Blind Landings: Low-Visibility Operations in American Aviation, 1918–1958

• 5 Instrument Landing Goes to War
• Chapter
• Johns Hopkins University Press
• pp. 104-122
• View Citation
• Additional Information

CHAPTER FIVE
Instrument Landing Goes to War

President Roosevelt’s May 1940 approval of the National Academy of Science’s report on blind landing systems had established a “standard” interim system and a research program to develop a final system. That approval allowed the army and CAA to continue supporting Edward Bowles’s microwave project while CAA began installation of the allowed ten copies of the Indianapolis system. But U.S. entry into World War II caused the aviation community to abandon its carefully negotiated agreement. The army, suddenly needing a system sooner than later, embarked on three different development programs to produce a better glide path, while also supporting an MIT Radiation Lab project to produce a radarbased system. That radar project, called ground-controlled approach, is the subject of Chapter 6.

The three glide path projects the Army Air Forces embarked on were a straight very high frequency (VHF) glide path, for which it contracted with International Telephone Development Corporation, CAA’s contractor; a ten-centimeter continuous wave (CW) microwave system based on Bowles’s work, carried out under a Signal Corps contract to Sperry Gyroscope Company; and a ten-centimeter pulsed glide path (PGP), which the MIT Radiation Lab began on its own initiative. The VHF system, called SCS-51 during the war, was used by the U.S. Army Air Forces in Europe, North America, and in the Pacific theater by transport aircraft and bombers. It was functionally identical to CAA’s postwar instrument landing system (ILS), and differed from the prewar Indianapolis system only in being portable and in using a straight, equi-signal, VHF glide path.¹ The temporary system thus became a permanent fixture of the aviation infrastructure, while both microwave systems—the wave of the future in 1940—have vanished into the past.

There is no other explanation for the entrenchment of VHF at the expense of microwaves other than the war itself. Entry into the war caused the United States to provide the AAF far greater resources, both financial and scientific, with which to pursue technological development. At the same time, those projects were constrained by immediate needs. The AAF could no longer wait for the future. It needed a landing aid immediately to avoid the huge weather-related losses that its new ally, the Royal Air Force, was already experiencing. The VHF system was the obvious choice if its glide path could be straightened quickly. At the same time, Sperry Gyroscope’s microwave project ran afoul of the MIT Radiation Lab’s belief in the superiority of its own work. In an odd paradox, the lab’s devotion to magnetrons prevented the production of a microwave glide path during the war, leaving the field to the “low-tech” VHF system.

ENTRENCHING THE LOW-TECH SOLUTION, PART I:
SCS-51 AND WORLD WAR II

In the two years following Franklin D. Roosevelt’s blessing of the National Academy of Sciences report, CAA’s Indianapolis system reached its nadir and was then suddenly, and permanently, rescued. A mistaken cost estimate nearly undid CAA’s program, while the pressing demands of a world war provided a golden opportunity for CAA to salvage its shipwrecked plan and, in a stunning reversal, forced the Army Air Forces to adopt it too. The war thus put the allegedly temporary Indianapolis system firmly on the road to permanency.

In 1940, CAA found that it had greatly underestimated the Indianapolis system’s cost. CAA had based its budget request for funding of the ten systems based on estimates made by its own experiment station personnel. They, in turn, had assumed that production units could be had at the same cost as the prototype contract, $25,000. CAA had therefore made a $250,000 request in its 1940 budget estimate, submitted to the Bureau of the Budget in late 1938. It received the full amount and found it was not nearly enough when it opened the three bids submitted by Air-Track, Bendix, and the International Telephone Development Company (ITD). Instead of ten units, it had only sufficient funds for four. Although CAA did let the contract to ITD, which was the low bidder, realization that only four sets would be installed during 1940 threw its plans into disarray. The airlines decided that four installations were too few to justify the cost of installing receivers in their planes and suddenly balked at CAA’s plans. At least six installations were necessary to make back the cost of receivers before the whole system would be discarded in favor of a microwave one in three years or so, the airlines informed one of their Army Air Forces contacts.²

CAA’s contractor also ran into engineering difficulties on top of the financial problems. The extensive modifications CAA had made to the original prototype meant that it had to be reverse engineered in order to make accurate drawings that could be used as a basis for manufacturing. The glide path transmitter had undergone the most extensive changes and required the most engineering work. That, in turn, consumed time, further delaying production and increasing costs. CAA did not get any equipment before 1941 and still had not managed to convince the airlines to buy receivers.³

CAA’s mistaken estimate threatened to kill the UHF system before it reached deployment. With the temporary Indianapolis system delayed for at least a year while the microwave system work forged ahead, the airlines believed it made little sense to invest in UHF receivers. The airlines therefore abandoned the system they had been pushing for years, causing no little feeling of betrayal at CAA. Nevertheless, CAA went ahead with its purchase of the four installations, while trying to convince the airlines to return to the fold.

While CAA was having problems with its system, events in the larger world convinced the Army Air Corps that it could not afford to wait for a microwave system to emerge from Sperry Gyroscope’s production facilities in a couple of years. The contacts forged between the Royal Air Force and Army Air Forces during 1941 brought home just how severe a problem lack of a blind landing system could be. Although Bomber Command’s statistics do not break out landing accidents, it lost hundreds of aircraft to noncombat causes during 1939–1941. Shortly before the war, the RAF had arranged a license to produce the Lorenz system in England, under the name standard beam approach (SBA) system. Like its predecessor, the SBA system was unstable, and in 1942 physicist David Langmuir reported to Lee DuBridge at the MIT Radiation Lab that RAF pilots had stopped using it. Lack of an effective landing aid coupled with the extreme density of aircraft involved in landing several hundred bombers in an area of about a hundred square miles within forty-five minutes, it seems reasonable to infer that many of Bomber Command’s noncombat losses were approach and landing accidents. Hap Arnold, who traveled to Britain in April 1941 for conferences with his RAF counterparts and an audience with the king, could not have been unaware of RAF’s problems.⁴

The army thus continued to fund Sperry’s continuous wave glide path project, while dropping the microwave localizer and marker beacons. The army intended to graft a microwave glide path onto the UHF CAA system, just as Bowles had recommended to Bush as the fastest approach to a useable system. With this “mix and match” approach to a landing aid, the ten-unit restriction was no longer necessary. With the approval of the army, therefore, RTCA recommended that the ten-unit restriction in the agreed-upon plan be dropped. This was intended to allow civil and military aviation to install and use individual elements of the approved system (i.e., localizer and marker beacons) while work continued on a variety of approaches to solving the glide path problem.⁵ One could make a low approach using only the localizer and marker beacons, after all, and although this was not a perfect solution to the blind landing problem, it was very much better than nothing. A later RTCA decision added the Army Air Corps’ compass locator stations as optional equipment; after the war, commercial pilots insisted that they be required because they made interception of the very narrow localizer beam much easier. The army’s A-1 system hardware and CAA’s system thus became completely integrated, on paper at least, by the end of 1941.

The Japanese attack on Pearl Harbor precipitated a massive installation program. With active prosecution of a strategic air war targeted for late 1942, the newly renamed Army Air Forces had to construct a ferry route up the East Coast to Newfoundland and thence across Iceland and into Scotland to get its new bombers to England. With CAA’s system the only one immediately available, the AAF officially adopted all but the glide path part in December 1941. Recognizing that CAA was the expert on its own system, the secretary of war transferred more than $1 million to CAA to procure and install the marker beacons and localizer portions of its system throughout the United States, based on a prioritized list provided by AAF.⁶ Glide path transmitters were to be provided by the army once it had developed one to its liking.

With only the glide path standing in the way of a complete system, and flush with suddenly vast development resources, the army pursued three different glide path projects in the hope that at least one might work out, and do so quickly. It continued supporting Sperry Gyroscope’s continuous wave glide path project, based on Bowles’s work, it entered into a contract with ITD for a 330 MHz equal signal glide path, and it approved a Radiation Laboratory proposal to build a pulsed glide path (PGP) based on the cavity magnetron. It also established a contract with ITD to build a completely portable and militarized version of CAA’s system (minus the glide path, of course), which the company had begun to do with its own resources earlier in the year.⁷

The Signal Corps had taken over CAA’s Indianapolis Experiment Station, its personnel, and its equipment, in early 1942 in order to more rapidly bring about an army version of the CAA’s system. It brought aboard ITD, which had changed its name to International Telephone and Radio Manufacturing Company (ITRM), to do the production engineering work on a portable system that was to be designed around standard tube sets. The army intended the use of standardized, easily available tubes to prevent having the system’s production delayed by parts shortages, which were already becoming common. It also improved the system’s maintainability by eliminating the need for specialty tubes, which would not normally be stocked in the supply system. These were important considerations in a wartime system.

The army still insisted on a straight glide path, however, and while ITRM’s engineers reworked the localizer and marker beacons for portability and standard parts, they also had to deal with the army’s demand for a long, straight glide path. They did this by raising the frequency, changing the antenna, and adapting the equal signal method used in CAA’s localizer for use in the glide path.

CAA’s glide path had operated on 93.7 MHz, while the localizer worked on 110 MHz. The two different frequencies required two transmitters, consuming more tubes and extra space. ITRM’s engineers replaced the glide path transmitter with a frequency multiplier, which raised the localizer transmitter frequency to 330 MHz (a VHF frequency) for broadcast through the glide path antenna. This saved space and tubes, but more importantly, the higher frequency reduced the environmental sensitivity of the glide path while improving the predictability of its propagation. Predictability was important because the adoption of the equal signal method allowed the glide path to be controlled to some extent. (Figure 5.1 shows the propagation pattern of ITRM’s glide path.) By broadcasting a 150-Hzmodulated signal from the upper antenna of the new glide path antenna system, and a 90-Hz-modulated signal from the lower antenna, the system created an overlapping series of lobes that could be adjusted relative to the ground by altering the relative intensity of the signals transmitted from the two antennas. A technician could thereby adjust the glide path to compensate for changing ground conditions.

The major drawback of this method is that it produced several possible glide paths, most of which were unflyable. Each of the 150-Hz lobes could appear as the correct glide path to a pilot. The possibility of intercepting the wrong glide path was reduced by procedure: the approach chart for each field directed pilots to be at a particular altitude when they reached the outer marker beacon, which corresponded to the altitude of the correct glide path at that point. (Figure 5.2 is the 1946 approach chart for Newark.) The chart’s lower section shows an elevation view of the approach, specifying that an approaching aircraft be at 800 feet when passing through the outer marker beacon at Metuchen, thirteen miles from the field. An aircraft at that altitude, and on the proper course, would intercept the correct glide path.

Figure 5.1. The SCS-51 glide path antenna radiation pattern. Only the lowest 150-cycle lobe (light gray) is a useable approach angle for most aircraft. Because the two antennae were independently excited, by adjusting the amount of energy supplied to them, an operator could adjust the angle that the 150-Hz lobes made with the ground to a limited extent. This helped compensate for changes in moisture. A monitoring device sounded an alarm if the lobe patterns deviated more than a preestablished amount from the desired pattern. M. E. Montgomery, “Latest Type AAF Blind Landing Equipment,” Electronic Industries (January 1945): 101.

The use of systematic procedures to reduce the probability of accidents was already common practice in aviation, although its application to airfield approaches had not been standardized and codified, even for individual airfields, before World War II. Chaotic conditions due to the vast increase in air traffic during 1942 forced the AAF and CAA to establish a joint board to standardize and publish the approach procedure for each major airfield in the United States so that all pilots near an airport would be following the same procedure, reducing the probability of collisions while also increasing airfield handling capacity.⁸ That standardization was also necessary for successful use of the new glide path; however, the army’s newly completed landing system did not drive the standardization process. Standardization began as a response to vastly increased traffic.

International Telephone and Radio’s glide path was placed into a competitive flyoff in late 1942 against the two microwave systems, where it performed well enough to win a Signal Corps order for 350 units, which, combined with the localizer and marker beacons, were to bear the designation SCS-51 (see Figure 5.3). Early production models were subjected to extensive testing at various locations in the United States during 1943, including locations in Alaska. There, the joint air forces/navy command found that it worked well enough to recommend permanent adoption, and largely on the strength of the navy commander’s recommendation, the navy agreed to accept SCS-51 as its own new standard later that year.⁹ The Army Air Forces established SCS-51 as its standard instrument approach in 1943 as well, despite the system’s unsuitability for small aircraft.

Figure 5.2. The standardized 1946 approach chart for Newark airport. The top section shows the approach as it appears from above, while the smaller middle section shows the approach’s vertical aspect. The bottom section provides rate of descent for several landing speeds, and the length of time a descent will last at that speed and rate of descent from the inner marker (Elizabeth, 0.9 miles from the field). The minimum ceiling at this airport was 500 feet, as the chart shows. The approach manual required that pilots unable to see the ground from 500 feet immediately climb back to 2,000 feet and request to “go around” for another try. Army Air Forces Instrument Letdown Procedures, 1 September 1946, p. 91.

The portable SCS-51 was deployed domestically first, because a low priority rating hamstrung CAA’s program to make permanent installations of similar equipment. Equipment for domestic use was automatically placed low on the priority list, while equipment intended for use in the combat theaters was given higher priority. Because SCS-51 was designated vital tactical equipment, it received one of the highest priorities. By November 1943, less than a year after its selection, SCS-51 had been installed at fields along the northeast all-weather ferry and air cargo route to Europe, from New York (Mitchell Field and Newark), through Westover Field in Massachusetts; Presque Isle, Maine; and Harmon Field, Newfoundland. The vast number of receivers needed to equip all of the AAF aircraft to use the system, however, took many months despite the high priority. High volume production of receivers took until mid-1944 to achieve, with new production aircraft, particularly the B-29s, getting most of the first batches.¹⁰ Hap Arnold intended to deploy SCS-51 to the Pacific theater with the early Twentieth Air Force B-29 units, although I have found no records attesting to specific locations. Administrative records dealing with the Pacific theater are in general far less available than those for the European theater, and to describe SCS-51’s operational record, we will have to rely on Eighth Air Force experiences with it.

Figure 5.3. An illustration of the SCS-51 localizer path. Top of the image shows the localizer truck. The correct “path” is the surface of equivalent signal—when the signal strengths of the 90-cycle and 150-cycle lobes are equal. National Air and Space Museum (NASM A-4974-A), Smithsonian Institution.

All U.S. air bases in Britain had been built with RAF equipment, and they were equipped with SBA systems. Accordingly, Eighth Bomber Command had adopted RAF’s system as its landing aid when it first arrived in Britain. Bringing in a new system made little sense, and in any case, the AAF did not have one available. Eighth Bomber Command ordered several thousand sets of SBA equipment in 1942, which were to be shipped back to the United States and installed in the aircraft at the factories. These were never built. Although it did receive small numbers of SBA receivers, the Eighth never obtained enough to equip most of its aircraft, and therefore had no blind landing system at all until late 1944.

Therefore, like the RAF, which by 1943 had begun abandoning SBA due to lack of receivers, Eighth Bomber Command did not fly missions if the ceiling for the return flight was expected to be less than 500 feet, and if forecasts turned out to be wrong, bomb groups broke up into squadrons, which descended to one hundred feet or so above the English Channel and then flew by visual landmark back to their fields. Various bomb groups worked out their own approach procedures based on their home field and aircraft equipment. Aircraft equipped with Gee, a navigation system designed primarily for bombing, could make approaches using that system, with the plane’s navigator feeding the pilot directions.¹¹ Aircraft without Gee sometimes could use compass locator stations called “slashers,” and intended as raid-forming beacons to shoot radio compass approaches to their home fields, much like the procedure for the army’s Hegenberger system. In truly blind conditions, airfields also stationed men with flare guns at either end of the landing field. The soldier at the approach end fired a green flare when he heard a plane passing overhead; if the man at the departure end heard a plane flying overhead, he released a red flare. The gun crews in the planes were expected to watch for the flares and tell the pilot whether to land or not. There were, therefore, several field solutions to the blind approach problem, although all were dangerous for novice pilots.

With the establishment of SCS-51s along the ferry route in late 1943, the equipment started to earn a reputation for itself, and Gen. Carl Spaatz, commander of U.S. air forces in Europe (USSTAF), requested that a unit be sent to the United Kingdom for trial there. The equipment arrived in late January 1944 and in early February a two-week long series of tests took place at Defford, attended by senior commanders from Eighth and Ninth Air Forces, RAF’s various commands, and the joint chiefs. Based on these tests, the joint chiefs adopted SCS-51 for use by both RAF and USAAF in the European theater immediately. Initially, thirty sets of ground equipment were required, with enough receiver sets to entirely equip the Eighth and Ninth Bomber Commands’ aircraft.¹²

The major selling point for European theater commanders was SCS-51’s easy adaptability to automatic control. At the tests at Defford, Signal Corps Lt. Col. Francis L. Moseley, formerly an engineer for Sperry Gyroscope, had demonstrated a device that converted the system’s localizer receiver output into a signal that the aircraft’s autopilot could use. Although the device was not intended to land the plane, it significantly reduced the pilot’s workload during the approach by doing most of the flying. Essentially, the pilot’s job during an approach became monitoring the system’s performance, dealing with the throttles, and taking over when the field was in sight. Spaatz liked the idea enough to demand that the experimental device be procured immediately as an integral part of the system. The demand for this “automatic coupler” was driven, in part, by the legacy of the failed SBA system, which had left pilots distrustful of the very similar SCS-51.¹³ The automated approach removed much of the mental stress that blind landings entailed, while removing many of the skill requirements (and therefore training requirements) as well. In short, it fit the Army Air Force’s wartime need for a system that its relatively inexperienced pilots could use.

Spaatz’s demand for the automatic landing coupler caused the Signal Corps to embark upon a crash program to develop versions of it compatible with the variety of automatic pilots in use, including the Honeywell C-1 and Sperry A-5, which ran the duration of the war. The device never made it to combat theaters, however, because it could not be produced in time. Similarly, the Eighth Bomber Command did not receive enough glide path receivers for all of its aircraft until January 1945, despite the receiver’s design having been completed in late 1943.¹⁴ Even high priority equipment could not be manufactured so quickly in such large volumes.

USSTAF headquarters kept up the pressure, however, complaining in September 1944 that lack of glide path receivers was preventing C-47s of Air Transport Command from keeping supplies moving to the front, and that the automatic coupler was vitally needed.¹⁵ By early 1945, enough glide path receivers had reached Europe to allow equipping all bombardment and transport aircraft with them. This left the AAF with a training problem, since its pilots had not been trained to use the system.

The AAF had attempted to establish an instrument landing training program in 1942 using sets of YB equipment borrowed from the U.S. Navy.¹⁶ Because the YB system was very similar in operation to the SBA system pilots deploying to Europe were expected to use, pilots would be able to gain familiarity with the SBA’s landing procedures while still at the army’s training schools. The idea failed in application, however, for the same reason the navy abandoned the system the following year: it was simply too unstable, especially for training use. The training schools quickly dropped the idea, and pilots deployed to Europe with no training in instrument landing.

Instrument landing training therefore had to be done in the combat theater. To accomplish that, USSTAF pressed into service the Link trainer, which could be equipped to simulate instrument landing the same way it already simulated instrument flying.¹⁷ A cross-pointer instrument in the trainer’s cockpit was manipulated remotely by the trainer operator. Pilots in the trainers then used the indications to fly the trainer with, as a pilot would if the trainer were a real aircraft. Some pilots reported that they received no training at all, while others report having been able to practice on the Link equipment between missions. Hence the opportunity to train on the Link equipment was clearly not available to all pilots.

To help ameliorate the training problem, the AAF’s training manual suggested that pilots practice on the real system as much as possible, by making their approaches “under the hood” even in good weather, with the copilot completing the landing visually. With only thirty SCS-51s in the European theater, however, most fields did not have one to practice on.¹⁸ Instead, the sites for SCS-51 installation were usually chosen so that the equipped field could serve as an emergency field for several nearby bases. That was necessary due to the dense spacing of fields in England, combined with the limited number of channels available to SCS-51. Although more equipment could have been set up, the mutual interference would have rendered the sites useless. Hence training for SCS-51 proved a severe problem for the AAF, which had thousands of crews to train in Britain.

Even without the automatic approach coupler, however, and with a limited training program, pilots who had access to SCS-51 seem to have appreciated it. One pilot remembered the glide path was the most satisfying part of the system because it eliminated the altitude uncertainty remaining in the barometric altimeter. It greatly reduced the number of “go arounds” because aircraft broke through the clouds in consistently better position for landing. The AAF had long ago given up on the idea of routine blind landings, and this pilot’s recollection is exactly what the AAF’s leaders had hoped to achieve.

World War II, therefore, rescued the Indianapolis system from an oblivion virtually guaranteed by CAA’s mistaken cost estimate and the resulting defection of the airlines. The war also established the system’s reputation by demonstrating that a different VHF glide path would work. Finally, by instigating widespread deployment of this hybrid system, the war had made it easy to justify adopting the system permanently. Vannevar Bush had warned against the dangers of deploying a system too early, thus making far more difficult its replacement by a superior system, and as we will see, his foresight had been correct.

A BRIEF INTERLUDE: DEMISE OF THE HIGH-TECH SOLUTION

The army did not stop supporting Sperry Gyroscope’s work to construct a microwave glide path when it decided to purchase International Telephone’s VHF one. Instead, it pursued both projects. Success of the VHF glide path was therefore not solely responsible for the failure of the microwave one. The MIT Radiation Lab undid Sperry’s work by challenging Sperry’s glide path with one of its own.

Sperry Gyroscope, which had long been a contractor for both the U.S. Army and U.S. Navy for navigation equipment, entered microwave work when it contracted with Stanford’s physics department to fund the development of the klystron into a commercial product. Chapter 4 details Edward Bowles’s work on a prototype forty-centimeter blind landing system at MIT, and the Signal Corps’ award of a contract to Sperry to build a complete blind landing system based on Bowles’s work. That system proved to have one major problem: size. The radiator horns needed to broadcast efficiently at forty centimeters were too big to be portable, and were also a significant collision hazard for aircraft. The obvious solution was to use a shorter wavelength, which would require smaller radiators. The Signal Corps, CAA, and Sperry chose to stop work on a forty-centimeter system, and put their development effort into a ten-centimeter system.¹⁹ They began testing that system in 1942 (see Figure 5.4).

Figure 5.4. The Sperry Gyroscope microwave glide path transmitter. Note that the original horn radiators have been replaced by what is essentially a vertical “slice” through them. With microwave transmission, the width of the beam is inversely related to the transmitting antenna’s dimensions. In this case, the antenna produced a pattern that was wide horizontally but very narrow vertically. Courtesy of Hagley Museum and Library.

Stanford’s researchers and Sperry’s engineers had always intended to use klystrons for continuous wave transmission, and the Sperry Gyroscope ten-centimeter glide path was built in that tradition. In two works examining the development of the two early forms of radio transmission, radiotelegraphy (based on interrupted wave signaling) and radiotelephony (based on continuous wave transmission), Hugh Aitken has argued that the two different forms of transmission represented two different intellectual traditions.²⁰ Practitioners of one could not easily adapt to the other, because the mental tools needed to work in one tradition were inappropriate for the other. Although Aitken relied on radiotelephony narrowly defined—we would call it radio broadcasting—continuous wave transmission with, or without, voice modulation, utilizes the same techniques. Sperry’s engineers thus worked within a tradition, or technological frame, of continuous wave transmission. When faced with pulsed transmissions, they adapted poorly.

The pulsed system, which was supposed to supersede Sperry’s work, was a product of a new organization, the MIT Radiation Laboratory. It had been founded in 1940 by Vannevar Bush’s National Defense Research Committee to pursue the development of microwave techniques for use in the war effort.²¹ The lab’s foundation had been provided by the arrival of British physicist Henry Tizard in the United States with a new device for generating microwaves: the cavity magnetron. The magnetron, like the klystron, relied upon electrically resonant cavities to produce microwaves. The internal structures of the two tubes, however, was completely different, with the result that the magnetron was better suited to providing high power output if operated intermittently. Because higher power translated directly into longer range, clearly a benefit for weapons systems, the Radiation Lab focused exclusively on designing equipment for pulsed use, and the lab’s entrant into the glide path competition was no different. Instead of broadcasting continuously, the Pulsed Glide Path (PGP) transmitted a train of pulses that aircraft could receive.

The lab decided to produce a glide path after Vannevar Bush assembled a committee in November 1941 to revisit the blind landing system progress obtained since his investigation two years earlier. Alfred Loomis, a former industrialist turned amateur physicist, chaired the committee. The group reinvestigated the blind landing work being done by CAA, army, navy, and private companies around the country. It reviewed nine systems in all: Hegenberger, Air Track, the CAA’s VHF system, the CAA-MIT microwave system, Sperry’s microwave system, PGP, GCA, the new 330-MHz glide path being made by ITRM for the army, and the Lorenz system. The committee recommended that the lab develop the PGP for production due to the rapid accumulation of pulsed techniques and the increasing availability of equipment designed for pulsed operations.²² Perhaps unsurprisingly, it thought little of the possibilities for any of the non–Radiation Lab projects, however, and other documents suggest that it thought Sperry’s system too complex to work. Hence, the lab decided to make a full-scale push toward getting the PGP adopted for production by the Signal Corps.

The PGP project had begun informally earlier in 1941, administered under the lab’s Group 73, the “Landing Group.” J. S. Buck was the project engineer. His goal was to produce a glide path operated on three centimeters to minimize the radiators’ size. The PGP, which the lab developed in partnership with General Motors’ Delco Electronics subsidiary, operated by sending a train of pulses through a horn radiator, producing a narrow beam. The receiver employed an averaging circuit so that it provided a continuous indication on a cross-pointer type instrument, making the pulses invisible to the pilot. Unfortunately, there is practically no other information available on the PGP project. I have been unable to locate either photographs or a detailed description of its operation. Fortunately, there are documents relating to flight testing and the acceptance of the two microwave systems still available, despite the lack of technical documentation.

A series of tests held in late 1942 between PGP, the Sperry continuous wave system, and the SCS-51 system from the ITRM held at Pittsburgh, Cincinnati, Indianapolis, and Wright Field, resulted in an army decision to buy both PGP and the ITRM system. The Sperry continuous wave system was to be abandoned. The reports do not make clear why PGP was chosen over Sperry’s system, as each appeared to perform equally well (or poorly, as neither put in particularly encouraging performances.) The evidence suggests, however, that the AAF had already come to rely heavily on the advice of the Radiation Lab’s physicists in its electronics procurement, and it is very clear that the Radiation Lab supported its own program over Sperry’s.²³ The Signal Corps chose to assign production of PGP to Sperry Gyroscope, a decision apparently based on an analysis that suggested Delco Electronics did not have sufficient production resources to manufacture PGP in addition to its other obligations. Sperry, in turn, never put PGP into production.

PGP did not reach production for two reasons, related directly to the company’s previous work with the klystron and Bowles’s continuous wave glide path. Sperry’s engineers were not experienced in “pulse techniques,” as the Radiation Lab’s project supervisor put it.²⁴ Sperry’s engineers were unable, or at least very unwilling, to adopt pulse techniques over the continuous wave design that they had spent several years developing. The company’s managers also preferred to leave their engineers assigned to Sperry’s own continuous wave system. The Radiation Lab’s official historian, Henry Guerlac, attributes the nonproduction of PGP to Sperry’s having lost interest in the project.²⁵ It seems fair to say, however, that Sperry was never very interested in the first place. Sperry had spent a great deal of time, energy, and money developing the klystron, upon which the continuous wave system was based and to which it also owned the rights. It could not recoup that investment if the klystron were supplanted by the cavity magnetron. The PGP was not in Sperry’s long-range financial interests, any more than it was within the realm of Sperry’s existing technical experience. Hence, Sperry’s management chose to keep its engineers at work on the klystron-based continuous wave glide path throughout the war. This combination of lack of expertise and lack of management interest spelled the end of PGP, and ultimately of microwave glide paths.

The Radiation Lab’s intrusion into microwave glide paths thus prevented the production of a microwave glide path during the war, but Sperry continued to work its microwave system, hoping that it could convince the AAF and CAA to replace the temporary VHF system. In 1946, it launched a substantial marketing effort to get its microwave system adopted as the U.S. and international standard. Yet the substantial number of SCS-51 and fixed CAA ILS systems installed during the war, and the tens of thousands of receivers for it produced during the war, proved to be an insurmountable obstacle. To suggest how fully the war had biased the aviation community against a rapid replacement of the “temporary” system, we examine the decision of the infant Provisional International Aviation Organization (PICAO) to rely upon the SCS-51 as the basis for an international landing aids standard. That decision served as the next nail in the microwave glide path’s coffin.

ENTRENCHING THE LOW-TECH SOLUTION, PART II: PICAO

The Provisional International Civil Aviation Organization had been formed as the result of a series of conferences between Britain, the United States, Canada, and a host of newly restored governments, governments in exile, and governments of the few noncombatant states in late 1944 and early 1945.²⁶ PICAO’s function was to establish the framework for a permanent organization to regulate international civil aviation. One of the powers granted to the organization by the member states was the ability to set technical standards for navigation and communication equipment, so that aircraft flying between nations did not have to carry different equipment for each of its destinations. The delegations all recognized the financial absurdity of that situation.

The problem of landing aids was only one of a number of technical issues that PICAO needed to resolve and probably the least controversial. The real “battle between systems” at the international level took place over short- and medium-range air navigation systems and was between the British delegation, led by the famous radar physicist Robert Watson-Watt, and the American group, led by career bureaucrat Charles Stanton, who had been administrator of civil aeronautics until fired by President Roosevelt in 1944.²⁷ This fight suggests some of the reasoning that likely underlay the less controversial, and therefore less welldocumented, landing aids decision.

Watson-Watt demanded the adoption of the wartime Gee system as the future international short- and medium-range air navigation system. Gee operated via a series of ground stations laid out in a network. These stations generated timebased signals that an airborne receiver detected and decoded as a navigational grid, which could then be used by a navigator to determine the receiving aircraft’s position. It was operationally similar to the long-range Loran system but more precise. Its major advantage, according to Watson-Watt, was that it allowed aircraft to operate anywhere within the area of broadcast coverage, meaning that aircraft were not confined to specific, narrow “airways” like those provided by the four-course ranges in the United States. This, he contended, would allow higher traffic densities than an airway-based system, which he believed would be necessary to serve Europe’s dense population.²⁸

Stanton, in contrast, promoted the U.S. airways model. Instead of producing a network of lines from which to derive a position, the U.S. system of radio ranges simply produced a line of bearing. Pilots then simply flew from range to range along well-defined airways. These airways were essentially highways in the sky. Fliers could leave the airways, but outside them pilots had no effective navigational references other than landmarks on the ground. Initially, the ranges were the four-course type discussed in Chapter 3, but during World War II a new sort of omnidirectional range had been devised that provided a reasonably accurate course in all directions. This was called a visual omni-range (VOR), the visual meaning that the information was displayed to the pilot on an instrument instead of aurally. Distance to the range was to be provided by another wartime innovation, called Distance Measuring Equipment (DME). The great advantage of this VOR-based system was that it was relatively simple and inexpensive, and Stanton openly ridiculed Watson-Watt over the cost of installing and operating the hundreds of short-range Gee transmitters necessary to cover North America.²⁹ VOR’s biggest disadvantage, Watson-Watt correctly noted, was that it confined fliers to specific routes, greatly reducing the volume of airspace useable by instrument fliers.

Neither man presented a rigorous argument or substantial data to back up his claims, however, and the delegates chose the U.S. system. Most nations simply took over the airways stations that the U.S. Military Air Transport Service had installed to facilitate its worldwide wartime operations, and later expanded upon them.³⁰ It made little sense immediately after the war for nations to spend a lot of money on navigational aids when a system sufficient for their immediate, basic needs had already been installed.

A similar sort of reasoning made the adoption of the USAAF’s SCS-51 as the international standard completely uncontroversial. SCS-51 had also already been installed, again by the U.S. Military Air Transport Service, at a number of major European airfields. It made sense to adopt a system that was available immediately and relatively cheap. New installations, at $70,000 apiece, also seemed a less budget-busting solution than the other technologies presented during demonstrations held in late 1946 for PICAO’s benefit, which included the Sperry microwave system and the MIT Radiation Laboratory’s Ground Controlled Approach system, expected to cost $200,000.

Further, as the Belgian delegate pointed out, other nations did not want to be forced to buy all of their equipment from a single source, and many countries wanted to be able to manufacture it for themselves.³¹ He raised this as a condition for the acceptance of all PICAO standard equipments, and both Watson-Watt and Stanton quickly agreed. The chosen standard thus could not be proprietary or contain military secrets. It also had to be fairly easy to manufacture, which was manifestly not true for microwave equipment.

The most widely reported reason for PICAO’s selection of ILS as the international standard, however, was its ease of adaptability to automatic landing. The automatic approach coupler that General Spaatz had demanded for Eighth Air Force in 1944 was very important to the plans of PICAO’s technical committees, which sought greater automation of flying overall. Why they sought to automate landings was not made explicit, but it seems reasonable to assume that their goal was the reduction of missed approaches and landing accidents by elimination of “pilot error.” No one seems to have expected the elimination of pilots themselves, since the control of a plane’s attitude during landing, which was all the approach coupler did, was only one part of pilots’ jobs. Landing was the highest workload phase of a flight, and the delegates no doubt believed that automating part of that workload would result in safer flights and improve the all-important regularity of service. When the first president of ICAO, American aeronautical engineer Edward P. Warner, discussed the technical work being done by the organization, he focused on the need for automation to reduce workload and errors.³²

PICAO’s delegates thus had several good reasons to adopt the low-tech SCS-51 over its competition. They had the evidence of its wartime performance, they were able to fly the systems during demonstrations held in England and at Indianapolis, and they had specific goals of economy, national self-interest, and automaticity to help sway their decisions. The war, however, had provided the conditions that informed the choices of PICAO’s national delegations.

CONCLUSION

From a scenic overview of the technoscape of aviation dated 1940, the future of Bowles’s microwave work seemed assured. Microwaves were clearly the future. World War II altered that technoscape of aviation dramatically, however, by driving the Army Air Forces to adopt the CAA’s previously “inadequate” system as a “good enough” expedient to support its wartime operations. At the same time, by causing Winston Churchill to dispatch Henry Tizard to the United States with one of the United Kingdom’s most valuable possessions, the cavity magnetron, the war established the basis of the MIT Radiation Lab, and thus of the Pulsed Glide Path system. Without the pressing needs generated by the war, the AAF would not have contracted with International Telephone and Radio for the VHF glide path, which the Civil Aeronautics Administration incorporated into its postwar ILS, and the Radiation Lab would never have been founded to produce its challenger to Bowles’s work. In sum, neither challenger to the Bowles-MITSperry Gyroscope continuous wave system would have existed. Hap Arnold would have adopted Sperry’s glide path when it was ready, and the airlines, as their sudden reluctance to adopt the CAA’s VHF system in 1940 suggests, would happily have followed the AAF’s lead. Without the war, finally, the conditions under which PICAO was established and made its technological choices would also not have existed. The pressing demands of World War II, then, radically altered the outcome of the landing aids development process.

The exigencies of war did not merely reverse the fortunes of the two 1940 glide path projects, however. The MIT Radiation Lab produced yet another challenger to the CAA’s system, and this one was not merely a different way to produce a glide path. Instead, this radar-based system replaced the entire model pioneered by the National Bureau of Standards system. Instead of using radio beams to activate a cockpit instrument, the ground-controlled approach system used radar to inform operators on the ground of a plane’s position. This system thus challenged the community’s very conception of how a landing system should work, shattering the consensus that the National Bureau of Standard’s pilot-centered model was the one best way for a landing aid to function.