Analysis of Radio Direction Finder Bearings in the Search for Amelia Earhart

(Revised August 2006)

by Bob Brandenburg TIGHAR #2286

This paper presents an analysis of radio direction finder bearings obtained by Pan American Airways (PAA) direction finder (DF) sites, at Wake Island, Midway Island, and Mokapu Point at Oahu, Hawaii, and by a temporary U. S. Coast Guard DF site at Howland Island, during the search for Amelia Earhart in July 1937. All bearings were taken on signals heard at night, on or near 3105 kHz, the Earhart night frequency.

The Ionospheric Communications Enhanced Profile Analysis and Circuit Prediction Program¹ (ICEPAC) was used to model all aspects of high frequency (HF) signal propagation for this analysis.

The Low Frequency/Medium Frequency (LFMF) model²was used to model propagation in the medium frequency AM broadcast band.

The radiation pattern of the dorsal antenna on the Earhart aircraft, NR16020 was modeled with the Numerical Electromagnetics Code version 2 (NEC2),³ using physical configuration details of the antenna obtained from Lockheed Electra 10E configuration drawings,⁴ and photographs⁵ of the Electra cabin interior.

International Telecommunications Union (ITU) Recommendation F.339-6 specifies the required SNR values for various signal types and grades of service. An amplitude-modulated double-sideband signal, such as emitted by the NR16020 transmitter, is classified by F.339-6 as an A3E emission. The lowest acceptable grade of service specified in F. 339-6 for an A3E emission is 90 percent understandability of sentences (“just usable”) for non-diversity⁶ reception in fading conditions. The F.339-6 required input SNR for this grade of service is 51 decibels (dB) in a 1 Hertz band, which yields a 6 dB audio output SNR in a receiver with a 6 kHz noise bandwidth.

However, the F.339-6 results for analog systems are based on white Gaussian noise,⁷ and Spaulding⁸ has presented results showing that a given voice understandability can be achieved with a 6 dB smaller SNR in atmospheric noise, which is impulsive in nature. Applying this reduction would change the required SNR to 45 dB in a 1 Hertz band, for a 6 dB audio output SNR. This implies that a 39 dB input SNR would produce a 0 dB audio output SNR.

This analysis assumes that the background noise at the DF sites was predominantly atmospheric, and that the 6-dB SNR reduction for voice signals applies, based on the following considerations:

• Radio noise in the 3 MHz to 30 MHz band is a combination⁹ of galactic, man-made, and atmospheric noise, the characteristics of which are discussed in ITU Recommendation pp. 372-8, “Radio Noise.”
• Galactic noise did not affect reception of signals near 3105 kHz because the F2 layer critical frequency, the lowest frequency at which galactic noise energy can penetrate the ionosphere at night, was 4.7 MHz or greater during the period of interest.
• Manmade noise, which is produced by electrical equipment, was assumed to be minimal at the DF sites. All four sites were designated as “quiet rural,” the lowest of the manmade noise levels specified in P.372-8. Howland Island had no human population other than the small shore party stationed there during the search period. Wake Island and Midway Island were PAA Clipper bases, with small resident staffs and basic accommodations for passengers. PAA flight operations at each base consisted of one eastbound flight and one westbound flight each week. The Mokapu Point site, about 12 miles northeast of Honolulu, was in a sparsely populated farming region and was shielded from Honolulu electrical noise by the Koolau Mountain Range, more about which later.
• Atmospheric noise, which is generated by lightning discharges and can travel very long distances via the ionosphere, was the dominant noise factor, as indicated by comments in the radio logs of the Coast Guard cutter Itasca, and in the post-search summary reports of the PAA DF sites. This is consistent with the fact that atmospheric noise is most intense in a latitude band extending about 20 degrees north and south of the equator, particularly in summer, as shown by the worldwide atmospheric noise contour maps in ITU-R pp. 372-8.

Receiver selectivity, the ability to reject unwanted signals, must be considered when deciding whether signals can be heard on a frequency other than the one to which the receiver is tuned.

The radio equipment at each mid-Pacific Pan Am DF site¹⁰ included an RCA type AR-60 state of the art superheterodyne communication receiver. It is assumed that the AR-60 was used to listen for Earhart signals. Although the AR-60 selectivity characteristics are not available, communication receivers of the day typically operated at a bandwidth of 6 kHz when listening for voice signals. This analysis assumes that the AR-60 selectivity was essentially the same as that of the Hammarlund SP-110 superheterodyne receiver,¹¹ which was introduced in 1936 and is considered to be representative of the state of the art at the time. The SP-110 featured user-selectable bandwidth which, at 6 kHz, had a selectivity response that attenuated the output SNR of a signal on an unwanted frequency signal, relative to the SNR of a signal on the desired frequency, by 6 dB at a 3 kHz frequency difference, 23 dB at 5 kHz, 60 dB at 8 kHz, 90 dB at12 kHz, and 103 dB at 15 kHz. Accordingly, only signal source frequencies within 15 kHz of 3105, i.e., in the range 3090 kHz to 3120 kHz, were considered in this analysis.

The characteristics of the receiving antennas at the DF sites are not known, so this analysis assumes isotropic antennas. Similarly, it is assumed that the transmitting antennas of the potential signal sources, under consideration were isotropic, with the exception of the dorsal antenna on NR16020, and the broadcast band vertical tower radiators. The broadcast band antenna gain characteristics are provided in the LFMF model.

Aircraft

3105 kHz (A3) was an air-to-ground calling frequency for U.S. civilian aircraft.¹² Ground stations responded on 278 kHz.¹³

Since all DF bearings were obtained at night, areas where U.S. civilian aircraft were known to operate were evaluated for inclusion in this analysis, on the basis of ability to support night flight operations. Those areas were: the west coast of the United States, Hawaii, the Panama Canal Zone, and Venezuela.

Airfields in the vicinities of Seattle, San Francisco, Los Angeles, and San Diego supported night flight operations, so aircraft along the entire west coast of the United States were included in the analysis.

U.S. civilian aircraft operated in Hawaii in 1937, but no details of civilian airfield ability to support night operations have been discovered thus far. However, a December 7, 1941 post attack photo¹⁴ of Wheeler Field, a U.S. Army airfield on Oahu, shows the field was an open grass surface, without paved runways. This suggests that the field was not lighted for night flying operations, and further suggests that civilian, smaller, airfields were similarly unable to support night operations. Furthermore, local aircraft transmissions on 3105 kHz would have been heard by Navy radio personnel in Honolulu listening for Earhart signals, or by the PAA DF site at Mokapu. No such signals were reported, so it is concluded that there were no night flight operations in Hawaii, and aircraft in Hawaii were ruled out as potential signal sources.

The 1939 International Telecommunications Union (ITU) list¹⁵ of frequencies showed aircraft of the Caribbean Petroleum (CP) company as operating on 3105 kHz (A1,A3). However, neither the station location nor the date of commencement of operation was listed, suggesting that the capability was planned, but not yet operational in 1939. The 1938 edition of the list has no entry for Caribbean Petroleum , which implies that CP aircraft were not using 3105 kHz in 1937. Accordingly, C P aircraft are ruled as potential signal sources for this analysis.

The history of U.S. Army air defense preparations in Latin America,¹⁶ which began in 1938, provides strong evidence that there were no night flight operations in Latin America in 1937. Pan American Airways had by 1938 become the dominant commercial air service throughout the West Indian, Central American, and South American regions, and the Army considered using the Pan Am airfields as air defense bases. However, they were too small and not equipped for night flying, and thus unsuitable for military operations. Given that the Pan Am airfields were not equipped for night flying, it is reasonable to conclude that other civilian airfields did not support night flying either. This is consistent with anecdotal information in the TIGHAR archive¹⁷ from the former Chief Pilot of Creole Aviation, who operated in Venezuela from 1935 to 1958 and flew to wildcat oil well locations, landing on runways created by clearing off vegetation with a road grader and then oiling the surface. Accordingly this analysis assumes there were no night flight operations in Latin America.

U.S. civilian aircraft operated in the Panama Canal Zone (CZ) from the 1920s, and by 1930 all commercial air services into the CZ used Army Airfields.¹⁸ France Field, near the Atlantic side of the CZ, was the principal Army airfield until the 1930s¹⁹ when its inferior landing surface, lacking a paved runway, resulted in the field being deemed unsafe for operation of the large military and commercial aircraft of the day. France Field served as Pan Am’s primary flying field until 1936,²⁰ when commercial service moved to Albrook Field, at the Pacific side of the CZ. Albrook also lacked a paved runway.²¹ A photo²² of Albrook taken on 21 August 1937 shows the field had only an open grass surface, and thus did not have a lighted runway. Construction of a paved runway was funded by Congress²³ in 1938 and was completed in April 1939. Accordingly, it is assumed for this analysis that there were no night flight operations in the CZ during July 1937. This assumption is supported by the July 15, 1937 timetable²⁴ of Panama Airways, showing 4 daily flights between France Field and Balboa – the site of Albrook Field – during daylight hours only. However, some signals heard at the DF sites occurred during daylight in the CZ, so aircraft there were considered as potential signal sources of those signals.

U.S. Army aircraft operated in the CZ, and 3100 kHz was listed²⁵ as a frequency used by Army aircraft, mobile stations, and portable stations. The U.S. Army station at Quarry Heights, at the Pacific side of the zone, was listed²⁶ as a multi-function station, serving as an aeronautical services station, a coast station, and a point-to-point communications station, on 3100 kHz. The coast station frequencies listed²⁷ for Quarry heights did not include 3100 kHz, which leaves communications with aircraft and mobile/portable stations as possibilities. Therefore, this analysis considers Quarry Heights as a potential source of signals heard at the DF sites, subject to the previously stated assumption that there were no night flight operations in the CZ.

Ships

The ITU listed²⁸ a total of 13 ships worldwide, 1 U.S. and 12 Soviet, capable of voice (A3) transmission on frequencies within 15 kHz of 3105 kHz. The U.S. ship and 11 of the Soviet ships operated on 3120 kHz. These ships were ruled out as possible DF signal sources, under the procedure described in Appendix A. The twelfth Soviet ship, the Magnitogorsk, had 3105 kHz among its assigned frequencies, but was ruled out as a possible DF signal source, under the procedure described in Appendix B.

Coast Stations

Coast stations provided two-way communications with ships at sea and also with ships on inland waterways. There were numerous coast stations,²⁹ but few were capable of A3 transmission on frequencies within 15 kHz of 3105 kHz.

Alaska.

There were 21 on Gulf of Alaska, 13 on operating on 3090 kHz and 8 on 3092.5 kHz. All were capable of A3 transmission. The King Cove station (3092.5 kHz, 100 watts), at 162.19 W, 55.04 N, had the highest power among these stations and was the closest to the DF sites, and is used in this analysis as proxy for the other Gulf of Alaska coast stations.

U.S. Pacific Coast, Hawaii, and Guam.

There were 12 coast stations in this group, 9 on the U.S. Pacific Coast, 2 on the island of Oahu in Hawaii, and 1 in Guam. All were assigned 3105 kHz as a calling frequency, and 3120 kHz as the primary working frequency. All operated in A1 (Morse code) mode, none in A3. All were ruled out as potential sources of A3 signals.

Soviet Union.

There were three A3-capable ocean coast stations: Anadyr Mys and Navarin Mys on the Soviet Pacific coast, and Billings on the Northern Maritime route. There also were 4 A3-capable coast stations on the inland waterway system, at Oust-Kiakhta, Oufa, Nijne-Angarsk, and Oulan-Oude.

Other Land Stations

Aeronautical Service.

There was an A3-capable aeronautical service station on 3088 kHz at Winslow, Arizona.

Point-to-Point Communication.

There was an A3-capable Soviet point-to-point communication station on 3090 kHz, at Voronej.

AM Band Broadcast.

There were two AM band broadcast stations operating on frequencies with harmonics near 3105 kHz. Station 3AR at Melbourne, Australia, operated on 620 kHz, at 4,500 watts. The fifth harmonic of 620 kHz is 3100 kHz. Station RW26, at Stalino, Soviet Union, operated on 776 kHz at 10,000 watts. The fourth harmonic of 776 kHz is 3104 kHz.

The output power spectrum of a transmitter designed without harmonic suppression circuitry can be estimated by assuming that the final power amplifier is designed according to the method of Terman³⁰ and Roake, and by numerically integrating the Fourier integral equations³¹ for the amplitudes of the output harmonic components.

Using this procedure, the 4th harmonic output of RW26 is estimated to be approximately 6 percent of the fundamental, or 600 watts. Similarly, the 5th harmonic of 3AR is estimated to be approximately 1.55 percent of the fundamental, or 70 watts.

It is highly doubtful that the Soviet government or the Australian government would have allowed significant harmonic radiation. RW26 was about 600 nmi from the Caspian Sea coastal station at Jilaia Kosa, which operated on 3105 (A1) at 70 watts. The 600-watt harmonic on 3104 kHz from RW26 would very likely interfere with Jilaia Kosa's operations at night. The 3,150-watt second harmonic output of 3AR on 1240 kHz would very likely interfere with station 6IX in Perth, about 1500 nmi to the west, which operated at 500 watts on that frequency.

But since no evidence of harmonic suppression in either transmitter has been found thus far, this analysis assumes unrestricted harmonic radiation. However, any consideration of either station as a plausible signal source should be tempered by the practical realization that it is highly doubtful that either government would tolerate significant harmonic radiation.

The potential signal sources are compiled in Table 1.

Each source in Table 1 was tested using the method described in Appendix A, to determine the feasibility of reception at each DF site under hypothetical ideal propagation conditions. This test provides a conservative basis for deciding whether the source can be ruled out as the potential origin of signals at a given DF site. If the source signal could not produce the required receiver input SNR under the ideal conditions of the test, then it could not do so under realistic propagation conditions. The test results are shown in Table 2.

The receiver selectivity attenuation corresponding to the source frequency is shown, together with the source distance from each DF site, and the free-space SNR resulting from the test. Recalling that the threshold input SNR, for detecting the presence of an A3 signal without distinguishing any words is 41 dB for this analysis, the results shown in Table 2 can be sorted into three categories. The SNR for signals from some sources robustly exceed the threshold at some or all DF sites; some marginally exceed or fall below the threshold; and some fall far below the threshold.

Retaining the first two categories and deleting the third yields the following table of potential sources retained for further consideration in the analysis.

During the search the Itasca deployed an experimental high-frequency direction finder on Howland Island manned³² by Coast Guard Radioman Second Class Frank Cipriani. This equipment was used on July 5 to obtain an approximate bearing on a signal near 3105 kHz.

The PAA DF System

Sandretto’s description³¹ of the PAA Adcock DF system illustrates the importance of signal strength and duration. A signal bearing was indicated by an aural null. But instead of measuring the null bearing directly, the operator observed bearings on each side of the null where the signal level was high enough for accurate measurement, over a period of 2 to 3 minutes, and averaged those bearings to obtain the null bearing. The accuracy of the bearings on each side of a null, and thus the accuracy of the average bearing, would be adversely affected if a signal was weak or of short duration.

Bearing Ambiguity

The Adcock DF system produces a figure-eight antenna beam pattern, which is rotated via a goniometer to obtain a null, or signal minimum. The antenna pattern is symmetrical about the null axis, which allows bearing ambiguity because it is possible to get two null bearings 180 degrees apart. Adcock DF system designs of the 1930s usually incorporated a vertical sense antenna to eliminate bearing ambiguity. But some remarks in the PAA DF site logs indicate the possibility of bearing ambiguity, suggesting that PAA might have simplified their system design by omitting the sense antenna, since each DF site would know the general bearing of a Clipper flight. This is consistent with the recollection of Captain Almon A. Gray, USNR (ret), who was the Assistant Communications Superintendent for the PAA Pacific division during the Clipper era. Gray said³² of the PAA DF system:“The system was bi-directional hence one had to be ever mindful of the possibility of reciprocal bearings. That was not much of a problem however as one usually knew the general location of the aircraft”.This analysis assumes the PAA DF system was bi-directional and examines the possibility of a reciprocal bearing source in each case.

Distances

The distance from Wake Island to Midway Island is approximately 1,080 nautical miles (nmi), and the distance from Midway to Honolulu is approximately 1,150 nmi. Thus a flight on the route between Wake and Honolulu would never be more than about 600 nmi from the nearest DF site. In contrast, the distances from Gardner Island to Wake, Midway, and Honolulu are 1,825nmi, 1,990 nmi, and 1,850 nmi respectively.

Signal Strength Considerations

High frequency direction finding sites take bearings on skywaves, which travel long distances via refraction from the ionosphere. Short term variations in the ionosphere can cause signal strength fading, reducing the SNR and increasing the difficulty of getting an accurate bearing. If fading occurs on an already weak signal, the SNR can drop below the reception threshold, making it impossible to obtain a bearing. On the other hand, increasing atmospheric noise can cause the SNR of an already weak signal to drop below the threshold even without fading.

The PAA DF system³³ site locations, radio equipment, DF equipment, and operating procedures clearly were chosen to ensure accurate bearings would always be available for support of PAA flights .

The 70-watt output power of transmitter³⁴ on the Pan Am Clippers was roughly comparable to the 50-watt output of the transmitter on NR16020. However, as noted earlier, the distances from Gardner Island to the PAA DF sites were about 3 times the maximum distance from a Clipper to the nearest DF site on the route between Wake Island and Honolulu. Hence, the propagation loss for signals from Gardner Island was at least 10 times greater than for signals from a Clipper to the nearest DF site, making the SNR for any signals from Gardner Island correspondingly lower. And signals from the other potential sources, which were even further away, were likely to be much weaker than those expected from Clipper aircraft. The result, not surprisingly, was that bearings on potential post-loss signals were tenuous and approximate at best. That the DF site personnel were able to get any bearings at all in such conditions is a tribute to their skill and dedication.

Each PAA DF site³⁵ used two concentrically arranged sets of Adcock antennas, covering the frequency band 200 kHz to 6,000 kHz. Adcock antennas were used³⁶ because, although they are electrically equivalent to loop antennas, they are immune to the effects of polarization shifts³⁷ in radio waves refracted by the ionosphere.

A downcoming skywave, will have both vertically and horizontally polarized components which fade³⁸ independently of each other. The presence of a horizontally polarized component in a radio wave causesnight effect³⁹ in a loop antenna by inducing voltages that do not cancel out when the plane of the loop is perpendicular to the bearing of the radio wave. This results in an inaccurate bearing, or an indistinct minimum, or both. The Adcock antenna, being insensitive to horizontally polarized waves, avoids this problem⁴⁰ .

Notwithstanding its ability to obtain usable skywave bearings, the performance of the Adcock antenna is limited⁴¹ by the fact that its ability to extract energy from a passing radio wave is the same as that of a loop antenna with a single turn of wire, and hence is quite small. Consequently, even with the Adcock antennas at the DF sites, it was difficult to obtain usable bearings on weak signals.