Chapter 3: Transonic Wind Tunnel Development (1940 -1950)
[88] On a spring morning in 1940, Stack and I left the office and drove to the remote beach at the easternmost tip of the Virginia Peninsula to watch the first attempt to obtain supercritical aerodynamic data on an airplane in free flight. A Navy fighter, the Brewster XF2A-2, was to be dived vertically over Chesapeake Bay to its terminal velocity, about 575 mph, and then make a pullup at its design load factor. The Brewster had been instrumented to measure the pressure distribution at an inboard wing station by the Langley Flight Division. We were most apprehensive as we watched the dive through binoculars. This was before the possible consequences of compressibility effects on the buffeting and control of diving airplanes had been highlighted by the P-38 tragedy of 1941; nevertheless, our knowledge of shock-stalled flows in the wind tunnel left little doubt about the dangers of this dive. Happily, the flight was completed successfully without any undue difficulties for the Navy pilot, but we were both left with the strong feeling that a diving airplane operating close to its structural limits was not an acceptable way to acquire high-speed research information. This experience undoubtedly contributed to Stack's later advocacy of a special research airplane capable of supercritical speeds in level flight.
Tests of the NACA 230-series section used on the Brewster were made in the 4 x 18-inch high-speed tunnel and the results arc compared with the flight data in fig. 21. The principal differences (in shock location) were due primarily to irregularities in the airplane wing, some of them distortions under air loads (ref. 101). In general, we were satisfied that the wind tunnel had been validated at least up to Mach 0.75, but we could see that future flight testing would be much more valuable if the surface distortions could be eliminated by use of thicker skins.
By 1942, it was apparent that the diving speeds of advanced fighters would penetrate more deeply into the supercritical region, equalling or exceeding the choking speeds of the wind tunnel test configurations then in use. We considered it unlikely at that time that the wind tunnel could ever be used at speeds beyond about Mach 0.8, and we therefore increasingly leaned toward the idea of a specially configured and instrumented test airplane capable of safe operation in this speed range. The....

[89] FIGURE 21.-Comparisom of flight and wind tunnel pressure distribution measurements for Cn = 0.4, airplane wing , NACA 230-series section 141/3 percent thick, wind tunnel model, NACA 23015 section.

[90] recollection of those of us who were involved is that this idea did not appear full-blown as a visionary new proposal of any single individual. Rather it took form gradually, manipulated and developed in innumerable lunchroom conversations and other contacts. Stack was a central figure in these discussions, and became the chief Langley promoter of the idea, but he was vague in regard to a specific origin. In a talk given in 1965 at a history session of the AIAA devoted mostly to the X-1 research airplane (ref. 10), he said:
After some deliberation, free flight with a manned instrumented airplane seemed the best and most direct way. Now, of all the people who contributed to this effort, it seems to me the two most noteworthy were General Arnold of the Air Force and Dr. Lewis of NACA. And as I noted, it was just about 23 years ago to the day (the summer of 1942] when word was given that we ought to go on something like this, with the caution that we couldn't spare many men because there was a war on.
This verbal authorization to "go" by no means implied general approval to design and procure a research airplane; it was simply permission for a limited preliminary study of the problems and desirable design features. Milton Davidson and Harold Turner, Jr., were logical choices to make preliminary layouts and performance estimates because they had done some work of this kind for Jacobs. Under Stack's direction, Davidson and Turner first concentrated on designs capable of high subsonic speeds up to about Mach 0.9. It is important to note here that Stack, in that period, did not consider or advocate pushing through Mach 1 to supersonic flight speeds. My firm recollection on this point is substantiated by those of several others including Soule (ref. 110). It is also supported by documentation (e.g., ref. 107) which gives Mach 0.8 to 1.0 as the range of NACA interest for a research airplane. The possible performance of prospective turbojets was uncertain at that time but it appeared likely that an engine would emerge which might marginally provide Mach 0.9 in a small airplane. The idea of rocket propulsion was quite beyond NACA thinking at that time; however, the Army with its background of JATO rocket development was willing to consider it. The first Army proposal for a high-speed research airplane by E. Kotcher in 1939 listed rocket propulsion as an alternative, and in the Army study of the "Mach 0.999" research airplane in early 1944 a principal objective was to compare performance of rocket and turbojet versions (ref. 111). It was [91] obvious that the only hope at that time for pushing through Mach 1 in level flight lay in rocket propulsion.
Navy interest in a possible high-speed research airplane also began to stir in the 1942-44 period (ref. 112). However, no direct action toward procurement was taken by any of the interested, parties prior to the March 15, 1944, seminar-type meeting at Langley of Army, Navy, and NACA personnel (ref. 113). Some significant differences of opinion relating to the design features and goals of a transonic research airplane surfaced at this meeting. NACA tended to think of the airplane as a device for collecting aerodynamic data unobtainable in the wind tunnel at high subsonic speeds. But the Army thought of it more as a major developmental step toward higher operating speeds extending upward through Mach 1. The Navy view inclined toward dispelling the myth of an impenetrable barrier and providing needed high-speed data. These differences could rather easily be accommodated in a single vehicle concept except for the Army's interest in demonstrating transonic and low-supersonic speeds which led to their advocacy of a rocket engine for propulsion, a feature which NACA considered too risky. Except for propulsion, the configurations of the airplanes being studied by NACA and the Army were similar; both agencies considered very simple conventional unswept designs.
Undoubtedly, one of the main values of this meeting was a stirring-up of the competitive natures of the participants to the point where actual procurement activities would soon be initiated. Further discussions took place with the Army on May 15 and 16, 1944, and on July 18, 1944, the final NACA turbojet-powered design produced by the small Langley group was transmitted to the Army personnel who by now had declared themselves expressly interested in funding a research airplane (ref. 113). A critique of the NACA design was presented by Army personnel at a Langley meeting on December 13 and 14, 1944, centering on the inadequate performance achievable with the turbojet. NACA emphasized the supposed safety aspects and relatively long-duration data-gathering flights possible with the conventional power plant. Furthermore, the turbojet would have obvious applicability to future military aircraft while the rocket propulsion system did not. This apparently unreconcilable difference was easily resolved; the Army was putting up the money [92] and they decided to do it their way. In late December they started negotiations with Bell Aircraft to procure a rocket airplane.
When it became clear at the meetings in early 1944 with E. Kotcher and his cohorts that the Army was likely to be insistent on a rocket airplane, Stack renewed his efforts to interest the Navy in procuring the kind of airplane NACA wanted. Almost all of his contacts were by telephone, personal visits, or through M. Davidson who had been detailed to the Navy. Stack's view then was that the rocket approach was so risky that the Bell airplane would probably not survive many flights and in any event would not get enough air time to collect much data. The Navy, in the persons of E. Conlon, W. S. Diehl, and I. Driggs, was receptive. Nothing had been done in the Navy in the way of research airplane studies and they were ready to accept the NACA general guidelines. Belatedly, in September of 1944, they started to consider details of such a vehicle within the Bureau of Aeronautics, developing a philosophy not inconsistent with NACA's that the aircraft should be designed with some potential for militarily useful follow-on versions. Douglas was some potent selected to build the airplane in early 1945. It was designated the D-558-1, and was almost exactly the airplane Stack desired (ref. 110). Throughout the development period, he displayed a strong preference for the Navy airplane and we extended ourselves in every way to assist in its development.
During Stack's absence on his first European trip, I was sent to Wright Field on March 15, 1945, to represent NACA at the first design review of the X-1 (then designated XS-1). Prior to leaving, I examined recent drop-body drag data in the vicinity of Mach 1, visited the Flight Division, and talked to Davidson to get their views on performance, operations, and instrumentation. According to my notes, Mel Gough, Langley's chief test pilot, condemned the rocket airplane. "No NACA pilot will ever be permitted to fly an airplane powered by a damned firecracker" was his ultimatum. (Ironically, it was the turbojet-powered D-558-1 which killed a NACA pilot due to engine failure while the X-1's had a good safety record at Edwards. The D-558-1 barely exceeded Mach 0.83 in level flight and was limited to Mach numbers below 1.0 in dives. With further irony, it was the transonic and supersonic flight achievements of [93] the rocket-powered X-1 which brought NACA and Stack a share of the Collier Trophy for 1948.)
At Wright Field, I found Bell's design to be basically similar to the simple arrangements of the Army Mach 0.999 study and the NACA studies. In general, NACA recommendations other than power plant and speed range had been accepted (refs. 107, 114). Almost two-thirds of the takeoff gross weight was in rocket propellants-an unheard-of fuel fraction for military aircraft of that day. Were it not for the fact that a major part of the propellants were used up in takeoff and climb, the X-1 as then defined could have reached projected speeds far in excess of Mach 1.2, the "cruise" speed required by the Army. It was apparent that a cruise speed of Mach 1 could certainly be reached from ground takeoff even with more conservative drag estimates based on the body-drop data, and I pointed out that this made the airplane acceptable from the NACA viewpoint which suggested Mach 0.8 to 1.0 as the desired region for flight research (ref. 107).
Later Bell's considerations of safety and performance with a less energetic propulsion system led in May 1945 to a major change from ground takeoff to air launch. NACA strongly opposed air launch. Not only did it violate the NACA notion that a research airplane should operate as conventionally as possible, but it also meant that in all probability the airplane could never be operated out of Langley Field. Langley managers thus feared they would lose control of an air-launched X-1 flight program (ref. 110). The NACA protests were of no avail because air launch was now the only remaining option if low supersonic speeds were to be achieved as required by the Army.
Concurrently with, but unrelated to, the X-1 and D-558-1 research airplane activities of 1944 and 1945, M. C. Ellis and C. E. Brown of Langley's 9-inch supersonic tunnel section studied the feasibility of a small supersonic airplane powered by a hypothetical ramjet engine at Mach 1.4. As was appropriate in a rough preliminary assessment of this kind, their airplane was a primitive assemblage of basic elements-straight sharp-edged wings and tail, and simple propulsive-duct fuselage with the pilot sitting in a small enclosure in the middle of the duct (fig. 22). The results showed that a ramjet of practical proportions could....

[94] FIGURE 22.-Ramjet-powered configuration analyzed by Ellis and Brown, 1944-1945. From a Langley Conference chart.
....indeed provide the necessary cruise propulsion for a 60-mile range at Mach 1.4; however, other means of acceleration through the transonic region (rockets) would be required, and airplane tow (later, air launch) was envisioned for takeoff (ref. 115). At Stack's instigation, there was brief local consideration of this vehicle as a possible addition to the X-1 and D-558-1 stable of research airplanes. However, because of the lack of any ramjet engine, the problems of acceleration, and more particularly the fact that transonic flight testing of the simpler X-1 and D-558-1 was still several years away, it was quite obvious that such a vehicle [95] could not logically be undertaken at that time, and Stack wisely did not attempt to mount a real crusade for it.
It is unlikely that the many innovations and rapid progress in the transonic ground facilities would have happened as they did without the stimulus and focus provided by the X-1 and D-558. Clearly there was a most important two-way flow of benefits: stimulated by the problems of the research airplanes, new ground facilities and techniques were developed which, in turn, produced vitally needed new data in time for the design and safe operation of the aircraft.
The primary postulate justifying the transonic research airplanes was the supposed impossibility of useful wind tunnel operations in the speed range above Mach 0.8. And yet before the research airplanes were operated transonically the fallacy of this justification had been demonstrated; the wind tunnel choking problem had been circumvented in a variety of ways. Thus the early concept of research aircraft providing unique new data otherwise unobtainable became obsolete. Instead, a principal value of the transonic flights lay in evaluation and validation of the ground-based techniques. The fact that the first transonic flights showed no unexpected occurrences was also of great value. The most basic value, however, was the liberation of researchers and aircraft designers from their fears and inhibitions relative to the "sonic barrier." The awesome transonic zone had been reduced to ordinary proportions, and aeronautical engineers could now proceed with the design of supersonic aircraft with confidence.
During the course of this review of the first research airplanes, I turned up a number of apparent misconceptions and inaccuracies in the records which are worth noting. One should expect, of course, that the offhand and undocumented remarks recorded in interviews of NACA old-timers will contain inadvertent inaccuracies, distortions, and oversimplifications. I am concerned here with larger issues, in which questionable NACA party-line versions of what happened seem to have gained general acceptance, establishing a sort of agency mythology or folklore.
Myth: That NACA deliberately planned for two complementary [96] vehicles, one (the X-1) to be a unique special design for pushing through Mach 1, and the other (the D-558-1) to be representative of advanced military service types with turbojet propulsion for studying flight problems in the Mach range up to about 0.95. This view is specifically stated in certain of the interviews conducted by Bonney in the early seventies, and both Keller and Hallion gained the same impression from their interviews (refs. 112, 116). Actually, as previously documented, NACA had argued strongly against a rocket vehicle like the X-1, and even after it was in procurement NACA stated that the subsonic speed range from 0.8 to 1.0 was the area they desired to explore (refs. 107, 114). The D-558-1 was the research airplane NACA wanted (ref. 110).
If the D-558-1 could have been promoted in the early forties, it would have been timely. But coming into the flight picture as it did in 1947, it was unnecessary. Contemporary service airplanes with equal or better performance became operational in the same period and they could have been instrumented and used for most of the work conducted by the D-5581. For example, the F-86 Sabre began to exceed Mach 1 regularly in dives in the summer of 1948, some time before the D-558-1 inadvertently slightly exceeded Mach 1 for the only time on September 29, 1948. The world's speed record of 650 mph briefly held by the D-558-1 also fell to the F-86 on September 15, 1948, when 671 mph was recorded.
Nevertheless, it was the D-558-l's and not the advanced service aircraft that were used for extensive flight research at high subsonic speeds by NACA, complementing coverage of the higher transonic speeds by the X-l's. It is quite understandable how some NACA managers by hindsight can see a logic in the way these two vehicles were used that did not really exist when they were promoted in 1944 and 1945.
NACA always chose to emphasize the positive factors of the program as it finally evolved, passing over early controversies. An example is seen in Stack's 1951 paper (ref. 54) in which he said, "The research airplane program has been a cooperative venture from the start.... The extent of the cooperation is best illustrated by the facts that the X-1, sponsored by the Air Force, is powered with a Navy-sponsored rocket engine, and the D-558-1, sponsored by the Navy, is powered with an Air Force-sponsored turbojet engine."
[97] Myth: That a lack of knowledge or misunderstanding of the effects of wing thickness ratio on transonic performance led to the major differences of opinion in NACA as to what thickness ratio should be used on the research airplanes. Actually, from the earliest works of Dryden and the NACA high-speed airfoil group, a major conclusion was that the severity of shock-stall effects could be minimized by using thin sections. Ferri's airfoil work (ref. 45), extending to Mach 0.94, edited in report form in January 1945 and published in June 1945, listed as a primary conclusion: "Airfoils of large thickness ratio should not be used at high Mach numbers because of radical adverse changes in their characteristics at supercritical speeds. "Gilruth's secret wing-flow data of 1945 extended the test speed range beyond Mach 1and it served to underscore the existing understanding of the problem. It did not provide pivotal new revelations of the advantages of thin wings as has been implied (ref. 112).
The real argument was over whether the research airplanes should be designed deliberately to encounter severe shock stalls well below Mach 1 for correlation with the wind tunnel data. Stack argued vociferously for a 12-percent-thick wing (an "average" rather than a "thick" wing according to 1945 practice) which would start to encounter flow changes at Mach numbers of about 0.75. This was one of the first major crusades into which he put the full force of his unusual talents. The main thrust of his argument was that there would be far less risk with this over-strength airplane with a 12-percent wing in level flight than Army test pilots had accepted repeatedly in pullups from high-speed dives. Gilruth, however, took the more conservative view that the first aircraft to penetrate deeply into the supercritical zone should have every known feature which would contribute to a safe operation-and a thin wing was indisputably one of the most important features for minimizing supercritical buffeting, lift loss, and control problems. Thompson sided with Gilruth. The first X-1 was flown with an 8-percent-thick wing of very low camber. However, the pressure distribution measurements, which were of prime importance for comparison with the wind tunnels, were made on a 10-percent-thick wing-not much thinner than Stack had wanted.
At it turned out, the most important region for comparison of flight and tunnels was from Mach 0.9 to 1.1, and the thinner wings served as well as a thicker one would have. The region of deep shock stall, Mach [98] 0.75 to 0.9, which Stack advocated, proved relatively unimportant from the correlation standpoint. Twenty years later, accepting the teachings of in history, Stack acknowledged the correctness of the thin-wing decision in remarks made at the AIAA history meeting of 1965 (ref. 10) where he said, "We knew it should have a thin wing."
Myth: That NACA made a substantial effort to promote a supersonic ramjet-powered research airplane in 1945. The unusual emphasis with which Stack recalled the exploratory study of Ellis and Brown in his 1965 history talk and interviews with Hallion and others (refs. 10, 112) seems to have created an exaggerated historical view of the importance of this concept in the research airplane picture of 1945. There was no ramjet engine then in existence to power such a vehicle; the X-1 and D-558 were still in the early stages of procurement; rather obviously, any proposal for such a vehicle was premature and had virtually no chance of support. Neither Ellis nor I have any record of the proposal. H. A. Soule believes he recalls a Stack memorandum which was either lost or withdrawn (ref. 110). In any case Stack's effort was brief and in no way comparable to his vigorous and long-standing promotions of the transonic airplanes. There is no doubt that Stack had a strong personal interest in supersonic flight in 1945-in addition to his better-known interest in flight research at high subsonic speeds. Perhaps this is the point he wished to make in his talks with the historians.