Category Archimedes

Before a new engine or airframe achieves its first flight much prior ground testing has been done in wind tunnels or engine test cells. Ground and flight tests are run to establish performance characteristics and to aid in design development and refinement. This requires collection and analysis of relevant test data from test runs in specially instrumented engines, scale models, and aircraft.

The fundamental task of such tests is collecting performance and reference data. What data are collected depends upon the purpose of the tests:

Development testing to refine the final production design;

Type or endurance testing as precursor to military or civilian acceptance of the basic design;

Flight tests demonstrate aircraft or engine ability to operate under realistic circumstances, uncover design difficulties, and establish maintenance schedules for production aircraft or engines;

Acceptance tests to show that individual production engines meet minimum contractual performance characteristics.[2]

Some development, acceptance, and engine endurance testing can be done in wind tunnel and engine test stand ground facilities; the others invariably are airborne.

Instrumentation, which is the source of data from tests, tends to be most extensive in development testing and flight test – which are my focus. Airborne tests technically are the most demanding. For data to be useful, they must be recorded and processed into interpretable forms.

Instrumentation, recording, and processing of aircraft data have evolved substantially since the latter 1800s. These developments do not sort themselves into nice periodizations, but can be construed as three contrasting testing styles, overlapping for as much as 40 years, but each dominating different periods.

In the first style the primary airborne instrument is the test pilot’s subjective judgments augmented by notes on a knee pad and whatever readings of basic flying instruments could be jotted down. This style is important from the beginning of flight until about 1945, though a remnant today is the test pilot controlling what parts of the test flight are recorded at what data density.

67

P Galison and A. Roland (eds.), Atmospheric Flight in the Twentieth Century, 67-105 © 2000 Kluwer Academic Publishers.

The second style emphasizes enhanced instrumentation recorded by something/ someone other than the test pilot, where recorded data, not the pilot’s reactions, are the primary data. Instrumentation can include an observer taking manual readings, gun cameras recording duplicate instrument panels, recording barographs, photopanels, transducer fed oscillographs, and telemetering to ground stations. Another defining characteristic is that data recording does not allow direct computerized analysis of the data. This style begins in the 1920s and dominates in the 1950s and 1960s.

The third style emphasizes very extensive automated instrumentation using transducers and probes, automatic pre-processing of data and digital data recording for computer analysis. This first comes in with the XB-70 and dominates high-end flight test subsequently – though style two continues in lower-end testing today where oscillographs continue to be used.

My story concerns evolution of instrumentation and data from style one to three. On the eve of World War II pilot reports, limited recording of data, and hand-analysis of recorded flight test data were typical. During the next three decades, automated data collection and digital computer reduction and analysis of data became the norm,2 with as many as 1200 channels of data being recorded and analyzed. The transformation essentially was complete with the instrumentation and data handling systems of the XB-70.1 will discuss the main transformations and changes in flight – test and ground-test instrumentation, data reduction, and analysis during that pivotal thirty-year period. Consideration will be given to wind tunnel testing, engine test-cell investigations, and flight-testing of both engines and airframes. I focus on turbojet – powered aircraft.

I also make some systematic philosophical remarks on data and modeling and offer concluding observations. 1

far enough to ensure reliable performance, the now proven new engine could be put into the unproven airframe.6

Before an engine ever is taken aloft, it undergoes a great deal of testing on the ground in test cells. Similarly, before a new airframe is taken aloft, it has undergone extensive aerodynamic testing in wind tunnels. A general rule of thumb is to use flight test primarily for what cannot be studied in ground testingfacilities.

The NACA research on transonic and supersonic compressors remained classified until the late 1950s. (Even the design “bible,” which focused on more conventional stages, was classified until 1958.) Consequently, the results of the research were not generally disseminated to those outside the United States, and even in this country they were not readily accessible. Moreover, unlike the “bible,” the reports themselves were aimed more toward providing a record of what had been done than toward instructing those outside NACA how to exploit the results. Even today, when read from the perspective of our far greater knowledge of transonic and supersonic stages, the reports are not always easy to assess. A large fraction of the knowledge that the NACA had gained on high Mach number stages remained in the heads of the engineers who had conducted the research.

This knowledge diffused out of the NACA through more than publications, however. Many engineers who had worked on high-Mach-number stages throughout the decade left NACA in 1955 and 1956. The Committee curtailed compressor research when Lewis, believing no fundamental problems remained in air-breathing engines, turned its attention to nuclear and rocket propulsion.46 Langley’s Jack Erwin and Lewis’s John Klapproth, Karl Kovach, and Lin Wright moved to General Electric. Kovach and Wright joined the company’s axial compressor aerodynamic design group, headed by Richard Novak, where they shifted their primary focus from research on airfoil shapes and parameters to design.

In some respects this timing was opportune. NACA research had produced the compressor design bible and had achieved sufficient success with transonic stages to turn the future over to the engine companies. The decade of research on supersonic compressors, the promising results in the last years notwithstanding, had yet to yield flight-worthy designs, making it hard to argue for continued funding. General Electric proved the beneficiary of the NACA’s change in focus, for GE offered the NACA engineers the chance to apply their experience with advanced, experimental designs to real engines. The knowledge Kovach and Wright brought from the government research establishment into the industry immediately began having an impact on the advanced designs GE was then developing. Wright’s knowledge, in particular, proved crucial to GE’s development of a radically advanced fan that formed the basis of their first flight-worthy turbofan engine, to which we now turn.47

Wood was the dominant structural material for airplanes from the pre-history of flight until the early 1930s. By the late 1930s, however, wood was rapidly disappearing, especially in the structures of high-performance military aircraft and multi-motored passenger airplanes. Metal succeeded as a result of intense efforts to develop all-metal airplanes, efforts that began in Germany during World War I and quickly spread to Britain, France and the United States after the Armistice.4

In both Europe and the United States, national aeronautical communities maintained a powerful commitment to developing metal airplanes between the world wars. As I have argued elsewhere, this commitment cannot be explained by the technical advantages of metal. The technical choice between wood and metal remained indeterminate between the world wars; wood had advantages in some circumstances, metal in others. Claims for metal’s superiority in fire safety, weight, cost, and durability all proved equivocal throughout the 1920s.5

Despite the questionable advantages of metal in the 1920s, national governments and private firms concentrated their research and development programs on improving metal airplanes, while shortchanging research and development on wood structures. This bias was especially strong in the United States, where the Army Air Service began shifting research funds from wood to metal as early as 1920. Nevertheless, successful metal aircraft proved quite difficult to design, and the U. S. Army remained heavily dependent on wooden-winged aircraft until the mid-1930s. After about 1933, however, new all-metal stressed-skin structures proved competitive with wood, especially in larger airplanes. Even with the substantially increased production costs required by the new all-metal stressed-skin structures, wood quickly disappeared from most high-performance airplanes in both the United States and Europe.6

One cannot, however, invoke metal’s eventual success to explain why this path was chosen in the first place. Metal’s success resulted from years of intensive development before the predicted advantages of metal became manifest. Proponents of metal advanced no clear-cut technical arguments to justify continued support for metal in the 1920s, when experience with metal failed to corroborate claims for its superiority to wood.7 In the United States, at least, the embrace of metal was driven not so much by technical criteria as by the symbolic meanings of airplane materials.

Metal’s supporters openly articulated these symbolic meanings in the 1920s. They insisted that the shift from wood to metal was an inevitable aspect of technical progress, arguing that the airplane would recapitulate the triumph of metal in prior wood-using technologies, such as ships, railroad cars, and bridges.

Advocates of metal drew upon pre-existing cultural meanings to link metal with progress, modernity and science, while associating wood with backwardness, tradition and craft methods. These symbolic associations gained their evocative power from the ideology of technological progress, a set of beliefs deeply embedded within the aviation community. By linking metal with progress, advocates of metal were able to construct a narrative of technological change that predicted the inevitable replacement of wood by metal in airplane structures. This narrative provided more than rhetoric; it also inhibited expressions of support for wood while insuring that metal received a disproportionate share of funds for research and development.8

By 1940 virtually all flight test involved some form of instrumentation and means for recording test data. Today instruments typically are electrical or electronic, and are built out of three basic kinds of units:

Input transducers convert physical quantities of interest into electrical signals.

Examples: pressure transducers, thermocouples.

Modifiers change those signals from one form to another. Examples: filters,

amplifiers, analog-to-digital converters.

Output transducers convert modified signals into a non-electrical quantity.

Examples: Meters, digital read-outs, X-Y plots.7

Early mechanical instruments such as recording barographs and manometers can be analyzed similarly. Whatever their form, transducers and modifiers realize various mathematical transforms or transfer functions.8

There are two main kinds of data: Analog data represent information by continuous changes in signal frequency or amplitude. Digital data code information via sequences of signal pulses.

Information gathered by instruments may not be in usable form for analysis. Data reduction is the process of converting data from the recording format (e. g., pilot notes, photos of instruments, oscillograph traces) into the format required for data analysis.9 It includes calibration corrections for systematic and dynamic instrument errors as well as environmental influences.10 Although data are collected against time, the data analysis almost always is against some performance characteristic. For example, engine performance most often is plotted against engine speed.11

General Electric’s interest in the turbofan concept goes back to the mid-1940s, when they obtained a preliminary design from Frank Whittle.48 GE had undertaken a substantial effort to develop a fan engine in the early 1950s. The combination of the fan’s larger diameter and tip Mach number restrictions requires its rotational speed to be much lower than the rotational speeds of single spool compressors. To meet this requirement, GE adopted an approach that was standard in turboprop engines, namely using a speed reduction gear to drive the fan off the core engine rotor. Where Whittle had sought to “gear down the jet” aerodynamically, GE did it literally. The first step in the development of such a geared turbofan was to develop an efficient, high specific-power core engine that could drive a highly loaded fan through gears. GE designated the core engine they designed for this purpose, using internal funding, the D-2. On test, it proved to be worse than disappointing. In pursuing exceptionally high efficiency at design speed, GE had compromised off-design operation to such a degree that the overall engine was not self-sustaining until it had nearly reached full RPM. Correcting this fault was going to require an extensive redesign of the gas generator. Instead, GE abandoned the D-2 project.49

Peter Kappus, the principal advocate of the turbofan engine within GE, then began pushing the concept of an aft fan. The idea was to install an independently rotating fan rotor behind the gas generator. The exhaust from the gas generator would drive turbine blades mounted on this rotor, and fan blades would extend from the tips of the turbine blades. This idea dates at least as far back as a Whittle patent50 and the Metro-Vick counter-rotating fan discussed above. One of the leading academic experts on turbomachinery, G. F. Wislicenus of Penn State University, had promoted its advantages in a talk entitled “Principles and Applications of Bypass Engines” presented at the Society of Automotive Engineers Golden Anniversary Aeronautical Meeting in April 1955.51 The most obvious advantage of an aft fan from GE’s point of view was that a new core engine would not have to be developed. GE could use the J-79, or what amounted to almost the same thing, its commercial counterpart, the CJ805. This engine had the specific-power required for a viable turbofan engine. Because the fan component was to be aft of the core engine and not mechanically connected to it, the performance of the core engine could be taken as a given. Only the turbofan component would require development funding.

GE committed funds for the development of an experimental aft fan engine in 1956.52 The responsibility for designing the fan component was assigned to the Flight Propulsion Laboratory. John Blanton (see Figure 9) had responsibility for the overall performance of this component. Blanton, a graduate of Purdue, had joined GE in 1956 after a distinguished career at Bell Aeronautical, where he had risen to Assistant Chief Design Engineer. The detailed aerodynamic design of the fan itself fell under Dick Novak’s compressor aerodynamic design group. Novak, a graduate of MIT, had been with GE since the mid-1940s, starting as a field test engineer in the Mohave Desert, but subsequently coming to focus on the aerodynamics of axial compressors, placing great emphasis on analytical design. Novak assigned the aero­dynamic design of the fan to Lin Wright. A graduate of Wayne State University, Wright had joined GE in mid-1956 after a ten year career as one of the central figures in the NACA supersonic compressor research program, starting at Langley and then moving to Lewis. His last project at NACA-Lewis had been the design and test of a highly loaded 1260 ft/sec tip-speed transonic rotor intended to be the first stage of a two-stage counter-rotating compressor.53

Americans were not alone in the shift to metal in the interwar period; aviation technology had little respect for national boundaries. Although there were distinct design styles in particular firms, all the major industrialized powers followed the same general pattern with regard to airplane materials. By 1939, the air forces of Germany, France and Britain had all converted to metal structures, with aluminum alloys preferred. Italy and the Soviet Union lagged behind somewhat, continuing to use wood for some combat airplanes, but the trend in those countries was clearly towards metal as well.

Nevertheless, in the late 1930s, wood was poised for a significant revival in aircraft structures. Despite the apparent triumph of the all-metal airplane, wood construction had not remained static. In Germany, Britain and the United States, a few aviation researchers and airplane designers began exploring new construction techniques during the 1930s using synthetic resin adhesives. These new adhesives, which were based on common phenol-formaldehyde thermosetting plastics, eliminated the worst problems of traditional wood glues, especially the tendency to deteriorate when damp. In addition, the synthetic resins made possible significant improvements in the strength properties of laminated wood products, while permitting the use of various molding techniques that promised substantial savings in labor.9

In the United States, interest in the new adhesives was driven by the high skill levels and labor inputs required to manufacture all-metal airplanes. For metal airplanes, the key problem lay with the lowly rivet, a fastener required by the difficulty of welding heat-treated aluminum alloy. A small training airplane could require 50,000 rivets, and a large bomber nearly ten times as many; riveting accounted for some 40 percent of the costs of a typical airframe.10 According to Virginius E. Clark, a prominent American aeronautical engineer, “any type of structure which demanded such a multiplicity of reinforcing parts and so many thousands of rivets did not constitute the best final answer for rapid and inexpensive production.”11 In addition, rivets made it very difficult to obtain the extremely smooth external surfaces needed by high-speed airplanes. Although engineers developed various methods of flush riveting to deal with this problem, smooth riveted surfaces remained difficult and expensive to manufacture.12

Around 1935, Sherman Fairchild, president of the Fairchild Engine and Airplane Corporation, began to have doubts about the suitability of riveted all-metal construction for quantity production and high-speed flight. Fairchild assigned the task of eliminating the rivet to Clark, who was then Fairchild’s vice president for engineering. Clark turned his attention to resin-bonded wood veneers, which could be molded into large curved panels to produce a well-streamlined airframe. Clark began working with the Haskelite Manufacturing Corporation, formerly a major supplier of aircraft plywood. The Fairchild and Haskelite companies jointly developed a bag-molding technique for producing airplane parts of resin-bonded plywood, termed “Duramold” by Clark. In 1937 Clark designed a five-place commercial airplane with a Duramold fuselage, the Fairchild F-46, which completed its first flight on Dec. 5, 1937.13

Clark faced tremendous practical difficulties in developing manufacturing techniques using the new adhesives. The Duramold process represented a synthesis of two lines of development in wood products: molded plywood and resin-bonded “improved” wood. Bag-molding techniques were not new to airplane construction, having been used on the Lockheed Vega, the most successful high-speed airplane of the late 1920s. But in contrast to the casein-glued Vega fuselage, the thermosetting resins in Duramold required molding pressures as high as 100 psi and temperatures up to 280 deg. F, which made the molding equipment much more complicated and expensive.14

Although Duramold started as a civilian project, Clark almost immediately turned to the Army for development and production contracts. Clark, who had been chief engineer for Army aviation in World War I, promised the Army rapid production at low cost. In his correspondence with the Army in early 1938, Clark did his best to disassociate Duramold from wood. Duramold was based on wood, Clark admitted, but “we prefer, insofar as possible, to avoid the use of this word because of the unpleasant associations resulting from most unhappy experiences with ‘wooden’ airplanes in times past.” Instead, Clark attempted to link Duramold with plastics, which in the 1930s carried the aura of a progressive, science-based technology.15

The Army was not fooled. J. B. Johnson, the Army’s chief expert on airplane materials and a metallurgist by training, had no time for wood in any form. Duramold, insisted Johnson, was “simply” plywood glued with a synthetic adhesive.16 Johnson’s assessment of Duramold was shared by other engineers and officers at Wright Field, home of the Materiel Division, the Army Air Corps’ organization for aviation research, development, and procurement. Despite opposition from Wright Field, Clark was able to gamer some support from Army Air Corps officials in Washington, notably General H. H. Arnold, then assistant chief of the Air Corps. Nevertheless, in February 1938 the Secretary of War rejected a request to fund the development of Duramold and other “plastic” materials, arguing that “the present highly satisfactory all-metal airplane is the result of a long period of development at considerable expense. We should concentrate on the perfection of metal airplanes.”17 Clark never obtained an Army contract, and later left the Fairchild company to work with Howard Hughes on his large wooden flying boat.

These negotiations illustrate a struggle to define the symbolic meanings of “plastic” plywood. Clark sought to emphasize the symbolic link to plastics, a progressive technology ripe with manifold possibilities, while Johnson insisted on identifying Duramold with wood, a discredited material already rejected by the Army.18 Soon, however, interest in wood airplanes would be revived, not by its link with the modernity of plastics, but rather due to the threat of war.

More than anything else, it was the threat of war that revived American and European interest in wood airplanes. By itself, the technical promise of synthetic adhesives could not overcome the opposition rooted in wood’s symbolism as a traditional material. Proponents of synthetic adhesives did get some attention by invoking the symbolism of plastics, but this strategy could not prevent critics from pointing out that materials like Duramold consisted mainly of wood veneers. The prospect of war, however, brought problems of production to the foreground. Wood offered potential solutions to some of these problems, in particular shortages of metals, labor, and production facilities. Furthermore, the issue of production gave defenders of wood an opportunity to air a whole range of technical arguments concerning choice of materials.

Renewed interest in wood first emerged in Europe, where the growing threat of Nazi Germany was most keenly felt, especially after the Munich crisis of September 1938. In November 1938, the British journal Aeroplane published an article defending wood by F. G. Miles, a designer of small commercial airplanes and military trainers. Miles insisted that metal airplanes had not “not lived up to early expectations” for quantity production. Wood airplanes, he claimed, offered a number of advantages over metal in design and production. They could be designed more quickly, and they could take advantage of skilled labor in the wood-working trades. Miles predicted that costs would be lower and the supply of material greater. He insisted that, except for large aircraft, wood airplanes could meet the same demanding specifications as metal airplanes with regard to speed and durability. Similar arguments were presented in French and Dutch aviation journals.19

Beginning in 1939, the American aviation press also published a flurry of articles highlighting the new opportunities created by resin adhesives and plywood molding techniques. Most of these articles stressed advantages for war production, even before the German invasion of Poland. For example, in an article in the Scientific American, journalist Forest Davis pronounced molded plywood airplanes of “tremendous wartime significance.” Airplanes were “a machine-age paradox,” argued Davis, still largely made by hand while “automobiles roll off the assembly line like shelled peas into a basket.” Duramold provided the solution, making possible “a practically unlimited supply of stout, cheap, fast airplanes.”20 H. O. Basquin of Haskelite provided a similar but more sober assessment, pointing to the 170,000 workers in the furniture industry who could be shifted to wooden airplane production in wartime.21

Despite the interest in wood generated by the threat of war, proponents of wooden airplanes still had a long way to go to translate promise into practice. As Donald MacKenzie has pointed out, the inherent potential of a technology, which he terms the “intrinsic” properties, are ultimately irrelevant in choices between competing technologies. Most engineers and managers base their choices on extrinsic properties, that is, what the technology achieves in practice. But what a technology achieves in practice depends heavily on the resources devoted to its development. Beliefs about intrinsic properties can influence the allocation of resources to competing technologies, becoming in effect self-fulfilling prophecies, promoting the success of the technology that people believe has the most potential to succeed.22

Proponents of the new wooden airplanes understood this process. Through their interventions in the technical press, they hoped to convince the aeronautical community to devote its resources to solving the considerable development problems that stood between the promise and reality of wood construction. And the problems were indeed daunting. After more than a decade of neglect of wood, metal had a vast advantage in available design data, accumulated experience in manufacturing, and lessons learned from commercial and military service. Metal was in a similar position in the early 1920s, when wood framework structures were dominant. As with metal construction in the 1920s, only the military had the resources to compensate for this disadvantage.

Engine evaluation tends to focus on duct and compressor (turbine and nozzle) efficiencies, pressure ratios, turbine fluid dynamic flow resistance, rotational speed, engine air intake amounts. Typically one measures variables such as fuel flows, engine speeds, pressures, stresses, power, thrust, altitude, airspeed – and then calculates these other performance parameters through modeling of the data. The results usually are presented as dimensionless numbers characterizing inlet ducting, compressor air bleeding, exhaust ducting, etc.12

a. Engine Test Cells

There are two main types of engine test cells:

Static cells run heavily instrumented engines fixed to engine platforms under standard sea-level (“static”) conditions;

Altitude chambers run engines in simulated high-altitude situations, supplying “treated [intake] air at the correct temperature and pressure conditions for any selected altitude and forward speed the rest of the engine, including the exhaust or propelling nozzle, is subjected to a pressure corresponding to any selected altitude to at least 70,000 feet.”13

In earlier piston-engine trials an engine and cowling were run in a wind tunnel to test interactive effects between propeller and cowling. Wind tunnels specially adapted to exhaust jet blasts and heat sometimes are used to test jet engines.14

Test cells measure principal variables such as thrust, fuel consumption, rotational speed, and airflow. In addition much effort is directed at solving design problems such as “starting, ignition, acceleration, combustion hot spots, compressor surging, blade vibration, combustion blowout, nacelle cooling, anti-icing.”15

Test cell instrumentation followed flight-test instrumentation techniques, yet was a bit cruder since miniaturization, survival of high-G maneuvers, and lack of space for on-board observers were unimportant. Flight test centers usually had test cells as well, and performed both sorts of tests. Test cell and flight-test data typically were reduced and analyzed by the same people, and similar instrumentation was efficient. Thus test-cell instrumentation tended to imitate, with lag, innovations in flight test instrumentation. Here we only discuss instrumentation peculiar to test cells.

Test cell protocols involve less extreme performance transitions, and thus are more amenable to cruder recording forms such as observers reading gauges. The earliest test cells had a few pressure tubes connected to large mechanical gauges16 and volt-meter displayed thermal measurements. Thrust measurements were critical. Great ingenuity was expended in thrust instrumentation using “bell crank and weigh scales, hydraulic or pneumatic pistons, strain gauges or electric load cells.”17 Engine speed was the critical data-analysis reference variable, yet perhaps easiest to record since turbojets had auxiliary power take-offs that could be directly measured by tachometer.

By the late 1940s electrical pressure transducers were used to record pressures automatically. Since they were extremely expensive, single transducers would be connected to a scanivalve mechanism that briefly sampled sequentially the pressures on many different lines, with the values being recorded. The scanivalve in effect was an early electro-mechanical analog-to-digital converter,18 giving average readings for many channels rather than tracking any single channel through its variations. Data processing, however, remained essentially manual until the late 1950s.

Test cells operate engines in confined spaces, and engine/test chamber interactive effects often produce erroneous measurements. For example, flexible fuel and pressure lines (which stiffen under pressure) may contaminate thrust measurements. This can be countered by allowing little if any movement of the thrust stand – something possible only under certain thrust measurement procedures. Other potential thrust-measurement errors are:

• air flowing around the engine causing drag on the engine;

• large amounts of cooling air flowing around the engine in a test cell have momentum changes which influence measured thrust;

• if engine and cell-cooling air do not enter the test cell at right angles to the engine axis, an error in measured thrust occurs due to the momentum of entering air along the engine axis;

• pressure difference between fore and aft of the engine may combine with the measured thrust.19

These can be controlled by proper design of the test cell environment or by making corrections in the data analysis stage.

b. Engine Flight Testing

In the 1940s and early 1950s, photopanels were the primary means for automatic collection of engine flight-test data. Early photopanels were mere duplicates of the

test-pilot’s own panel. Later the panels became quite involved, having many gauges not on the pilot’s panel. Large panels had 75 or more instruments photographed sequentially by a 35 mm camera.

After the flight film would be processed. Then using microfilm readers data technicians would read off instrument values into records such as punched cards. Data reduction was done by other technicians using mechanical calculators such as a Friden or Monroe. On average, one multiplication per minute could be maintained.20 This and the need to hand record each measurement from each photopanel gauge placed severe limitations on the amount of data that could be processed and analyzed. Time lag from flight test to analyzed data often was weeks.

The development of electrical transducers such as strain gauges, capacitance or strain-gauge pressure transducers, and thermocouples enabled more efficient continuous recording of data. They could be hooked to galvanometers that turned tiny mirrors reflecting beams of light focused as points on moving photosensitive

paper, giving continuous analog strip recordings of traces. Mirrors attached to 12-24 miniaturized pressure manometer diaphragms also were used.21 Such oscillograph techniques for recording wave phenomena go back to the 19th century,22 but by the 1950s had evolved into miniaturized 50-channel recording oscillographs (see Figure 10). By 1958, GE Flight Test routinely would carry one or two 50-channel CEC recording oscillographs in its test airplanes.

With 50 channels of data, one had to carefully design the range of each trace and its zero-point to ensure that traces could be differentiated and accurate readings could be obtained. Fifty channels of data recorded as continuous signals on a 12” wide strip posed a serious challenge. A major task for the instrumentation engineer was working out efficient and unambiguous use of the 50 channel capacity. Instrumentation adjusted the range and zero-point of each galvanometer to conform to the instrumentation engineer’s plan.

Another aspect of the instrumentation engineer’s job was to design an instrumentation package allowing for efficient adjustment of galvanometer swing and zero point. This amounted to the design of specific Wheatstone Bridge circuits to control each galvanometer. The 1958 instrumentation of the GE F-104 #6742, designed by George Runner, had a family of individual control modules hand-wired on circuit boards with wire-wound potentiometers to adjust swing. (Zero point was adjusted mechanically on the galvanometer itself.) Each unit was inserted in an aluminum “U” channel machined in the GE machine shop, with plug units in the rear and switches and potentiometer adjustments in the front end of the “U”. These interchangeable modules could be plugged into bays for 50 such units, allowing for

Figure 7. Typical Oscillograph trace record; only 16 traces are shown compared to the 50 channel version often used in flight test. [Source: Bethwaite l%3. p. 232; background and traces have been inverted.)

speedy and efficient remodeling of the instrumentation. A basic fact of flight test is that each flight involves changes in instrumentation, and 742’s instrumentation was impressively flexible for the time. 100 channels of data could be regulated; two recording oscillographs were used.

Oscillograph traces are, of course, analog. By the latter 1950s data analysis was being done by digital computer. This meant that oscillograph traces had to be digitized. Initially it was done by hand. Later special digitizing tables let operators place cross-hairs over a trace and push a button sending digital coordinates to some output device. At GE Flight Test initially a modified IBM Executive typewriter would print out four digits then tab to the next position. Later output was to IBM cards.23 Typewriter output only allowed hand-plotting data, whereas IBM cards allowed input into tabulation and computer processing.

GE Flight Test-Edwards did not have its own computer facilities in the late 1950s and early 1960s. Some data reduction used computers leased from NASA/Dryden, which only ran one shift at the time. Sixteen hours a day, GE leased the NASA computer resources – initially an IBM 650 rotating drum machine, later replaced by an IBM 704 and then by an IBM 709 in 1962. Most data reduction was done, however, on the GE IBM 7090 in Evandale, Ohio. Sixteen hours a day, IBM cards were fed into a card-reader and teletyped to Evandale where they were duplicated and then fed into a data reduction program on the 7090; the output cards then were teletyped back to GE-Edwards for analysis and plotting. Much plotting was done by an automated plotter placing about 8 ink-dots per IBM card.

Electronic collection of data with computerized data reduction and analysis radically increased the amount of data that could be collected, processed, and interpreted. Numbers of measurements taken increased at the rate of growing computing power – doubling about every 18 months. The very same computerized data collection and processing capabilities were incorporated into sophisticated control systems for advanced aircraft such as the X-15 rocket research vehicle, the XB-70 mach 3 bomber, and the later Blackbird fighters. As aircraft and engine

control systems themselves became more computerized and dependent on ever­more sensors, engine flight test likewise had to sophisticate and collect more channels of data. Fifty channels was the upper limit of what could reliably be distinguished in 12” oscillograph film, and analyzing data from two oscillograph rolls per flight stretched the limits of manual data processing.

The only hope was digital data collection and processing. When data are recorded digitally, many inputs can be multiplexed onto the same channel. Multiplexing is a digital analog to use of a scanivalve that avoids problems of overlapping and ambiguous oscillograph traces by digital separation of individual variables. A further advantage of digital processing is that the data are in forms suitable for direct computer processing, thereby eliminating the human coding step in the digitization process.

GE got the contract to develop the J-93 engine for the B-70 mach 3 bomber. Instrumentation on this plane was unprecedented, exceeding its North American predecessor, the X-15. GE geared up for the XB-70 project, building a new mammoth test cell for the J-93 in 1959-60 with unusually extensive instrumentation (e. g, 50-100 pressure lines alone), introducing pulse-coded-modulation digital airborne tape data recording, developing telemetering capabilities, and contracting for a half-million dollar Electronic Data Processing (EDP) unit.

The EDP unit primarily was a “modifier” in the instrumentation scheme, filtering signals through analog plug-in filters, doing analog-to-digital conversions, and performing simple scalings. Data recorded on one 2” digital tape could be converted into another format (2” tape or punched card) suitable for direct use on the IBM 704,

709, and 7090. It also had limited output transducers that produced strip or oscillograph images for preliminary analysis. The surprising thing about this huge EDP unit is that it had no computer – not if we make having non-tape “core” memory the minimal criterion for being a computer. The decision was to build this device for data reduction, then “ship” the data via teletype to GE-Evandale for detailed processing and analysis.

In telemetry signals collected by transducers are radioed to the ground, as well as sent to on-board recorders, where a ground station converts them to real-time displays – originally dials, meters, and X-Y pen plotters, but today computerized displays, sometimes projected on large screens in flight-test “command centers.” Test flights are very expensive, so project engineers monitor telemetered data and may opt to modify test protocols mid-flight.24 Telemetry provides the only data when a test aircraft crashes, destroying critical on-board data records. GE Flight Test developed telemetering capabilities in preparation for the XB-70 project, trying them out in initial X-15 flights.

The X-15 instrumentation was a trial run for the XB-70 project (both were built by North American), although the X-15 relied primarily on oscillographs for recording its 750 channels of data.25 With the XB-70 project, the transition from hand-recorded and analyzed data to automated data collection, reduction, and analysis is completed. The XB-70B instrumentation had about 1200 channels of data recorded on airborne digital tape units. Data reduction and processing were automated. Telemetry allowed project engineers to view performance data in real time and modify their test protocols. Subsequent developments would sophisticate, miniaturize, and enhance such flight-test procedures while accommodating increasing numbers of channels of data but do not significantly change the basic approach to flight test instrumentation and data analysis.

A supersonic test-bed was needed for flight test of the XB-70’s J-93 engines. Since the J-93 was roughly 6’ in diameter – larger than any prior jet engine – no

established airframe could use the engine without modification. GE acquired a B-58 supersonic bomber which it modified by placing a J-93 engine pod slung under the belly. Once the aircraft was airborne the J-93 test engine would take over and be evaluated under a range of performance scenarios.

The aft fan required the mechanical design of a new type of blading, with relatively high temperature turbine blades – or, as GE called them, turbine buckets – in the inner portion and relatively cold fan blades of the opposite camber in the outer portion; GE dubbed these blades “bluckets” (see Figure 10). The aerodynamic

design of the turbine blading fell within the state of the art, whether the fan component consisted of one stage or two. But the same was not true of the fan blading. Considerations of weight and simplicity strongly favored a single stage fan. As is always the case, the thermodynamic cycle design of the fan component in­volved a complex set of trade-offs. The CJ805 turbojet produced 11,000 pounds of take-off thrust at a specific fuel consumption – i. e. pounds of fuel per hour per pound of thrust – around 0.70. Blanton found that a 1.56 bypass ratio aft fan behind the CJ805 could increase the take-off thrust to 15,000 pounds at a specific fuel consumption as low as 0.55, a quantum jump in both parameters! The one issue was the thrust-to-weight ratio of the engine, which depended on the weight of the fan component. The fan would have to pass 250 lbs/sec of air at a pressure-ratio of 1.6 with an installed efficiency no less than 0.82. Could this be achieved in a single stage? It was far beyond any single compressor stage GE had ever designed before, or for that matter any stage that had ever been in flight. Wright was nevertheless insistent that it could be done.54

The detailed aerodynamic design of the fan was predicated on two crucial decisions. The first was to set the tip Mach number of the fan at 1.25. Klapproth’s 1400 ft/sec tip-speed design had shown that the losses in appropriately designed blading correlated continuously with those in conventional blading up to a Mach number of 1.35. Wright’s 1260 ft/sec transonic rotor, which had a design tip Mach number of 1.25, had been predicated on the 90 percent speed results of Klapproth’s design.55 In effect, based on his experience at NACA, Wright decided that losses associated with shocks would not become obtrusive so long as the tip Mach number did not exceed 1.25. His high confidence in the design, which was questioned by several of GE’s experienced compressor designers, came in large part from the safety margin he believed he had introduced in choosing the 1.25 tip Mach number.

Fan Aerodynamic Design-A New Computer Method The second crucial decision was to adopt a novel analytical design approach. A distinctive feature of both Klapproth’s 1400 ft/sec and Wright’s 1260 ft/sec NACA

rotors “was a fairly elaborate three-dimensional design system which allows both arbitrary radial and axial work distributions”56 within the blade row. Just as the annulus or flow area must contract in a high-pressure-ratio, multistage compressor, the flow area within a high-pressure-ratio blade row must contract between the leading and trailing edges far more than in conventional blade rows. Furthermore, high Mach number airfoil profiles are very sensitive to incidence angle. As a conse­quence, radial equilibrium effects, redistributing the flow radially, becomes important within blade rows in this type of stage, and not just from stage to stage as in more conventional compressors.

One of the first computer programs GE had developed after delivery of its IBM – 704 digital computer in 1955 solved the radial equilibrium problem in multistage axial compressors. The program employed the so-called streamline-curvature method, an iterative procedure for solving the inviscid flow equations. Specifically, an initial guess is made on where the streamlines lie radially in the spaces between each blade row throughout the compressor, and the flow along these streamlines is calculated; the streamlines are then relocated iteratively until the continuity equation is satisfied.57 When used in design, the work done and losses incurred across each blade row are specified as input along the streamlines, and the flow analysis results are then used to select appropriate standard airfoil profiles for the blades.58 Such an iterative approach was out of the question without digital computers, for the total number of calculations required is immense. Even with an IBM-704, the solution for a single operating point of the 17-stage J-79 compressor would take two or three hours, depending on the initial streamline location guess. The advance in analytical capability, however, justified this. The radial redistribution of the flow throughout a multistage compressor could be calculated with reasonable confidence for both design and off-design operating conditions. Streamline-curvature computer programs revolutionized the analytical design of axial compressors.59

Novak’s strong advocacy of streamline-curvature methods had been one of the chief reasons GE had developed this program in the first place. In the original program radial equilibrium was imposed only in the open spaces on either side of each blade row. Novak now proposed that GE’s streamline curvature program be specially modified to allow radial equilibrium to be imposed at select stations within blade rows. The streamlines and calculation stations for the fan are shown in Figure 11. In effect, the modified procedure “fools the IBM computer into thinking it is going through a series of stators with no swirl in the inlet of the compressor, through a series of rotors with small energy input through the rotor proper, and a series of stationary blade rows when it actually computes through the complete stator.”60 The second key decision in the design of the fan was to modify the streamline-curvature computer program and use it in designing the rotor and stator airfoils.

Specifically, the following parameters were specified as input and the requisite shape of the airfoils was inferred from the flow solution: (1) loss or entropy change distributions, both radially and along stream surfaces through the blades and annulus; (2) energy or work distribution radially and along stream surfaces; (3) blade blockage – i. e., a reduction in flow area within the blade rows; and (4) an allowance for boundary layer thickness along the casing wall. The solution determined (circumferentially average) velocities and pressures at each station along the streamlines. Blade surface velocities could then be inferred by assuming a linear cross-channel variation in static pressure; and blade contours were inferred from the (circumferentially average) relative flow angles at each station by assuming a blade thickness distribution and a distribution of the difference between air and metal angles within the blade row.

The choice dictating the values of the aforementioned parameters is based on judgment, prior test data and on a knowledge of the probable mechanical re­quirements of blade-thickness distribution, and so on…. It is to be recognized from the start that the entire procedure presupposes an iteration, with many variables, to a selfconsistent solution. Hence, each input parameter itself was considered as subject to change.61

No cascade airfoil contours had ever been designed by means of such an elaborate procedure before. The “Arbitrary Blade Contour” Program, as Wright called the modified program, gave him good design control in a design that stood well outside the established state of the art.62

Its complexity and sophistication notwithstanding, this analytical method fell far short of providing a scientifically rigorous or exact calculation of the flow in the fan. First of all, the program was solving the inviscid equations of motion, with viscous losses simply estimated and superposed numerically at calculation stations. In particular, the viscous boundary layers on the blade sur­faces were ignored, their effects represented by superposing on the inviscid flow a stipulated sequence of thermodynamic losses distributed linearly with axial distance.63

Second, the actual rotor blades and stator vanes indicated in Figure 11 were not literally included in the analysis. The streamlines shown in the figure were really axisymmetric stream surfaces in the calculation – not just between blade rows, but within them as well. The physical presence of the blades was represented by a numerically superposed blockage of the flow within the blade rows. The velocities and pressure calculated at each axial station within a blade row were accordingly treated within the analysis as if they were uniform around the circumference. The velocities at the blade surfaces were then calculated, in a subsidiary program, by stipulating the number of blades and assuming a linear variation in pressure from one blade surface to the next. The method thus replaced the actual three-dimensional geometry and flow by a highly idealized model; it did not include even a two­dimensional blade-to-blade flow solution of the sort that had been promoted by Chung-Hua Wu at NACA.64

Third, no effort was made to determine the precise locations of the shocks, much less to determine their interaction with airfoil boundary layers. The American Society of Mechanical Engineers’ paper by Wright and Novak describing the design of the fan and the method followed never mentions shocks. Yet shocks were surely present, for the design relative incident velocity was supersonic over all but a small fraction of the blade span, ranging from a Mach number of 1.25 at the tip to 0.98 at the hub. The shocks were taken into account only in the input distributions of losses and work specified within the rotor blade row; the shock structure assumed for this purpose was based on two-dimensional Schlieren photographs of the sort shown in Figure 8.

In short, what the analytical method did was to provide a highly idealized analysis of radial equilibrium effects within the blade rows. High Mach number blading is sensitive to deviations in incidence angles. The principal source of such deviations was thought to be radial migration of streamlines within the blade rows and blockage caused by the casing wall boundary layers. The method yielded blade contours in which radial equilibrium effects within the blade rows were consistent, under the assumptions of the analysis, with the computed incidence angles. The inputs assumed in the design were based on judgment and previous test data, reflecting Wright’s experience at NACA. A large number of passes through the design procedure (each requiring more than 20 minutes of IBM 704 computer time) were made, with these inputs changing, before a result emerged that was deemed adequately “selfconsistent.” The analytical method, for all its sophistication, was a tool in a design that remained essentially a product of judgment.

A central element of this judgment was to maintain the diffusion factors across the blade rows within the established limits, subject to Klapproth’s proviso (quoted earlier) that the velocity distributions within the blade rows not depart too radically from those of conventional airfoils. The computer program served to define the radial relocation of the stream surfaces within the blade rows, across which the diffusion factor was calculated, and it helped assure that the design would fall within the regime Klapproth had singled out. How much the blades designed on the basis of it differed from blades that might have been obtained, exercising the same judgment, from the computationally less intensive methods followed by Klapproth and Wright at NACA is an open question.65

Despite the arguments advanced by proponents of wood, the mere threat of war did little to stimulate renewed development of wooden airplanes by potential belligerents. Germany, the main source of renewed military tensions, showed little interest in wooden airplanes. The expansion of the German air force, begun soon after the Nazi seizure of power, was also accompanied by a huge expansion of Germany’s aluminum capacity; by 1939 Germany had surpassed the United States and become the largest aluminum producer in the world. This expansion was dictated more by National Socialist Autarkiepolitik than by projected needs of the Luftwaffe, but this vast capacity no doubt dampened German interest in developing wooden airplanes.23

The raw material situation was quite different in Britain, where serious rearmament began in 1936. The expansion of the RAF occurred simultaneously with the shift to aluminum stressed-skin construction, yet British aluminum production in 1939 amount to only 15 percent of Germany’s. British strategy was to rely on Canadian and American production to supply its needs. Already in April 1939, the British Air Ministry estimated that imports would have to supply two – thirds of British requirements; these estimates proved low.24 Although the Air Ministry appeared confident of its ability to obtain the necessary aluminum supplies, this dependence apparently made the British more willing to continue the use of wood in non-combat airplanes, primarily trainers. In the late 1930s, the RAF stepped up purchases of wooden training aircraft, and by 1943 all British-made training aircraft in production used all-wood or wooden-winged construction.25

Yet Britain’s use of wood was not confined to non-combat aircraft, due largely to the efforts of a single major British aviation firm, the De Havilland Aircraft Company. This firm designed the most famous wooden airplane of the war, the de Havilland Mosquito, a twin-engine bomber, fighter-bomber, night fighter and reconnaissance airplane. The Mosquito was conceived by the de Havilland firm shortly after the Munich crisis in 1938. Geoffrey de Havilland, the company’s founder, proposed building a fast unarmed bomber, protected only by its speed and maneuverability. The de Havilland design dispensed with the anti-aircraft guns standard for bombers at the time. Without defensive armament, claimed de Havilland, his design would fly faster than the opposing fighters. He also noted the advantages of wood for production in wartime, when it would not compete for resources with the metal-using industries. The de Havilland proposal was presented to the Air Ministry in October 1938, but the unconventional design generated little interest. After the declaration of war the following September, the de Havilland firm pressed its case for the design before the Air Ministry, and in December de Havilland received an order for the Mosquito prototype. The Mosquito first flew in November 1940, a mere 11 months after serious design work began. Performance exceeded expectations, and the Air Ministry placed large production orders for the airplane.26

Production deliveries began in July 1941. The airplane soon proved itself in combat, becoming “one of the most outstandingly successful products of the British aircraft industry during the Second World War.” The Mosquito excelled in speed, range, ceiling and maneuverability, making it useful in a variety of roles. Even before the prototype flew, De Havilland began developing reconnaissance and night-fighter variants.27 With a range of over 2000 miles, the original reconnaissance version could photograph most of Europe from bases in Britain at a height and speed that made it practically immune to enemy attack. Later modifications extended the range of the reconnaissance version to over 3500 miles.28 Studies of the Allied air offensive against Germany showed the Mosquito to be far more efficient at placing bombs on target than the large all-metal bombers that formed the backbone of the bombing campaigns. Compared to the heavy bombers, the Mosquito was cheaper to build, required a much smaller crew, and suffered a much lower loss rate, only two percent for the Mosquito compared to five percent for the heavy bombers. One British study calculated that the Mosquito required less than a quarter of the investment to deliver the same weight of bombs as the Lancaster, the main British four-engine bomber.29

The Canadians also got involved in wooden aircraft production. The Canadian case is particularly instructive because of its similarity with the United States in technology and availability of materials. During the interwar period, Canada had built up a small aircraft industry, though its design capabilities remained limited. For armaments, Canada remained largely dependent on Britain, and the Canadian armed forces followed other Commonwealth countries in standardizing on British materiel. As the British rearmed in the late 1930s, they looked to Canada as a possible source of aircraft and munitions, in addition to Canada’s traditional role as a supplier of raw materials. The Canadians, however, were loathe to finance expansion of their production capacity without guaranteed orders from Britain. In November 1938, the British finally placed a significant order with the Canadian aircraft industry for 80 Hampden bombers and 40 Hurricane fighters, accepting a 25 percent higher cost as the price for creating additional aircraft capacity.30

Canada remained a reluctant ally even after joining Britain in declaring war on Germany. Nevertheless, Canada did agree to host the British Commonwealth Air Training Plan, an ambitious program that eventually provided nearly 138,000 pilots and other air personnel for the British war effort. This program would require an estimated 5,000 training airplanes. The Canadian subsidiary of De Havilland was already producing the Tiger Moth, an elementary biplane trainer of wood construction that would be used for the training program. But Britain discouraged Canadian production of the more sophisticated training airplanes, insisting instead on supplying these types from their own production or American purchases.31

This situation changed radically with the fall of France. In the bleak summer of 1940 a British defeat seemed a very real possibility. Britain cut off shipments of aircraft and parts to Canada, and no replacements seemed likely from the U. S. for quite some time. It appeared that Canada might be forced to depend on its own resources for defense.32

One key Canadian resource was timber. In a report dated May 1940, J. H. Parkin proposed a program for developing wooden military airplanes in Canada. Parkin, director of the Aeronautical Laboratories at the National Research Council (NRC), presented strong technical arguments in favor of wood structures. Parkin also stressed Canada’s large timber resources, which included large reserves of virgin Sitka spruce. Parkin proposed that “the design and construction of military aircraft fabricated of wood should be initiated in Canada immediately.”33

These proposals helped launch a major Canadian program for producing wooden airplanes. Air Vice-Marshall E. W. Stedman, the chief technical officer in the RCAF, strongly advocated the construction of wooden airplanes. In London, the Ministry of Aircraft Production sought to discourage “inexperienced Canadian designers” from developing their own airplanes. Nevertheless, the RCAF continued to urge production of a wooden combat airplane in Canada; these efforts eventually led to Canadian production of the Mosquito.34 De Havilland Canada built a plant with a mechanized assembly line for Mosquito production; this plant reached a production rate of 85 airplanes monthly by mid-1945.35

Despite British skepticism, the Canadian government strongly supported the development of innovative wooden airplanes of Canadian design. In coordination with the RCAF, the NRC launched a substantial research program to develop molded plywood construction. In July 1940 RCAF and NRC staff traveled to the U. S. to investigate the latest techniques in wooden aircraft construction. They were especially impressed with Eugene Vidal’s process. Vidal was former Director of Civil Aeronautics at the Commerce Department and an enthusiast of the “personal” airplane. Vidal had started research on molded plywood after his unsuccessful attempt to develop a $700 all-metal airplane while at the Commerce Department.36

By the fall of 1940, Vidal had become the leading American developer of plywood molding techniques, due to Clark’s failure to secure military support for Duramold.37 The Canadian government asked Vidal’s company to build an experimental fuselage for the Anson twin-engine training plane, a British design then being built in Canada. The fuselage was a success, and in 1943 a Canadian company began manufacturing the fuselages under license to Vidal. From 1943 to 1945 over 1000 of the Vidal Ansons were built in Canada. A rugged, reliable airplane, the Vidal Ansons found wide use as civil aircraft after the war. The Vidal Ansons provided one of the largest and most successful applications of molded plywood to airplane structures during the war.38

The United States also launched a major wooden airplane program during World War II, but not until severe aluminum shortages threatened to curtail aircraft production. But unlike the British and Canadians, American support for wooden airplanes remained highly ambivalent.

American rearmament did not begin in earnest until after the German invasion of the low countries in May 1940, when President Franklin Roosevelt startled Congress with his 50,000-airplane program, which called for roughly a ten-fold increase over current production. Before FDR’s dramatic announcement, military planners had repeatedly insisted that aluminum supplies were ample to meet any emergency. Although Air Corps planners had given some attention to increasing the capacity of airplane plants, they had “virtually ignored” possible shortages of aircraft materials and accessories. Conditioned by interwar parsimony, the planners had little inkling of the numbers of airplanes that the President and armed forces would demand, especially when the U. S. became involved in a shooting war.39

American aircraft manufactures began publicly reporting serious aluminum shortages in late 1940 as production accelerated to meet British as well as American needs. In early 1941, the U. S. Office of Production Management finally acknowledged that the country was facing a serious aluminum shortage, and began restricting civilian consumption of aluminum. The federal government responded by financing a massive increase in aluminum refining capacity.40 But in the meantime the military would have to find other materials if it hoped to meet the President’s production goals.

With the onset of the aluminum shortage in early 1941, the Army rushed to expand the use of wood in non-combat airplanes. Early in the year, Wright Field began asking some of the Army’s largest suppliers to establish programs for converting aluminum airplane parts to plywood or plastics, and by the mid-1941 these programs were well under way. North American Aviation had an especially active substitution program for the AT-6, the most widely used advanced trainer in the war. Three major manufacturers were developing all-wood bombing trainers for the Army; two of them, Beech and Fairchild, received production contracts before the end of the year. The Air Corps also accelerated orders for its wood primary trainers already in production; by August 1941 Fairchild was building four PT-19 trainers a day. Cessna began building a twin-engine trainer for the Army based on its commercial light transport. Wooden airplanes appeared poised to play a major role in American mobilization.41

Plans for wooden airplanes grew even more ambitious after Pearl Harbor. By March 1942, Wright Field had plans to order some 16,000 wooden airplanes,

28.0 wooden propellers and 3,000 wooden gliders. Wright Field staff estimated that the substitution program would save some 45 million pounds of aluminum in the production of existing airplanes, largely by using plywood for non-structural parts. The Army Air Forces also ordered 400 Fairchild AT-13s, a new all-wood crew trainer, a fivefold increase over the original order. In an even more ambitious project, the Army launched plans for quantity production of a large wooden transport to be developed by Curtiss-Wright, the C-76. By December 1942 Curtiss-Wright had received orders for 2600 C-76 airplanes at a total estimated cost of over $400 million, including the construction of two huge new factories.42

At first glance, the American wooden airplane program appears almost as successful as those of Britain and Canada. The Army and Navy purchased some

27.0 airplanes that used wood for a significant part of the structure, along with nearly 16,000 gliders built largely of wood. These figures imply that wood airplanes made a significant contribution to the U. S. war effort, amounting to some 9 percent of the 300,000 airplanes produced for the military from July 1, 1940 to Aug. 30, 1945 43 But a closer look reveals this contribution to be less than it seems. With one exception, none of these wooden types were for combat, and the one combat airplane never entered production. The vast majority were relatively light-weight, low-performance training airplanes, mostly based on designs from the 1930s that did not take advantage of synthetic adhesives or molding techniques. In terms of airframe weight, a more reliable index of manufacturing effort, wooden airplanes accounted for only about 2.5 percent of the total44 Furthermore, most of the models produced in quantity used wood for just a small part of the total structure, such as the wing spars.

When it came to developing new designs, the American wooden airplane program was almost a complete failure. Problems occurred in design, production and maintenance. One of the most disastrous design failures was the Curtiss-Wright C – 76, a large twin-engine transport designed to carry a 4500-pound payload. Curtiss-Wright was one of the Army’s leading suppliers, but it had no recent experience designing wooden airplanes. The project began in March 1942 with an order for 200 airplanes; the Army ordered an additional 2400 before the first prototype was completed. When the C-76 prototype was finished a mere 11 months later, it proved overweight, under strength, and difficult to control in flight. In June 1943, repeated failures in static tests led Wright Field to reduce the permissible gross weight to 26,500 pounds pending successful strengthening of the structure, leaving the airplane with a pitiful payload of 549 pounds. The project became the subject of a Congressional investigation, and in July Gen. H. H. Arnold canceled the project at a loss to the Army of $40 million 45

By the summer of 1943, aluminum had become plentiful in the United States, and Wright Field began canceling production of wood designs in favor of proven metal models. In September, J. B. Johnson reported to a NACA committee that the Army was “discouraging the use of wood construction” due to “disappointing results” with wood airplanes.46 The Army also cut off funding for wood research being conducted at the Forest Products Laboratory in Madison, Wisconsin.47 The momentum of some projects carried them forward for almost another year, but in time they too were canceled. Howard Hughes was able to continue working on his giant flying boat despite military opposition, since his funding came from the Defense Plant Corporation rather than the military. But Hughes was engaged in an act of technological hubris that was bound to fail, despite the tremendous technical skill that he brought to the project.48

Airframe performance ultimately is a function of forces, moments, and pressures produced as the aircraft moves through the air. The six most fundamental dimensions are lift, drag, side force, and moments around the three axes of rotation – pitch, roll, and yaw.26 These contribute to, but do not fully determine performance, handling qualities, stability, and control including stall and spin behavior. In development of an aircraft design, modifications may be made to improve these characteristics, reduce drag, improve structural loadings, or modify systems operations.27 One function of flight test is to reveal prototype design limitations or flaws so that they can be corrected prior to production.

Wind tunnels give essential information about the six fundamental performance dimensions. Their ability to reveal handling characteristics are limited to what can be learned by flying tethered powered models in the tunnel.28 Flight test is required to put the airframe through its full paces and reveal design problems and weaknesses.

Flight testing typically begins with taxi tests, followed by lift-offs to no more than half the wingspan before settling down to assess handling characteristics at crucial take-off and landing transitions where the plane is held aloft by ground effect while below stall speed. Finally take-off is attempted, leading to a succession of flights performing the panoply of test protocols.29

a. Wind Tunnels

There are three main types of wind tunnels: Low speed or subsonic, transonic, and supersonic. The latter two types essentially are modifications of basic low-speed tunnel design and techniques. Although NACA Langley inaugurated a full-sized wind tunnel in 1929,30 wind tunnels typically insert scale models of aircraft into a fast-moving streams of air and take measurements of the six fundamental dimensions and of pressures at various airframe surface locations.

Wind tunnel tests are valuable because (i) “there is absolutely no difference traceable to having the [airframe] model still and the air moving instead of vice versa”31 and (ii) the data collected can be converted into “nondimensional coefficients that have meaning with regard to the full-scale aircraft” despite the fact that the actual “forces and moments measured on a scaled model in a wind tunnel will be considerably different from those of an aircraft in flight.”32 When the equations governing a model can be put in dimensionless form – as can the Navier Stokes equation governing wind tunnels – the observations are scale invariant and can be applied to different-scaled phenomena.33

Figure 14: GLMWT wind source, a B-29 propeller connected to a 4,300 volt electric motor, achieves wind speeds of up to 235 mph (200 knots). Below, a scale model is attached to a pylon connected to a balance platform. In the foreground is an array of manometer tubes used to measure pressures at various points on the model’s airfoil surface. [GLMWT]

The balances are labeled as follows: Roll; Drag; Yaw moment; Pitch moment; Side Force. Rolling moment, being a Function of differential lift on the two wings, is calculated From the lift. [Rae and Pope 1984, Figures 4.2, 4.3.]

Giant propellers accelerate air movement directed at instrumented test models. In typical subsonic configurations, the model is attached to pylons connected to a balance turntable. Mechanical linkages connect turntable movements to separate beam balances that record each of the six fundamental performance dimensions. Each beam balance is similar to those found at a doctor’s office, although automated. In the original 1949 Glenn L. Martin Wind Tunnel (GLMWT) configuration, the weight was on a threaded screw, and a set of contacts off the wing end of the balance send signals causing the weight to move towards a limit cycle about the balance point. A Selsen servo transmitter connected to the screw counted revolutions. Its output was sent to a Selsen receiver driving a rotary switch giving digital readouts to six significant digits. If loads exceeded what the traveling weight could balance, a cam device automatically added weights to the end of the beam and adjusted the Selsen output. Digital outputs were fed to control panel displays and IBM card punch and gang printer. Measurement accuracy was one part in 100,000. Today the Selsen units have been replaced by links from the beam balances to digital readout laboratory scales. Accuracy is unchanged.

Pressure measurements are made at various places along wing and fuselage. 200 tube manometers connected by hoses to openings on the model’s surface were used at GLMWT. Suitably configured, the adjacent manometer tubes created performance “curves” that could be photographed for more detailed analysis. By the 1960s strain-gauge pressure transducers replaced the manometers. They were so expensive that one was connected to a 48 channel scanivalve. Oscillographs were experimented with, but were not very usable except for dynamic load displacement experiments because they have too high a frequency response and most of what was recorded was noise. Wind tunnel measurements are average values, and readings need to be passed through low-pass filters to eliminate noise. To filter oscillograph data required excessive data processing. Similar problems were experienced when digital tape recorders were tried. The most efficient and reliable way to obtain average values was for tunnel technicians to eyeball meter or digital outputs and record the average value. Only about 8 readings per second are required.

Data processing in 1949 consisted in draftsmen plotting data off digital read-out displays during runs and a bevy of ten women “computers.” Early attempts were made to use IBM card tabulating equipment for data processing, but proved inferior

Figure 20. Two means оГdigital data read-out of the six fundamental measurements plus wind speed, ca 1949: An IBM gang primer (right) and digital readout (left) where six rows of ten lighted numbers would give digital readouts of each of the seven measurements. Since average values were desired, one could lock – in a value for recording by use oflhc banks of push buttons below the displays. An IBM card punch (not shown) records data on IBM cards for computer processing. Detail on the left shows two draftsmen hand plotting from the digital read-out in mid-run. [GLMWTJ

to what women could do. Around 1962 IBM introduced a Mid-Atlantic regional computing center in Washington, D. C., having one IBM 650 rotating-drum computer. The GLMWT began nightly transfers of boxes of IBM cards there for processing. Most universities acquired their first computing facilities around 1961 through purchase of the IBM 1620. When the University of Maryland installed its first UNI VAC mainframe, the original 1620 was moved to the GLMWT where it served as a dedicated machine for processing tunnel data until replaced by an IBM 1800. In 1976 a HP 10000 was installed for direct, real-time data processing. Today it is augmented by five Silicon Graphics Indigo workstations with Reality Graphics.

None of these advances increased precision or accuracy of tunnel measurements. The primary benefit was greater data-processing efficiency. Ideally one wants real­time data analysis so that one can alter test protocols as things proceed. On-line

computing capabilities provided the same testing economies that telemetry provides in flight test environments. Computer processing also improved control of systematic errors in the data gathering process. For example, the six fundamental measurements are supposed to be independent of each other. But the mechanical linkages often transferred influence from another of the components onto a given beam balance. Various Rube Goldbergs involving screw mechanisms driven by motors to counteract such effects were attempted, but never were that satisfactory. Measurement of and data correction for such contaminating influences proved far more reliable.34

Unlike normal flying conditions, the wind tunnel test chamber is enclosed and “the longitudinal static pressure gradient usually present in the test section as well as the open or closed jet boundaries in most cases produce extraneous forces that must be subtracted.”35 These lateral boundary effects corrupt virtually every measurement. Corrections are made for “additional effects due to the customary

Figure 21. Wind tunnel test section calibrations typically reveal asymmetric dynamic pressure distributions with variations above acceptable limits. Even more troublesome are problems of excessive angular variation. These are minimized by adjustment of guide vanes, screens, and propeller pitch. Remaining affects are subtracted from the data using calibration curves. [Pope 1947, p. 83.]

failings of actual wind tunnels – angularity, velocity variations, and turbulence” which are known from calibration of the tunnel itself.36

Wind tunnels such as GLMWT cannot achieve transonic or supersonic wind speeds. Early means for achieving such speeds included attaching scale models to very fast subsonic planes and diving them or dropping instrumented bodies for free – fall from very high altitudes.37 Eventually transonic (Mach 0.7-1.2), supersonic (Mach 1.2-5), and even hypersonic (Mach > 5) wind tunnels were created by a combination of beefing-up the drive system (using multiple propellers or even multistage turbine compressors) and drastically narrowing down the throat to produce a Venturi effect that briefly achieves the desired wind speeds.

The main differences between high speed and low speed tunnels are placement of the model downstream from the narrowest part of the throat; the need to change nozzle shape since a unique nozzle shape is required for each mach number; much higher magnitude of energy losses induced by tunnel walls, models, apparatus, etc.; the need to keep the air free of water vapor and small particles of foreign matter such as dust that disturb supersonic flows; and the need to mount the model on beam balances in a manner that does not use pylons and the like which cause excessive disruption of airflow near the model. This is accomplished by mounting the model from behind with a strain-gauge balance measuring device inserted in the interior of the model. The strain gauges measure torsional and other deflections of the mount.38