WHEN it came to experience and demonstrated ability in building aerodynamic research facilities, NACA had no peer, and some of NACA's most experienced designers and builders came to Moffett Field to build and operate the facilities of the new Ames Aeronautical Laboratory. True, they had a rather limited amount of money to work with, but this was an old and familiar condition which merely whetted their ingenuity. One could scarcely have chosen a better man to build a new laboratory than Smith DeFrance. He and John Parsons worked together as an extremely effective building team. But there were many others at Ames who contributed greatly to the design of the new facilities, men such as Don Wood, Carl Bioletti, H. J. (Harvey) Allen, James White, Walter Vincenti, Manley Hood, and J. S. W. (Sam) Davidsen. Indeed, in the first years of the Ames history, the research men and everyone else turned to and helped with the design of new facilities. Bioletti labored not only on the design of facilities for Ames but also found time to assist in the design of an altitude wind tunnel for the Engine Research Laboratory at Cleveland. Even after research work at Ames had gotten well under way, certain research men were still dealing with the steel and concrete of which the major structures of the Laboratory were composed. Surprisingly perhaps, even the staff of the Theoretical Aerodynamics Section were so involved. Hence, though it was intended as humor, there was plenty of truth in Harvey Allen's greeting when he answered the telephone, "Theoretical Concrete and Reinforced Aerodynamics Section!"
The builders of the Ames Laboratory faced a real challenge as well as an opportunity to demonstrate imaginative design. The main theme at Ames was to be research in high-speed aerodynamics, yet the military had a need for facilities that could be built in a hurry to perform tests at conventional  speeds of around 250 mph. 1 To satisfy the latter need, the two 7- by 10-foot tunnels were built. The pair cost a little under a million dollars. These two facilities involved no important extension of the design art. They were conventional closed-throat tunnels designed to operate at atmospheric pressure and at speeds up to about 280 mph. They were designed to measure forces and moments on relatively simple models that could easily and inexpensively be modified.
Of the wind tunnels originally planned for Ames, the 16-foot tunnel had perhaps the highest priority. It had been assigned this precedence because it was to have a higher speed than any other major wind tunnel in the NACA and would provide aerodynamic data at the speeds at which future military airplanes were expected to fly. There was an 8-foot-diameter, 600 + -mph tunnel at Langley; but the new tunnel at Ames was to operate at speeds up to 680 mph, about 0.9 of the speed of sound, and was to have four times the cross-sectional area of the Langley tunnel. Its cost was nearly $2 million. The huge 27,000-horsepower drive motors of the 16-foot tunnel would generate so much heat that an air-exchange tower had to be provided. The function of the tower was to replace the heated air in the tunnel gradually with cool air from the outdoors. This was a device originated by NACA and first used in the 8-foot high-speed tunnel at Langley.
The motors of the 16-foot tunnel also produced a deep rumble which in the quiet of the night could be heard for miles and sounded like an approaching fleet of bombers. This characteristic caused some trouble when the tunnel was being calibrated during the early months of the war. The military was pretty jumpy at that stage and when the tunnel began to operate in the middle of the night, they were sure the Japanese were coming in as they had at Pearl Harbor. In accord with air-raid plans, all major power absorbing equipment was ordered shut down and this included the 16-foot tunnel. Happily, the enemy raid appeared to have been a false alarm. The all-clear sounded and the tunnel innocently resumed its tests. But oh those foxy Japs! There they came again and the tunnel was again shut down. "All clear!" and the tunnel start button was pushed. Moments later the sirens shrieked their warning of yet another attack. This was war in its most exasperating form. After a few cycles of on-and-off operation, it began to dawn on someone that there was a connection between the tunnel operation and the suspected air raid. Disillusioned, the military called off their air raid for the night and arrangements were made to avoid such confusion in the future.
Another bit of excitement at Ames came one evening when a hydrogen-filled barrage balloon, used to protect certain installations in the upper Bay region, broke loose from its moorings. Like a homing pigeon, it headed....
 ....down the Bay straight for the 110,000-volt substation which Ames had built to provide power for its new wind tunnels. The rendezvous was accomplished with a brilliant flash that could be seen for miles. Confusion again reigned. No one had seen the balloon approach and thus no one knew what had happened. Was it sabotage? Jim White rushed out to find the substation in what, in the twilight, appeared to be shreds. He had trouble in restraining the military guards from crawling over the fence into the danger area. When the scene was illuminated and calm restored, the substation was found to be largely intact. The shreds were the remains of the balloon.
The power requirements of the Laboratory were growing quite rapidly, and the Pacific Gas & Electric Co. was taking steps to meet the demand. The 40- by 80-foot tunnel required 36,000 horsepower to turn six motors driving six fans all mounted in the same plane and all synchronized in their turning. This power, "Teat though it was, moved a lot of air and generated a speed of only 225 mph. Far larger than any previously existing tunnel, the 40-by-80 was 180 feet tall, covered eight acres, and together with the huge airship hangars became the eye-catching trademark of Moffett Field. Its cost was just over $7 million.
The throat of the 40-by-80 was made of steel, of course, but the huge return passages were constructed largely of corrugated asbestos-cement sheet supported on the outside by a steel trusswork. Originally the tunnel plans called for an air-exchange tower to remove the heat and exhaust gases of airplane engines which would be run in the tunnel. But the air exchanger would add a million dollars to the cost and this respectable increment, it was felt, might better be spent on a much-wanted supersonic wind tunnel. Further calculations showed that, because of the huge volume of air enclosed in the tunnel, the air leakage occurring at the purposefully unsealed joints, and the large thermal capacity of the tunnel and its concrete base, the air exchanger was not essential. If, after a considerable period of engine running, the airstream became excessively contaminated and heated, the great doors in the throat could be opened and the hot contaminated air pumped out.
Don Wood and Harvey Allen did much of the design work on the 40-by 80. According to Harvey, they each started at the throat and worked in opposite directions, hoping to meet in some consistent fashion on the opposite side. And this they did. For the joining, Harvey designed a special section which he was pleased to call his "Rumanian joint." To this fancy the others smiled indulgently. Allen's first initial "H" stood for Harry, but the pseudonym "Harvey," applied originally in fun, fell into such popular usage that his real name was soon forgotten. His family, however, called him Julian.
Although construction on the 40-by-80 had begun late in 1941, the task was large and the tunnel was not completed and ready to operate until June....
....1944. The completion, it was felt, provided an auspicious occasion for holding a formal dedication of the Ames Aeronautical Laboratory. A number of research facilities were in operation and these, together with the 40-by-80, would make quite a show for the invited guests. The administration building also had been completed, thus providing a suitable place for hanging a picture of Dr. Ames. Unfortunately, Dr. Ames would not be present, for he had died a year earlier.
The dedication was held on June 8, 1944. High-level representatives from the aircraft industry, from local universities, and from civil and military branches of the Government were present. The dedicatory address was given by the beloved and respected Dr. William F. Durand of Stanford who, along with Dr. Ames, was one of the original members of NACA. Dr. Durand, though 85 years of age at the time of the dedication, had recently been reappointed to the NACA for the specific purpose of guiding the Committee's work in the new field of jet propulsion.
As a fillip to the dedication ceremony, the group was taken on an inspection tour of the 40-by-80 in which the new XSB2D-1 dive bomber, pride and joy of designers Ed Heinemann and Gene Root of the Douglas El Segundo plant, had been mounted. While the visitors peered through the windows of the test chamber, Dr. J. C. Hunsaker, Chairman of NACA, threw the switch and the six fans, after kicking into synchronism, began to turn. The sight was impressive. The largest wind tunnel in the world was in operation.
Shortly before the Ames Aeronautical Laboratory was founded, Eastman Jacobs, John Stack, Harvey Allen, Ira Abbott, and others at Langley had developed a new class of airfoils having a less arbitrary and more scientific design rationale than earlier airfoils. The new design theory concerned the....
....boundary layer, the thin layer of air adjacent to the wing the velocity of which has been retarded by skin friction. It had long been known that the air as it passed over a wing would produce much less drag if the flow in the boundary layer was smooth and laminar rather than turbulent; but unfortunately it was generally the case that the airflow over a wing quickly became turbulent, remaining laminar for only a short way along the airfoil chord.
What the Langley men found out was that, if the airflow over the airfoil could be kept in a continually accelerating state by properly shaping the pressure distribution along the chord, the flow would remain laminar. The desired pressure distribution could be achieved by carefully shaping the airfoil-choosing the proper thickness distribution along the chord. The method of calculating the shape (thickness distribution) of an airfoil to achieve the desired pressure distribution was tedious, but the results were rewarding The viscous drag of a wing might thereby be cut in half. Later the new technique proved useful in designing wings for high-speed airplanes,
John Stack had put some test wings in one of the Langley high-speed air jets and had used a schlieren apparatus to directly observe and photograph any disturbances that might occur. These tests revealed that as the jet airspeed reached a certain value, vicious-looking shock waves would suddenly appear in the flow over the wing. These shock waves, usually one or more on each surface, would occur when the air velocity in the accelerated flow  over the wing reached the speed of sound, and this point was reached when the general airflow ahead of the wing was considerably less than the speed of sound.
The airspeed at which these dramatic effects appeared was dubbed the "critical speed" and there was a different critical speed for different configurations. The shock waves sometimes remained fixed in position, but more often danced back and forth causing violent disturbances in the airflow. This phenomenon caused by the compressibility of the air represented a very serious and discouraging problem with which designers of future high-speed airplanes would be faced. But it was found that the onset of the compressibility phenomenon and the violence of its disturbance could be reduced by properly shaping the thickness distribution, and thus the pressure distribution, of the airfoil. The Clark Y and other old types of airfoil having maximum thickness well forward were very bad. New designs that were thinner in front and attained their maximum thickness farther back were much better. There was, indeed, a good deal of similarity between airfoils designed for high speed and those designed to promote laminar flow.
By the time the Ames Laboratory came into being, the subject of airfoil design, which once had seemed fairly well in hand, had opened up into a promising new field of research. Of particular interest to the Ames staff was the design of airfoils having high critical speeds and mild compressibility effects, but a special facility was really needed for the experimental testing involved in such work. Of course the 16-foot tunnel was available, but it was far too large and expensive to be used for that purpose. It would have to be reserved for complete wing or model tests. What was needed was a small inexpensive high-speed tunnel in which whole series of inexpensive airfoil models could be tested. Out of this need came the 1- by 3 1/2-foot tunnel conceived and largely designed by Harvey Allen and manufactured out of material mostly scrounged from various sources around the Laboratory. It was built on a small bit of bare ground within the loop of the 16-foot tunnel. The cost was indeterminate, but the actual outlay was surprisingly little, about $50,000. The two 1000-horsepower motors used to power the tunnel were some the laboratory had acquired for propeller tests in the 40-by-80.
The power used in the 1-by-3 1/2, was sufficient to drive the tunnel airspeed up to the speed of sound, at which point a tunnel is said to be "choked." The phenomenon of choking had seldom been encountered before, but it was understood in theory and the theory was further elaborated in TR 782 authored by Harvey Allen and Walter Vincenti.
In subsonic tunnels the test model is always located in a constricted portion (throat) of the tunnel where the airspeed is a maximum. But as the wind-tunnel airspeed is increased, the airspeed in the throat remains a maximum only up to the point at which the speed of sound is reached. With further applications of power, the airspeed downstream of the throat becomes  supersonic but the speed in the throat remains unchanged-exactly sonic. Thus for subsonic tunnels the term "choking" is descriptive of a point in the speed range where the throat, or test section, of the tunnel is effectively throttled and at which the tunnel loses its ability to simulate free-flight conditions A much more complicated type of device is necessary for the simulation of supersonic flight conditions.
As the air in a wind tunnel has to accelerate in passing around the test model and its supports, a tunnel will obviously choke earlier (at a lower tunnel airspeed) with a model in place than when empty. Moreover, the larger the model and its supports relative to the tunnel cross section, the Sooner the tunnel will choke. Thus while the choking speed of an empty tunnel might approximate the speed of sound, the choking speed of the tunnel with a model in place might be no more than 80 percent of the speed of sound.
In the 1-by-3 1/2, Ames engineers wanted to study compressibility effects on airfoils at the highest possible speed, so the tunnel, and the model-support system, were designed to achieve the highest possible choking speed. Toward this end, the throat was made deep (3 1/2 feet) and narrow (I foot) and the sidewalls of the tunnel were used to support the model. So designed, the tunnel was able to provide reliable data at airspeeds up to 90 percent of the speed of sound.
Inasmuch as the airfoil model was attached rigidly to the sidewalls of the 1- by 3 1/2-foot tunnel, something special in the way of instrumentation had to be provided if, as desired, the lift, drag, and pitching moment of the model were to be measured. Allen solved this problem ingeniously by designing a unique integrating manometer that quickly determined the desired forces and moments from pressures measured along the top and bottom walls of the tunnel and in the wake of the test model. A brief description of this device is given in some of the test reports later prepared by the tunnel staff. 2
The advantage of operating a wind tunnel under pressure to simulate flight conditions more faithfully had been established at Langley long before the Ames Aeronautical Laboratory was born. To be more precise, what was really needed for faithful simulation of subsonic flight conditions was that the Reynolds numbers of the simulating and simulated conditions be the same. Since Reynolds number is defined as pVL/µ (where p is air density, V is air velocity, µ is air viscosity, and L is some characteristic dimension of the model or airplane being tested), aeronautical engineers had a number of factors to juggle to obtain true simulations in wind tunnels.
Any inaccuracies in the test data resulting from an inadequate simulation in wind tunnels were called "scale effect," since they could generally be  related to the scale of the model. Compensation for the size of small-scale models, such as were required in all but the largest and most expensive wind tunnels (e.g., the 40-by-80), could be obtained by increasing either the velocity of the wind-tunnel airstream (costly in terms of power required) or its density. Density could be increased by increasing the operating pressure, but this procedure involved designing the tunnel as a pressure vessel and also introduced a number of operational problems. Nevertheless, NACA had long felt that, to achieve the proper test Reynolds number, the construction and use of pressure tunnels were justified despite all the complications involved.
Aside from pressurization, there were other desirable features a wind tunnel should have. There was a need for high speed, since it had been found that in the near-sonic range of speed certain effects appeared that were significantly different from those produced by Reynolds number. This additional speed effect could be stated in terms of Mach number-the speed of the tunnel airstream, or airplane in flight, expressed as a fraction of the speed of sound. Thus for reliable simulation of high-speed flight, the wind tunnel conditions should reproduce both the Reynolds number and the Mach number of flight.
There was still another wind-tunnel characteristic that engineers at Langley had recently found to be very important-important at least for tests of the new laminar-flow wings. This characteristic was freedom from turbulence, particularly the fine-grain turbulence that is not found in the atmosphere in which airplanes fly but which unhappily is found in wind tunnels as a result of the churning of the fan and the disturbances of the corner turning vanes. Unfortunately the prime virtues of the laminar-flow wings and bodies were largely obscured by the turbulence existing in the air streams of ordinary wind tunnels. It had been found that such turbulence could in a large degree be eliminated, but not very easily.
The possibility occurred to NACA engineers that all three of these desirable wind-tunnel features-high Reynolds number (pressure), high Mach number (speed), and very low turbulence-could be incorporated in the design of a single tunnel. Langley engineers thought the idea was worth trying, as did the designers at Ames. Both Langley and Ames wanted this unique tunnel. Langley got started on the design while the fate of Ames was still being debated by the congressmen but, once the establishment of Ames had been confirmed, NACA management, strongly encouraged by Admiral Towers, decided that the tunnel should be designed and built at Ames. The eventual result of this decision was the 12-foot low-turbulence pressure tunnel which became known as the 12-foot tunnel.
The 12-foot tunnel called for design and construction techniques well beyond the state of existing experience or knowledge and represented a tremendous challenge for the Ames designers. Many people contributed to the  design, but Carl Bioletti took the brunt of the load. The tunnel was to operate at up to 6 atmospheres of pressure and, to achieve low turbulence, its shell had to be larger in diameter and longer than a conventional 12-foot tunnel. As the welded steel plates of which the tunnel was fabricated had to be very thick (more than 2 inches in some places) to resist the high internal pressure, the total mass of the tunnel shell was tremendous, amounting to over 3000 tons. Moreover, this massive shell had to be supported on the ground in such a way that it was fixed at one point but otherwise free to move in any direction to accommodate the thermal expansions and contractions that were expected. This freedom of motion necessitated the use of a special flexible coupling in the fan drive shaft, as the drive motors themselves were rigidly mounted on the ground.
The stress analysis of a thick-walled vessel of the size and odd shape of the 12-foot tunnel was very difficult, requiring extensions of existing theory. In developing the stress-analysis methods for the tunnel, Walter Vincenti made some major contributions, and in this matter the counsel of Dr. S. Timoshenko, the famous structures scientist at Stanford, was sought on several occasions. An unusual feature (one of many) of the 12-foot design was the absence of corners in the tunnel loop. The almost universal way of designing wind tunnels, until then, had been to incorporate four right-angle corners around each of which the air was guided by a venetian-blind-like array of guide vanes. But in the 12-foot, owing to the high internal pressure, the stresses calculated for right-angle elliptical corner sections were intolerably high. To avoid these high stresses, it was necessary to turn the corner in small angular steps, thus maintaining a more nearly circular cross section throughout the tunnel loop.
There was no way of eliminating the sources of turbulence in the tunnel. It was generated by the fans, by the turning vanes, and by the friction on the sidewalls. The only way to obtain a low-turbulence flow in the test section was to eliminate, or attenuate, the turbulence produced by the disturbing elements. Also it was obvious that the attenuation had to be accomplished just before the air entered the test section. NACA had earlier found that considerable attenuation of turbulence could be obtained if a large contraction in the diameter of the tunnel was incorporated just ahead of the test section. Such a contraction would cause a rapid and large increase in velocity in the throat and thus the small random increments of velocity representing the upstream turbulence became a much smaller part of the increased throat velocity. All wind tunnels had some contraction at the throat, but in most cases this amounted to only 6 or 8 to 1. If the contraction ratio in the 12-foot tunnel were increased to 25:1, the turbulence would be considerably attenuated but its level would still be unacceptably high. Something else was needed-screens.
It was known from British research that, if a turbulent airflow was....
 .....passed through a fine screen having a drag coefficient of about 2.0, the turbulence downstream of the screen would be largely eliminated. Ames designers calculated that if they could mount not one but eight fine-mesh screens across the 12-foot flow channel, these screens together with the 25:1 contraction ratio would reduce the stream turbulence to the low level required. It was decided to mount the screens in a 63-foot-diameter bulge, or settling chamber, that would be built into the flow passage just ahead of the entrance to the test chamber. In accomplishing this objective, the Ames designers learned the difference between theory and experiment. It was easy to draw a picture of eight screens stretched tightly across a 63-foot settling chamber, but quite a different thing to fabricate such large screens, made of phosphor bronze and each weighing 1600 pounds, and actually mount them snugly in the tunnel. But the feat was accomplished and how it was accomplished would be a long story in itself.
Before the screens and motors were put in the tunnel, however, the integrity of the shell as a pressure vessel had to be confirmed. Filled with air at 6 atmospheres pressure, the 12-foot tunnel was a bomb-a big bomb. Any rupture of the walls while the tunnel was under pressure could cause vast devastation and much loss of life in the local community. It was estimated that the energy of the compressed air in the tunnel was, if properly directed, sufficient to blow the whole tunnel half a mile high. The integrity of the pressure shell was a matter of grave concern, especially considering the uncertainties of the stress analysis and welding techniques. The welding of steel pressure vessels was by no means new, but there had been little if any experience in welding pressure vessels made of plates as thick as those used in the 12-foot, or in welding pressure vessels of complex configuration having many points of stress concentration. To make matters worse, there was no reliable way of inspecting the welds in such thick plates.
To be assured of the soundness of the 12-foot tunnel, the only thing to do, Ames engineers felt, was to run a hydrostatic test-fill the tunnel shell with water and apply hydraulic pressure corresponding to 6 atmospheres. Little energy would be required to compress the water; thus, if the tunnel walls failed during the test, there would be no damage. But it would take a whale of a lot of water, 5 million gallons, it was calculated, and what about the weight of the water? Also, what would happen if an earthquake occurred when the vessel was full? These matters were considered ahead of time and the weight of the water was taken into account in the design of the tunnel shell and the foundations. Indeed, the water weight had dictated the design of some parts of the support system. The earthquake stresses, however, could not be dealt with. The vessel would fail in such an eventuality. In further preparation for the test, 600 wire strain gages were attached to the exterior Surfaces of the tunnel at points where, it was believed, stresses might be critical.
 When the tunnel shell was ready, 5 million gallons of water were pumped in. A calculated additional 6000 gallons were added to make up for the stretch in the shell. The filling took a week. The water came from an old well on the field. The well must have been approaching exhaustion, for the water was brackish and a certain amount of sediment came in with the water unnoticed. The tunnel, filled with water, had no apparent leaks, and the time had arrived for applying hydraulic pressure. The pressure, it was planned, would be applied on and off in cycles, thus simulating the pressurizing loads to which the tunnel would be subjected in normal service. The applied test pressure, however, would be 120 pounds per square inch which, combined with the weight of the water, would subject the tunnel shell to bursting stresses 50 percent greater than those expected in normal service. This overload was intended to account for the effects of vibration, fatigue, and other unpredictable stress factors present in tunnel operation.
Staff engineer George Edwards was standing near the tunnel as the pressure was being applied. Everything seemed to be going well when, as the pressure reached 107 psi, a terrific report came from a part of the structure just a short distance away. Somewhat jittery, as were all the other people present, George jumped with the sudden noise, whirled about, and saw water pouring out of a rupture in the tunnel wall. The failure had occurred at one end of the tunnel at a point of stress concentration where two shell plates of widely differing thickness had been welded together. As might have been expected, there was no strain gage at that point, so the actual stresses at rupture were undetermined. Before the failure could be repaired, the whole vessel had to be drained. The released water flooded the surrounding field. The nature of the structural problem was not difficult to see and the fault was soon corrected. Five million more gallons of water, this time from the city water supply and costing about $5000, were let into the tunnel. This time it held. The tunnel came through the pressure test with flying colors. The only thing left was to drain out the water-and to clean out the muck. The dirty water from the first test had left the interior of the tunnel a frightful mess, and days were spent in scrubbing and drying it out.
A few more things should be said about the 12-foot-tunnel design. Somehow the heat generated by the 12,000 horsepower transmitted to the coaxial fans had to be removed from the tunnel. In unpressurized tunnels this had been accomplished by an air exchanger, but some other arrangement was obviously required for the 12-foot. The solution adopted was one that had been developed and successfully used at Langley. It was in the form of a water pipe which was mounted on top of the tunnel and which produced a flow of cooling water over the tunnel walls. In addition, a canopy was placed over the top of the tunnel to protect the shell from solar rays and the stresses and strains resulting therefrom.
The requirements for pressurizing a vessel as large as the 12-foot tunnel  necessitated the construction of a considerable amount of air-handling equipment This equipment, which included pumps, air coolers, dehumidifiers, and several electric motors, was installed in an auxiliary building located adjacent to the tunnel. The total cost of the tunnel and auxiliaries was about $3 1/2 million.
An outstanding operational feature of the 12-foot tunnel, as earlier noted, was its low-turbulence airstream. Also, the tunnel could operate at high Reynolds numbers or high, subsonic, Mach numbers but, owing to lack of power, not at both at the same time. Its highest Reynolds number was to be achieved at the highest pressure, 6 atmospheres, but the pressure would have to be lowered to attain high Mach numbers. Indeed, to reach choking speed -the highest speed a subsonic wind tunnel can attain-the tunnel would have to be evacuated to a fairly low pressure. In the original design no evacuation was contemplated, but Harvey Allen had insisted that the lower end of the pressure range be dropped to one-sixth atmosphere. As a result choking speeds were readily attainable. Thus, with much design ingenuity, was produced a unique and extremely versatile wind tunnel, a suitable monument to this great building period at Ames.
But now the age of supersonic wind tunnels was beginning. Although the basic principles of supersonic flow had been developed years before, no one really knew very much about designing supersonic wind tunnels. There were a couple of blowdown jets at Langley that reached slightly supersonic speeds, and Arthur Kantrowitz and others at Langley had designed a 9-inch continuous-flow supersonic tunnel which was completed in 1942. Nevertheless, existing knowledge regarding practical means for designing supersonic tunnels was rudimentary.
Supersonic tunnels not only required fabulous amounts of power and high pressures for their operation but also introduced a very difficult problem in the matter of effecting desired variations in their airspeed. In the first place, as earlier mentioned, the maximum airspeed in a supersonic tunnel occurs not at the constricted throat but in the enlarged portion of the tunnel downstream of the constriction. And it is here, in what might be called the supersonic throat, that the test model is located.
But that is not all. In a subsonic tunnel the airspeed through the test section can be varied in a continuous fashion by merely changing the rpm of the tunnel fan, while in a supersonic tunnel speed changes are accomplished only by altering the basic geometry of the throat, or, more specifically, by changing the relative cross-sectional areas of the constricted and the supersonic throats. Moreover, the changes in geometry must be made very precisely if uniform flow in the test section is to be obtained. Thus in the de- -sign of the first supersonic tunnel at Ames, the provision of means for varying the speed represented a most challenging problem.
Ames engineers had wanted a supersonic wind tunnel from the start; but Dr. Lewis, while agreeing with the eventual need for such a facility, felt that the other tunnels planned for Ames should have a higher priority and that construction of a supersonic tunnel should be delayed awhile. Nevertheless, in 1943 serious plans were being made for the construction of such a tunnel. Ames engineers had wanted to build a 4- by 4-foot supersonic tunnel, 3 but the 40-by-80 was costing so much that their rather ambitious plans had to be pared down considerably; and Dr. Lewis, still questioning the advisability of embarking in wartime on such a far-out and costly project, was presumably not displeased that the size of the supersonic tunnel was reduced. After all, the dimensions of 1 by 3 feet finally adopted for the new tunnel were a rather large step beyond the 9-inch size of the Langley tunnel and the cost, $1.2 million, was all that could be afforded.
A major decision in the design of the 1- by 3-foot supersonic tunnel was whether or not the tunnel should be pressurized to achieve a wider Reynolds number range. Pressurizing would add complications and cost, but the most convincing argument against it was that Dr. Theodore von Karman, world-famous aeronautical scientist, then at Caltech, had suggested that the effects of Reynolds number might be expected to disappear at supersonic speeds. 4 Disagreement with von Karman was, of course, a position to be adopted with caution. Dr. Lewis and certain people at Langley were inclined to agree with the famous scientist, but Harvey Allen was not. He argued the point so effectively that NACA Headquarters, if not fully convinced of the rightness of his contentions, nevertheless yielded to the urgency of his request to pressurize the 1-by-3.
Once the pressurizing decision was made, a group headed by Allen and Vincenti got busy and produced a design for the first supersonic wind tunnel to be built at Ames. The tunnel, having a test section I foot wide and 3 feet deep, would be a closed system powered by four compressors each driven by a 2500-horsepower electric motor. The pressure in the system was to be variable from 0.3 to 4.0 atmospheres, giving a Reynolds number range of from 0.5 to 10x 106 per foot of model length; the speed (Mach number) range of the tunnel would be from 1.4 to 2.2 times the speed (about 760 mph) of sound. To prevent moisture condensation in the low-pressure region of the test section, the air introduced into the system would first be dehumidified. This precaution had also been taken in the design of the 12-foot tunnel.
The 1- by 3-foot tunnel was to be immediately adjacent to the 12-foot tunnel. This location was chosen as part of a scheme developed by Ames designers for acquiring a second 1- by 3-foot supersonic tunnel for very little more than the million dollars the first one was costing. They reasoned that the compressed-air energy in the 12-foot, which through cruel circumstance could blow the tunnel a half mile high, might with profit be used to operate a supersonic tunnel or jet. If the supersonic jet were conveniently located beside the 12-foot tunnel, the dry air with which the tunnel was filled could be discharged through the jet. The pressure and capacity of the 12-foot tunnel were such as to operate a 1-by 3-foot supersonic jet for several minutes -the higher the speed, oddly, the longer the period. This line of reasoning appealed to Ames and NACA management; accordingly, plans were laid to build both a continuous-flow 1- by 3-foot supersonic tunnel and a 1- by 3-foot blowdown jet, or tunnel, the latter having somewhat higher Mach number capabilities than the continuous-flow tunnel.
The same type of variable-geometry test section was to be used on both the continuous-flow and the blowdown tunnels. The top and bottom walls of the test section would be made of flexible steel plates that could be deflected to the required curvature and configuration for any desired Mach number. Such deflections as were required in the plates would be produced by a number of precisely controlled, motor-driven screw jacks. In the continuous-flow tunnel, it would be possible to change the throat shape, and Mach number, while the tunnel continued in operation; but in the blowdown tunnel the operating periods would be too short to allow any change of throat setting during a test.
The flexible-throat scheme used in the 1- by 3-foot tunnels was fairly simple in theory, but in actuality it was a very complex device with demanding performance requirements. The task of building the throats for the two Ames tunnels was undertaken by the Baldwin Southwark Division of the  Baldwin Locomotive Works. 5 Baldwin had also contracted to build a similar throat for a supersonic wind tunnel which Allen Puckett and his colleagues at Caltech had designed for the Aberdeen Proving Ground. 6 The Aberdeen tunnel, together with the 1- by 3-foot tunnels at Ames, represented perhaps the earliest ventures into the field of large supersonic wind tunnel design.
Construction work on the Ames 1- by 3-foot tunnels began in 1944 and continued into the next year. Smitty DeFrance, driving his men with his customary explosive vigor, laid down the law asserting that the 1-by-3 would begin operation on August 1, "or else." This dictum was not taken lightly. Ames designers had recognized the difficulties of building the supersonic throats to meet the rather stringent specifications which they had established and were not surprised when Baldwin Southwark began to encounter trouble. Cannily, they had taken the precaution of building two simple fixed geometry throats either of which could be fitted into the space reserved for the flexible throats. Each of the fixed throats was designed for a different Mach number and thus, if worse came to worst with respect to the flexible throats, the tunnels could still be operated at either of two Mach numbers. In the end it was the use of these fixed throats that enabled Ames engineers to meet DeFrance's operational deadline.
One of the first things discovered in the 1-by-3 tests was that, in supersonic flows, Reynolds number continues to be an important factor, perhaps even more important than in subsonic flows. Thus Harvey Allen's contention, maintained in the face of contrary opinion by the famous Dr. von Karman, was confirmed. 7 This finding, later substantiated in other supersonic tunnels, supported the later decision of the Langley Laboratory to increase the operating pressure and power of a 4- by 4-foot tunnel it had built.
As the war progressed, the notion of supersonic flight by manned aircraft became increasingly plausible to people engaged in research and military activities. There was a growing feeling that supersonic tunnels were needed immediately to obtain the design information that would be required in the approaching supersonic era. Moreover, it was thought that such tunnels should be large so that the test models could be realistic representations of practical airplane designs rather than the tiny, highly idealized models to which testing in currently planned supersonic facilities would be restricted.
The urgency of building large and expensive supersonic tunnels was not clearly evident, however, to some of the old-timers in the aeronautical world who had watched our rather slow and labored progress toward that speed  at which the drag of airplanes leaps upward, seeming like a brick wall to block further advance. And the doubts of these men were further raised by schlieren photographs revealing the alarming patterns of shock waves that form over the wings and tails of airplanes operating at high subsonic speeds. Did not the exigencies of war, and common sense, dictate that NACA devote its limited research energies to solving the problems of current military airplanes rather than expend them recklessly in a premature attempt to breach the sound barrier?
Ames engineers were among those who felt that NACA should proceed immediately to build a large supersonic wind tunnel. It should be, they thought, large enough so that a man could walk into the test section to mount and service a test model. In 1944, they made a rather sketchy design and cost estimate for such a facility and presented them to Dr. Lewis. Normally very progressive with respect to the development of new NACA facilities, Dr. Lewis believed this idea to be a little premature. 8 He may also have felt that the prospect of getting the $4 1/9 million required for the tunnel was not very bright, particularly in view of the watchfulness of economy-minded Albert Thomas, who then was chairman of the House Independent Offices Appropriations Subcommittee. In any case, Dr. Lewis received the Ames proposal with little warmth and, on returning to Washington, was said to have buried it in his lower desk drawer. What happened then is known only from the scattered recollections of a number of NACA people who heard about it later.
According to these recollections, the Ames tunnel design was quickly forgotten by Lewis; but into his office a few weeks later strode Rear Adm. D. C. Ramsey, Chief of the Bureau of Aeronautics. Ramsey spoke of the Navy's need for supersonic aerodynamic data and rather pointedly questioned Lewis on what NACA was doing about building a large supersonic wind tunnel. He gave the impression that he felt NACA had perhaps been a little slow in giving thought to such a project. Somewhat taken aback by this implied criticism, Dr. Lewis hesitated a moment and then, with some relief, remembered the Ames tunnel study buried in his desk. "Well, Admiral," he said, as he searched for and found the Ames study, "we have given this problem a lot of attention. In fact, we have designed a tunnel of the kind you mentioned but don't know where we will get the money to build it." Ramsey examined the design. "How much will it cost?" he asked. "Four and a half million," Lewis sighed, as if he were speaking of a gold mine on the moon. Ramsey pondered a few moments and then declared, "The Navy will give you the money to build the tunnel." 9 Lewis, of course, was both aston- ished and gratified. Never before had NACA obtained so much money so easily. What a haggling there had been in getting the first $4 million for the Ames Laboratory! What a blessing it was to have wealthy friends!
The transfer of funds from the Navy actually took place on January 27, 1945. 10 By the end of May, the first contracts for the tunnel were let. Clearly the tunnel would not help win the war, but its construction was, nevertheless, very timely and it would certainly fill a very important need.
The new tunnel was to have a test section 6 feet square and would become known as the 6- by 6-foot tunnel. It would be the first of the really large supersonic tunnels built by NACA. The tunnel would be driven by a huge compressor turned by two electric motors delivering a total of 60,000 horsepower-more power by two-thirds than was used in the giant 40-by-80. Despite this great expenditure of power, the tunnel would attain its maximum speed of Mach 1.8 only when the tunnel was partially evacuated. It was designed to operate at stagnation (air at rest) pressures of from 0.3 to 1.0 atmosphere. Higher operating pressures might have been desirable, but the required power would then have been prohibitively high.
The complete air system of the 6-by-6 was to include air pumps, air storage tanks, dehumidifiers, and a circulating-water air cooler built into the tunnel airstream. The tunnel design originally submitted to Lewis had called for the use of centrifugal compressors. An axial-flow fan, it was agreed, might be more compact and efficient, but there were difficult design problems and few if any companies had experience in constructing axial-flow fans of the size required for the tunnel. Once the Ames engineers were able to give the matter serious attention, however, it was found that an axial-flow fan could probably be constructed at a considerably lower cost than the centrifugal compressors. Thus the final design of the 6-by-6 incorporated a huge multistage axial-flow fan.
Before the design of the 6-by-6 had gotten under way, Ames engineers had become aware of problems Baldwin and Southwark was having in the construction of flexible throats for the 1- by 3-foot tunnels. Indeed, Ames designers had from the first been apprehensive about the design of these throats. Harvey Allen felt there must be a simpler, more reliable way of varying throat geometry to accomplish speed changes. At the Engine Research Laboratory, Abe Silverstein and his staff were experimenting with a circular tunnel having a central plug that could be pushed into the throat, thus decreasing the constricted throat area and increasing the speed of the tunnel. Unfortunately the central-plug scheme had some serious faults, one being that the test model would necessarily lie in the disturbed wake of the plug.
Allen reasoned that the wake problem could be solved if the central....
 ....plug were replaced by a properly configured sliding block representing the bottom wall of a rectangular supersonic throat. The block could be positioned very precisely by a motor-driven screw. Of course, the flow through the throat would not be symmetrical about the horizontal plane as it had been in all other wind tunnels, but the peculiarities of supersonic flows were such that the flow asymmetry might cause no trouble if the contours of the sliding block and of the fixed top wall of the supersonic throat were carefully designed. Allen first tested his design theory by building and testing a 2- by 2-inch asymmetric, sliding-block throat; this throat was sufficiently successful that he decided to build a larger, 8- by 8-inch model which could be installed in the 12-foot tunnel auxiliaries building and operated with air from the compressors of that tunnel. The experiment proceeded swiftly and the 8- by 8-inch tunnel, as it was called, was put into successful operation late in 1945. In the meantime, design of an asymmetric nozzle for the 6- by 6-foot tunnel was under way.
The asymmetric sliding-block nozzle had numerous advantages but also a few faults. For one thing, it was not quite so efficient, aerodynamically, as a symmetrical flexible-wall throat and required a somewhat higher pressure ahead of it to achieve the same Mach number as a symmetrical throat. Also, since the walls of the throat were rigid, there was no way, such as provided by a flexible throat, to compensate for small discrepancies in the contours of the sliding block and top wall. Moreover, the available methods for calculating the contours were tedious and did not precisely account for factors such as boundary-layer growth. As a result, it might not be possible to completely avoid small flow distortions in the test section. But the simplicity and reliability of the sliding-block throat were very attractive and were expected to overbalance any minor faults. The invention was original with Allen and might well have been called the Allen Throat. Harvey, however, pursuing his fancy of Balkanizing Ames wind tunnels, chose to call it the "Bulgarian throat." In formal literature, the name "asymmetric" or "sliding-block" throat or nozzle was used; but by whatever name, the device represented a notable achievement. Its development was later described by Allen in TN 2919.
Despite the forbidding complications of supersonic tunnels, there was at least one good thing to be said about this type of device. In such a tunnel, with suitable auxiliary equipment, it was possible actually to see important features of the airflow about the test model. And what was seen in the first transonic wing tests at Langley was frightening. It lent encouragement to the notion that the supersonic flight of manned airplanes was a long way off, if not a complete pipe dream. Nevertheless, the technique of making supersonic-flow patterns visible to the naked eye was scientifically interesting and very useful. The use of the technique was certainly a "must" in all supersonic wind tunnels.
 The technique for viewing supersonic flows was possible because of the variations in air density found in such flows-variations relating of course to the so-called "compressibility effect." There were actually two important techniques for revealing the density patterns in the wind-tunnel flow, both requiring that a very uniform, collimated beam of light be passed transversely through the tunnel walls and through the airflow in the region of the model. These techniques necessitated the installation of large windows in the sidewalls of the test section. Any distortion of the light beam caused by density variations in either the airstream or other parts of the light path could, by proper focusing, easily be made visible to the naked eye and thus subject to recording by photography. The simplest arrangement of this kind was called a "shadowgraph." It worked best when the density gradients were strong.
A more sensitive device revealing more details was called a "schlieren system." Here the light beam,-before it was viewed or projected on a screen, was brought to a sharp focus and then, by cutting the focus halfway in two by a razor-sharp knife edge, the density variations in the visual field would be revealed with great contrast. The schlieren technique was not new. It was the same in principle as the knife-edge test developed by Jean Foucault in the mid-19th century for figuring telescope mirrors. In this application, variations in the curvature of the telescope mirror were observed rather than distortions in the intervening air. And if the light beam was passed through a telescope lens, it would also reveal, with great clarity, all the flaws, striae, and optical distortions of the glass. Therein, as a matter of fact, lay one of the major problems in applying the schlieren system to supersonic wind tunnels. The problem was to avoid having the flow pattern of the airstream confused by flaws and optical distortions in the wind-tunnel windows.
The windows of a supersonic tunnel had to be large, to encompass a suitably large area of the flow field, and they had to be thick, to withstand the pressures in the tunnel. Moreover, their surfaces had to be ground perfectly flat and parallel and their interior had to be free of flaws and optical distortions. Indeed, they had to be made of similar materials and by the same slow painstaking processes as were used in the production of high-quality telescope lenses. But the difficulty of securing high quality in a telescope lens increases very rapidly with the diameter of the lens, and the biggest lens that had ever been built was the 40-inch-diameter lens, made from glass cast in France, of the Yerkes telescope at Lake Geneva, Wis. And the making of this lens was considered an extremely difficult task. 11 It is not surprising, therefore, that many very troublesome technical problems were encountered in applying the schlieren system to large wind tunnels.
The first attempt to use the schlieren system at Ames was in the 1- by....
.....3 1/2-foot tunnel but, owing to the excessively cramped quarters and lack of solid foundation, it was not very successful. The less-sensitive shadowgraph system proved to be more useful in this instance. The first really successful use of the schlieren system at Ames came with its application to the 1- by 3-foot supersonic tunnel. The most difficult and costly application, however, was in the 6- by 6-foot tunnel where windows 50 inches in diameter and more than 4 inches thick-larger than the Yerkes telescope objective-were required.
For Ames, the 6-year period from the Laboratory's beginning in 1940 to the end of the war in 1945 was characterized by construction. During that period, the funds committed to construction at Ames, about $21 million, were five times the amount NACA had spent on construction in the 24-year period from the date of its founding in 1915 to the time in 1939 when the Ames Laboratory was approved by Congress. Indeed, it was 25 percent greater than the total appropriations received by NACA during that period. While the growth of NACA facilities was undoubtedly accelerated by wartime requirements, this growth did, nevertheless, reflect in some quantitative fashion the increasing significance of aviation to the Nation.
1 This need was expressed to Dr. Lewis by Maj. A. J. Lyon of the Air Corps and recorded in Lewis memorandums for the Chairman NACA, one dated Dec. 14, 1938.
2 E.g., NACA TR 832.
3 According to recollections of Harvey Allen.
4 According to Lewis. It is possible that von Karman in conversations with Lewis put forward this suggestion only tentatively in any case, there is no question but that Lewis took it seriously or that it had repercussions at Ames as noted.
5 Baldwin was a subcontractor. Prime contractor, for the whole tunnel shell, was the Pittsburgh Des Moines Steel Co.
6 According to Harvey Allen.
7 Allen surmises that von Karman's belief was based on the results of sphere tests which proved inapplicable to airfoils.
8 Dr. Lewis was not unprogressive for, according to a memorandum for the files he wrote On Nov. 8, 1938, he listed a 3- by 5-foot, 1400-mph tunnel as one of the facilities that should be built at the proposed new West Coast station.
9 Recollections: see above.
10 As indicated in letter from James E. Webb, Director of the Bureau of the Budget, to the President [Truman], dated Apr. 18, 1947.
11 See David 0. Woodbury. The Glass Giant of Palomar, New York: Dodd, Mead & Co., 1941, p.75.