DEVELOPMENT OF THE LOCKHEED SR-71 BLACKBIRD

• Document Type: CREST
• Collection: General CIA Records
• Document Number (FOIA) /ESDN (CREST): CIA-RDP90B00170R000100080001-5
• Release Decision: RIPPUB
• Original Classification: K
• Document Page Count: 27
• Document Creation Date: December 22, 2016
• Document Release Date: September 27, 2011
• Sequence Number: 1
• Publication Date: July 29, 1981
• Content Type: LETTER

Mr. Leslie Dirks
Deputy Director for Science
and Technology
Central Intelligence Agency
Washington D.C. 20505
Some time ago, it was suggested that I record the development
aspects of the U-2 and the Blackbirds stemming from the
original A-11 program.
I have just completed the paper, which I hereby submit for
clearance. When it is cleared, I would plan to submit it to
the Lockheed Horizons Technical Journal for publication, and
I would also plan to discuss it in a series of seminars which
I am committed to at the U.S. Naval Academy in September of
this year.
You will note that I am including, in total, Mr. Bill Brown's
discussion on the development of the J-58 engine. This was
given at an AIAA meeting in Long Beach in May of this year.
Obviously, it was totally cleared, and I am also enclosing a
copy of a letter from Pratt & Whitney indicating their permission
to use the paper as I have.
A few comments on some of the things included in the paper:
1. I have carefully avoided any reference to performance
data not already approved by our security groups.
2. No information on any of the payload packages or electronic
gear has been included.
3. In regard to the A-11 airplane, I wanted to be sure to give
the CIA credit for their sponsoring the original development.
The fact that the aircraft ended up being stored was not only
well known and published in various CIA documents, but we
were instructed several years ago to roll out all the aircraft
and put them on display at the Palmdale facility. This was
done and strangely enough, there were no comments on where
did this aircraft come from. They were then put back in their
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-2-
4. The photographs of the D-21 were published in a West German
newspaper. The aircraft were displayed at Davis Montham in
such a way that the Germans had no difficulty in securing the
photographs. I believe the aircraft are still in that position.
I have talked to Lt/Colonel Mark Smith of Det 6 in regard to
getting their approval for publication of the article and I am
submitting an identical letter of this to him.
I would appreciate receiving your approval as soon as possible
so that I might prepare for my Navy speeches in the Fall.
I am also enclosing a copy of the Lockheed Horizons so that you
can see what type of journal it is. Mr. Bob Fuhrman's article
on the Polaris is an excellent example on what data was cleared
in his case.
#/ 4 11~
 PRATT& WHITNEYAIRCRAFT GROUP
Government Products Division
4818 Lincoln Blvd., Suite 212
Marina Del Rey, California 90291
Mr. Kelly Johnson
Lockheed Corporation
P.O. Box 551
Burbank, California 91520
Dear Kelly:
In response to your telephone call of
June 1st, be advised that P&WA has no
objection to your incorporating some, or
all, of Bill Brown's recent AIAA paper in
your planned paper on the SR-71. We are
pleased to cooperate with you in this
effort.
0 1
tnam
TICFNIOLOOItS
 LOCKHEED SR-71 BLACKBIRD
Clarence L. Johnson
Senior Advisor
Lockheed Corporation
INTRODUCTION
This paper has been prepared by the writer to record the development
history of the Lockheed SR-71 reconnaissance airplane. In my
capacity as manager of the Lockheed Advanced Development Division
(more commonly known as the "Skunk Works") I supervised the design,
testing, and construction of the aircraft referred to until my
partial retirement five years ago. Because of the very tight
security on all phases of the program, there are very few people who
were ever aware of all aspects of the so-called "Blackbird" program.
Fortunately, I kept as complete a log on the subject as one
individual could on a program that involved thousands of people,
over three hundred subcontractors and partners, plus a very select
group of Air Force and Central Intelligence Agency people. There
are still many classified aspects on the design and operation of
the Blackbirds but by my avoiding these, I have been informed that
I can still publish many interesting things about the program.
In order to tell the SR-71 story, I must draw heavily on the
data derived on two prior Skunk Works programs -- the first Mach 3+
reconnaissance type, known by our design number as the A-12, and
the YF-12A interceptor, which President Lyndon Johnson announced
publicly 1 March 1964. He announced the SR-71 on 24 July of the
same year.
BACKGROUND FOR DEVELOPMENT
The U-2 subsonic, high-altitude reconnaissance plane first flew
Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-2-
in 1955. It went operational a year later and continued to make
overflights of the Soviet Union until 1 May 1960. In this five-
year period, it became obvious to us who were involved in the U-2
program that Russian developments in the radar and missile fields
would shortly make the U-Bird too vulnerable to continue overflights
of Soviet territory as indeed happened when Francis Gary Powers was
shot down on May Day of 1960.
Starting in 1956, we made many studies and tests to improve
the survivability of the U-2 by attempting to fly higher and
faster as well as reducing its radar cross-section and providing
both infrared and radar jamming gear. Very little gains were
forthcoming except in cruise altitude so we took up studies of
other designs. We studied the use of new fuels such as boron
slurries and liquid hydrogen. The latter was carried into the early
manufacturing phase because it was possible to produce an aircraft
with cruising altitudes well over 100,000 feet at a Mach number of
2.5. This design was scrapped, however, because of the terrible
logistic problems of providing fuel in the field.
Fear for the safety of our orbiting reconnaissance satellites
in a hot war made it apparent that we still would need a manned
reconnaissance aircraft that could be dispatched on worldwide
missions when required. From vulnerability studies, we derived
certain design requirements for this craft. These were a cruising
speed well over Mach 3, cruising altitude over 80,000 feet, and
possessing a very low radar cross-section over a wide band of
frequencies. Electronic counter measures and advanced communications
gear were mandatory. The craft should have at least two engines for
safety reasons.
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-3-
GETTING A GRASP ON THE PROBLEM
Our analysis of these requirements rapidly showed the very
formidable problems which had to be solved to get an acceptable
design.
The first of these was the effect of operating at ram-air
temperatures of over 800?F. This immediately ruled out aluminum as
a basic structural material, leaving only various alloys of titanium
and stainless steel to build the aircraft. It meant the development
of high-temperature plastics for radomes and other structures, as
well as a new hydraulic fluid, greases, electric wiring and plugs,
and a whole host of other equipment. The fuel to be used by the
engine had to be stable under temperatures as low as minus 90?F in
subsonic cruising flight during aerial refueling, and to over 350?F
at high cruising speeds when it would be fed into the engine fuel
system. There it would first be used as hydraulic fluid at 600?F
to control the afterburner exit flaps before being fed into the
burner cans of the powerplant and the afterburner itself.
Cooling the cockpit and crew turned out to be seven times as
difficult as on the X-15 research airplane which flew as much as
twice as fast as the SR-71 but only for a few minutes per flight.
The wheels and tires of the landing gear had to be protected from
the heat by burying them in the fuselage fuel tanks for radiation
cooling to save the rubber and other systems attached thereto.
Special attention had to be given to the crew escape system to
allow safe ejection from the aircraft over a speed and altitude
range of zero miles per hour at sea level to Mach numbers up to 4.0
at over 100,000 feet. New pilots' pressure suits, gloves, dual
oxygen systems, high-temperature ejection seat catapults, and
Darachutes would have to be developed and tested.
Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-4-
The problems of taking, pictures through windows subjected to
a hot turbulent airflow on the fuselage had to be solved.
HOW THE BLACKBIRD PROGRAM GOT STARTED
In the time period of 21 April 1958 through 1 September 1959, I
made a series of proposals for Mach 3+ reconnaissance aircraft to
Mr. Richard Bissell of the CIA and to the U.S. Air Force. These
airplanes were designated in the Skunk Works by design numbers of
A-1 through A-12.
We were evaluated against some very interesting designs by the
General Dynamics Corporation and a Navy in-house design. This
latter concept was proposed as a ramjet-powered rubber inflatable
machine, initially carried to altitude by a
rocket boosted to a speed where the ramjets
on this aircraft rapidly proved
The carrying balloon had to be
lift the unit which had a proposed wing
Convair's proposals were much more
balloon and then
could produce thrust.
it to be totally
a mile in diameter
area of 1/7 of an acre'
serious, starting out with a
ramjet-powered Mach 4 aircraft to be carried aloft by a B-58 and
launched at supersonic speeds. Unfortunately, the B-58 couldn't go
supersonic with the bird in place, and even it if could, the
survivability of the piloted vehicle would be very questionable due
to the probability of ramjet blow-out in maneuvers. At the time of
this proposal the total flight operating time for the Marquardt
ramjet was not over 7 hours, and this time was obtained mainly on a
ramjet test vehicle for the Boeing Bomarc missile. Known as the
X-7, this test vehicle was built and operated by the Lockheed
Skunk Works'
 The final Convair proposal, known as the Kingfisher, was
eliminated by Air Force and Department of Defense technical experts,
who were given the job of evaluating all designs.
On 29 August 1959 our A-12 design was declared the winner and
Mr. Bissell gave us a limited go-ahead for a four-month period to
conduct tests on certain models and to build a full-scale mock-up.
On 30 January 1960 we were given a full go-ahead on the design,
manufacturing, and testing of 12 aircraft. The first one flew
26 April 1962.
The next version of the aircraft, an Air Defense long-range
fighter, was discussed with General Hal Estes in Washington, D.C.
on 16 and 17 March 1960. He, along with Air Force Secretary for
Research and Development, Dr. Courtlandt Perkins, were very pleased
with our proposal so they passed me on for further discussions with
General Marvin Demler at Wright Field. He directed us to use the
Hughes ASG 18 radar and the GAR-9 missiles which were in the early
development stages for the North American F-108 interceptor. This
we did, and when the F-108 was eventually cancelled Lockheed
worked with Hughes in the development and flight testing of that
armament system. The first YF-12A flew 7 August 1963.
In early January 1961 I made the first proposal for a strategic
reconnaissance bomber to Dr. Joseph Charyk, Secretary of the Air
Force; Colonel Leo Geary, our Pentagon project officer on the YF-12;
and Mr. Lew Meyer, a high financial officer in the Air Force. We
were encouraged to continue our company-funded studies on the
aircraft. As we progressed in the development, we encountered very
strong opposition in certain Air Force quarters on the part of those
trying to save the North American B-70 program, which was in
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-6-
considerable trouble. Life became very interesting in that we were
competing the SR-71 with an airplane five times its weight and size.
On 4 June 1962 the Air Force evaluation team reviewed our design and
the mock-up -- and we were given good grades.
Our discussions continued with the Department of Defense and
also, in this period, with General Curtis LeMay and his Strategic
Air Command officers. It was on 27 and 28 December 1962 that we were
finally put on contract to build the first group of six SR-71
aircraft.
One of our major problems during the next few years was in
adapting our Skunk Works operating methods to provide SAC with
proper support, training, spare parts, and data required for their
special operational needs. I have always believed that our Strategic
Air Command is the most sophisticated and demanding customer for
aircraft in the world. The fact that we have been able to support
them so well for many years is one of the most satisfying aspects of
my career.
Without the total support of such people as General Leo Geary
in the Pentagon and a long series of extremely competent and helpful
commanding officers at Beale Air Force Base, we could never have
jointly put the Blackbirds into service successfully.
BASIC DESIGN FEATURES
Having chosen the required performance in speed, altitude, and
range, it was immediately evident that a thin delta-wing planform
was required with a very moderate wing loading to allow flight at
very high altitude. A long slender fuselage was necessary to
contain most of the fuel as well as the landing gear and payloads.
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90BOO170R000100080001-5
-7-
To reduce the wing trim drag, the fuselage was fitted with lateral
surfaces called chines, which actually converted the forward
fuselage into a fixed canard which developed lift.
The hardest design problem on the airplane was making the
engine air inlet and ejector work properly. The inlet cone moves
almost three feet to keep the shock wave where we want it. A
hydraulic actuator, computer controlled, has to provide operating
forces of up to 31,000 pounds under certain flow conditions in the
nacelles. To account for the effect of the fuselage chine air
flow, the inlets are pointed down and in toward the fuselage.
The use of dual vertical tails canted inward on the engine
nacelles took advantage of the chine vortex in such a way that the
directional stability improves as the angle of attack of the
aircraft increases.
AERODYNAMIC TESTING
All the usual low-speed and high-speed wind tunnel tests were run
on the various configurations of the A-12 and YF-12A, and continued
on the SR-71.
Substantial efforts went into optimizing chine design and conical
camber of the wing leading edge. No useful lift increase effect
was found from the use of wing flaps of any type so we depend
entirely on our low wing-loading and powerful ground effect to get
satisfactory takeoff and landing characteristics.
Correlation of wind tunnel data on fuselage trim effects was
found to be of marginal value because of two factors: structural
deflection due to fuselage weight distribution; and the effect of
fuel quantity and temperature. The latter was caused by fuel on
the bottom of the tanks, keeping that section of the fuselage cool,
Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90BOO170R000100080001-5
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-8-
while the top of the fuselage became increasingly hotter as fuel
was burned, tending to push the chines downward due to differential
expansion of the top and bottom of the fuselage.
By far the most tunnel time was spent optimizing the nacelle
inlets, bleed designs, and the ejector. Figure shows a quarter-
scale model on which over 250,000 pressure readings were taken. We
knew nacelle air leakage would cause high drag so an actual full-
size nacelle was fitted with end plugs and air leakage carefully
measured. Proper sealing paid off well in flight testing.
shows the nacelle test set-up.
With the engines located half way out on the wing span, we were
very concerned with the
should an inlet stall.
very high yawing moment that would develop
fuselage that immediately
rudder booster to apply 9
of 0.15 seconds. This device worked
very often couldn't tell whether the
They knew they had had a blowout, of
that occurred with a "popped shock."
installed accelerometers in the
yaw rate and commanded the
correction within a time period
so well that our test pilots
right or left engine blew out.
course, by the bad buffeting
Subsequently, an automatic
restart device was developed which keeps this engine-out time to a
very short period.
POWERPLANT DEVELOPMENT
Mr. Bill Brown of Pratt & Whitney presented a fine paper on this
subject 13 May 1981 to the American Institute of Aeronautics and
Astronautics in Long Beach, California. Mr. Brown's paper is
reproduced herewith.
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-9-
J58/SR-71 Propulsion Integration
William H. Brown
Pratt & Whitney Aircraft Group
See Attached
Paper and Figures
A thru K
I have little to add to Mr. Brown's fine paper except to
record an interesting approach to the problem of ground starting
the J-58. We learned that it often required over 600 horsepower to
get the engine up to starting RPM. To obtain this power, we took
two Buick racing car engines and developed a gear box to connect
them both to the J-58 starter drive. We operated for several years
with this setup until more sophisticated air starting systems were
developed and installed in the hangars.
STRUCTURAL PROBLEMS
The decision to use various alloys of titanium for the basic
structure of the Blackbirds was based on the following considerations:
1. Only titanium and steel had the ability to withstand the operating
temperatures encountered.
2. Aged B-120 titanium weighs one half as much as stainless steel
per cubic inch but its ultimate strength is almost up to
stainless.
3. Conventional construction could be used with fewer parts involved
than with steel.
4. High strength composites were not available in the early 1960s.
We did develop a good plastic which has been remarkably
servicable but it was not used for primary structure.
Having made the basic material choice, we decided to build two test
units to see if we could reduce our research to practice. The first
Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-10-
unit was to study thermal effects on our large titanium wing panels.
We heated up this element with the computed heat flux that we would
encounter in flight. The sample warped into a totally unacceptable
shape. To solve this problem we put chordwise corrugations in the
outer skins and re-ran the tests very satisfactorily. At the design
heating rate, the corrugations merely deepened by a few thousandths
of an inch and on cooling returned to the basic shape. I was accused
of trying to make a 1932 Ford Trimotor go Mach 3 but the concept
worked fine.
The second test unit was the forward fuselage and cockpit, which
had over 6,000 parts in it of high curvature, thin gages, and the
canopy with its complexity. This element was tested in an oven
where we could determine thermal effects and develop cockpit cooling
systems.
We encountered major problems in manufacturing this test unit
because the first batch of heat-treated titanium parts was
extremely brittle. In fact, you could push a piece of structure
off your desk and it would shatter on the floor. It was thought that
we were encountering hydrogen embrittlement in our heat treat
processes. Working with our supplier, Titanium Metals Corporation,
we could not prove that the problem was in fact hydrogen. It was
finally resolved by throwing out our whole acid pickling setup and
replacing it with an identical reproduction of what TMCA had at
their mills.
We developed a complex quality control program. For every
batch of ten parts or more we processed three test coupons which were
subjected to the identical heat treatment of the parts in the batch.
One coupon was tensile tested to failure to derive the stress-
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-11-
strain data. A quarter-of-an-inch cut was made in the edge of the
second coupon by a sharp scissor-like cutter and it was then bent
around a mandrel at the cut. If the coupon could not be bent 180?
at a radius of X times the sheet thickness without breaking, it
was considered to be too brittle. The value of X is a function
of the alloy used and the stress/strain value of the piece. The
third coupon was held in reserve if any reprocessing was required.
For an outfit that hates paperwork, we really deluged ourselves
with it. Having made over 13 million titanium parts to date we can
trace the history of all but the first few parts back to the mill
pour and for about the last 10 million of them even the direction of
the grain in the sheet from which the part was cut has been recorded.
On large forgings, such as landing gears, we trepanned out 12 sample
coupons for test before machining each part. We found out the hard
way that most commercial cutting fluids accelerated stress corrosion
on hot titanium so we developed our own.
Titanium is totally incompatible with chlorine, flourine,
cadmium, and similar elements. For instance, we were baffled when we
found out that wing panels which we spot welded in the summer,
failed early in life, but those made in the winter lasted indefinitely.
We finally traced this problem to the Burbank water system which had
heavily chlorinated water in the summer to prevent algae growth but
not in the winter. Changing to distilled water to wash the parts
solved this problem.
Our experience with cadmium came about by mechanics using
cadmium-plated wrenches working on the engine installation
primarily. Enough cadmium was left in contact with bolt heads which
had been tightened so that when the bolts became hot (over 600?F) the
hnlt hpnrls just dropped off' We had to clean out hundreds of tool
Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-12-
boxes to remove cadmium-plated tools.
Drilling and machining high strength titanium alloys, such as
B-120, required a complete research program to determine best tool
cutter designs, cutting fluids, and speeds and feeds for best metal
removal rates. We had particular trouble with wing extrusions
which were used by the thousands of feet. Initially, the cost of
machining a foot out of the rolled mill part was $19.00 which was
reduced to $11.00 after much research. At one time we were
approaching the
net dimensions,
$30,000,000 new
learn how
which was
numerical
were able
ability at our vendor's plants to roll parts to
but the final achievement of this required a
facility which was not built.
Gordon was given $1,000,000 for a research program to
to forge the main nacelle rings on a 50,000-ton press
successful. Combining their advances with our research on
controls of machining and special tools and fluids,
to save $19,000,000 on the production program.
parts from going under-gage while in the acid bath,
series of metal gages two thousandths of an inch
thicker than the standard gages and solved this problem. When we
built the first Blackbird, a high-speed drill could drill 17 holes
before it was ruined. By the end of the program we had developed
drills that could drill 100 holes and then be resharpened
successfully.
Our overall research on titanium usage was summarized in reports
which we furnished not only to the Air Force but also to our vendors
who machined over half of our machined parts for the program. To
use titanium efficiently required an on-going training program for
thousands of people -- both ours in manufacturing and in the Air
 Force in service.
Throughout this and other programs, it has been crystal clear
to me that our country needs a 250,000-ton metal forming press --
five times as large as our biggest one available today. When we
have to machine away 90% of our rough forgings today both in
titanium (SR-71 nacelle rings and landing gears) and aluminum (C-5
fuselage side rings) it seems that we are nationally very stupid'
My best and continuing efforts to solve this problem have been
defeated for many years. Incidently, the USSR has been much smarter
in this field in that they have more and larger forging presses than
we do.
FLUID SYSTEMS
Very difficult problems were encountered with the use of fuel tank
sealants and hydraulic oil. We worked for years developing both of
these, drawing as much on other industrial and chemical companies as
they were willing to devote to a very limited market. We were
finally able to produce a sealant which does a reasonable job over a
temperature range of minus 90?F to over 600?F. Our experience with
hydraulic oil started out on a comical situation. I saw ads in
technical journals for a "material to be used to operate up to 900?F
in service." I contacted the producer who agreed to send me some
for testing. Imagine my surprise when the material arrived in a
large canvas bag. It was a white powder at room temperature that
you certainly wouldn't put in a hydraulic system. If you did, one
would have to thaw out all the lines and other elements with a blow
torch' We did finally get a petroleum-based oil developed at the
University of Pennsylvania to which we had to add several other
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-14-
chemicals to maintain its lubricity at high temperatures. It
originally cost $130 per gallon so absolutely no leaks could be
tolerated.
Rubber 0-rings could not be used at high temperatures so a
complete line of steel rings was provided which have worked very
well. Titanium pistons working in titanium cylinders tended to
gall and seize until chemical coatings were invented which solved
the problem.
THE FLIGHT TEST PHASE
The first flight of the A-12 took place 26 April 1962 or thirty
months after we were given a limited go-ahead on 1 September 1959.
We had to fly with Pratt & Whitney J75 engines until the J58
engine became available in January 1963. Then our problems really began!
The first one was concerned with foreign object damage (FOD)
to the engines -- a particular problem with the powerful J58 and the
tortuous flow path through the complicated nacelle structure. Small
nuts, bolts, and metal scraps not removed from the nacelles during
construction could be sucked into the engines on starting with
devastating results. Figure _ shows damage to a compressor blade
from an inspector's flash light used to search for such foreign
objects. Engine damage -- $250,000: Besides objects of the above
type, the engines would suck in rocks, asphalt pieces, etc. from
the taxi-ways and runways, (Figure _). An intensive campaign to
control FOD at all stages of construction and operation -- involving
a shake test of the forward nacelle at the factory, the use of
screens, and runway sweecinp with double inspections prior to any
engine running -- brought FOD under reasonable control.
The hardest problem encountered in flight was the development
Sanitized Copy Approved for Release 2011/09/27 : CIA-RDP90B00170R000100080001-5 throw out the
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-15-
initial pneumatic design after millions of dollars had been spent
on it and go to a design using electronic controls instead. This
was very hard to do because several elements of the system were
exposed to ram-air temperatures over 800?F and terrific vibration
during an inlet duct stall. This problem and one dealing with
aircraft acceleration between Mach numbers of 0.95 to 2.0 are too
complex to deal with in this paper.
Initially, air temperature variations along a given true altitude
would cause the Blackbird to wander up and down over several thousand
feet in its flight path. Improved autopilots and engine controls
have eliminated this problem.
There are no other airplanes flying at our cruising altitude
except an occasional U-2 but we were very scared by encountering
weather balloons sent up by the FAA. If we were to hit the
instrumentation package while cruising at over 3,000 feet per
second, the impact could be deadly'
Flight planning had to be done very carefully because of sonic
boom problems. We received complaints from many sources. One such
stated that his mules on a pack-train wanted to jump off the cliff
trail when they were "boomed." Another complained that fishing
stopped in lakes in Yellowstone Park if a boom occurred because the
fish went down to the bottom for hours. I had my own complaint when
one of my military friends boomed my ranch and broke a $450 plate
glass window. I got no sympathy on this, however.
OPERATIONAL COMMENTS
The SR-71 first flew 23 December 1964. It was in service with the
Strategic Air Command a year later.
In-flight refueling from KC-135s turned out to be very routine.
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-16-
Over eighteen thousand such refuelings have been made to date by
all versions of the Blackbirds and they have exceeded Mach 3 over
11,000 times.
The SR-71 has flown from New York to London in 1 hour 55 minutes
then returned nonstop to Beale Air Force Base in 3 hours 48 minutes
for the round trip.
It has also flown over 15,000 miles with refueling to
demonstrate its truly global range. It is by far the world's fastest,
highest flying airplane in service. I expect it to be so for a long
time to come.
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001L5 JULY 1981
DEVELOPMENT OF THE LOCKHEED SR-71 BLACKBIRD
FIGURE TITLES
FIGURE 1 Lockheed SR-71 at Altitude
FIGURE 2 Fuselage Cockpit Section in Oven for Testing Effects of
High Temperatures on Structure and Systems
FIGURE 3 Full-Scale Fuel System Test Rig to Test Various Angles
of Climb and Descent on Fuel Feed Capability
FIGURE A Pratt & Whitney J58 (JT11D-20) Engine
FIGURE B Comparison of J58 Development Objectives with Then-
Current Production Engines
FIGURE C J58 Flight Temperatures
FIGURE D Inlet and Engine Air-Flow Match
FIGURE E Net Thrust Comparison -- Bleed Bypass Cycle versus
Turbojet Cycle
FIGURE F Heated Environment Test Stand
FIGURE G-1 Air Flow Patterns -- Static Aircraft
FIGURE G-2 Air Flow Patterns -- High Speed
FIGURE H Original Engine Mount
FIGURE I Modified Engine Mount
FIGURE J Exhaust Gas Temperature Vernier Control System -- EGT
Error Gage Operating Modes
FIGURE K J58 Engine Under Test
FIGURE 4 Provisional Engine Starter Cart Which Used Two Buick
Racing Car Engines Geared to a Common Shaft Drive to
Rotate the J58 Engine. This Rig Produced Over 600
Horsepower for Starting.
FIGURE 5 Final Air Starting System Built into Operational Hangars
FIGURE 6 Upper Wing Surface Showing Wing Chordwise Corrugations.
Photo Taken During Static Tests at Design Limit Load.
FIGURE 7 Wing Fuel Tank Showing Structure and Sealant After 100
Hours Flying at Mach Numbers Over 2.6
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
-2-
FIGURE 8 Machine Shop with Numerical Controlled Milling Machines
FIGURE 9 This Nacelle Ring was Originally Made from 30 parts.
These Parts were Machined on Old, Outdated Profiling
Machines and then Assembled into the Final Ring Segment
in 487 Hours. Today, the Ring Segment is Made from a
Single Forging (upper photo) Weighing 325 pounds and is
then Machined on Profilers of Advanced Design to Produce
the Finished Part (lower photo) in 150 Hours.
FIGURE 10 Titanium Wing Extrusion as Received from the Mill (left)
and After Machining to a Finished Part (right)
FIGURE 11 Objects Which Could be Sucked into Engines on Take-off
FIGURE 12 Engine Name Plate Sucked into an Engine on Ground Run Up!
FIGURE 13 First-Stage Compressor Blade After Hitting Inspector's
& 14 Flashlight -- Total Engine Damage $250,000
FIGURE 15 The Author About to Fly in an Early A-12 Flight Test
FIGURE 16 YF-12A Test Pilot in Full Pressure Suit with Walk-around
Oxygen Kit.
FIGURE 17 Surface Temperatures at Design Cruising Speed and
Altitude
FIGURE 18 Engine Nacelle Leakage Test Model. Tested to Over
50 psi.
FIGURE 19 Palmdale, California Production Test Facility
FIGURE 20 Southeast Asia Combat Missions for One SR-71 -- June 1971
FIGURE 21 A-12 Used as a Launch Platform for an Unmanned Ramjet-
Powered Target Drone
FIGURE 22 A-12 Test Aircraft Fleet in Storage After Development
Testing
FIGURE 23 A-12 Test Fleet -- April 1964
Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
 Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
ATTACHMENT
J58/SR-71 PROPULSION INTEGRATION
OR
THE GREAT ADVENTURE INTO THE TECHNICAL UNKNOWN
By William H. Brown, Retired Engineering Manager
Government Product Division
Pratt & Whitney Aircraft Group
United Technologies Corporation
Successful integration of the J58 engine with the SR-71 aircraft
was achieved by:
o Inherently compatible engine cycle, size and characteristics.
o Intensive and extensive design/development effort.
Propulsion integration involved aerodynamic compatibility,
installation and structural technology advances, development of a
unique mechanical power takeoff drive, and fuel system tailoring. All
four areas plowed new ground and uncovered unknowns that were
identified, addressed and resolved. Interacting airframe systems,
such as the variable mixed compression inlet, exhaust nozzle, and fuel
sys cam were ground tested with the J58 engine prior to and coincident
with flight testing. Numerous iterative redesign-retest-resolution
cycles were required to accommodate the extreme operating conditions.
Successful propulsion operation was primarily the result of
o Compatible conceptual designs.
o Diligent application of engineering fundamentals.
o Freedom to change the engine and/or aircraft with a minimum
of contractual paperwork.
o A maximum of trust and team effort with engineer-to-engineer
interchange.
 The centerline of the basic J58 engine was
laid down in late 1956. It was to be an
takeoff turbojet rated at 26,000 lb maximum
takeoff thrust and was to power a Navy attack
aircraft which would have a dash capability of up
to Mach 3 for several seconds. By the time Pratt
6 Whitney Aircraft, along with Lockheed and
others, began to study the SR-71 'Blackbird'
requirements several years later, we had completed
approximately 700 hours of full-scale engine
testing on the J58.'
In the "Blackbird" joint studies, the
attitude of open cooperation between Lockheed and
Pratt S Whitney Aircraft personnel seemed to
produce better results than if a more
"arms-length" attitude were adopted. This open
cooperation resulted in a more complete study
which identified the enormous advances in the
state-of-the-art and the significant amount of
knowledge which had to be acquired to achieve a
successful engine/airframe integration. The
completeness of this study was probably
instrumental in Lockheed and Pratt d Whitney
Aircraft winning the competition. The Government
stated that the need for the "Blackbird" was so
great that the program had to be conducted despite
the risks and the technological challenge.
Furthermore, the Government expected the risks to
be reduced by fallout from the X-15 and 8-70
programs. Unfortunately, there was no meaningful
fallout.
Figures 1 and 2 indicate some of the
increased requirements of the "Blackbird" engine
compared to the requirements for the previous J75
engine. As it turned out, even these requirements
didn't hold throughout the "Blackbird's" actual
mission. For example, the engine inlet air
temperature exceeded 800?F under certain
conditions. The fuel inlet temperature increased
to 350"F at times and the fuel temperature ranged
from 600?F to 700?F at the main and afterburner
fuel nozzles. Lubricant temperatures rose to
700'F and even to 1000?F in some localized parts
of the engine.
Because of these extremely hostile
environmental conditions, the only design
parameters that could be retained from the Navy
J58-P2 engine were the basic size and the
compressor and turbine aerodynamics. Even these
were modified at a later date.
Mach Number
Altitude
Compressor Inlet
Temperature
Turbine Inlet
Temperature
Maximum Fuel Inlet
Temperature
Maximum on Inlet
Temperature
Thrust/Weight Ratio
Military Operation
Afterburner Operation
J57 and J75 JTI1D-20
2.0 for 15 min (J75 OnM 3.2 (Continuous)
55.000 ft 100,00 ft
250?F (J75 Only) 800?F
1750?F (Tatteo0) 2000?F (Continuous)
1550'F (Cruise)
110.130?F 300?F
250?F 550?F
4.0 5.2
30.m1n Tlme limb Continuous -
IntamlittMlt Continuous
Figure 1. Comparison of J58 Development
Objectives with Then Current
Q7 Production-Type Engines.
[800?F
The extreme environment presented a severe
cooling problem. It was vital to cool the pilot
and aircraft electronics; but this left little or
no heat sink in the fuel available to cool the
rest of the aircraft or the engine. Because of
this, the only electronics on the engine was a
fuel-cooled solenoid which was added later and a
trim motor buried inside the engine fuel control.
To keep cooling requirements to a minimum, we even
had to provide a chemical ignition system using
tetraethyl borane (T.E.B.) for starting both the
main engine and the afterburner. A new fuel and a
chemical lubricant had to be developed to meet the
temperature requirements. Pratt & Whitney
Aircraft together with the Ashland, Shell, and
Monsanto Companies took on the task of developing
these fluids.
Early in the development, we found that a
straight turbojet cycle did not provide a good
match for the inlet nor the required net thrust at
high Mach number operating conditions. To
overcome these problems, we invented the bleed
by-pass cycle with which we could match the inlet
airflow requirements as shown in Figure 3.
Another advantage of this cycle was that above
Mach 2, the corrected airflow could be held
constant at a given Mach number regardless of the
throttle position. The bleep by-pass cycle also
provided more than 20 percent additional thrust
during high Mach number operation. See Figure 4.
Corrected
Airflow
Bleed Bypass Cycle
Mach N~o, y
 Percent
SLTO
Thrust
Flight Mach Number -
Figure 4. Net Thrust Comparison - Bleed Bypass
Cycle vs Turbojet Cycle.
Fabrication and materials technology
presented one of the greatest challenges. We had
to learn how to form sheet metal from materials
which previously had been used only for forging
turbine blades. Once we had achieved this, we had
to learn how to weld it successfully. Disks,
shafts, and other components also had to be
fabricated from high-strength,
temperature- resistant turbine blade-like materials
to withstand temperatures and stresses
encountered. I do not know of a single part, down
to the last cotter key, that could be made from
the same materials as used on previous engines.
Even the lubrication pump was a major
development. The newly developed special fuel was
not only hot, but it had no lubricity. A small
amount of fluorocarbon finally had to be added to
allow the airframe and engine pumps and servos to
work.
Fuel was used as the engine hydraulic fluid
to actuate the bleeds, afterburner nozzle, etc.
Because there was nothing to cool the fuel, it
just made one pass through the hydraulic system
and then was burned.
if the foregoing were not enough,
developmental testing problems also had to be
overcome. There were no test facilities which had
the capabilities to provide steady-state
temperature and pressure conditions required for
testing at maximum operating conditions nor could
they provide for performing transients. A partial
solution is shown in Figure 5. On this test
stand, the exhaust of a J75 engine was run through
and around the J58 to simulate transients of the
temperature environment.
Figure 5. Heated Environment Test Stand
U
In addition, there was essentially no
instrumentation rugged enough to obtain accurate
real-time measurements. As Pratt & Whitney
Aircraft developed more rugged instrumentation and
better calibration facilities, improved data were
gradually obtained. Lockheed, of course, was kept
up-to-date as we obtained better data. A good
part of the time Lockheed and Pratt & Whitney
Aircraft jointly ran fuel system rigs, inlet
distortion rigs, etc., as well as some engine
calibration tests and wind tunnel testing of the
ejector.
It's important to remember that this all
started nearly a quarter of a century ago.
Although Pratt & Whitney Aircraft had a very large
computer system for it's day (the IBM 710), it was
no more sophisticated than some of the hand held
calculators now available. Consequently, the J58
engine, in effect, was a slide-rule design.
Despite all of the testing and faired curves, we
knew we had to solve many of our mutual
integration problems through flight test.
Approximately three months before Pratt &
Whitney finished the Pre Flight Rating Test which
j was 3 years and 4 months after go-ahead (the Model
Qualification Test was completed 14 months later),
the first "Blackbird" took to the air. It was
powered by two afterburning J75 turbojet engines
to wring out the aircraft subsonically. As soon
as Lockheed felt comfortable with the aircraft, a
J58 was installed in one side. After several
months of subsonic flight tests, J58 engines were
installed in both sides, and we started flight
testing for real.
Naturally there were problems. Here are a
few notable ones and the solutions.
The first problem happened very early-the
engine wouldn't start! The small inlet wind
tunnel model did not show the inlet being so
depressed at the starting J58 airflows. In fact,
instead of air flowing out of the compressor
4th-stage through the bleed ducts into the
afterburner, it flowed the other way! As a
temporary fix, Lockheed removed an inlet access
panel for ground starts. They later added two
suck-in doors (see Figure 6) and Pratt & Whitney
Aircraft added an engine bleed to the nacelle.
These two changes eliminated the ground starting
problem.
Tertiary Doors Open
Ejector Flaps Closed
Originally, the blow-in door ejector or
convergent-divergent nozzle was built as part of
the engine. It was subsequently decided jointly
by Lockheed and Pratt & Whitney Aircraft that it
would save weight if it was built as part of the
airframe structure. This was deemed appropriate
particularly as the main wing spar structure had
 *esponsible for nozzle performance in conjunction
.ith the engine primary nozzle. In addition, we
would perform all of the wind tunnel testing. In
exchange, Pratt S Whitney Aircraft would build the
remote gearbox because Lockheed's gearbox vendor
had no experience with gear materials or bearings
and seals that would withstand the temperatures
required. As a matter et fact neither did we, but
we were already committe to learn.
A problem partially related to the ejector
was that the airplane burned too much fuel going
transonic. To help solve the problem, thrust
measurements were taken in flight, movies of
ejector operation in flight were made, local Mach
numbers were measured, etc. Two fundamental
mistakes were uncovered. The back end of the
nacelle (the ejector) went supersonic long before
the airplane did, and the fairing of the aircraft
transonic wind tunnel drag data was not accurate.
While we were puzzling, out the solution, some
pilot decided to go transonic at a lower altitude
and higher Kens. This for all intents and
purposes solved the problem. From this we learned
not to run pace'. le wind tunnel tests unless the
model contains at least a simulation of the
adjacent aircraft surfaces. We also learned to
take enough data points so that transonic drag
wind tunnel data does not have to be faired.
As flight testing increased to the higher
Hach numbers, new problems arose. One, which
today may be considered simple with our modern
computer techniques, concerned the remote
gearbox. The gearbox mounts started to exhibit
heavy wear and cracks, and the the long drive
shaft between the engine and the gearbox started
to show twisting and heavy spline wear. After
much slide-ruling, we finally decided that the
lore t ion of the gearbox relative to the engine was
unknown during, high Iva cli number trans l eats. We
resorted to the simple test of putting styluses on
the engine and mounted a scratch plate on the
gearbox. We found, to our astonishment, that the
gearbox moved about 4 inches relative to the
engine. This was much more than the shaft between
rite engine and the gearbox 'could take. The
problem was solved by providing a new shaft
containing a double universal joint.
Another problem arose when the aircraft fuel
system plumbing Immediately ahead of the engine
started to show fatigue and distortion.
Measurements with a fast recorder showed that
pressure spikes at the engine fuel inlet were
going off scale. This overpressuring was found to
be caused by feedback from the engine hydraulic
system. This phenomena did not show up either
during Lockheed's or Pratt A Whitney Aircraft's
rig resting nor during the engine ground testing
because of the large fluid volumes involved. To
solve the problem Lockheed Invented a
"high-temperature sponge" (promptly named "the
football") which they installed in an accumulator
ahead of the engine. This reduced the pressure
spikes to a tolerable level.
A mounting-related problem occurred under
certain conditions of down load on the Wing. At
at these conditions, the Outer half of the nacelle
would rotate into the engine and crush the engine
plumbing and -anything, else in the way.
Originally, the engine was mounted on a stiff rail
structure at the top of the nacelle with a
stabilizing link from the top of the engine rear
mount ring to the aircraft structure as shown in
Figure 7. To solve the crushing problem Pratt &
Whitney Aircraft redesigned the rear mount ring so
that a tangential link could be installed between
the engine and the outboard side of the nacelle.
This maintained a finite distance between
nacelle and engine under all conditions.
Figure B.
Figure 8, Engine Mount
As- mentioned previously, there was a minimum
of elecronics in the engine control system because
electronics would not survive the environment and
the fuel was already too hot to provide cooling.
Consequently, control adjustments normally made
automatically had to be made manually. For
example, the pilot operated a vernier trimmer to
make fine adjustments in the ECT (Exhaust Gas
Temperature) as conditions varied from standard
(one such device was used successfully in the
U-2). The pilot was provided with a curve of ECT
versus engine inlet temperature to make the
required manual adjustments. However,
4
Sanitized Copy Approved for Release 2011/09/27: CIA-RDP90B00170R000100080001-5
 unexpectedly sharp atmospheric changes were
encountered. These, in combination with the speed
of the aircrft, resulted in changes too fast for
the pilot to handle. By the time he read the
engine inlet temperature and adjusted the EGT, the
inlet temperature had changed. This caused some
inlet unstarts (highly reduced inlet airflow) and
other undesirable results. To correct this
unacceptable state of affairs, Pratt & Whitney
Aircraft proposed to revise the aircraft EGT gauge
by feeding in an engine inlet temperature signal
and adding some additional gadgetry to trim
automatically. The digital EGT readout was
retained as was an override manual trim in case of
failure. See Figure 9. This modification has
worked well ever since.
Figure 9. ECT Vernier Control System - EGT Error
Cage Operating Modes
The most sensational and most confusing
problem at the high Mach number condition was
inlet unstarts. These occurred without warning
and were seemingly inconsistent. To add to the
confusion, the pilots consistently reported the
unstart occurring on the wrong side of the
airplane. This anomaly was solved rather quickly
when Lockheed found that the Stability
Augmentation System (SAS) slightly overcompensated
for the sudden one sided drag. This led the pilot
to believe that the wrong side had unstarted, and
consequently, his corrective action usually
resulted in worsening the problem. Oddly enough,
the engine did not blowout. It just sat there and
overheated because the inlet airflow was so
reduced that the engine minimum fuel flow was
approximately twice that required. Worst of all,
the inlet would not restart until the pilot came
down to a much lower altitude and Mach number. A
great many tests and investigations were conducted
including the possibility of engine surge being
the initiator. This was not the case. Three
major causes were finally isolated:
1. Manual trimming of engine.
2. High, inconsistent nacelle leakage at
the approximately 40:1 pressure ratio.
3. Alpha signal (angle of attack from
noseboom) to inlet control subject to
C-loading.
The following improvements were incorporated
by Lockheed and Pratt & Whitney Aircraft
essentially as a package:
1. Improved sealing of the inlet and bypass
doors.
2. Auto-trimmer of engine installed (Ref.
Figure 9).
3. Derichment valve with unstart signal
installed on engine to protect turbine
(Ref. Figure 9).
4. Increased area inlet bypass doors and
addition of an aft inlet bypass door
which bypassed inlet air direct to
ejector.
5. Added a 'G" bias on Inlet control.
6. Automated inlet restart procedure on
both inlets regardless of which
unstarted.
The foregoing six items essentially
eliminated inlet unstart as a problem. An
additional benefit was also realized by the
ability to use the aft inlet bypass door in normal
flight instead of dumping all inlet by-pass air
overboard. As this air became heated as it passed
over the engine to the ejector instead of going
overboard, drag was substantially reduced. Also
better sealing of the nacelle reduced crag further.
As you have probably noticed, I have had
difficulty in differentiating between "we" Pratt 6
Whitney Aircraft and "we" Lockheed. But that is
the kind of program it was.
In any complicated program of this magnitude
we all do something dumb and we both did our
share. Here is one from each of us: "We, (Pratt
6 Whitney), became so obsessed with the problems
of hot fuel and hot environment that we neglected
the fact that sometimes the fuel was cold when the
environment was hot and vice versa. When this
occurred, the engine fuel control did not track
well. To correct this, we had to insulate the
main engine control body from the environment and
make all the servos, etc., respond only to fuel
temperature. Eventually, we had to make a major
redesign of the control.
Lockheed and Pratt & Whitney Aircraft spent
many hours coordinating the inlet and engine
arrangement so that doors, bleeds, air conditioner
drive turbine discharge, etc., would not affect
any of the engine control sensors in the engine
inlet. In fact, the air conditioner turbine
discharge was located 45 deg on one side of the
vertical centerline and the engine temperature
bulb was located 45 deg on the opposite side. To
save design time, Lockheed built one inlet as a
mirror image of the other. It is now easy to
conclude where the 1200'F air conditioner turbine
discharge turned out to be:: For a while the fact
that one engine always ran faster than the other
was a big mystery!
That this complex, difficult program was
successful is attributable, in large part, to the
management philosophy adopted by the Government
people in charge. Their approach was that both
the engine and airframe contractors must be free
to take the actions which in their judgment were
required to solve the problems. The Government
management of the program was handled by no more
than a cozen highly qualified and capable
 individuals who were oriented toward understanding
the problems and approaches to solutions, rather
than toward substituting their judgment for that
of the contractors. Requirements for Government
approval as a prerequisite to action were minimal
and were limited to those changes significant cost or operational impact. As involving
result, reactions to problems were exceptionally
quick. In this manner, the time from formal
release of engineering paperwork to the conversion
to hardware was drastically shortened. This not
only accelerated the progress of the program but
saved many dollars by incorporating the changes
while the number of units were still relatively
small.
On this program, the Government fully
recognized that many of the problems involving
either the engine or airframe manufacturer, or
both, could be solved most effectively by a joint
engineering effort and the contracts were written
to ellou this activity without penalties. As a
result, an extremely close working relationship
between the engineering groups was developed and
flourished until the SR-71 became fully
operational. This method of operation led to
prompt solutions of many problems which, under a
more cumbersome management system, could have
severely impeded the program by introducing very
costly delays or forcing inappropriate compromises
because of contractual interpretations.
In summary, the method of managing this
program by the Government resulted in shorter
development time, faster reaction to field
problems, reduced retrofit costs, and earlier
availability of production systems incorporating
corrections for problems uncovered by operations
in the field. The result was an operating system
incorporating a magnum step. in the
state-of-the-art at an earlier time and at less
cost to the Government than would otherwise have
been possible.