the Optical Physics Parameters of Laser Radiation"],
Moscow, 1976, p 6.
3. Artemov V.M., Zhelkobayev Zh., Kalendin V.V., et al., PTE [EXPERIMENTAL
ENGINEERING AND INSTRUMENTS], 1976, No 2, p 188.
4. Zhelkovayev Zh., Kalendin V.V., Kukhtevich V.I., et al., Patent No. 506755,
Published in Bulleting No. 10, 1976, p 97.
5. Mustel' Ye.R., Parygin V.N., "Metody modulyatsii i skanirovaniya sveta"
"Light Modulation and Scanning Techniques"], Moscow, Nauka Publishers, 1970.
6. Galakhova O.P., Koltik Ye.D., Kravchenko S.A., "Osnovy fazometrii"
_ ["Fundamentals of Phase Metering"], Leningrad, Energiya Publishers, 1976.
COPYRIGHT: Izdatel'stvo "Nauka", "Pribory i tekhnika eksperimenta", 1981
8225
CSO: 8144/0178
220
FOR OMCiAL USE 01dLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
NUR OFFICIAI. USN: ONLY
UDC 621.317.341 (088.8)
PANORAMIC MEASURING DEVICE FOR COMPLEX PARAMETEAS OF MICROWAVE DEVICES
Moscow OPISANIYE IZOBRETENIYA 765755 in Russian 23 Sep 80
[Descrl.ption of USSR Patent No. 765755 by O.I. Agafoniyeva, A.P. Baklykov, I.K.
, Bondarenko, Yu.B. Gimpilevich, B.A. Prigoda and A.I. Soboleva, filed 1 Jul 76,
class G 01 R 27/28, G 01 R 27/061
[Text] This invention is in the area of microwave instrumentation and csn be used
for measuring and monicoring the complex parameters of microwave devices.
There exists a measuring device for complex microwave device parameter$ which con-
, sists of a sweep generator, polarization converter, secondary circular waveguide
' and detector head, during the rotation of which the complex parameters of microwave
devices are measured [1]. The presence of rotating parts, vibration and lag are
shortcomings of this device.
There also exists a panoramic device for measuring the complex parameters of micro-
wave devices consisting of a sweep generator, polarization converter, circular
waveguide, capacitive probes placed inside a circular waveguide every 450 within the
plane of the polarization ellipse, detector heads, each connected to its own probe,
and a panoramic display [2]. Shortcoming$ of this device include the errars caused
by the different amplitude-frequency characteristics of the detectors and their
temperature and temporal instability, as well as polarization distortions intro-
duced by the probes located on the side wall of the circular waveguide.
The purpose of the invention is to improve measurement accuracy while eliminating
polarization distortions.
This is achieved by making the capacitive probes with gaps which contains switching
diodes, e.g., pin-diodes which are joined to the control unit and connected to a
common detector head which is co-axial with the circular waveguide; furthermore,
all of the probes and the connecting rod of the detector head are connected at the
center of the circular waveguide.
Figure 1 shows the functional diagram of the proposed instrument; figure 2 shows
two views of the microwave section of the instrument in sections A-A and B-B.
The instrument contains sweep generator 1 and polarization converter 2. Four
capacitive probes 4-7, connected via switching diodes 8-11 (e.g., pin-diodes) by
means of rod 12 to detector head 13, are placed 450 apart in the plane of
221
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFIClAL USE ONLY
the polarization ellipse inside circular waveguide 3. Switching diodes 8-11
are controlled by control section 14, which is connected between detector head
13 and display 15, which is made up of readout devices 16 and 17 and cathade-ray
tube 18; the microwave power passes through polarization converter 2 to loaC 19,
hte parameters of which are being measured.
- The instrument operates as follows.
The microwave oscillations pass from sweep generator 1 to polarizat on converter
2, which converts a type H16 waveguide wave or a type TEM coaxial wave to an ellip-
tically polarized H11 wawe of a generally unmatched load.' This wave propagates
through circular waveguide 3. Control section 14 sequentially enables probes
4-7 by applying voltage to switching diodes 8-11, which are installed in the
- gaps of these probes; the activated diode then has current applied, while
the disabled diodes have slight negative bias. This makes it possible to
achieve transient attenuation of 40 dB. Detector head 13 sequentially outputs
voltages defined by the expressions
U. � K(1 � u' loros; I,
U, � K I I � U� ' K (1 � p'
- � U, - K ( t � c' :aor+c 1.
where U1 - Uy are the detector head output voltages obtained from probes 4-7,
respectively; K is the coef�icient of proportionality, p is the absolute value of
the coefficient of reflection; ~ is the phase oP the coefficient of reflection.
Switching section 14 switches the output of detector head 13 to the inputs of
panoramic display 15 in synchronization wlth the switching of probes 4-7 such
that voltages Ue and U7 are applted to the corresponding inputs of readout device
16, and voltages U4 and Us are applied to the inputs of readout device 17.,Voltages
U 4K/rcirwp = KlOm*C.
t 1� - 4K,Nns: � Klxow
are output by the readout devices.
These voltages are applied to the vertical and hori.zontal plates of the cathode-
ray tube, which displays the panoramtc characteristic of the complex coefficient
- of reflection.
Because a single detector head is used, the device does not requixe calibration
and is simple to operate. The electronic switching system which connects the
probes makes it possible to automate the measurement of complex parametera of
microwave devices.
Patent Claims
_ A panoramic instrument for measuring complex parameters of microwave devices,
containing a sweep generator, polarization converter, circular waveguide con-
taining capacitive probes, detector head and diaplay diatinguiahed by the fact
that in order to improve measurement accuracy while eliminating polarization
distortions the capacitive probea contain gaps in which switching diodes are
222
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500020034-7
FOR OFFICIAL USE ONLY
placed, whicti are connected with an additional control aection; furthermore, theae
capacitive probes are joined at the center of the circular waveguide and connected
to a common detector head which is located coaxially on the circular waveguide.
Inforiaation Sources
Used in Evaluation
1. Bondarenko, T.K., Deynega, G.A. and Magxachev, Z.V. "Avtomatizats-ipa izmexeniy
parametrov SVCh-traktov." [Automation of Microwave Circuit Parameter Measurement].
Moscow, Sovetskoye radio, 1969, pp 19-21.
2. US Patent No. 2818546, class 324-58, published 31 Dec 57 (prototype).
~
223
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
t------ ~

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
~~,c7tt
B - B
6T,
6900
CSO: 8144/0218-B
224
FOR OFFICIAL USE ONLY
~i
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
UDC 621.317.382/384:621.375.826
WIDEBAND INTEGRATED CIRCUIT DIGITAL POWER METER FOR LASER RADIATION
Moscow PRIBORY I TEKHNIKA EKSPERIMENTA in Russian No 4, Jul-Aug 81 (manuscript
received 25 Apr 80) pp 197-199
[Paper by A.I. Kirillov, V.I. Kishko and V.F. Morskov]
[Text] A digital integrated circuit laser power meter is
described having a measurement range of 10-8 - 10-2 W.
The analog converter nonlinearity ia < 5 percent, the
digitization error of the digital converter is < 2 percent.
The spectral seneitivity region is 0.4 to 1.2 micrometers.
The temperature inatability of the meter in a temperature
range of -10 to +40 �C is no more than 6 pe,:cent.
Series produced laser power meters have a measurement range of 10'6 - 10-3 W[1].
A power meter was described in [2] for an He+Ne laser with a measurement range of
10-7--10-6 W and a threshold sensitivity of 10-8 (for a signal to noise ratio of 1),
- which employs a mechanical modulator and a FD-7K silicon photo diode, operating as
a detector.. The dynamic range of this meter is limited by the fact that the volt-
ampere characteristic of the photo diode operating as a detector is nonlinear with
a high impedance load [3J. For this reasan, it ia necessary to have calibrated
attenuators for measurements in a range of 10-6 to 10'2 W.
The meter considered here.operates on the well known principle of autocompen4ation
of the source current (a photo diode in our case) [4] with the subsequent convers-
ion of the autocompensator output voltage to frequency. Operating the photo diode
as a rectifier makes it possible to realize a high sensitivity, and automatic
compensation of the current, because of the 100 percent negative feedback, assures
qn almost zero load on the photodiodea, which is defined by the expression [5]:
Rload - Rfb/K'
(1)
where Rfb is the resistance of the feedback reeistor; K is the gain of the opera-
tinal amplifier. In our case, when Rfb = 470 Kohm and K= 5- 104 (a typical value
for a K544UD1A), Rload = 10 ohms. With such a load, hhe volt/ampere characteristic
of the photo diode is has good linearity [5]. The dynamic range of the power being
measured is extended up to 10-2 W by reducing the value of the feedbaclc resistor.
225
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500020034-7
FOR OFFICIAL USE ONLY
The analog converter (Figure 1) contains an av.tocompensator for the photo diode
- current designed around operational amplifier M1 (K544UD1A), the input current of
which is < 0.1 pA, and a voltage to frequency converter. Switch B1 serves for
manually selecting the measurement range in,a range of 10-4 to 10-2 W. The voltage
to frequency converter (v.f.c.), which employs a voltage proportional to the radia-
tion power, is similar to that described in [6] and for the selectdd nominal values
of the components has a conversion factor of 100 Hz/uW in a range of 10'$ to 10-4 W,
10 and 1 Hz/uW in a range of 10-4 to 10-3 and 10-3 to 10-2 respectively; the con-
version nonlinearity is < 0.02 percent, the temperature coefficient is < 0.025 %�C.
The meter is zeroed with potentiometer R1.
The output frequency of the v.f.c., which variea from 1 to 104 Hz, is measured by
a low frequency digital frequency meter using countera, in which the period is
_ measured, while the value of the frequency is determined as the inverse of the
period [7]. The measurement time in thia case does not exceed two periods. The
frequency meter configuration uaing countere [7] is simpler than the well-known ones
of [8], since it can be designed uaing standard digital integrated circuits. A
functional block diagram of the frequency meter, designed around series K155 inte-
grated circuits, is shown in Figure 2. The frequency being measured, fX, which is
proportional to the laser radiation power, is fed to the driver FZ for the period
TX and then to AND gate 13, to the second input of which one of the sample fre-
quencies is fed. The result of ineasuring the period, NT, is stored by counter
Sch2. With the signal for the end of the measurement period, a sample time interval
Tp which turns on AND gate I1 for the sample frequency pulses fp is produced fram
driver F3 at the output of F1. From the point in time of triggering the driver for
- Tp, the number Np, set by counter Schl, is compared by means of the code comparator
(USr) with the number NT, which is set in counter Sch2. At the moment the codes
D1
+15
Figure 1. Basic achematic of the analog converter: M1 = K544UD1A; M2, M3 =
= K153UD1A; T1KT312B; Dj = FD-7K, D2-D7, D9-D13 = KD503B;
Dg = D818Ye.
226
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R000500020034-7
(1)
FOR OFFICIAL USE 1
fo 4
(3
~ ~ H~
(la
(9
. an
)NLY
(5)
C4~
' No
NF
yCp Cy3
� � (12)
Figure 2. Block diagram of the fr.equency meter. GOCh is the master reference
_ frequency generator.
Key: 1. GOCh = master frequency generator;
2. DCh = frequency divider;
3. F1 = driver 1;
4. I1 = AND gate 1;
5. Schl = counter 1;
6. WP = meseurement range selector;
7. F2 = driver 2;
8, F3 = driver 3;
9. IZ = AND gate 2;
10. USr = code comparator;
11. Sch2 = counter 3;
12. Sch 3= counter 3.
are equal, a comparison pulse is generated which sets counter Schl to zero and feeds
the measurement results Nf to the input of counter Sch3. Thus, a number Nf propor-
tional to the frequency being measured fX, will be set in counter Sch3 after a time
of Tp.
The frequency meter circuitry ia supplemented with a measurement range selector
(WP) which connects one of the sample frequencies fp/lOn (n = 0-- 3) generated by
the frequency divider (DCh) to counter Sch2. With each overflow of counter Sch2,
the AND gate 12 generates a pulse which is uaed at the same time for switching the
subband of the measurement range aelector and writing the number NT/10 into the
counter Sch? [number illegible]. The presence of the measurement selector makes it
possible to extend the dynamic range of the frequency meter up to four decades with-
out increasing the number of digits in counters Schl and Sch2 or the cc.',,_-~ comparator.
227
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFRICIAL USE ONLY
A He + Ne LG-76 laser, and a atandard IMO-2-2 laser power meter were used to cali-
brate the meter at a wavelength of a= 0.63 u; in this case, the graduation error
did not exceed 3 percent. In a measurement ran e of 10-8 - 10-4 W, the meter has
a sensitivity of 0.1 V/uW, and in ranges of 10-t - 10-3 and 10-3 - 10-2 W, it hgs
a sensitiv ity of 0.01 and 0.001 V/uW respectively. The temperature instahility
does not exceed 6 percent in a temperature range of -10 to +40 �C.
BIBLIOGRAPHY
l. Anikina L., Vanyukov G., "Informatsionno-spravochnyy listok" ["Reference Data
Sheet"], Moscow, VIMI, 1971, No 001563.
2. Plotnikov V.A., Chastukhina L.N., PTE [EXPERIMENTAL EQUIPMENT AND ENGINEERING],
1971, No 4, p 189.
3. Soboleva N.A., Melamid A.Ye.,."Fotoelektronnyye pribory" ["Photoelectronic
Devices"], Moscow, Vysshaya shkola publishers, 1974.
4. Aleksandrov V.S., Pryanishnikov V.A., "Pribory dlya izmereniy malykh napryazheniy
i tokov" ["Instruments for Measuring Low Voltages and Currents"], Leningrad,
Energiya Publishers, 1971. .
5. Vendlend, ELEKTRONIKA, 1971, No 11, p 30.
6. Tychino K.K., "Preobrazovateli napryazheniya v chastotu" ["Voltage to Frequency
Converters"], Moscow, Energiya Publishers, 1972.
7. Yermolov R.S., "Tsifrovyye chastotomery" ["Digital Frequency Meters"], Leningrad,
Energiya Publishers, 1973.
8. Inozemtsev Ye.K., Babayeva O.S., PTE, 1979, No 5, p 156.
COPYRIGHT� Izdatel'stvo "Nauka", "Pribory i tekhnika eksperimenta", 1981
8225
CSO: 8144/0178
228
FOR OFF[CIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
:a
SUBNANOSECOND PYROELECTRIC INFRARED RADIATION DETECTOR
UDC 621.373.826:621.317
Moscow PRIBORY I TEKiNIKA EKSPERIMENTA in Russian No 4, Jul-Aug 81 (manuscript
received 10 Dec 79) pp 205-207
[Paper by V.P. Kuleshov and D.D. Malyuta]
[Text] Results of testing a lithium metaniobate LiNb03
pyroelectric detector are given for the recording of
short C02 laser pulses. The response lag of the detector
is less than 1 nanosecond and the sensitivity is about 0.15 V/MW.
Detectors which have a high operational speed, wide dynamic range and which are
not destroyed by irradiation are needed for recording the radiation power of!'pulsed
C02 lasers. The sensitivity of the detectors can be comparatively low because of
the high radiation power. At the present time, detectors are used for these pur-,
poses which are based on the effect of entrainment of free charge carriers in the
conductors [1], as well as detectors using u-conductivity semiconductors [2],
cooled semiconductor photoresistors [3] and pyroelectric detectors [4, 5].
3 ~
~
~
~
Key: 1. Pyroelectric crystal;
2. SR-50-73F high frequency
conneetor;
3. Brass housing with cap.
Pyroelectric detectors receive radiation in a range from the microwave to the
ultraviolet wavelength, have a time resolution of < 0.1 nsec and a wide dynamic
229
FOR OFFICIAL USE ONLY
Figure 1. Structural design of a detector
with edge electrodes.
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500020034-7
FOR OFF'ICIAL USE ONLY
rii~ige. I'yrc;rluctr[c dvtecture opariiCe nt room temperature nnd are easy to fabri-
cate und operFite. Ferroelectric cryatals with good pyroelectric properties are
usually eaiployed as the material for such detectors: TGS, SBN, LiTa03, LiNb03,
etc.
When pyroelectrics in the form of rectangular parallelepipeds are used with the
electrodes placed on the opposing end faces, two ways of orienting them are pos-
sible: the electrodes arranged in planes peependicular to the beam (a frontal
- detector) or parallel to it (a detector with edge electrodes). In the absence
of thermal losses to radiation and heat transfer, the volt-watt sensitivity of
detectors with frontal and edge electrodes is defined by the following expres-
- sions:
S1 = TCl'Y/ECV; S2 = bSl/a; T> RC
where C is the detector capacitance; t is the time resolution; a is the coefficient
or radiation absorption by the detector material; Y is the pyroelectric coeff ic-
ient; e is the dielectric permittivity and Cv is the bulk thermal capacity of the
pyroelectric material; a and b are the dimensions of the crystal along and trans-
verse to the beam; R is the detector load resistance.
.2
~
N
O
V.
-
1
e
`
a 1
(4
)
1
51
(bl
m
N
�
=
-
-
- -

-
~
.-O
~ ~
-
Z
-
-

/oic
The condition T> RC can be rewritten as
follows for the two cases under consider-
ation here: A1/al < T/eR; A2/a2 80,000 h.p.;
and in 8 LtSvessels Ne = 120,000 h.p. Plans are now being developed for container
vessels with Ne = 200,000 h.p. and more. A plan is being drawn up for a ship in
England capab.le of carrying 5,000 containers at a speed of 30 knots with Ne =
210,000 h.p. In Japan, a 3500-container vessel will have a speed of 35 knots
and Ne = 250,000 h.p. According to predictions of US specialists, the interna-
- tional fleet will have 140 vessels with power exceeding 100,000 h.p. in 1980,
and 500 such vessels in 1990. Similar prospects are foreseen for other types
of vessels as well (Chapter 1[16]).
The power plant is the most complicated system aboard a vessel; its development
is determined to a significant extent by changes in the basic parameters of tne
vessel. On the other hand, new achievements in electrical engineering, electrical
equipment development and power engineering make it possible to create vessels
based on new technical principles, e.g., ships using electromagnetic hydroreactive
engines. The prediction of the development of maritime electrical engineering must
allow for achievement in shipbuilding and the electrical engineering industry
(chapter 1 [27]).
~
The book develops two approaches to predicting the development of maritime elec-
trical engineertng. One of them involves developing methods and models of pre-
diction, whlle the other is a descriptive method involving expertise. Both of
these approaches complement one another. Prediction methods and models provide
an apparatus for quantitative estimation of the future development of maritime
electrical engineering. The methodology involved, especially the analytical
algorithms, can be used in studying a broad class of problems in the area of
maritime-electrical engineering, both those which are being solved at present
as well as those which may arise in the future.
In contrast to the firsC approach, with which it is now impossible, for the
reasons indicated below, to make specific numerical calculations for all tasks
involved in predicting the development of maritime electrical engineering, an
attempt is made using the descriptive-expert panel approach to answer all
basic problems which arise during the practical sclution of prediction problems.
Here are two facts *ahich are involved in using the first prediction approach:
first of all, the extrapolation models which are most widely used in prediction
theory require as initial information a precise and rather full retrospective
339
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
N'OR OHN7('IA1. USN: UNI.Y
series of values which characterize the predicted process. A special infor-
mation service must be created to obtain this information, which forms a data bank
in support of the methodology in question. The organization of such a service will
make it possible, over time, to have sufficiently complete and valid information
for solving practical prediction problems. Secondly, many important problems
~ (in particular, prediction of the appearance of new types of maritime electrical
equipment) does not involve the examination of extrapolation models. No effective
or workable models have yet been developed for such problems. We may hope that
objective quantitative methods of evaluating the future status of maritime
electrical engineering will be used increasingly as the problem of insufficient
information is overcome by gradually accumulating source data and creating new
prediction methods.
When predicting development of technology, authors generally make note of the
risk involved in solving the type of problem involved. The predictions made in
the present work can be considered as an expert estimate of the prospects of
devel.opment of maritime electrical engineering provided by the authors of the
book. It is not impossible that another group of experts would have a different
opinion from that presented in this book. In making their predictions, the authors
take a canservative approach when necessary. For example, the section on expert
panel-descriptive prediction is given without allowing for the time factor of
the development of individual areas of maritime electrical engineering. This
prediction is necessary in order to form the goals of development of the branch
of technology. In those cases in which the authors have sufficient source infor-
mation, prediction is done on the basis of developed models.
The book makes use of works published by the authors, domestic and foreign monographs
and articles, as well as reviews on maritime electrical engineering contained in
the collections "Sudostroyeniye za rubezhom" (Shipbuilding Abroad). Chapter 1
was written by Doctor of Technical Sciences Yu.A. Svetlikov and Candidate of
Tecilnical Sciences V.F. Samoseyko, chapter 2 by Doctor of Technical Sciences
G.I. Kitayenko and Doctor of Technical Sciences Yu.P. Kos'kin.
A number of the positions put forth in the book were discussed with specialists
of the shipbuilding and electrical engineering industries. The authors are deeply
grateful to all comrades who participated in discussing the materials in the book,
especially I.I. Adrianov, D.A. Gidaspov, A.N. Kaganovich and P.G. Meshcheryakov.
The authors express their gratitude to N.V. Belousova, R.I. Gladkikh and L.S.
Kos'kina for their great assistance in making calculations, as well as selecting
and processing the literature. '
All critical remarks and suggestions should be addressed to 191065, Leningrad,
ul. Gogolya, 8, Izdatel'stvo "Sudostroyeniye".
Table of Contents
Foreword
340
FOR OFFICIAL USE ONLY
5
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007142/09: CIA-RDP82-40854R040500020034-7
FOR OFFICIAI. I1SE ON1,Y
Chapter 1. Methods of Investigating Regularities in the Development of
Maritime Electrical Engineering 9
1.1. Methodological principles
1.2. Information support for research models 18
1.3. Investigation of developmental trends of determining parameters
of maritime electric power systems 37
1.4. Methods of ineasuring maritime electric power systems as
function of parnmeters of future vessels 60
1.5. Analysis of structured problems of development of maritime
electrical engineering gl
References to chapter 1 103
Chapter 2. Problems of Development of Maritime Electrical Engineering 105
2.1. Modern maritime electrical engineering and its problems -
2.2. Electric energy parameters 113'
2.3. Deep-water electrical engineering 125
2.4. New electric power sources 133
2.5. Electromagnetic hydroreactive engines 147
2.6. Cryogenic technology and superconducting electrical equipment 150
2.7. Trends of development of maritime electrical engineering 174
References to chapter 2 182
Conclusion 185
COPYRIGHT: Izdatel'stvo "Suuostroyeniye", 1981
6900
cs Q. 1860/105
341
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICiAL USE ONLY
I1DC 621.396.96.001(07)
RADAR DATA PROCESSIPlG AGAINST INTERFERENCF BACY�GROUi~ID
Moscow TEORIYA I TEKHNIKA OBRABOTKI RADIOLOKATSIONNOY INFORMATSII NA FONE POMEKH
in Russian 1981 (signed to press 20 Apr 81) pp 2-3, 410-416
[Annotation, foreword and table of contents from book "Theory and Practice of
Radar Data Processing Against an Interference Background", by Yakov Davidovich
Shirman and Vladimir Nikolayevich Manzhos, reviewed by t',ctors of the engineer-
ing sciences and professors V.Ye. Dulevich and M.B. Sverdlik, Izdatel'stvo '
"Radio i svyaz l", 10,000 copies, 416 pages]
[Text] Questions of optimizing multichannel and single channel detection systems
as well as radar signal measurement and resolution are generalized.
Primary attention is devoted to design principles and questions of the engineering
. realization of various analog and digital detectors and signal parameter meters
. for signals against a background of correlated nonsteady-state interferenc.e as well
as questions of adaptation to conditions of apriori ambiguity, etc. A large number
of illustrative examples is cited.
The book is intended for specialists engaged with the theory, design and operation
of radio electronic hardware and systems.
Foreword
The rapid development of radio electronics, and radar processing engineering theory
in particular, makes it diff icult to generalize the multifarious publications in
this field. Nonetheless, there is the insistent need for generalizing publications
which encompass the state of the art of the theory and the prospects for implement-
ing its conclusions.
An urgent but difficult task is therefore the theoretical generalization of new
questions and "those which have arisen" concerning radar processing against a
- background of interference from unified, sufficiently general procedural stand-
points. For example, the refinement of the component base has expanded the pos-
sibility for multichannel reception. Signals and interference are described by a
set of time functions or by time and coordinate F�inctions in this case. The object
- of the optimization becomes the space-time processing of the signals against a
- background of interference, which provides for effective interference suppression
- in many situations (not just in radar). The generalizatian of the new problems of
- the indicated processing comprises a weighty portion of tte problem p4sed above.
342
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407102109: CIA-RDP82-00854R000500020034-7
FOR OFFICIAI. USE ONLY
'rhis book is an attempt to solve this problem to some extent. The courses of
lectures on radar theory are taken as its basis: for engineers completing. their
training; for graduate students and degree candidates preparing to meet the candi-
data minimum. Special att-ention has been devoted to questaons which are not alwayg
_ accessible or fully treated in the literature: the theory of primary space-time
= processing of radar data against a background of correlated interference, which
j provides not only for the accumulation of the useful signals, but also for the can-
cellation of interfering onES; the theory of ineasurements of timewise constant and
- :.hanging signal parameters against a background of interference during primary and
secondary processing; adaptation theory; adaptive antennas and moving target indi-
- cation systems; new techniques of coherent processing of simple and complex space-
time signals: digital, optical and spin methods.
Serious attention has been devoted to the compactness, generality and accessibility
of the presentation of the comparatively complex theoretical material. Many years
of teaching experience and personal research in the field treated here are ref lected
in the book. A large number of examples is given. The book can prove useful not
only to engineers and graduate students, but also to students in the higher educa-
tional institutes. That which yesterday was the prerogative of a narrow group of
_ researchers is today becoming public property.
The authors would like to express their gratitude for the useful comments of V.Ye.
Dulevich, D.I. Lekhovitskiy and M.B. Sverdlik.
Table of Contents
Foraward 3
1. Generai Infor.mati.on on Rac'ar Data and its Processing
I. Fundamenta].s of Multiclidnnel Radar Signal Detection Theory
2. The Formulation ;;f Optimization Problems for Signal Detection and
~ Techn;_ques of Solving Them
2.1. The formulation of problems of signal detection optimization 7
2.2. The main indicators for two and three alternative ~detection efficiency 9
2.3. Detection optimality criteria lU
2.4. OFtimizing snlutions in the case of two alternative detection 11
2.5. Optimizing solutions in the case of three alternative detection 14
3. Optimal Detection of a Time Digitized Signal with Known Parameters
Against a Background of Correlated Gaussian Interference
3.1..
The formulation of the problem. Signal and interference
models
17
3.2.
Optimal detection algorithms for a time digitized signal
with known
parameters
21.
- 3.3.
The quality indicators and parameter of two alternative
detection of
a digitized signal sample
24
- 3.4.
Accumulation, compensation and interelement normalizatio
n with
respect to the interference level as component parts of
optimal
weighted processing (the example of a two element sample)
75
343
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFF[CIAL USE ON~Y
4. Optimal Multichannel Detection of a CW Signal with Known Parameters Against
a Background of Correlated Gaussian Interference
4.1.
The transition f rom digitized realizationsto continuous ones
27
4.2.
The integral matrix equation of a
weighting
vector
29
4.3.
The major results of multichannel
detection
theory for CW signals
and examples of its application
30
4.4.
Brief information from the theory
of linear
filters for CW signals
with constant parameters
36
4.5.
Matched f iltration as a detection
operation
against a background of
steady-state white noise
38
4.6.
Optimal f iltration as a detection
operation
against a background of
-
steady-state nonwhite noise
41
5. Specific features of the Ttultichannel Iletection of High Frequency
Signals
5.1.
The complex notation for narrow band high-frequency oscillations
42
5.2.
The approxima.tion of integrals from the products of narrow band high
_
frequencies
43
5.3.
The calculation of cross-correlation functions of random narrow band
-
high-frequency oscillations, M[a(t)b(s)], in linear systems with
constant parameters
44
_ 5.4.
Complex notation for received oscillations,.the useful signal oscil-
~
lation and the interference oscillations. The complex correlation
matrix of interference
45
5.5.
Complex notation for the main relationships of tne detection theory
for CW signals with known parameters
46
~ 5.6.
A model of white noise for the narrow band description of high
frequencies
49
5.7.
Examples of multichannel detector synthesis using complex notatiori
-
of high-frequency oscillationa
50
5.8.
Complex notation of filtration equations of high-frequency oscillations
53
6. Specific Features of Multichannel Detection of Coherent Signals with
Random Noninformational Parameters
6.1. A procedure to account for noninformatianal signal parameters and its
application to detection against a background of gaussian interference 54
6.2. The likelihood ratio and optimal detection algorithm for a signal with
a random initial phase 56
6.3. The likelihood ratio and optimal detection algorithm for a signal with
random initial phases and amplitudes 57
6.4. Structural conf igurations of detectors for signals with a random
initial pha.se and a w ith a random amplitude and initial phase 61
6.5. Quality indicators for a two alternative optimal detector of coherent
- signals w ith random parameters 62
7. Optimal Detection of the Simplest Incoherent Signals in Gaussian
Interf erence in the Case of Multichannel Reception
7.1. General information on incoherent signals 65
- 344
FOR OFF'ICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
7.2. Optimal incoherent signal detection algorithms for the simplest
incoherency models
66
7.3.
Methods of calculating the efficiency
of post detector incoherent
signal storage
70
7.4.
Examples of calculating the statistic.
characteristics of the output
voltages and efficiency indicators for
post detector storage devices
73
7.5.
Incoherent signal detection quality in
dicators in the case of a fixed
sample volume
76
- 7.6.
Incoherent signal sequential detection
quality indicators
78
7.7.
Quasi-optimal procedures of binary and
multilevel digital incoherent
storage
81
8. Specific Design Features of Detectors of Ra:tdom Gaussian Signals against
a Background of Gaussian Interference
8.1.
The general problem
of detecting a discrete gaussian random process
-
against a background
of discrete gaussian
interference
83
8.2.
Auxiliary mathematic
al propositions
84
8.3.
Expressions for the
logarithm of rhe likelihood ratio for digitized
and CW signals
87'
- 8.4.
Example of synthesis
of optimal detectors
for coherent gaussian signals
89
8.5.
Example of synthesis
af optimal detectors
for incoherent gaussian signals
91
8.6.
Example of synthesis
of optimal detectors
�or partially coherent
~
gaussian signals
97
II. Radar Signals and Modern Techniques of Processing Them
9. Mismatch and Resolution Functions of Coherent Space-Time Signals.
Signals without Intrapulse Modulation and Methods of Processing T.hem.
9.1.
General relationships for coherent signal mismatch functions
103
9.2.
Spatial (angular) mismatch functions
104
9.3.
Time-frequency mismatch functions
105
9.4
Mismatch functions and the ambiguity solid for single radar pulses
without intrapulse modulation
108
9.5.
Processing techniques for single radar pulses without intrapulse
modulation
111
9.6.
Mismatch functions and processing techniques for coherent radar pulse
trains ~
112
9.7.
Principles of correlation filter processing and specific features of .
its use in the case of quasi-continuous signals
117
9.8.
Specific featurE:: of matched and optimal resolution
119
9.9.
Weighted processing of coherent radar pulse trains
122
10. Frequency Modulated Signals and Methods of Processing Them
10.1.
Mismatch functions
for linearly frequency modulated radar pulses
123
10.2..
Compression of LFM
radar pulses in matched filters
127
10.3.
Correlation filter
processing with generalized heterodyning
131
345
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500020034-7
FOR OFFICIAL USE ONLY
11. Phase Keyed Signals and Methods of Processing Them
11.1.
Signals based on Barker codes and multiphase codes
135
11.2.
Signals based on linear recurrent digital sequences
138
11.3.
Continuous 0, ff signals,keyed M-sequences
142
11.4.
Continuous 0, 0 signals, keyed M-sequences
144
11.5.
Pulsed 0, w signals, keyed M-sequences
145
12. New Digital and Analog Coherent Processing Techniques
12.1.
Specific features of digital coherent processing
146
12.2.
Discrete (digital) processing in the time domain
149
12.3.
Digital filtering in the frequency domain
152
12.4.
The fast Fourier transform
156
12.5.
The realization and utilization of a fast Fourier transform
159
12.6.
The Walsh transform as a possible digital processing technique
160
12.7.
Numerical transformations as a possible digital processing technique
164
12.8.
New analog processing techniques. The dual pulse processing method
using spin echo
166
12.9.
A three pulse processing method using spin echo
171
12.10. The possibilities of using spin waves
173
12.11. Optical processing methods
175
12.12. Matched coherent optical processing in side-looking synthesized
aperture radars
.180
III.
Fundamentals of Multichannel Radar Measurement Theory
- 13. The Formulation and Procedure in the Solution of the Optimal Measurement
of Radar Signal Parameters. Basic Laws Governing a Regular Measurement
13.1.
The formulation of optimal measurement problems .
183
13.2.
The post-test probabiiity density in Bayes estimation theory
185
- 13.3.
Optimizing estimates based on a minimum average risk criterion.
Estimates of the post-test probability density maximum and the greatest
likelihood
186
13.4.
The post-test probability density and correlation matrix of vector
parameter regular measurement errors in the absence of apriori data
190
13.5.
The multidimensional error ellipsoid for a regular measurement in the
absence of apriori data
192
13.6.
Very simple examples of point and interval regular Bayes estimation
193
13.7.
The post-test probability density and.correlation matrix for vector
parameter regular measurement errors where apriori data are present
195
13.8.
Discriminator methods of optimal measurement
197
14.
Specific Features of the Optimal Measurement _ Timewise Constant
Parameters of Coherent Signals against a Background of White Noise
14.1.
The varieties of ineasurable parameters and the initial relationships
200
14.2.
Likelihood equations and inverse regular measurement error correlation
matrices for the nonpower parameters of coherent signals where their
ini*_ial phase is random and apriori data are lacking
201
346
FOR OFFICIAY. USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102109: CIA-RDP82-00850R000500020034-7
FOR OFFICIAL USE ONLY
14.3.
Equations oF generalized discriininators of nonpower parameters of
coherent signals where their initial phase is random
203
14.4.
Examples of optimal nontracking delay time and frequency meters
203
14.5.
The potential accuracy of the separate measurement of delay ti.me and
frequency
205
14.6.
The potential accuracy of the combined measurement of delay time and
frequency
207
14.7.
Time discriminators
209
14.8.
Frequency and time-frequency discriminators
212
14.9.
The potential accuracy of angular coordinate measurement
214
14.10.
Examples of dual channel phase angle meters
216
14.11.
An example of a multichannel phase angle meter designed around a
receiving antenna array
220
14.12.
An example of an amplitude angle meter
221
14.13.
Specific features of coherent signal power parameter measurements
223
15. Specific Features of the Optimal Measurement of Timewise Constant
- Parameters of Incoherent Signals against a Background of White Noise
15.1. General features of the measurement of timewise constant nonpower
parameters of incoherent signals
225
15.2.
Specific features of the utilization of a model for a rapidly
fluctuating incoherent signal in measurement theory
227
15.3.
A calculation procedure for the potential precision of the regular
;
measurement of nonpower scalar parameters of incoherent signals against
; .
a background of white noise
230
- 15.4.
Calculation of the potential precision of delay time and frequency
j
measurements.of rapidly fluctuating signals
233
15.5.
Specific features of time and frequency measurements using incoherent
signals
237
15.6.
An example of an angle of arrival meter for spatially incoherent
oscillations
239
15.7.
An example of an angle of arrival meter for timewise incoherent
oscillations
241
15.8.
An example of a noise signal delay time di�ference meter
243
15.9.
An example of an angular velocity meter for the motion of a rapidly
fluctuating signal source
247
16. Optimal Measurement of Parameters which Vary Discretely with Time.
Specific Features of Indirect Measurement
16.1.
Models of the timewise change in signal parameters
248
16.2.
A Gaussian-Markov model for a discrete change in a parameter
249
16.3.
Possibilities of accounting for the time interrelationship of the
random elements of a real time maneuver
250
16.4.
Examples of modeling changes in parameters
251
16.5.
A variant of a model for a discrete change in a parameter as applied
-
to the case of indirect measurement. The concepts of filtration,
prediction and smoothing estimates
253
16.6.
Linearized equations and structural configurations for the filtration
of discrete estimates in the case of direct measurement
255
- 16.7.
Lineariaed equations and structural configurations for the filtration
of discrete estimates in the case of indirect measurement
257
347
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICIAL USE ONLY
16.8. Examples of thc syntliesis and ttnalysie of ineters �or diacretely
timewise changing parameters 260
- 16.9. Cumulative optimal smoothing of estimates of a discretely changing
parameter 269
16.10. The general case of optimal measurement of a discretely changing 271
markov parameter .
17. Optimal Measurement of Parameters which Change Continuously with Time
17.1. A model of a continuous change in a parameter 273
17.2. The characteristics of a model of a continuous change in a parameter 275
17.3. Equations and structural configurations for filtration in the case
of continuous estimazion 276
17.4. Examples of the synthesis and analysis of ineters for continuously
timewise changing parameters . 279
17.5. Cumulative optimal smoothing and interpolation of estimates of a
continuously changing parameter 288
17.6. The inadequacy of models and the divergence of estimates of ineasured
289
parameters
IV. Detection-Measurement, Adaptation and Related Questions
18. Detection-Measurements and Its Anomalies
1
18
General considerations regarding detection-measurement
293
.
.
18.2.
Detection-measurement during secondary processing of information
295
18
3
Detection-measurement when combining information from several
.
.
different sources. The principle of equality.
29~
299
18.4.
Detection-measurement and measurement anomalies
18.5.
Anomalies in estimating the dispersion and espectation value for
samples from a normal set with a small number of elements in the
301
samples
19.
Various Ways of Taking Noninformational Parameters of Signal and
Interference into Account. Adaptation. Nonparametric Detection.
1
19
Varlous ways of taking noninformational parameters i.nto account under
.
.
apriori unknown conditions
304
2
19
Introducing and accounting for noninformational parameters under
.
.
condit:ions of apriori ambiguity
308
309
19
3.
Examples of accounting for noninformational parameters
.
313
19.4.
Adaptation
315
19.5.
The synthesis of automatic noise gain control
19
6
The simplest ways of taking the nongaussian nature of interference
.
.
317
into account
320
19.7.
Nonparametric sign detectors
322
19.8.
19.9.
Nonparametric rank detectors
The nature of nonlinear transformations when using ranking algorithms
324
348
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAL USE ONLY
20.
Specific Features of Parametric Signal Detection and Measurement against
a Background of Interference with Known and Unknown Spatial Correlations
20.1.
Madels of signals and interference. The formulation of the problem
327
20.2.
The inversion of a special kind of correlation matrices with large
dimensions
328
20.3.
The inversion of any kind of correlation matrices with large dimensions
130
20.4.
Variants of space-time coherent processing against a background of cor-
related interference with known correlation matrices
332
20.5.
Weighting vectors and directionality characteristics
336
20.6.
Gain and energy utilization factors
340
20.7.
Specific features of space-time processing in broadband systems
342
20.8.
Principles of estimating complex correlation matrices for spatially
correlatEd interference
344
20.9.
Discrete estimation of a correlation matrix for interference which
changes with time
345
20.10.
Continuous estimation of a correlation matrix for interference which
changes with time
348
20.11.
Estimation of a changing inverse correlation interference matrix
350
20.12.
Estimation of a weighting vector. The use of correlation feedback in
processors.
352
20.13.
Analog processors with correlation feedback loops
356
20.14.
Transient processes during adaptation
360
20.15.
Adaptation in the case of a high intensity useful signal
363
20.16.
Specific features of optimal measurement of signal parameters against
a background of correlated interference
365
20.17.
Examples of optimal measurement against a background of correlated
interference
366
21.
Specific Features of the Detection and Measurement of Signal Parameters
against a Background of Interference with Known and Unknown Time
Correlations
21.1.
Models of interference with a known correlation. The formulation of
the problem
368
21.2.
Variants of models of passive interference with known parameters
370
- 21.3.
Specific features of the utilization of a model for steady-state
nonwhite noise in velocity gating designs
376
21.4.
The possibilities for using a steady-state nonwhite noise model in
spatial gating
379
21.5.
Specific features of gating in the case of unmodulated CW and
quasicontinuous transmissions
381
- 21.6.
Velocity gating in pulsed raaars with unainbiguous range measurement
in each pulse train period
381
21.7.
Design examples of the simplest compensation devices applied to various
models of passive interference for the case of unique range measurement
384
21.8.
The velocity characteristics of the compensation circuitry for moving
-
target indication
387
21.9.
The coefficients of energy utilization,. gain and subnoise visibility ,
ef coherent signals against a background of time correlated interference
388
349
FOR OFFIC[AL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFF'ICIAL USE ONLY
21.10. The simplest methods of adapting MTI to individual unknown
interference characteristics 390
21.11. The possibilities of designing adaptive MTI circuits with the
estimation of the direct and inverse correlation matrices or the
optimal weighting vector 392
Appendices 395
Main Symbols Used 398
Bibliography 400
Subject Index 406
COPYRIGHT: Izdatel'stvo "Radio i svyaz 1981
8225
CSO: 1860/8
350
FOit OFFICIAL USE ONLY
�:r~:
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/49: CIA-RDP82-00850R040500020034-7
FOR OFFICtAL USE ONLY
UDC 621.396.664(024)
RADIO MEASUREMENTS
Moscow RADIOIZMERENIYA in Russian 1980 (signed to press 18 Nov 80) pp 2, 238-240
[Annotation and table of contents from book "Radio Measurements" by Aleksandr
Nikolayevich Pashkov, Nikolay Vasil'yevich Avenirov, Aleksandr Vasil'yevich
Kisilev and Valentin ivanovich Akimov, Voyenizdat, 16,000 copies, 240 pages]
[Text]
Annotation
The increase in the number and types of radio measurement instruments makes it
necessary to use a variety of inethods to test them. This book describes the
design principles of radio measurement instruments, rules for testing various
groups of instruments, techniques and singularities of use, as well as the
methodology for selecting test devices.
The book is intended for a broad group of readers interested in metrology, and
can be used to train and update specialists at testing organizations.
Table of Contents
Introduction 3
Chapter l. General Information on Radiotechnical Measurement Devices 5
1.1. Classification of radiotechnical measurement devices 5
1.2. Metrological characteristics of radiotechnical measurement
devices 18
Chapter 2. General Problems of Testing Radio Measurement Instruments 22
2.1. Basic concepts and definitions 22
2.2. Testing conditions 25
2.3. Documents regulating measurement device testing 25
2.4. Selection of standard measurement devices for making tests 26
2.5. How to prepare and conduct testing of a radio measurement
instrument 28
Chapter 3. Instrumentation Oscillators and Their Testing 31
3.1. Design principles of instrumentation oscillators 31
351
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FUR OFF[CIAL USE ONLY
1.
Function and classification
31
2.
Functional elements of instrumentation oscillators
33
3.
Technique of application of instrumentation oscillators
35
3.2. Low
frequency instrumentation oscillators
37
1.
General information '
37
2.
Features of functional sections of low-frequency instrumentation
oscillators
39
3.
Metrological characteristics of low-frequency oscillators
and basic testing rules
45
3.3. High-frequency instrumentation oscillators
47
1.
General information
47
2.
Features of functional sections of high--frequency oscillators
53
3.
Metrological characteristics of high-frequency oscillators
and basic testing rules
58
3.4. Pulse generators
66
66
1.
General information
2.
Features of functional sections of pulse generators
68
3.
Metrological characteristics of pulse generators and basic
testing rules
75
Chapter 4
. Instruments for Measuring Voltage and Current and Their
.
Testing
79
4.1. General information
4.2. Functional diagrams of voltmeters. Measurement of alternating
voltage
4.3. Digital voltmeters
4.4. Voltage measurement technique
4.5. Metrological characteristics of voltmeters and basic testing
inf ormation
4.6. Ammeter design principles
Chapter S. Instruments for Power Measurement and Their Testing
5.1. Design principles of low- and high-frequency power meters
5.2. Features of microwave power meters
5.3. Absorption-type watt meters
1. Calorimetric method of power measurement
2. Thermoresistor method of power measurement
5.4. Measurement of transmitted power
5.5. Measurement of pulsed power
5.6. Metrological characteristics of power meters and basic testing
information
Chapter 6. Instruments for Measuring Intensity of Electrical and
Magnetic Fields and Their TesCing
6.1. Design principles of radio wave field intensity meters
6.2. Field intensity meter antennas
352
FOR OFFICIAL USE ONLY
79
81
87
91
92
94
96
96
98
100
100
102
107
109
110
112
112
119
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OF'FICIAL USE ONLY
6.3. Instrur-entation receivers 121
6.4. Metrological characteristics of field intensity meters and
basic testing information 123
Chapter 7. Instruments for Investigating Wave Form and Their Testing 124
7.1. Design principles of electronic oscillographs 124
7.2. Features of functional sections of vertical beam deflection channel 126
7.3. Horizontal beam defLection channel. Types of scan and scan voltages 127
7.4. Oscillograph measuring instruments 131
7.5. Oscillographic measurement technique 132
7.6. Methodological characteristics of electronic oscillographs and
besic testing information 134
Chapter 8. Instruments for MPasuring Frequency and Time Intervals
and Their Testing 136
8.1. Design principles of instruments for measuring frequency and
time intervals 136
8.2. Resonant frequency meters 138
8.3. Heterodyne frequency meters 138
8.4. Counting electronic frequency and time interval meters 141
8.5. Standard-frequency reproduction devices 147
, 8.6. Metrological characteristics of frequency and time interval
, meters and basic testing rules 149
Chapter 9. Instruments for Measuring Phase Shift and Their Testing 155
9.1. Design principles of phase shift meters 155
9.2. Functional diagrams of needle arid digital phase meters 158
9.3. Metrological characteristics of phase meters and basic testing
information 166
Chapter 10. Instruments for Measuring Signal Spectrum Parameters and
Their Testing 169
i 10.1.
Design principles of instruments for spectrum analysis and
I
evaluation of spectral composition of oscillations
169
' 10.2.
Spectrum analyzers
175
10.3.
Nonlinear distortion meters
179
' 10.4.
Amplitude modulation percentage meters
181
, 10.5.
Frequency deviation meters
183
10.6.
Metrological characteristics of spectrum analyzer and basic
testing rules
184
10.7.
Metrological characteristics of nonlinear distortion meters
and basic testing information
188
10.8.
Metrological characterietics of amplitude modulation percentage
characteristics and basic testing rules
189
10.9.
Metrological characteristics of frequency deviation meters and
basic testing information
192
353
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
N'UK UN'FI(:IAL USE ONLY
Chapter 11.
Devices for Measuring Passive Circuit Parameters and
Their Testing
193
11,1. Measurement of lumped-constant circuit parameters
193
1.
Design principles of instruments for measuring parameters of
circuits with lumped constants
193
2.
Instr.uments using voltmeter-ammeter principle
195
3.
Bridge-type instruments
196
4.
Resonant instruments. Q-meters
198
5.
Metrological characteristics of instruments for measuring
parameters of lumped-constant circuits and basic testing
information
203
11.2. Measuring parameters of distributed-constant circuits
203
1.
Design principles of instruments for measuring parameters
of circuits with distributed constants
203
2.
Instrumentation lines and their testing
205
3.
Impedance meters and their testing
217
4.
Automation of ineasurement of parameters of circuits with
.
distributed constants
222
11.3. Measurement of Absolute Value of Gain (Attenuation). Attenuators
227
1.
Attenuator design principles
227
2.
Methods for measuring absolute value of gain (attenuation)
231
3.
Devices for attenuation measurement
233
4.
Metrological characteristics of attenuators and basic
testing information
235
Bibliography
237
COPYRIGHT: Voyenizdat, 1980
6900
CSO: 1860/99
354
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2407/02109: CIA-RDP82-00850R000500420034-7
FOR OFFICIAI. USF ON1.Y
UDC 537.312.62:621.372.834
SUPERCONDUCTING ACCELERATING MICROWAVE STRUCTURES
Moscow SVERKHPROVODYASHCHIYE USKORYAYUSHCHIYE SVCh-STRUKTURY in Russian 1981
(signdd to press 3 Feb 81) pp 2-4, 208
[Annotation, foreword and table of contents from book "Sugerconducting Accelerating
_ Microwave Structures", by Andrey Nikolayevich Didenko, Larisa Mikhaylovna Seveyukova
and A1'bert Aleksandrovich Yatis, Energoizdat, 1150 copiea, 208 paizes]
[Text]
Annotation
Superconducting microwave strucures are now used extensively in linear and cyclic
accelerators, as well as high-energy particle separators.
The goal of this book is to provide a presentation of the current status of the
problem of utilizing superconducting microwave structures in accelerator technology
and nuclear physics. Problems of developing methods for obtaining super-pure
metals and alloys for fabricating microwave structures are illuminated, and modern
methods for technological processing of the working surface of accelerating
structures are described in detail. A great deal of attention is devoted to atomic
- and nuclear-physical analysis methods and their utilization in superconductor
_ technology.
Results obtained by the authors, as well as the achievements of Soviet and foreign
scientists in the area of microwave etructure theory, technology and application,
are presented.
The book is intended for scientific workers and graduate students working in the
area of accelerated technology and applied superconductivity, and for students in
senior physics courses.
37 tables, 90 illustrations, 423 bibliographic references.
Foreword
Superconductivity is one of the most interesting physical phenomena. Superconduct-
ing magnets are in extensive use. However, superconducting microwave structures
are used to a significantly smallez degree, even though they are of great interest for
many branches of science and technology, primarily accelerator technology and
nuclear physics. This is mainly because the tremendous capabilities which such
systems have are poorly understood. To a certain degree, this is explained by
355 10-
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007102/09: CIA-RDP82-04850R000500020034-7
rvn vaG VNLY
the scattered nature of available publications and the lack of sufficient books
on this topic. Only two books have been published to present. A.N. Didenko's
monograph"Sverkhprovodyashchiye volnovody i rezonatory"(Superconducting Waveguides
and Resonators) (Moscow, Izdatel'stvo Sovetskoye radio, 1973), is devoted to super-
conducting waveguides and resonators in general. F.F. Mende, I.N. Bondarenko,
and A.V. Trubitsin's book "Sverkhprovodyatsiye i okhlazhdayemye rezonansnye sistemy"
(Superconducting and Cooled Resonant Systems) (Kiev, Izdatel'stvo Naukova Dumka
1976) examines low-puwer microwave systems, for example, high-stability micro-
wave oscillators, gravimeters, etc.
The goal of the present book is to provide a presentation of the current status
of the utilization of superconducting microwave structures in accelerator
technology.
Chapter 1 studies the fundamentals of hf superconductivity. Types of losses of
electrical energy in microwave structures which cause the occurrence of residual
resistance are examined. Technological and design methods for reducing losses
are explained in detail.
Chapter 2 describes electrophysical and thermo-physical characteristics of
suprconducting low- and high-power microwave structures. Problems occurring
in the creation of microwave structures for accelerator technology (thermal
instability, secondary electron effect, electron mechanical instability) are
considered, as is the relationship between the surface impedance of resonant
systems, as well as Q-factor, and the level of microwave power input. Special
~ attention is devoted to eliminating the multiplicatory efrect in microwave
structures. General requirements for the stagblity of parameters of superconduct-
ing microwave systems and instrumentation are given.
Chapter 3 presents methods of volumetric and surface analysis as applied to
supenconducting niobium and niobium-alloy materials. Mett;ods of neutron and
charged-particle activation analysis are studied. The most complete presentation
is that of the method for instantaneous analysis of near-surface layers of
_ superconducting materials based on Coulomb ion scattering, the use of nuclear
reactions, flourescent analysis spectroscopy, Auger electrons, secondary ion
spectroscopy, etc.
Chapter 4 examines methods for obtaining super-pure metals and alloys for making
superconducting sections of linear and cyclic acceleraCors. The possibility of
preparing superconducting resonators and waveguides of monocrystals, as well
as the prospects for utilizing such new materials as nitrides and carbides,
are discussed. Electrical instability in an electrochemical system is examined
for the first time with application to the electromechanical processing of ac
and dc microwave structures. A description is given of equipment for high tem-
pe.rature annealing in ultrahigh vacuum and for electromechanical processing of
superconducting structures.
Problems of removing heavy element impurities electrochemically from near-surface
Iayers of niobium and niobium-based alloys and of applying protective coatings
356
FOR OFFICIAL USE ONLY
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7
FOR OFF1ClAL USE ONLY
with good electrophysical properties on the surface of structures are examined.
Information is given on rhe condition of the surface of superconducting materials
during different stages of surface processing: this information was obtained by
5nviet and foreign scientists using metheds of nuc lear-physical analysis of
elementary composition.
Chapter 5 analyzes the application of superconduct ing microwave structures in
accelerator technology and nuclear physics. Questions of using hf-superconduct-
ing devices in linear electron, proton and heavy ion accelerators and in super-
conducting cyclic accelerators, as well as the pos sibility of using superconduct-
ing microwave systems as the basis f or creating electron microscopes, are 3is-
cussed. A great deal of attention is devoted to the use of superconducting micro-
wave structures in high-energy particle separators.
Materials from foreign and Soviet scientists which were dfscussed at conf erences
and meetings on applied superconductivity and char ged-particle accelerators over
the past five years which were published in the proceedings from these conferences
and various journals were used in writing tize book, as wEre original �indings
ob tained at the Scientif.ic Research Institute for Nuclear Physics at the Tomsk
Polytechnical Institute.
Chapters 1 and 2 were written jointly by A.N. Did enko and L.M. Seveyukova,
chapter 3 by A.A. Yatis, chapter 4 by L.M. Seveyukova, and chapter 5 by A.N.
Didenko.
Table of Contents
Foreword 3
Chapter i. Fundamentals of High-Frequency Superconductivity 5
I.I. Surface resistance of superconducting materi als at microwave
frequerLcies
1.2. Losses or microwave energy caused by impurities and defects in
crystal structure 12
1.3. Dielectric iosses in films and impurities 17
1.4. Energy losses caused by generation of sound waves by incident
electromagnetic radiation 21
1.5. Additional losses in superconducting microwave structures under
influence of magnetic fields 28
Chapter 2. Electrophysical Characteristics of Superconducting
Microwave Structures 35
2.7. Electrophysical parameters of superconducting waveguides and
35
resonators
2.2. Cr:Ltical magneric hf fields. Basic thermophysical characteristics
of superconduct.ing microwave structures 45
357
FOR OFFICIAL USE ONLY
~
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7

APPROVED FOR RELEASE: 2007/02109: CIA-RDP82-00850R004500020034-7
b'OR UP'H'I('IAL USE UNLY
2.3. Emission effect in superconducting microwave atructures and electrical
field gradients 53
2.4. General requirements for stability of parameters of superconducting
microwave systems 62
Chapter 3. Methods of Analyzing Superconducting Materials
76
3.1.
Basic information
70
3.2.
Volumetric activation analysis
73
3.3.
Instantaneous nuclear analysis methods
83
3.4.
Mass-spectrometric analysis
95
3.5.
Low-erier gy electron spectroscopy
99
3.6.
X-ray analysis methods
104
3.7.
Comparative characterization
109
Chapt
er 4. Superconducting Materials and Technology of Superconducting
Microwave Structures
116
4.1.
Fabrication of superconducting microwave structures from lead
and niobium
116
4.2.
Electrochemical polishing of working surface of superconducting
resonators and structures
124
4.3.
High temperature annealing of niobium superconducting resonators
-
and waveguides in high vacuum
134.
4.4.
Anode oxidation and its use for protecting and cleaning
surf ace layer
144
4.5.
Technology for obtaining superconducting alloy-based microwave
structure coating
156
4.6.
Prospects for utilization of monocrystals, nitrides and carbides
for fabricating microwave structures
172
Chapter 5. Application of Superconducting Microwave Structures in
Accelerator Technology and Nuclear Physics
176
- 5.1. Superconducting Linear accelerators 176
5.2. Utilization of superconducting accelerating systems in cyclic
accelerators 181
5.3. Superconducting separators for high-energy charged particles 188
5.4. Creation of electron microscopea based on superconducting
accelerating systems 192
Conclusion 196
Bibliography 197
COPYRIGHT: Energoizdat, 1981
6900
CSO: 1860/104
- END -
358
FOR OFFICIAL USE ONLY
~
APPROVED FOR RELEASE: 2007/02/09: CIA-RDP82-00850R000500020034-7