By J Leland Atwood, former chairman, North American Aviation and Rockwell International -
The Supermarine Spitfire and it's stable mate, the Hawker Hurricane, are probably the most appreciated defensive weapons in the history of civilization for a very good reason. These airplanes -- with their elegant Rolls-Royce engines -- enabled the determined RAF to stand off certain defeat and occupation. The legendary Reginal Mitchell, leader of the Supermarine design team, worked to the end of his life perfecting the Spitfire, and Sidney Camm of Hawker brought in the Hurricane, largely with private financial backing of T.O.M. Sopwith, himself a World War I airplane designer. These events were surely high on the list of the accomplishments in England's "Finest Hour" and no later achievements in this category can be classified as in the same degree of effectiveness or timeliness. The Spitfire, in particular, with somewhat more performance, is especially memorable and symbolic.
But the RAF (the "Few") made its Thermopylae stand in 1940, and the war lasted for nearly five more long and bitter years. The United States was able to mobilize its capabilities, including massive air power and some very good airplanes in great numbers. I participated as an aeronautical engineer and manager and would like to describe the origin and some aspects of the P-51 "Mustang" fighter as one of those airplanes. I do not intend to elaborate on its capabilities as a first line warplane with the speed and range to carry air combat successfully to the heart of Germany -- which is a well known matter of record -- but rather on some interesting sequences leading to its origination and some technical aspects of its design which hopefully, I can describe in reasonable language without integral signs or complex equations -- although I can not eliminate mathematics entirely and still explain the design rationale.
To begin with, the Mustang had a large British component. In 1940, it was underwritten by England with their very scarce U.S. dollars (12 million of them), utilized Farnborough research in design, and in its final and best configuration it used the incomparable Rolls V-1650 "Merlin" engine. It was, of course, taken over by the U.S. Army Air Corps, which eventually purchased and financed its large wartime production and supervised its specifications and utilization.
Briefly the British and French both began to buy airplanes and engines in the U.S. in 1938, and shortly orders were being issued for this equipment, including engines from Pratt and Whitney, Allison, and Wright Aeronautical, planes from Lockheed, Douglas, North American, Curtiss (including P-40 fighters), and others. The British shortly established a Purchasing Commission, first under Mr. Arthur Purvis, who was replaced in 1939 by Sir Henry Self. Offices were taken at 15 Broad Street, New York, and staff was assigned.
North American Aviation (now Rockwell International) operated in Inglewood, California, adjacent to Mines Field, which is now Los Angeles International Airport (LAX), and or British orders were for advanced trainer planes, a version of the Air Corps AT-6, which the British named the "Harvard." These, for the times, were relatively large orders -- eventually involving several hundred planes -- and were naturally very important for North American, leading to greatly increased employment and additional buildings.
As we went into 1939, concerns about the possibility of war increased and our military received larger appropriations and began to place orders. In mid 1939, North American received a large order for our B-25 medium bomber and expansion continued at a rapid rate. This was true of the rest of the industry also, and capacity was getting to be the problem.
North American Aviation, though a derivative of some antecedent organizations, was a relatively young company. Through earlier investments and involvement, the General Motors Corporation owned about 30% of the stock and effectively controlled the company. Ernest R Breech, a General Motors vice president, was designated chairman -- although he never served in an operating capacity. In 1934, he recruited James Howard ("Dutch") Kindelberger, a vice president of Douglas, as president and chief officer. I worked for Douglas at that time under Dutch as a mathematical analyst and component designer, and he recruited me to come with him as chief engineer at North American. Dutch was just 39 year old and most of the rest of us were some 10 years younger.
We were quite successful with the advanced trainer line, including the Harvard, and had built a couple of medium bomber and attack plane prototypes, which didn't really get anywhere until the B-25 order. So in 1939, we were booked up and expanding, and a transition in the organization was shaping up. Dutch had a lot of balls in the air with contracts, building plans, machinery orders, financial requirement, personnel expansion, government interfaces, etc and after trying and failing a couple of times to get a competent deputy , he began to move me into that position. In my place, Raymond Rice became chief engineer and I ended up with the title of first vice president. This transition was somewhat gradual, but by the latter part of 1939, it was rather complete -- although I kept in very close touch with the engineering work.
At about this time, we first heard about the possibility of taking an order for supplementary production of the Curtiss P-40 fighter planes. Of course, fighters were and obvious requirement and in 1939 the P-40 was considered a good contemporary plane in this country, but it had some drawbacks. The Allison V-1710 engine had only a single-stage supercharger, and its critical altitude for maximum speed was only about 12,000 feet. While not in the high altitude interceptor class, it could be used for low altitude combat and ground attack missions. To me however, the radiator and cooling system seemed to be most inefficient and poorly located -- with the glycol and oil radiators under the rear of the engine and partially cowled. Also Dutch felt we were heavily loaded as far as tooling and production were concerned and we would have a hard time coping with Curtiss drawings, manufacturing standards and tooling, at least for some time.
As chief engineer, I had regularly reviewed the NACA (Nation Advisory Committee for Aeronautics, later NASA) reports on aerodynamics and related subjects, and in 1939, one came to my attention that was a review of some British experimental radiator work at Farnborough, a research establishment in England. An investigator named Meredith had experimented with energy recovery from airplane radiators. This, of course, was not anything new conceptually, since energy recovery in steam and heat engines was common, as in triple expansion cylinder engines and in turbine applications, but these all started at relatively high temperatures. In reciprocating internal combustion engine, the "coolant-out" temperature cannot be allowed to exceed something like 250 deg Fahrenheit, which is at about the temperature of the end of a heat recovery cycle in a steam engine. However, Meredith experimented with fully ducted radiators and showed that substantial recovery was possible.
Aircraft radiators had been generally treated like those in automobiles, using the speed of the airplane to force air through the radiators (ram air) and dissipating the heated air at random. The ram air pressure is proportional to the square of the speed(V2), but only directly proportional (1 to 1) for changes in density of the air with altitude. This is expressed as Pressure = 1/2 rV2 where r is air mass density [Bernoulli's classic equation for uncompressed flow], Mass of an object, or a quantity of air, is intrinsic and does not change with gravity, but on earth its measure is in pounds or kilograms. So we divided weight by the acceleration of gravity (32.2 fps/s) to obtain mass which will be the same on the moon or in space as it is on earth. Using the weight of the air and the earth's gravity constant, r at sea level is 0.00237 mass units per cubic foot in English measure, and 0.001606 at 25,000 feet.
Now airplane cooling has to be effective for various speeds and power settings, so the conventional radiator had to be able to cool and engine at full power in a climb at perhaps half its speed at full-power in level flight. So an airplane climbing at full power at 150 mph would require about four times the radiator exposure area as the same plane at full power in level flight at twice the speed, 300 mph. This fixed radiator exposure, of course, led to an unnecessarily high drag at high speed and absorbed a great deal of the engine's power. It also had to cool on the ground -- but only at idle or taxi speed power output.
Radiators constructed of tubing and metal fins considerably restrict airflow and so Meredith's experimented with ducting it out to the airstream. By making the outlet variable, he could restrict the air passing through the radiator to just that amount needed for cooling. Pressure ahead of the radiator, P1, is determined by the speed and air density (altitude) and is approximately 1/2 rV2. By closing the outlet partially, the pressure behind the radiator, P2, is maintained to the level that permits just enough air to pass thorough for cooling purposes. It is also apparent that the intake opening can be much smaller than the radiator size and that the drag is much less.
In passing through the radiator, the air is heated and expands in volume. A 200 deg Fahrenheit temperature rise expands the air some 40%, so it can be seen that the discharged air -- although having the same mass as the incoming air -- has a larger volume and for a given pressure requires a larger discharge opening providing some forward thrust. This trust is roughly the pressure behind the radiator, P2, times the area of the discharge opening . This, incidentally, is the principle of the ramjet engine with, of course, much higher temperatures.
With this insight from the Meredith report, I began to gradually think about some way it might be applied to the P-40. However, with a little more consideration I began to believe that in spited of the extra plumbing and probable weight increase, the radiator should be in the fuselage with only the duct openings exposed. The P-40 had the cooling system forward under the rear of the engine, and to balance the plane properly for stability, the pilot was rather far back -- somewhat compromising his view and limiting fuselage space.
The idea of a re-design, or even a new design, looked attractive, but the thought of such a possibility seemed somewhat fanciful since I had never seen any government buy a production plane without a set of requirements in detail, some kind of competition and/or flight test approval and a formal appropriation of money.
In my position as vice president, I had responsibility for contract administration, among other things, and so had occasion to go to 15 Broad Street rather frequently to negotiate contracts, prices, spare parts, equipment, and support services. In January 1940, I told Dutch that I would like to try to get some kind of a fighter authorization and that I hoped my ideas on reduced cooling drag might be a vehicle. He was generally supportive, but skeptical, as I was myself. My best hope was perhaps a contract to modify a single P-40 or possibly to build an experimental airplane.
The British Purchasing Commission, in addition to Sir Henry, had as principal personnel Air Commodore Baker, Colonel William Cave, and J.C.B. (Tommy) Thomas. Thomas was a senior technical man, and I used some occasions to talk to him about the cooling drag subject, making the point that my confidence in the possibilities of a major improvement was based on the Farnborough papers as well as the natural technical logic of the application.
I made a point of visiting Tommy and also Bill Cave when I could, both on direct business and from Dayton and Washington, which I visited frequently. Coast-to-coast was just a long overnight trip then in DC-3's and I could cover quite a bit of ground. I could see that my suggestion had been taken seriously after two or three visits, and I believe Thomas established some communication with Farnborough on the subject. I used only some free hand sketches, but Tommy was very astute and technically qualified. The questions about implementation got more concrete, but no company engineering work was started -- it seemed a long shot. I had discussed my concept with Ed Schmued, preliminary design supervisor, who, though not technically educated, had a real talent for shapes and arrangements and mechanical components, but the first work authorization, denominated NA-73, was not until April 1940.
Finally, early in that month, I was invited into Sir Henry's office and was advised approximately as follows: that they had decided to accept our proposal; that I should prepare a letter contract for his signature; that it should provide for the purchase of 320 aircraft of our design; that it provide a schedule and a not-to-exceed price per airplane; that the British supplies equipment, including engine, would be specified; and finally, that a definitive contract would be negotiated on the basis of this letter contract. Furthermore, he told me that since we had never produced a fighter airplane, he considered is desirable that we have some P-40 data as a helpful guide. He specified the P-40 wind tunnel report and the flight test report. He suggested that I attempt to obtain these data. I told him I would immediately try to do so and took the night train to Buffalo, home of the Curtiss plant. Parenthetically, this was on April 10, 1940, the day Hitler seized Denmark and the Norwegian ports. I remember on that day Colonel Bill Cave told me that this was just one of a number of obvious moves.
In Buffalo, Burdette Wright, general manager of Curtiss Airplane Company, was reasonable enough, considering the competitive aspects. Colonel Ben Kelsey of the Air Corps is reported to have said that the Air Corps encouraged him to sell me the data. This I didn't know, but it could have been the case. Later, Dutch Kindleberger quipped that we didn't even open the package, although I am sure that some of our technical staff did examine the reports. I gave Burdy a marker for $56,000 for the copies, went back to New York, and as soon as I could, presented the letter contract. After staff review, Sir Henry signed it, and I went to the LaGuardia Airport. Work Order NA-73 was issued shortly after.
Dutch Kindleberger put a lot of effort and talent into increasing the efficiency of airplane production. Even at high wartime rates of production parts were made in batches, and it was most unusual to have a machine tool dedicated to making one part, or even to one operation. Many tools, especially for sheet metal parts, were "soft" tooling, using masonite, plywood, or low temperature casting materials rather that tool steel, and were much cheaper -- if not as durable. However they were adequate for the purpose, were made very much more quickly, and were adaptable to the inevitable changes that came along. Dutch made many contributions to the cutting forming, and stretch-forming techniques, but his greatest improvement came from a rationalization of assembly and installation processes.
It was common practice to finish the structural elements, wing, fuselage, etc, and then begin installation of equipment -- electrical, hydraulic, armament, instruments and other items -- in the nearly completed structure. In large airplanes, with plenty of access room, this worked reasonably well with few bottlenecks, but in the smaller planes, such as fighters and trainers, the final assembly stage was crowded, hectic and in efficient. Starting with the T-6 series, Dutch required the fuselage and wing structures remain open in sort of half-shell condition until all wiring, tubing, and permanent equipment installations were made and that they be inspected and tested before joining into complete structures. This naturally required that the engineering design provide for this construction process -- so it became part of house practice in all models.
During the war, the War Production Board kept production statistics and the principal comparative parameter was labor hours per pound of airframe (airplanes less engines and equipments). North American's record was consistently about 20% below industry average. Noting this Jake Swirbul, production chief for Grumman, came out to Inglewood during the war and spent a couple of days looking at the process. On departing he visited Dutch and made approximately this remark: "Dutch, I don't believe you have better people or machinery or buildings or production control than we do, but how in hell do you get your engineers to design a plane so that the workers can get to the work?" The final 5,000 P-51 airplanes were built for 4/10ths of an hour per pound and sold for $17,000 each, less government furnished equipment: engine, armament, etc.
In 1940, the science of aerodynamics was largely empirical and much depended on actual tests. Even today, this is true to some extent as far as some fine points are concerned, and wind tunnels are still used. The Mustang first flew in October 1940, with an Allison engine, and soon some problems with the radiator ducting arose. The upper edge of the intake duct had been made flush with the bottom surface of the wing, and we soon found that the air flowing along the surface in front of the duct became a turbulent irregular pattern as it entered the duct and caused an audible rumble and vibration which was unacceptable. Also, it was thought that the opening should be larger for cooling on the ground at low speed, so a fold-down front panel was provided to admit more air for ground operation. This leaked pressurized air and caused considerable drag.
Both these problems required that some re-design and refinement be made. Some very capable aerodynamics people worked very diligently on the problems, using round-the-clock wind tunnel duct models and flight test measurements to arrive at the optimum configuration of a fixed intake with rounded lip edges. Also the intake was moved down some two or three inches to provide a gutter or scupper for the thin layer of turbulent air to bypass the intake. This has been common practice for such ducts ever since.
The Mustangs went to England and began to participate in reconnaissance and low-level rhubarb sorties over enemy territory, although they were not considered for high-altitude combat because of the single stage supercharging. However, the RAF began to note that the Mustangs were faster than the Spitfires at the same altitude, and interest was increased. The Rolls Royce factory actually installed a two-stage Merlin in a Mustang on an experimental basis, first flown in October, 1942. Also, the Army Air Corps had put the Mustang into production as an attack plane, the A-36, which saw service in North Africa, Sicily, and Italy.
The crucial part of the air war was clearly shaping up as an air superiority battle over Germany and occupied Europe, and our bomber losses were becoming insupportable. Unescorted daylight formations were badly cut up and it was becoming clear that formation flying with machine gun turret protection might be losing rather than gaining in this contest., as the Luftwaffe had in 1940 over England.
At about this point, I became aware that Rolls-Royce Merlin production had been established in the United States. Packard Motor Car Company was selected and it's chief engineer was none other than Colonel Jesse Vincent who had designed the Liberty engine of World War I fame. Colonel John Sessums called me from the Pentagon one day in late May, 1942. His message was terse and electrifying: We are sending you a pair of Series 61 Rolls V-1650 engines and we want you to install them in a couple of P-51's. North American engineers worked at top speed to make the structural, aerodynamic, and cooling adaptations -- and in November one was in the air. Excitement was high when the speed results came in -- over 440 mph, or Mach 0.65, at about 25,000 feet. A considerable portion of the potential of the Meredith Effect was being realized as advertised in April 1940.
Soon American made Merlins were flowing, and the P-51's so equipped were deployed on airfield in England in 1943. Although their range was more than that of the Spitfire, it was still inadequate for effective bomber protection to most of the key targets, and the need for "longer legs" was acute. Responding in a way that left flight test stability engineers aghast, Raymond Rice and his team designed and 85-gallon puncture-sealing fuel tank to fit behind the pilot's seat and in front of the ducted radiator. This tank, weighing some 600 pounds when full of fuel, moved the center of gravity of the plane backward several inches and made it longitudinally unstable, meaning that it would not fly hands off, and went into a sharp dive or climb, it would either pitch up or dive out of control if not corrected continuously by the pilot. It was manageable at least, until some of the fuel was used up.
Colonel Mark Bradley was strongly sponsoring this change, and in his assessment for General Arnold and General Spaatz, he took a fully loaded P-51 with wing-drop tanks and 85 gallon fuselage tank and flew a trial mission equal to London-to-Berlin round trip, engaging in a 20-minute full-power simulated combat at maximum distance out and returned. He dropped the wing tanks first, then used the fuselage fuel and finally the internal wing tank supply. The high command accepted this test and the tanks started to go to England.
Lieutenant Colonel Thomas Hitchcock was air attaché in London and was pushing hard for the long-range P-51. He was over 45, a member of a prominent New York family, an athlete, pilot, and reserve officer. He eagerly took a long-range P-51 for a test flight, and in pulling out of a high speed dive the plane failed and he was killed. I was not there of course, and do not know exactly what happened, but it would not be surprising that the stick force reversed and it came back in his lap, over stressing the airplane at high speed. If he had kept under some three times the indicated stalling speed, the wing loads would have been within a 9G limit and well within the wings strength capability. However, that was not his way -- he apparently wanted to do what the combat pilots would have to do. However, with some practice, the Air Corps pilots successfully flew and fought these long-range planes. The appearance of the little friends was a welcome sight for the battered bomber crews.
One of the most important activities in the production of the P-51's was a major coordinated effort to make changes and improvement on a continuous basis -- both on the production line and in the field -- with kits of parts and technical orders to operating units. We fielded a large group of qualified technicians who assisted the engineering and maintenance officers and reported back on problems, deficiencies, and recommendations. Also Colonel Ben Kelsey was very effective in implementing improvements. He would visit combat wings, sometimes flying combat missions, and monitoring problems. He would then make a circuit of Dayton, Inglewood and Washington -- both recommending and authorizing changes and improvements.
This kind of activity was not well understood by some higher-ups, and I had an interesting experience. The chief of the Aircraft Production Board at one point was Charles E. Wilson ("Electric Charlie"), who had been chief officer of the General Electric Company. Mr. Wilson was making the rounds of industrial companies, generally stressing the importance of war production and looking for ways to improve it. After touring the Inglewood plant, he gave a bit of a speech to a group in a conference room, consisting of some 50 or 60 of our leading engineers and production supervisors. He as talking production, production to the limit, and when he finally paused, I spoke up and said something like this: "Mr. Wilson, if we just produce all we can we are not doing our best for the war effort." He seemed surprised and almost affronted and asked what I could possibly mean. I tried to explain that we had a large backlog of changes that would improve the safety and effectiveness of the planes and that we must take some time to fit them in. He didn't say much more, but that evening, he spoke in downtown Los Angeles to a civic industrial group, and I had the satisfaction of hearing him say that in all the need for production, we must do all we can to make improvements in efficiency and serviceability as we go along.
One might ask where is the Meredith Effect today? It is alive and well and was applied on radial engines in the form of cowl flaps, but in modern jets, there is little requirement for direct cooling of fluids or air in the 200 deg Fahrenheit temperature range. Jet engine bleed air is hot and high pressure, so for cooling purposes, some heat is extracted by ducted heat exchangers in the Meredith manner, and the high pressure air is then rapidly expanded. The snow flakes sometimes seen in the jet passenger cabin ventilators are a result of expansion cooling, the reverse of compression heating.
To summarize, the Mustang cooling system provided just enough, but no excess of cooling air. Ideally, the back pressure P2 should be minimum, close to zero, in a low-speed, high-power climb and maximum at high speed and in long-range cruise, resulting in the lowest net drag where it is most needed.
An objective assessment of the Mustang is probably unavailable or inconclusive. During the war, the Collier Trophy Committee, certainly no military authority, passed over the Mustang repeatedly. The Congress, no center of airplane technology, in postwar assessment, declared the plane to be the most "aerodynamically perfect" plane of the war. Perhaps we should simply say that it was just one facet of the effort of million of people doing their best for the war effort with varying degrees of capability and effectiveness. After all, the front line personnel deserved the honor in war.
In recent years, there has been much introspection and analysis of the U.S. manufacturing establishment, and much has been written about product quality, employee participation, more choice for customers, quick change of models, rapid correction of deficiencies, flexibility in tooling and methods, and cost in general. I have seen many notable and almost heroic efforts in engineering and production and great efforts at product improvement. In the Apollo lunar program, I have seen responsible people work to an almost unbelievable degree to make some very difficult goals. However, while war production was a massive struggle against shortages of every description, I doubt that I shall ever see again such a degree of product improvement, employee participation, relative product value, economic production and generally superior results as I experienced in Dutch Kindelbergers airplane production complex during the period of 1939-1945.