PROJECT KEMPSTER FINAL REPORT VOLUME II EQUIPMENT DEVELOPMENT PROGRAMS

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP67B00341R000800050001-7
• Release Decision: RIFPUB
• Original Classification: K
• Document Page Count: 287
• Document Creation Date: December 23, 2016
• Document Release Date: August 21, 2013
• Sequence Number: 1
• Publication Date: June 30, 1965
• Content Type: REPORT

Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R000800050601-7 7
Copy / of
/NI -/
P-1
1.
PROJECT KEMPSTER
FINAL REPORT
VOLUME II
EQUIPMENT DEVELOPMENT PROGRAMS
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
PROJECT ICEMPStut
Final Report
Volume II
Equipment Development Programs
Prepared by
Westinghouse Research Laboratories
WESTINGHOUSE ELECTRIC CORPORATION
Pittsburgh, Pennsylvania
June 30, 1965
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
FOREWARD
While the primary aim of the Kempster program was to
ascertain the extent of the radar cross-section reduction attainable
by an electron cloud, a retrospective examination of the program
clearly showed that the development of the specialized equipment to
prove the concept in flight required by far the larger portion of the
contract effort. During the preparation of the final report, then,
it was not surprising to find that there was sufficient mateihial to
be documented on the equipment effort alone to warrant an entirely
separate volume on that subject. The order of the chapters is
generally chronological, beginning with theory and experiments for
electron-beam-transmitting windows, following into the window tube
work and its more successful successor) the dynamically pumped gun,
and concluding with a description of the operation of the equipment
in the field.
Every attempt has been made to make the report complete.
However, the length of the program and the absence of any formal
reporting requirement during its first 80 percent make it entirely
possible that those preparing the final report have overlooked import-
ant contributions by others, some of whanare no longer even associated
with this organization. Where such omissions come to light, the
editor will appreciate being so informed) in order that he might
arrange for the necessary revisions in the document.
_
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
TABLE OF CONTENTS
Chapter 1 - Electron Beam Windows
1.01 Introduction
1.02 Assessment of Power Lost by Beam in Al Foil
1.03 Figure of Merits of Different Materials as
Electron Beam Windows
1.04 Spencer Thick Target Theory
1.05 Estimated Thermal Gradients for Thin Foil
Windows
1.06 Mechanical Stresses
1.07 Be as a Window Material
Be Foil
Pinhole Leaks
Improved Materials
1.08 The Be Joining Program
Welding
Spot Welding
Electron Beam Welding
Diffusion Bonding
Brazing of Beryllium
Ultrasonic Welding
1.09 Testing of Experimental Be Windows
Methods and Apparatus
Test Results
1.10 Aluminum Windows
Comparisons with Be
Joining Problem
Coined Al Window Approach
Testing of Al Windows
1.11 Miscellaneous Window Materials and Techniques
Lockalloy
Silicon
Ag-Mg
Pyrolytic Graphite
Vacuum Evaporated Be Foil
Cyrogenic Approach to High Power
Conclusions
Chapter 2 - Window Tubes
2.01 Introduction and Summary
2.02 Machlett Cathodes
2.03 Adjustable Cathode
Design
Fabrication
Testing
Modifications Recommended
Adjustable Optical System
Pages
1.01-1 to 1.11-4
1.01-1
1.02-1 to 6
1.03-1 to -7
1.04-1 to -13
1.05-1 to -33
1.06-1 to -41
1.07-1 to -4
-1
-2
-3
1.08-1 to -6
-1
-1
-1
-2
-3
-6
1.09-1 -5
-1
1.10-1 to -4
-1
-1
-4
-4
1.11-1 to
-1
-1
-1
-2
-2
-3
-4
2.01-1 to 2.09-1
2.01-1
2.02-1 to -2
2.03-1 to -5
-1
-1
-1
-3
-3
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
- 2 -
2.04 Demountable Tubes
2.05 Testing Windows with
2.06 Prototype Tubes
Introduction
Assembly
Vacuum System
Automatic Baking
Insulation System
2.07 Tube Shipment
2.08 Flight Tests of Dummy Tubes
2.09 Alternate Approaches
Demountable Tubes
Chapter 3 - Dynamic Pressure Staging
3.01 Introduction
3.02 Pumping Design
3.03 Design Diagram for Dynamic
3.04 Pumped Gun Application
Introduction
Modifications
Installation
Chapter
4.01
4.02
4.03
4.04
4.o5
4 - C-Gun and Bolt Cathode
Introduction
and Heraeus Gun
Pressure Staging
Pages
2.04-1 to 2.04-3
2.05-1 to
2.06-1 to - -7
-1
-4
-5
2.07-1
2.08-1
2.09-1
3.01-1 to 3.04-8
3.01-1
3.02-1 to -4
3.03-1 to -4
3.04-1 to -8
-1
-1
-7
4.01-1 to 4.08-4
4.01-1
-15
-12
-6
The Bolt Cathode
13 ..02-1 ?to
Introduction
-1
General Design Considerations
-3
Assembly Procedure
-6
Performance
-10
Vibration Testing of Bolt Cathode
-13
Electron Optics
4.03-1 to
Summary
-1
Laboratory Tests
C-Gun Laboratory Experiments
-7
Insulator Development
4.04-1 to
Vibration Testing of Cone and Cathode
Assembly
-6
C-Gun'Vacuum System
4.05-1 to
Refinements over
Sorb Pump
Design Diagram
Control Valves
Ejector Valve
4.06 Beam Losses
Calculation of Single-Scatter Losses in Gun
External Beam Scattering Measurements
Heraeus
-8
-2
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
3
Pages
4.07 Installations 4.07-1 to 4.07-9
,4.08 Operating Procedure 4.08-1 to -4
Introduction -1
Removing Transition Section -2
Exhaust and Outgas -2
Seasoning -3
Anode Tuning -3
Catcher -3
Chapter 5 - Power Supplies and Controls 5.01-1 to 5.06-8
5.01 Introduction 5.01-1
5.02 Comparative Weight Study 5.02-1 to -2
5.03 Q-Supplies 5.03-1 to -4
5.04 C-Supply Circuit Selection 5.04-1 to -10
Resonant Transformer Circuit -1
Higher Frequencies for Conventional Circuits -2
Voltage Multiple Circuits -4
Phase Balancing Network -6
5.05 C-Supply Description 5.05-1 to -10
Experience -7
C-Supply Weights -9
5.06 Control and Telemetry 5.06-1 to -8
Chapter 6 - Field Support Equipment and Procedure 6.01-1 to 6.o6-
6.01 Introduction 6.01-1
6.02 Ground Support Equipment 6.02-1 to -3
6.03 Flight Test Operating Techniques 6.03-1 to -12
Preflight Preparation -1
Procedures during Flight -2
Postflight Items -12
Recommendations -12
6.04 Maintenance and Modifications 6.04-1
Records and Systems -1
Maintenance Philosophy -1
-6.05 X-ray Protection of Personnel 6.05-1'
6.06 Recommended Modifications 6.06-1-to
? Introduction -1
Gun -1
Power Supply -2 -
Cooling System -3
? Mounting ?? -3
Environmental Test, Loads and Temperatures -4
Support Equipment 14.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
CHAPIEH 1
ETECiRON BEAM WINDOWS
1.01. Introduction
The design of a window device for transmitting electron beans
into an atmosphere having pressures from several torr to atmospheric
pressure depends on several technical factors. The thin foil which
constitutes the window should not present an excessive absorption
cross-section to the beam. It must be capable of dissipating the
power lost by the beam as it traverses the material without excessive
thermal gradients and, most important, must be able to withstand the
mechanical forces produced by thermal stresses and the pressure of the
external atmosphere.
The problem of theoretically computing whether a given window
configuration will stand up when transmitting the desired amount of
power requires determining the energy losses within the window as a
function of the transmitted power, calculating the thermal gradients
and mechanical stresses associated with this loading and with the
pressure differential across the window and comparing the result with
data on mechanical strength of the window material at the temperature
in question.
In the sections which follow, the losses associated with the
passage of high speed electrons through different materials will be
discussed, the thermal gradients associated with different configura-
tions will be computed, the mechanical stresses discussed, and a
comparison made with experimental data.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21 : CIA-RDP67-B00341R0008000-50001-7
-
1.02 - 1
-1.02. Assessment of Power Lost by Beam in Al Foil
Bethe(1) has given the following expression for the average
energy loss per unit path length for a charged particle traversing
matter:
dE2J1= f(E) - Nt4 (ergs/cm)
(1.02.01)
mv2 2T
? where N = nuMber of electrons per cm3 of matter
k 6 = charge of electron
in = mass of electron
v = velocity of incident electron
e.= 2.718
average excitation energy of the atoms being traversed.
The energy lost by a given particle in traversing a thin section
of window material in which electron scattering maybe considered neg-
ligible can be obtained by integrating Eq. 1.02.01.
? /frEo? dE to
TO.
E1
dx=
fEE
dE
7:1-ET?dx
(1.02.02)
where Eo is the energy of the incident particle and E is its energy at
a distance x within the material. The ( 1 ) function easily lends
dE dx
itself to graphical integration methods, so that a convenient solution
is available once the dE/dx function has been determined.
1. Bethel H.A.1 Handbuch der Physik (Julius Meyer, Berlin, 19663)
vol. 24.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.02 - 2
Lane and Zaffarano(2) have evaluated Eq. 1.02.01 for electrons
traversing aluminulifoil and obtained the following equation:
dE _
(,307) -115: [in (51.1 p ) + 3.66] key/mg/cm2 (1.02.03)
dx
P
where Z is atomic number, A is atomic weight, and p= v/c is the ratio
of the particle velocity to the velocity of light.
In evaluating Eq. 1.02.01 for Al, they employ the value of 150
ev for I, and have solved for N as follows:
N - riPZ
A
(1.02:04)
where n is Avogadro's number, and p is the density of the material.
Fig. 1.02.01 gives a plot of d_1 ax vs. kv as derived from
Eq. 1.02.03. The upper curve represents the actual curve obtained
directly from Eq. 1.02.03. The lower curve incorporates a 1.45 factor
reduction obtained by observing the departure of theory from experiment
as plotted for Al foil in Figure 19 of Reference 2. Part of the dis-
crepancy between theory and experiment was due to the fact that the
electron energy losses determined from the Bethe equation are taken
along the actual path of the electrons in their zig-zag course through
the material, whereas the desired result relates to the attenuation of
beam power with thickness in a given direction through the material.
In section 1.04 a more sophisticated theory which takes electron
scattering into consideration will be discussed. Use of the corrected
curve to obtain the energy loss from the beam as a function of window
thickness is exemplified by the cross-hatched area. Determination of
this area, of course, represents the integration of Eq. 1.02.02 -
referred to previously. The area of the cross-hatched section thus
2. R. 0. Lane and D. J. Zaffarano, Transmission of 0 to 50 kv
Electrons by Thin Films with Applications to Beta Ray
Spectroscopy, (U.S. Atomic Energy Commission ISC-439, Dec. 1953.)
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
0.5
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1
Fig. 1.02.01 dE dx in Reciprocal Key/mg/cm2
versus Primary Beam Kv
T I I
0.1
25
50
II II I
75 loo
125 150 175 200 Kv
Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R000800050001-7 ?
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.02 -
represents the mg/cm2 of Al foil which will reduce by 10 key the
average energy of electrons which are incident at 110 kv. From the
curve we see that this area turns out to a close approximation to be
(.22)(10) = 2.2 mg/cm2 which corresponds to a thickness of 8.2? or
.32 mil. Thus a 110 kv beam will lose approximately 10 key in
traversing a .32 mil thickness of Al foil. The calculations in
section 1.04 on the heat dissipation problem are based on a 10 key
loss in a 1/3 mil foil, corresponding to an 8.5? thickness. A 110
kv primary voltage is assumed.
On a more formal basis, it is evident that the cross sectioned
area of Fig. 1.02.01 represents the window height per unit area M
necessary to produce an average energy loss El per electron in a given
electron beam. It is given approximately by the product dE1dx at the
appropriate location of the curve and E1 the energy loss per electron.
Thus the average energy loss per electron can be expressed by the
following equation:
dE dE
El = M crx = pD
(1.02.05)
where D is the thickness of window material and dE/dx is the energy
loss per unit weight per unit area. For beam current I, the energy
per second expended by the electron beam in traversing the window is
given by the following equation:
dE dE
W = 1E1 = IM? = IpD
dx dx
(1.o2.o6)
It is apparent that electron range, which is defined as the
amount of material necessary to reduce the energy of the incident
particles to zero, is obtained by integrating Fig. 1.02.01 over the
whole area from the desired incident energy to zero energy.
As a means of obtaining the order of magnitude of the energy
i
losses involved when an electron beam traverses a material, the mg/cm2
required to achieve a given energy loss in Al foil is plotted in
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.02 - 5
Fig. 1.02.02 for 100 kv, 125 kv and 150 kv electron beams. These data
are obtained using the corrected curve of Fig. 1.02.01 assuming
it-is linear in the region of interest. The thickness of Al foil
necessary to produce a 10 key energy loss in Al foil is thus 2.10 2.4
/
- and 2.75 mg/cm2 0 corresponding to thicknesses of 7.8, 8.9 and 10.1
microns for 1000 125 and 125 key electrons respectively.
It is important to note that the loss in the window increases
as voltage is decreased. The maximum loading of the window will be
determined by the maximum allowable temperature gradient required to
dissipate the window heat losses incurred by the beam to an assumed
heat sink adjacent to the window. In section 1.05 this maximum
loading is assumed to be 30 ma for a primary beam of the order of
110 kv. For 8.5 micron Al foil, the loss per electron is about
10 key. The heat loss in the window is thus approximately (10,000)
(.030) = 300 watts. Table 1.02.01, below, shows how the key energy loss
varies with primary beam voltage, as well as how beam loading must
be varied to keep the energy losses at the window constant.
Table 1.02.01 - Energy Loss in 8.5 Micron Al Window, Also Tube
Maximum Allowable Current for 300 Watt Loss
Primary
kv
Anticipated Energy Loss
in Window in key
Maximum Allowable Current
in Tube in ma
149
7.9
38
110
10
30
81.5
13
23
58.3
16.6
18
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R000800050001-7 - 6
12
No
11
0
4?
10
9
0
8
cH
0
6
5
3
2
1
1 t
1 1 1
1
1111
1
10
Energy loss in KV
I 1 1 1 1 1 1 1 1 1 1 1 1 1 1
4o
10 20 30
150 kv
125 kv
0 kv
50
Fig. 1.02.02 Weight of Al Foil which will Produce
Given Energy Loss for 100 kv, 125 kv and 150 kv Beams
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4-1
0
18.5
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.03 - 1
1.03. Figure of Merits of Different Materials as Electron Beam Windows
Theoretical values of dE/dx and range for a number of different
materials have been determined by Nelms(3) as a function of the energy
of the incident electrons. Using these data, curves similar to
Figs. 1.02.01 and 1.02.02 can be plotted and the energy loss determined
as a function of the thickness of the material employed. Theoretical
data obtained in this fashion may be subject to the same order of
correction which has been made for Al foil in section 1.02. From
Fig. 1.02.02, particularly for higher energies, it is evident' that
the energy loss in traversing a given material can, to a reasonable
approximation, be assumed to be linear with the thickness of the
material.
The thermal figure of merit can be derived from the following
considerations if we neglect all heat losses from the window except
thermal conductivity losses. From heat theory the temperature
gradient across a window having thermal conductivity k, separated
by a distance A from a heat sink, can be expressed by the following
equation:
WA
Lai=
kLD
(1.03.01)
where W is the energy/sec absorped in the window, L is the length of
the beam and D is the thickness of the window.
Substituting Eq. 1.02.06 in Eq. 1.03.01, we obtain
LT =
kLD
AID (;i.) p
(1.03.02)
The thickness of the window DI which is present in both denomi-
nation and numerator of Eq. 1.03.02, cancels out. Thus the temperature
of the window is substantially independent of window thickness.
Essentially, this means that the improved thermal conductivity obtained
,by increasing the thickness of a given window is cancelled out by
iproportionate increase in the energy losses experienced by the beam
as. it traverses the thicker window.
3. A. T. Nelms, "Energy Loss and Range of Electron and Position,"
NBS Circular 577 (1952)
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.03 - 2
If the thermal figure of merit is taken as being proportional
to lidikT, then for a given beam size, current and window geometry,
we can define an electron beam window thermal figure of merit F by
the following equation:
F = dE
/3(;E)
(1.03.03)
where k is thermal conductivity, 0 is the density of the material and
dEidx is the energy loss per unit weight per unit area incurred by
the beam in traversing a given window material.
The relevant physical parameters of a number of window materials
are listed in Table 1.03.01 including the thermal figure of merit F
for 100 kv electrons-as defined by Eq. 1.03.03. Since F is proportional
to 1/LT, a knowledge of F permits ready assessment of the maximum
temperature associated with different window materials once the tempera-
ture gradients associated with one given material have been determined.
The values listed in Table 1.03.01 represent uncorrected
values of the electron beam energy loss parameters. For this reason
the energy losses associated with aluminum are somewhat lower on
this table than the corrected values determined from Fig. 1.02.01.
For very thin windows it is believed that the use of Table 1.03.01
to compare the theoretical performance of different materials with
the theoretical performance of Al will give reasonably accurate
answers when referred to thermal data for Al for which corrected values
have been obtained. Since the calculations of section 1.04 are based
on corrected values of Al, comparisons of different materials using
temperature data taken from section 1.04 will give "corrected" thermal
numbers for other materials.
Comparing, for example, the figure of merits of Al vs. Ag, we
obtain a ratio of 5.9/3.9 = 1.52 in favor of Al. Thus, if, as discussed
, in section 1.04, the window has been designed to limit the window
\temperature gradient to 150?C for aluminum with a 300 watt heat input
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
?
Thble 1.03.01 - Figures of Merit for Metal Windows for 100 key Electrons
14.
5
Material
and k
Atomic1 cal/cm2/ dE/dx in
Number gm/cm' cm/?C/sec k/p key cm2/mg
6
F - dE
P
Thermal Figure
of Merit
Be 4
, 1.85
.385
.21
3.4
6 x 10-2
c 6
2.22
.057
.026
3.71
6.9 x 10-3
Al 13
2.7
.52
.19
3.24
5.9 x 10-2
Ti 22
4.5.
.04
.0089
2.9*
3.1 x
Fe 26
7.87
,.19
.024
2.83
8.6 x 10-3 '
Ni 28
8.9
.14
.016
2.8*
5.6 x
Cu 29
8.94
.923
.10
2.72
3.8 x 10-2
Mb 42
10.2
.35
.034
2.46
1.4 x 10-2
Ag 47
10.5
.974
.093
2.40
3.9 x 10-2
W 74
19.3
.476
.025
2.05
1.2 x 10-2
Pt 78
21.45
.166
.078
2.01
3.9 x 10-3
Au 79
19.3
.707
.037
2.01
1.8 x 10-2
*By interpolation
**
Appreciable radiation
loss is neglected
7
1
F -
8
9
Window
10
11
P (gx)
Figure of
Thickness
in microns
Merit for
6T Gradient
to Equal
Range
Range
Absorption
for Window
Output
at
at
Losses for
Compared to
of 8.5
100 ky
100 ky
Equal.
= 150?C
micron Al
in
in
Thicknesses
for Al
Window
mg/cm
microns
.16
148?C
12.4
17.3
94
.12
l280?C
9.3
15.9
72
.11
150?C
8.5
18.3
68
.077*
2860?C
5.95
.01+5
1030?C**
3.48
21.2
27
.04*
1580?C**
3.1
.041
233?C
3.17
22.1
25
632?C
3.1
24.6
24
.01+
227?C
3.1
25.3
24
.025
714.0?C
1.93
30.1
16
.023
2270 C**
1.78
30.7
14
.026
492?C
2.0
30.6
16
tri
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.0
to the window, then an Ag target would operate with a temperature
gradient of (150)(1.52) = 227?C under the same beam and window
geometry conditions. As indicated previously, the temperature condi-
tions derivable from the thermal figure of merit are independent of
window thickness. For thicker foils with fractional ranges in excess
of 0.2 or 0.3, Spencer's theory (section 1.04) showed give more
accurate results.
It is evident that the maximum temperature experienced by
a window is only one of several parameters which must be evaluated
in selecting the window for an electron beam tube. One of the major
factors which must be considered relates to the window thickness
required to withstand the pressure differentials across the window
at the temperatures associated with operation of the tube under
maximum loadings. Still another important parameter relates to
the output power which is transmitted through the window at the
thickness in question, which is, of course, the total beam power
minus the energy losses incurred in the window. An even more import-
ant factor are the thermal stresses incurred as related to the physical
properties of the window at the temperatures in question.
Taking subscript 1 to apply to Al, and subscript 2 to apply
to Ag? the ratio of the window absorption losses in Al to those in
Ag from Equation (5) are as follows:
(E1)1p /dE\
k-)
1D 1 dx 1
(L.03.04)
dE
1 2 - p D i N
22 dx 2
It is evident that we can define still another electron beam
window figure of merit which will permit direct comparison of the
absorption losses incurred by the beam in traversing equal thicknesses
of different materials. From Eq. 1.03.04 for a given value of D, this
, figure of merit which applied only for very thin parts is given by
ithe following equation:
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.03 - 5
1
E = (1.03.05)
p(g)
Values of this figure of merit are also listed in Table 1.03.01.
It is important to bear in mind in using E to compare the electron
absorption losses of given materials that window thickness must be
taken into consideration, whereas, the use of the F figure of merit
for thermal gradient comparisons is independent of window thickness.
Returning to the comparison of Al and Ag, we observe El/E2 =
.11/.04 = 2.75. Thus, if a 10 key loss is incurred by a 110 kv electron
.beam in traversing an 8.5? thickness of Al, a 27.5 key loss will be
experienced by the beam in traversing an equal thickness of Ag for
thicknesses where scattering is not excessive. In. order to obtain
. the same power output through the Ag window as is obtained through
the 8.5? Al window, the Ag window thickness according to the simple
theory must be reduced by a 2.75 factor to 8.5/2.75 = 3.1?. The 3.1?
window must be able to withstand the required pressure differentials
at 227oC compared to similar requirements on the 8.5? Al window at
150?C.
Column 8 of Table 1.03.01 lists the relative temperature
gradients computed for different window materials on the basis of the
thermal factors F assuming that the beam and window geometry will
produce a 150?C temperature gradient for Al foil. This column is
independent of window thickness for the range of thickness for which
the linear approximation holds. Column 9 gives the window thickness
of different materials required to produce electron beam power out-
puts equivalent to that obtained for 8.5 micron Al foil for which a
10% power loss has been calculated.
Materials having low figure of merits F will have temperature
gradients which are so high that other means of heat dissipation such
as radiation cooling can no longer be considered negligible. The
total area of the window bombarded by the beam is .3 cm2 on the basis
of assumptions made later. ' A black body having this area at
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.03 - 6
- 1000?C will radiate 4.5 watts. At 2000?C a black body having this area
will radiate 45 watts. Thus, heat dissipation by radiation does not
get to be a Very major factor for window inputs of the order of 300
watts.
Column 10-gives the total 100 kv range of different window
materials computed by Nelms,(3) and column 11 gives these range data
in microns.
While data in this report are restricted to metal foils, pro-
mising materials are by no means limited to metals. A major 'problem
associated with non-metallic materials relates to the low thermal
conductivities frequently associated with them. However, some such
materials have .relatively high thermal conductivities. For example,
the thermal conductivity of Be0 at room temperature is comparable
with Be. Unfortunately, the thermal conductivity of Be0 drops from
.63 at 20?C to .07 at 800?C making it more difficult to capitalize
on the superior high-temperature properties of this material.
Data of the type shown in Table 1.03;01 can be used to make
a direct comparison of materials for thin foil windows. Column 9
of the table represents the approximate thickness of a given material
which will produce a 10% loss in an electron beam operating in the
100 to 110 kv range. Column 8 indicates the temperature gradient
from the center of the window to its edge which the window will
assume under operating conditions. As noted previously, this tempera-
ture is substantially independent of window thickness.
Columns 8 and 9 in a sense may be considered as defining the
requirements on the material imposed by the conditions of service.
It is then only necessary to consider the strength of the material
as a function of its thickness at the required operating temperature
for the thermal loading stresses produced in order to compare different
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.03 - 7
window materials. An extra premium results if a material will meet
the operational conditions in a thickness which is less than the 10%
loss thickness defined by column 9 of the table. Thus, if 1.5 microns
of silver will meet operating conditions, then it is evident that
the power loss to the window will be reduced by a factor of 2 from
10% to 5% of the total beam power.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.04 Spencer Thick Target Theory
spencer(45) has developed a theory from which energy
losses incurred by high speed electrons as they traverse a plane
perpendicular target of infinite thickness can be calculated as a
function of their fractional range. It was felt that computations
based on his data could be used with a fair degree of confidence
for thicker targets (in excess of 0.2 to 0.3 fractional range)
since fairly good correlation with experimental results were
obtained. A brief review of the calculations based on Spencers
data follows using the following definitions:
Let I(z) dz be the energy dissipated per square centimeter
in a plane layer between z and z + di. I(z) dz represents the
average energy per electron dissipated in the plane layer between
z and z + dz, ro = r (Eo) is the residual range of the electrons',
_-_as measured along their path.
4f)E = the stopping power of electrons at the initial
energy of the source E0.
x = ? the fractional range or thickness z transversed
ro
through mateirial divided by residual range for initial energy E0.
,dE
4x) = I(z)/k?)- is the de-dimensionalized energy
dr
Eo
distribution for the plane perpendicular case.
To obtain the energy loss (4W) incurred in penetrating
a given material to a thickness zl, it is necessary to integrate
I(z) dz
4.
L. V. Spencer, "Energy Dissipation by Fast Electrons,"
NBS .Monogram No. 1, Sept.
10,
1959.
5.
L. V. Spencer, Phys. Rev.,
98,
1597
(1955).
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
(a)L =1 I(z) dz
zl
By substitution; recognizing that dz = ro dx, we obtain:
o
(a)L .'* J(x4f.)E ro dx.
I
x1 o
,dE
For a given ro and k.--) are constants. The curves
dr
E
of AE versus fractional range for Pb, Sn, Cu, Al, C and Au shown on
Fig. 1.04.01 were obtained from the tabulated values for J(x)
published in Spencer's report using Simpson's rule for accomplishing
the integration. Spencer gives his data as a function of increments
Aix of .025. Solite loss in accuracy was incurred in the interest of
saving time by making the calculation using ix = .05 as the basic
increment for the integrations. When (CIT) is expressed in mg/cn,
dE
Eo
(601, is determined directly in kv.
The concept Of fractional range is used to eliminate beam
voltage as a parameter on the graphs. For voltages of present inter-
est, say 50 to 200 kv, there is little change in the data represented
in Fig. 1.04.01 with change in beam energy. The decrease in the
fraction of beam energy absorbed with increase in voltage is deter-
mined by the decrease in fractional range which occurs for a given
thickness material as the voltage is increased due to the increase
in range with voltage. The dashed lines in Fig. 1.04.01 represent
the fraction of the total energy reflected at high values of fractional
range.
In interpreting Fig. 1.04.01 it is important to keep in
mind that Spencer's "thick" target assumption gives larger values
of AE for a given value of.x than would be expected for a foil of
comparative thickness. For example, the 31% of electron beam energy
absorbed in traversing 20% of the range is based on the assumption
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.0
O. 9
O. 8
0.7
O. 6
O. 5
0.4
O. 3
0.2
0.1
1.04 - 3
CURVE 567591-8
Cu
sp,
04.-diu,
rr
rA?Ar
/
:!,
? RE (Sn)
?
r 4r
-A
--
RE-Clu
--7
----R.
0.1
0.2 0.3
04
O. 5
O. 6
0.7
?AE and RE versus fractional range x for thick materials
08
O. 9
Fig. 1.04.01
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.04 - 4
that the actual material is thicker than the fractional range thick-
ness under consideration. Thus the 31% absorption figure includes
some backscatter from sections of the material which are thicker than
the 0.2 fractional range for which AE is determined. A foil having a
thickness corresponding to fractional range of 0.2 would lose less
energy than would be expected on the basis of Spencer's thick target
theory since it does not have thicknesses beyond 0.2 of the frac-
tional range to cause scattering back into the thickness of interest.
Berger has applied a Monte Carle method in the application
of the transport of fast charged particles to compute reflection
and transmission characteristics of thin aluminum foils. His data
which has fairly good agreement with experiment, include quantitative
Information on electron beam energy changes as will numerical infor-
mation on changes in numbers of electrons as a function of the direction
of the incident electron beam and its initial energy. Fig. 1.04.02
gives a direct comparison of Berger's results for electrons have a
900 angle of incidence with Spencer's results which also assumes
normal incidence.
The following definitions are used in Fig. 1.04.02:
RN =
TE =
T =
N
fraction
fraction
fraction
fraction
fraction
fraction
of beam energy absorbed
of total number of electrons
of beam energy reflected
of total number of electrons
of beam energy transmitted
of total number of electrons
absorbed
reflected
transmitted
As would be expected, the fraction of energy absorbed,
predicted by Berger for thin foils is significantly less than is
obtained by Spencer on the assumption of electrons impinging as a
material of infinite thickness. The fractional absorption loss for a
foil having a fractional range thickness of 0.2 is only about 22%
as compared to the 31% obtained from Spencer's thick target theory.
AE)
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
a
Fractional Number or Fractional Energy
-n
rN)
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
C?N
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
dr
-Comparison Bercier and Spencer data for Al
AH curves Berger data except as noted
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7 ?
1.0)4
From Berger's data in Fig. 1.04.02 about 90% of the
electrons are transmitted through a thin foil having a thickness
of 0.2 of the fractional range, but only about 72.5% of the energy
is transmitted. About 8% of the electrons are reflected. These
reflected electrons take with them about 5% of the total incident
energy. Two percent of the electrons are neither accounted for by
transmission or reflection and thus are assumed to be captured
within the boundries of the foil.
/
In Fig. 1.04.03 the fraction of beam energy absorbed, AE,
is plotted as a function of atomic number for different values of
fractional range for a target of infinite thickness. The fractional
reflected energies corresponding to large values of fractional range
are shown at the top of the figure. As shown in Berger's data in
Fig. 1.04.02, full reflection takes place once the thickness of a
given material approaches about 30% of the fractional range.
In order to simplify obtaining data for different materials,
.range data obtained by Nelm's ate _plotted in Fig. 1.04.04 as a
function of atomic number for different values of kv. Nelm's data
on range translated in thickness of foil in mils as a function of
kv in the voltage range of interest. are shown in Fig. 1.04.05. ?
Spencer's data we applied to foils of different materials
to give the sum of AB as a function of foil thickness for the 125 kv
electrons plotted on Fig. 1.04.06. -Since the ranges of different
materials have roughly the same variation with change in kv, the thick-
ness of foils corresponding to the AB values shown in the curves can
be estimated for 150 kv and 100 kv by multiplying the thickness of
foil as given in the absence of Fig. 1.04.06 by 1.3 for the 100 kv
case and by .7 for the case of 150 electrons.
Fig. 1.04.06 can be used for comparing absorption raises
for Al foil with those obtained in the earlier discussion. The com-
parison tabulated below for .32 mil Al indicates that the estimate
first obtained for Al was reasonable.
Declassified and Approved For Release 2013/08/21: 'CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.0
0.9
O. 8
0.7
0.6
o.5
0.4
0:3
0.2
0.1
1.04 ? 7
CURVE 567593-8
O. 9
O. 8
x-0.
7
/
R Fractional Energy of Electrons
E'Reflected from Thick Target
-
6
x-0.5
/

x-0.4
r
4' /A
x0.3
TE
x-0.
AE + RE + TE ? 1
(See Example x ? 0.2)
.,
'
I
I
i
I I
I
I
i I
I
II
I
I
12
24
3-6
48 60 72 84 9i
Atomic Number Z
AE. fractional energy absorbed and Tr fractional energy transmitted as function of
er
fractionange x
Fig. 1.04.03
Declassified and Approved For Release 2013/08/21: 'CIA-RDP67B00341R000800050001-7
0.1
RE
0.2
0.3
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
2
Range in Mg/cm as Function of Atomic
Number for Various Values of Kv
Atomic Number Z
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
250 Kv
200 KV.
150 Kv
125 Kv
100 Kv
80 Kv
60 Kv
^
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
2
10
30
50 60 70
100 110 120 130 1110 150 160
Kv
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Fractional Absorption, AE
90*110'T '2Ta
12
0
C)
114
6.50
0.40
0.30
0.
0.10
Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R000800050001-7
r
Traational AbsorptibniLoss. for 15 Kv Electrons as Function
of Foil Thickness in Mils for Different Materials - Solid Lines
Thick Target Theory (Spencer)
I 1 [
Ag
.2
.8
10
1.2
1.4
Foil Thickness in Mils'
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.6
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.04 - 11
Source of Information
KV
AE
Fig. 1.02.01
110
.091
Spencer
125
.116
Spencer
100
.150
Berger
125
.08
Berger
100
.102
In this report unless specified otherwise, Berger's data
will be used for Al. Data for all other materials will be based on
Spencer's thick target theory. For law atomic number materials
where back reflection is minimal, excessive loss of accuracy would
not be entailed% However, for high atomic materials such as Cu and
Ag, serious errors may be incurred for thickness corresponding to
law values of fractional range. For very thin materials at high
atomic number it is possible that more accurate estimates can be made
from eq. 1.02.06.
In order to apply the consideration on losses just discussed
usefully, it is desirable to have available the constant of propor-
tionality between the power lost in a given window, Q, and the useful
transmitted power U. This constant of proportionalities is derived
as follows:
The total input of the beam WB = Q + U + WR where W is
the power reflected from the window and Q is the power lost in the
window proper. From consideration of Fig. 1.04.01 and 1.04.03 the
absorbed energy becomes prohibitively high as atomic number is in-
creased. For lower atomic number to which our attention is restricted,
the fraction of electrons reflected is less than 10% and is neglected
in these considerations.
WB = Q U
Q = AE WB
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
By substitution for WB in eq. 1.04.01
(1 - AE)
where f - is the constant of proportionality between
AE
useful and absorbed power.
In applying the data derived in this section and practical
windows it is important to make sure that electrons scattered in
traversing a given window are transmitted. as a result of the ,
particular window configuration employed.
-7
Values of f as .a function of kv are plotted in Fig. 1.04.07
for different thicknesses of Be and Al foil. It is interesting to
note that the pilfer transmitted through a 0.5 mil Be foil at 150 kv
is 19 times the power absorbed. Thus a window design having 0.5 mil
Be which will dissipate 200 watts should transmit almost 4 kw of
power. A half mil Al window will transmit about half this payer for
the same absorbed power.
(1 - AE)
U = Q -AE
-Q(f)
1.04 -12
1.04.03
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Factor f
18
16
12
8
4
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Factorf (b.---t:Thich Power Absorbed Must be
Multiplied to Obtain Useful Power) as
Function of Kv for Different Thicknesses
of Be and Al Foil
10.5
/
/
/
mil Be
/
/
/
/
?
mil Al
/
/
/
/
/
/
,
..
../
,, 1.0
mil Be
.75
mil Al
t
/
7-
7.
."
---
,
/
/?
/
...-'--
.--

..,
,
....-
mil Be
.0
mil Al
_
-
---
__
70
90
110
130
150
170
190
210
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
230
KV
250
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.05. Estimated Thermal Gradients for Thin Foil Windows
Assumption of Negligible Thermal Gradient Across Beam:
.03 am
Beam
1
A
.05
cm
1.05 - 1
= 8.5 x 10-4 am
Assume 0.3 mm wide beam landing on the Al foil window shown.
The power absorbedin window will dissipate to the wall according to the
equation:
W A
62 = -2- kLD
(1.05.01)
where W is power in calories/sec absorbed in window, k is thermal con-
ductivity, AT is the thermal gradient across window, L is the length- of
spot and D is thickness of window. Substituting in Eq. 1.05.01 on the
assumption that 300 watts = 72 calories per second is absorbed in the
.windowl A = .01 cm, D = 8.5 x10+ cm, k = 0.5 cal/cm2/sec, and solving
for L for AT = 150?C, we obtain L = 5.6 cm = 2.2". The assumption of 300
watts dissipation is,_of course, equivalent to a tube loading of 30 ma for
a 10 key energy loss per electron. In the previous sections, it was
determined that a 10 key loss is reasonable for the assumed thickness of
foil for primary beams of the order of 110 key.
The power dissipated by radiation and other mechanisms for
removing energy from the window have been neglected in these calcula-
tions.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.0 -2
Determination of Thermal Gradient From Center to Edge of Beam for the Condition
Above:
Let Qo = 2LC = power density in calories/sec/cm2, assumed uniform, being
absorbed, by window in the region of direct beam landing, where C = .015 cm
is the 1/2 width of beam, and Q = xgoL is the absorbed power which is.
flowing
flowing through differential element dx at distance x from center of beam
toward the edge of window. From heat conduction theory:
and
dT'
.dx kLD
(1.05.02)
-
ji
JT r o Q Lxdx ro QokD xdx C2 Q0
6X = dT = 0kLD - - 2kD (1.05.03)
C 1
where Tm and T are steady state temperatures at x = o and x = C resped-
72
tively. For the example cited first, %=
(5.6)(.015)(2) cal/cm2/sec.
Substituting in Eq. 1.05.03, we obtain 6T = 114?C.
-Total Thermal Gradient From Center of Beam to Edge of Window:
From the results above, the total temperature gradient 6T
from the center of window to edge of spot is equal to 150?C 4- 114?C =
261?C.. for a beam that is .03 cm by 5,6 cm long which impinges on an Al
'window 8.5 x 1074 cm thick. 'To-drop AT from 264?C to 150?C requireS
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.05 - 3
lengthening the beam from 5.6 cm to 9.9 cm or 3.9". For a 10 cm beam,
the thermal gradient from the center to edge of beam will be 65?C, and
the gradient from edge of spot to edge of window will be 85?C or 150?C
total temperature gradient. It will be observed that temperature gradient
from center of beam to edge is about 43% of the total 'gradient from center
to edge of window, and thus the temperature gradient across area of
window irradiated by beam is by no means negligible.
Effect of Thermal Heat Capacity on Window Loading:
Assume that the bombarded area of an Al window, which is 10 cm
long by .03 cm wide by 8.5 microns in thickness, has negligible heat
dissipation, what exposure time at 300 watt loading will bring the window
to 100?C? The heat capacity of Al is taken as .225 calories/gram.
The volume of window affected by this assumption is
(10)(.03)(8.5 x 10- ) = 2.55 x 10- cm3, and weighs (2.7)(2.55 x 10-4) -
6.9 x 10-4 grams. Heat input is 300/4.18 = 72 calories/sec.
The calories required to heat the window to 100?C are given by
(.225)(6.9 x 10-4)(100) = 1.55 x 10-2 calories. The time in seconds
required to equal 1.55 x 10-2 calories at an input of 72 calories/sec is
-4
1.55 x 10-2/72 = 2.16 x 10 seconds = .22 milliseconds = 220 microseconds.
Thus, significant increases in loading are possible for exposures
in the microsecond region. A 3000 watt heat input to the window should
be permissible for exposures of 22 microsecond duration having a 10% duty
cycle
Temperature Gradients Across Different Window Configurations:
(a) Slotted Configuration:
Let QB be the power absorPed by a given beam width 2C impinging
on a window of width w. As shown in Fig. 1.05.01a the clearance between
beam and window is given by dimension A. The beam density in watts/cm2
is assumed to be uniform over the cross sectional area of the line
focus and is defined as Qo. As outlined above the total temperature
Declassified and Approved For Release 2013/08/21: ICIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
E 2C A
w
?r
(a)
(c)
ci 144?
A
E 8
_t
-1
H20-
r
(e)
Window Configurations for Which
Temperature Gradients were Computed
Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R60680.0050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.05 - 5
gradient (LT)Tfrom the center of the window of length L to its edge is
given by
Since Q
wo
-
QBA C2, Q0
(AT)T = 2kLD 2kLD
? (1.05.04)
By substitution using Eq.
QB t, CN U(A+C/2)
(619T = 2/ F(2kDL)
where K is the thermal conductivity in cal/cm2/, ,?C/sec/cm* and the
,
power is expressed in cal/sec, and D is window thickness. All dimensions
are in ems. The first term in the above equations represents the
temperature gradient from the edge of beam to the window; the second
term the gradient from the center of beam to edge.
(b) Circular Windows
Referring to Fig. 1.05.01b let us assume a beam which loses a
total power of Qb calories in traversing a thin foil having a circular
cross section area of diameter 2R o. The tital diameter of the window is
2R A heat sink is assumed to be attached to the periphery of the
1.
window. The temperature gradient in the area directly loaded by the
beam is derived as follows: The temperature drop across a circular
strip dr located at a distance r from the center is given by
(1.:05.05)
dT Qdr
2nrkD
where Q is total loading in a circle of radius r?
but Q = nr2Qo0
Note that k is in general a function of T.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.05 - 6
where Qo again represents the power density in calories/cm2, assumed
uniform. Substituting for Q and integrating, one obtains for the
temperature gradient from the center of the beam impingement location
to its periphery.
clor02 QB
(AT)1 = TIZIT = 41(10
(1.05.06)
It is interesting to note that the temperature gradient from
the center of the beam to its periphery is independent of beam diameter.
Thus the temperature gradient across a circular window assumed to have
the same size kts the beam depends only on the total loading, the thick-
ness of the window, and its thermal conductivity, and not on the window
diameter. This provides an opportunity for obtaining a high total
loading by using many small diameter windows at relatively high loading.
If the beam diameter is smaller than the diameter_of the window,
we must also compute the temperature gradient 0492 between the periphery
of the beam and the edge of the window. We assume that the area being
loaded corresponds to a circle of diameter 2R0 for a window diameter
2R1
R
1 QB 0 dr -B
2nrkD = '2nkD
, --'R
o
r R
? a a / .1%
2kD R
Thus the total thermal gradient LT is given by
+
(1i05..07)
(1.05.08)
IT =
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Qor2
(12)T = 2kD
+
R,
Q13(+ 1n -A)
2okD 2
o
1.05 -7:
(1.05.09)
(1.05.10)
Terms involving power in any of the above equations can be
converted to watts by substituting the loading in watts and multiplying
1
the value of watts by 4.18 -
- (c) The Registration Problem:
Eq. 1.05.05 and Eq. 1.05.10 demonstrate the nature of the
registration problem. If the beam is made significantly smaller than
the window, the, extra- (6,1)2 term in these equations must be used to
compute the additional gradient associated with the fact that the
window is larger than the beam. If the window is diminished in size
until it coincides with the window, this additional temperature gradient
goes to zero. Exact coincidence between the window and the beam
cross-sectional areas represents the ideal case. It is also possible
to make the beam larger than the window to make the registration problem
less critical. This of course is accomplished at the expense of
requiring additional power, which serves no useful purpose, except to
simplify the problem of registry between beam and window.
The registration problem must be considered not only in the
light of accomplishing registration at the beginning of service, but
also of maintaining registration under conditions of useful life.
(d) Multihole Configurations:
Referring to Fig. 1.05.01c, let us assume a line focus from
having uniform power density impressed into a single line of holes.
In the initial assessment of temperature gradients we assume the main
heat sink is relatively thick and ignore temperature gradients due to
the width of the web t between individual holes. We also assume that
a beam of width d and length L is impressed on this series of holes
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
1.05 -8
in exact registration with them. An exactly registered configuration
in which the beam exactly matches each hole is of course much more
efficient and more difficult to realize in practice.
The fraction of the beam which impinges directly on the windows
will define the "optical" transmission, T of the window geometry which
is given by:
4 (A) = .785 (k)
, (1.05.11)
Strictly speaking this value of transmission assumes that the
beam is equal to L + t. Since t is 3`5'5,
5-1.
kliiq_i_c"kri,
),..v?
fio re, :
Iri K ur
-# ? 12.,
ZAL Set
IL ?e)
Ab4 dr5t-3
7JtlftCOILe4rai-
4.?1-PY
.N17 Ty
31-
Ay^ ?4;t. 13, t.A1..
...7 Irt. %
C poi- /3
, b et ,
4,p&(/ 4/ 11,0-41 Art02 91 h A(
'}.. ,,0111711
re_ ST 1 5 im &IC
4;17-1+
Tpe 5eld E.y--,...,45,43.
'\
A s ea
Tt air A :)T.5
ilfii u.s.--
ci ..,,v ms a 0,0
13 ,
A AI j. '/,".?,"
1?
1--N___.
_
- 1
....,_
,..-
,
4, -
?
)
,-._........,
, G
\
- . ,
I . ,) ,..., ?
.' \
.

?
ad--44,
/q II " /1 Nti,. 31-2/3/41c 341, _
4' 75YA..* / 70 w,--O
Fig. 4.03.02
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
;-/e3,14,,`;
-flet L A , ? n Di ..4
Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R000800050001-7 _ 6
o 5)ocs-rt,/,1 )( 114 Ii."1 "
. 0.4.. .... 3 ?
v ? THE SPACE ABOVE THIS LINE IS FOR FILING AND MUST NOT BE
),VLA-Q 7.ea- -0 s =
SUBJECT IYA144;:-'
7?
C STOMER
Vf too isv V - V 1 p ??="- awe.- to % SERIAL
ii
S. 0. OR I II SHAFT
xo<
/ e.
TEST NO. NO. CiL
TO DETERMINE
4CC , 3
WRITTEN ON
K 1674348
57.= 6/''
ni i
AVM
' ?
2 NI 0
=5t
5_
,
_
_
rffi.. IM
"""*"" ?
....4?i ./.4.
...f LY-
(1 4:3-:i
' '
?
.4.4.4.4...?t
INIIMIne
-5-0.,
14,3-L-1::
Min
.,,
1 ,
.
=WPM
ii....zi:
4,-,,,...?.4_,A1111-!..,
Kam
-
. . ei-ti
,
11111111,-0.,
.,,, .,
....c.32....(L(.,
,
,,,-- 74
, -
:______
Iiii-mimg
---.--=
- -----
?Meassi
....,_
_
_.,...?..
_ ? , ....._ ?
... ..
Pno,,,
MU
ov I
7. 30o
,,,..-
*
Igin
-I-in
SL,2----
lor,:.'2,
ES
0
IIIIIfrljMl
..
_
I/ II
S 41c
-.;

?.=;11"Me=
V -=./0
kV
I.
0 ck
Ir7-
0.)
MEM
NEMEININNIM
N
MI
55-s= 10
...J.,
.
........
my
wr
'
..,
go
..
MEE
2 5/t t
5;?
6pIA "
qo
) 1 Aj4
0 V
Mt
0 v. --7.-
rommom..............
IIMMeaninding
Normor.....
,
ro .),,,
'V 7. -
-.
,c,
--- s.,,
3
:._ 1 4
"%MP
'
11111
s,:. i
?......t ?
? _.?
.
,
:5-.'.1
,,
% .y.
......
IIIIIIIMIIIII?
5 T.
. MEMO
12-5
y.,:i
-www-----
Y\ al) 4Lct ear.- CLA
Fig. 4.03.03
Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R000800050001-7 ?
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.03 - 7
Grid bias was varied from -180 volts to zero volts in 4 steps
and 5 mil aluminum discs were used at anode distances of 2-3/4" and 6-3/4".
As the bias was decreased the beam aperture in the disc location of
2-3/4" became smaller. It was the smallest at zero bias and the beam
crossover apparently was closer to this point at xero bias than any other.
This data combined with beam visual data indicates that the crossover
point approaches the cathode as the bias is increased.
By the use of such methods as these, the design converged to
a cathode configuration which could be used in the new C-gun with
reasonable transmission through the orifices.
C-Gun Laboratory Experiments:
Some means of monitoring the performance of the beam was
needed in the cOmpact C-gun, whose beam transfer section could have discs
inserted for burnout measurements only with great difficulty. Viewing
ports in the pumping ducts were finally adopted as superior to electrically
insulated probes around the apertures or temperature instrumentation on
the apertures, since the ports changed the performance of the gun in no
way. The viewing ports were installed in the ducts of two chambers and
gave a view of the beam as well as of the apertures, which quickly became
incandescent when struck by a beam. This warning was the most direct
warning to reduce power and trim up the beam direction. Another periscopic
arrangement of mirrors was used to permit the observer to see around his
x-ray shielding with a telescope for a very satisfactory view. The beam
transfer section connected to a 6" dia., 12" long Pyrex cylinder. This
chamber was observed through a shield of lead-glass in the form of a
rectangular brick 9" x 5" and 3" thick. With this system, the beam could
easily be observed at any pressure desired, from microns to 40 torr.
Beam cross section as a function of molecular species and pressure of the
ambient gas could also be observed and measured through this window.
Tests in the Heraeus gun usually involved no more than 2 or 3
minutes running at any time. From these series of tests were determined
the: optimum dimensions of the electron optics to be used in the C-gun.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.03 - 8
However, when longer running times were required in the C-gun a time
dependent decrease in the emission current was experienced.
This decline of beam current with time was checked for
repeatibility. As the cathode came up to terminal temperatures the
current was noted and its rate of decline recorded. After terminal
temperatures were reached (about 40 minutes after cathode turned on
to space charge limit operation), the cathode was then shut off and
left over night to cool. When the cathode was again turned on, beam
current versus time was noted, and the plot repeated itself. ,This
would indicate that the drift was due to temperature changes in the
cathode and not due to any permanent change such as distortion, bom-
bardment filament poisoning, or bolt emission face poisoning. An
examination of the cathode individual components that might change
enough in dimension with temperature to cause an appreciable change
in the electron optics was then undertaken.
The first and most obvious component was the tungsten emitting
bolt itself. It was fastened in the cathode by set screws 1-1/2" from
.itsi emitting face, where the approximate temperature was 2450?C. A
teperature at the clamp of 500?C or less was assumed until terminal
temperatures in the cathode were reached. Knowing the coefficient of
thermal expansion of tungsten and using an average value for these two
temperatures over a 1-1/2" length, the axial expansion of the bolt was
calculated to be .008". Of course, most of this expansion would occur
as soon as emission temperatures were reached at the emission face.
Therefore, the change in length of this bolt over a period of 1/2 hour
would be fairly insignificant. In any case, the change in dimension
would be in the wrong direction to account for a decrease in beam current.
The second component considered was the focusing cup. If it
were to increase in length, thereby causing an increase in-S , the
ggg
perveance of the system would decrease. Since the temperature of this
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.03 - 9
extension grid was unknown, the decision was made to construct one
out of Invar instead of stainless steel. The coefficient of expansion
of Invar is considerably less than that of stainless steel, and it was
hoped that a test would show an appreciable change in the rate of decline
of, beam current,with time. The tests Showed no noticable difference in
the current drift with cups of the two metals.
A third consideration was the body of the cathode itself.
This, too, was stainless steel, so one structure was made of Invar and
tested. Again, no change in drift was noted. Further investigations
in-Volved grinding down the main support stem of the bolt, clamping the
bolt closer to the filament area, and removing the grid from the cathode.
All of these efforts proved futile in controlling the drift. Since there
was a schedule to be met, zero bias was relinquiShed, and the current
control used in the Heraeus gun was adopted. Cathode bias resistors
were added in the auxiliary power supply.
At this time in the program the requirement was advanced to
12 kilowatts of power or 80 ma at 150 kv, per gun. The first step was
to run the cathode fully space charge limited and note the maximum current'
obtainable from this particular cathode geometry at 150 kv. As previously
mentioned, the beam current, I , was 64 ma for the first few minutes of
cathode operation and then began to deteriorate at an initial rata of
approximately 1 ma/min. It eventually stabilized in around 40'min at
an:I of 40 ma. From this data, it was obvious that higher emission
currents and a drastic improvement in stability were needed.
There are three basic adjustable dimensions (specified by
large letter S in Fig. 4.03.04) that can be varied in the cathode to
produce a change in perveance; S ) Sand S Experience has shown
gb'
hv ggg)
that high voltage arcing becomes a major problem if Shy is adjusted to.
less than 1-1/8" so this dimension was fixed at 1-1/4". Variations in .
had also been tried but it was noted that transmission efficiency
decreased if the bolt were adjusted any closer to the anode than its
present setting. The remaining parameter Sggg was then selected as the
one with which to experiment.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
SHV
Dgga
1/8" Radius
Heat ShielcL
or Lid DBolt
Sbg,
g
Grid
4.03 - 10
Extension
Grid
MA KW Sb,t, Sgb Dga Sggg Dgga SHV Dpa DBolt R-BIAS
9K
7K
5. 2K
50
7. 5
.014"
.001"
. 250"
. 578"
1. 180"
1. 594"
.312"
.062"
70
10. 5
. 014"
. 001"
. 250"
. 578"
1180"
1. 156"
. 312"
. 062"
93
14. 0
.014"
.. 001"
. 250"
. 398"
.961"
1. 250"
. 312"
. 062"
Note: Min. Shy 1 5/32"
Anode 0. 2" per rev.
E-gun bolt nomenclature and settings
Fig. ?4. 03. 04
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
It was apparent that S dimension alone could not be changed
ggg
without causing loss of transmission, since tests on the Heraeus gun had
revealed the dependency of the beam crossover location on the spacing
Of S ggg. This crossover point then was a function of the shape of the
equipotential lines surrounding the emission face of the bolt. As a
first approximation, the Sgsg and D combinations that exhibited 100%
gg
transmission were compared and found to fall on a rough parabola that
intersected the emission face of the bolt(called the origin) at one point
and the perimeter of the extension grid at the other point. The equation
of this parabola with dimensions in inches was y = 1.71 x 2. Two more
extension grids that would fall along this same parabolic envelope
were then constructed but with the added advantage that the total distance
from anode to emission face would be reduced. That is, Shy
fixed, D was reduced, thereby causing a reduction in S
gg ggS
the perimeter of the extension grid to intersect this envelope. The
results were quite rewarding in that 100% transmission was maintained
and higher emission currents produced.
The intermediate extension grid with a Dgg of 960 mils was set
at Sggg = 462 mils. The emission current was approximately 90 ma at 5 min
of cathode running time.and 70 ma at 15 minutes.
Transmission became 100% after 14 minutes. The other extension
grid had a D of 706 mils and S was set at 300 mils. Since we were
gg ggg
set up to read only 100 ma of I at maximum, we could not determine the
true I versus time for the first 16 minutes because the beam current was
more than 100 ma. However, at 22 minutes the beam current was 87.5 ma
and transmission was 9814. The same drift problem in I was apparent.
Even though these power levels were attainable it became obvious that a
further step was required to produce a satisfactory electron gun.
Cathode self bias was selected as the most expedient method
available to obtain more stable beam operation. Since biasing the present
cathode only led to defocusing of the beam, some dimension had to be
.changed in the cathode in conjunction with the addition of bias. The
was kept
in order for
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.03 - 12
decision was made to change D from .100 inch to 1/4" as an attempt to
pattern the area immediately surrounding the emission face after that
of!an. Heraeus, filament type cathode because the filament cathode was
knOvn to operate with satisfactory stability.
The results of this redesign are summarized in Fig. 4.03.04.
The beam stability became satisfactory, and the perveance was so increased
by the change in D that the standard extension grid (Dgg = 1.18") could
be used.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.o4 Insulator Development
Because of the temperatures expected in the cathode area,
a high alumina insulator was desired. Simplicity of shape was
emphasized, for there was not time to retrace the insulator develop-
ment if troubles appeared, and the performance of the simpler shape
could be predicted with more confidence. The shape which combined
r6quired surface leakage path length with mechanical strength was
a cone, convex toward the pressure) Figure 4.04..01. Initial models
made of Vycor 'borosilicate glass cracked when the cathode was heated,
but the alumina units proved as effective as had been hoped.
When subjected to the design maximum vibration environment, a cone
and cathode assembly developed resonant displacements which would
have severely degraded performance, but all joints remained sound
and leak-tight.
The defects which did appear in operation fall into two
classes: braze failures and electrical failures. In time, some
cones began to leak at the ceramic-to-metal seal, apparently
through failure of the metalized joint. Service life could be
extended in these cases by coating the area with Dow Corning
silicone varnish number 994 and baking over 150?C. The elec-
trical failures divide in turn into two classes. A yellowish
coating on the vacuum side of the insulntor built up over a
long period of time, apparently through decomposition of diffusion
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
7.40 Rer
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
7.00 REF
4.041 - 2
3.50 REF
P?1?-.66
REE
SKS
ir*
r/A
1 ire
. ...., uuuu
- m .N,.. N ...- ? ,
filstanntt isonnun mitinim
u,............. ra ukuvuum,
aq1.111.1r0I-
111041,..
% 1
k? k 7,4 1:,-& V
N ;IV F CIO
$NSk 4; 0
Ilk 001r A
Or01 Al
,,,.-4.....= --i,----------
/147iite
/ /
//
// //'
1
1
t
II'
MI im
' win
_ ..er...........-;,..,c-
-,,,-.7? ..-Iso?
-------,
ifilgainfAVINV,litictivwykar elimin
;ill
i-e.1111
/ /
,5?Nosovkh. ..10K-4014,1,1,...=
1
?1110:--. .
, , 4?.? __.
.
!,. ...?
- gun section
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Fig. ?4. 04.01
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.o4 - 3
pump oil. Unless removed with scouring powder and elbow grease,
this layer eventually flashed over and developed low resistance
tracks. The other kind of failure was an actual perforation of the
insulator, which occurred in a region of high electrical stress.
In an effort to control voltage distribution and charge
accumulation, the outside of one insulator was coated with a silicon
carbide paint chosen for its nonlinear volt-ampere characteristic.
The effects on steady dc test were good, but the operating exper-
ience was no better than with bare cones, and poorer when the paint
began to come off.
An insulator in the form of a flattened cone with annular
convolutions has been built Figure 4.04.02. Laboratory tests of
the insulator alone and operating tests of the insulator and cathode
iri a gun have proved the suitability of this design for extended
service. No vibration tests have been run with this structure, but
the shortened cathode stem permitted by this design raises the
natural frequencies of the structure and greatly reduces the moments
generated by shock loading. The wavy surface was necessary to pro-
vide the length of surface path 150 kv operation with a bare surface
required.
Along with the more compact insulator, a new plug-in
cathode (Figure 4.04.02) was designed to replace the existing
complicated heavy and sensitive configuration. Replacement in the
field was a major effort, which could greatly be simplified with
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Plug in cathode and new insulator
Fig. 4.04.02
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.o4 - 5
the plug-in feature, allowing replacement through a handhole instead
of by disassembling the gun. All set screw type adjustments were
replaced by proper dimensional tolerance and control thereby making
all assemblies readily interchangeable. The new cathode is also much
lighter in weight, raising the natural frequency of the cantilivered
system.
Mechanical centering of the bolt is automatic once,the stem
and insulator are centered at initial build up. The bolt is centered
with respect to the pilot diameter during manufacture, insuring auto-
matic centering when the cathode assembly is plugged in place.
A new stem was designed to accommodate the plug-in cathode.
A copper sleeve was incorporated in the stem to help transfer the
heat generated by the bolt back to the gas cooling of the power
supply.
The overall diameter of the new cathode assembly is much
smaller than the old and can conceivably reduce the tendency for
electrical arc-overs. The outside surface has no sharp edges or
joints which also can minimize arc-overs.
Approximately one month of continual testing on two
cathode assemblies has confirmed the practicality of the plug-in
features and also indicated that performance was as good or better
than the previous design.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.04 - 6
Vibration Testing of Cone and Cathode Assembly:
While the alumina insulator looks extremely rugged and stress
resistant, the weak spot is the brazed joint to the metal cup which
begins the cathode structure. In addition, the means adopted to
isolate the ceramic-to-metal seal from the cathode heat contributed
a large compliance to the cantilevered cathode stem. The possible
effects of vibration, both on the bolt cathode and on the insulator,
were a source of concern. To put quantitative limits on these effects,
a series of vibration tests was run. The cathode was not hot, nor
was the space around the cathode evacuated, but the atmospheric
damping was estimated as low enough to be neglected, and not nearly
powerful enough equipment was available to shake the system needed
to keep a cathode in operation.
The first step was a slow sweep from very low to very high
frequencies to identify the natural frequencies of the vibrating
structure. Two were found with low dynamic amplification. Then at
each of these, the assembly was vibrated for two minutes at an
excitation level defined by Figure 4.02.05, which was the guide in
all questions of environmental vibration. After the vibration
exposure, the assembly was retested for natural frequency and leak
checked to discover any failure. None was discovered.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.05 C-C-un Vacuum System
Refinements over Heraeus:
The C-gun system has several modifications in its pumping
components from those in the Heraeus gun, although the pressures
maintained at each comparable stage have been similar. These modi-
fications can be listed as follows:
1) Use of MCF-60 diffusion pump or P4CS-2B diffusion
pump and BCRU-20 baffle.
2) Completely redesigned sorb pump.
3) Redesigned ejector for compactness and lightweight.
4) Shorter dump connections.
5) Use of discharge tube through access door.
Sorb pump:
The redesigned sorb pump for the C-gun system is shown
in Figure 4.05.01. It consists of three concentric tithes or con-
tainers, each with its own function. The center tube, of copper,
contains the molecular sieve material and is, in turn, immersed in
the second container, of stainless steel, holding the liquid
nitrogen. The third tube, also of stainless steel, surrounds the
first two, and contains the vacuum insulation, reducing the loss of
liquid nitrogen to a minimum. The high polish on the surfaces
surrounding the vacuum space is intended to reduce radiant heat loss
through it. Two tubes, a fill and vent line, transfer the nitrogen
into the second container.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
MOLECULAR
5/ EVE
Li Qu/O -
/V / TROGE /V
CARTR/DGE
HEATER
Co/v4x Firr/Na
PROTECT/VE
LAP
VAcuunei
51-qinILES5 5TEEL
SCREEN
COPPER PINCHOFF
SFCT/ON OF AS?S'EMBLED .SORB PUMP
Fig. -4. 05. 01
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.05 - 3
As one of the features of the design was the liquid nitrogen
container insulating vacuum, the attainment of this vacuum was very
important. Several methods were available:
(1) A direct connection from system vacuum to container
This method as employed in standard liquid nitrogen traps
is inconsistent with the use of the final vacuum system,
since the high vacuum portion of the system is not alWays
in operation.
(2) A closed system employing a final sealant
The vacuum chamber is to be pumped down and sealed off
by either a valve or pinchoff. The final vacuum attained
is limited by the cleanliness of the container as well as
the pumping system used. In addition, outgassing of the
valve or pinchoff upon sealing would increase the pressure
of the sealed system.
(3) A closed system employing an internal absorbent and
final sealant
A molecular sieve material is physically placed inside
the vacuum chamber, which is pumped down and sealed off.
As used on many commercial metal dewars, the material is
in contact with the liquid nitrogen container and by its
own pumping action would further reduce the pressure in
the chamber. If proper assembly and pumping facilities
were available, this method would be best, but its use
demands prior baking of the molecular sieve quick
assembly, and pumpdown.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.05 - 4
(4) A closed system employing bake-out and final sealant
Though long and relatively involved, this method is the
best available at present to produce maximum results.
It involves clean assembly procedures and pumpdown on an
ion pump bake-out system. A 400 C bake produces ultra-
high vacuum in the chamber and a copper pinchoff seals it.
The sorb pump must periodically be regenerated by heating
tO drive out the gases adsorbed in its active material. All the
atmospheric gases but water will be desorbed at room temperature.
Complete regeneration requires that the sorbent bed reach a tempera-
ture of 3650C. An electric heater was built into the pump for this
purpose, but a series of failures of the heater lead wires prompted
the adoption of a hot air baking. Air, heated to about 500?C, was
blown into the fill pipe of the liquid nitrogen chamber, escaping
from the vent, and the sorbent reached 270?C in 45 minutes,
starting from liquid nitrogen temperatures. This proved to be hot
enough for a satisfactory bake, even though it fell somewhat below
the manufacturer's recommended temperature.
In the final system for which the pump is intended, an
.059 in. diameter by .187 in. long aperture is used in the sorb
pump's stage. The preceding (higher pressure) stage is maintained
at .5 torr, or less, during operation of the system, and the pump
must maintain at least .010 to .040 torr for the operation of the
Declassified and Approved For Release 2013/08/21 : CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
4.05 - 5
following (lower pressure) stage. Fig. 4.05.02 shows the minimum
characteristics the pump must maintain. The throughput presented is
.04 torr liter/sec.
The performance of the sorb pump on test is summarized in
Fig. 4.05.03.
Design Diagram:
Fig. 4.05.02 is the vacuum design diagram for the C-gun.
A water cooled baffle on the oil diffusion pump was added to control
oil backstresming.
Control Valves:
The function of the control valves is to close the vacuum
system when a beam is not required to permit maintaining a vacuum in
the cathode space and avoid upsetting the pump cascade. There are two
valves which perform this task together. One is the shutter valve which
seals the P2 chamber from ambient and the other is the ejector valve
which seals the P2 chamberfrom the ejector. Together they completely
isolate the system.
The shutter valve seals the water cooled N2 nozzle with a
two-way pneumatically operated plate sliding on a stationary 0-ring
surrounding the nozzle opening. An insulated electrical contact at the
top of the air cylinder indicates when the valve is opened, thereby
disabling a high voltage lockout.
In use there have been several malfunctions of the valve,
mostly attributable to assembly techniques. One of these is the use
of a large amount of vacuum grease for lubrication) which eventually
solidifies from the high temperatures caused by the beam, making the
valve stick. For use at high power levels, the design will need modi-
fication, such as removing the 0-ring completely'from the orifice area
when the valve is opened.
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
Declassified and Approved For Release 2013/08/21: CIA-RDP67B00341R000800050001-7
CONDUCTANCES OF NOZZLES
AS FUNCTION OF SIZE
CONDUCTANCES OF NOZZLES AND PUMPS AS FUNCTION OF PRESSURE AND THROUGHPUT
pOF -DIAGRAM
lid
lid
too 1
lid
I 0 100
ImF(l/d)=Kx'F?
t `F(I/d) 'F(I/c1). V To
10'
S,F
10
100
CC
10
S,F
0
10-
10-
,/ \c,
4/ ,4
./ Tarr sifter ??/.