In this review, "pliotron" is the name which General Electric used for its vacuum-tubes.
General Electric Review, October, 1920, pages 807-812:

Duplex  Radiophone  Receiver  on  U.S.S.  George  Washington


    In arranging for duplex or two-way communication between the George Washington and land stations, one of the major difficulties encountered was the prevention of each receiver from being affected by the powerful interference of its own transmitter. This problem as applying to the land station at New Brunswick was solved by separating the transmitting and receiving stations a distance of four miles. The solution employed on the George Washington was of necessity totally different. A description of the schemes used, with particular reference to receiving, is given in the following article. Extracts from the ship's log are included to show the successful performance of the equipment.--E

Introduction Fig. 1: Receiving set
    On February 22, 1919, Secretary of Navy Daniels, sitting at his desk in Washington, picked up his telephone and spoke a few words of greeting to President Wilson, then 800 miles at sea on the U.S.S. George Washington.
    Secretary Daniels' voice was carried over the regular toll line from Washington to the Naval Radio Station at New Brunswick, N. J., where the voice currents were amplified to such an extent as to modulate the output of an Alexanderson alternator. This modulated energy was radiated from the New Brunswick antenna and was picked up on the George Washington.
    Previous to this demonstration, the operators on the George Washington had heard New Brunswick's radiophone while the ship was lying at anchor in Brest Harbor, France.
    The results of these tests were so encouraging that the Navy Department decided to install a powerful radiophone on the George Washington, to enable the ship to talk back to the shore. The General Electric Company was asked to furnish the radiophone equipment, both transmitting and receiving.

Requirements for Two-way Radiophone Conversation
    In order to make a two-way conversation possible over a radiophone, it is necessary either to shut off the transmitting set when receiving, or to so arrange the receiving apparatus that it will be unaffected by the powerful interference from the local transmitter.
    On small radiophone sets the first method is often used, the transmitter being started and stopped by a convenient push button located on the microphone support. More or less confusion is likely to result from this method of control, as it is impossible for the party talking to hear the other party until he releases the push button, thereby shutting off his transmitter. This method of operation is not readily adapted to remote control over a long-distance toll line.
    On high-power radiophones, particularly where an alternator is used to supply the radio frequency energy, the push button method of control is impractical, if not impossible in many cases. For these reasons, the second or duplex method for two-way conversation was chosen for both New Brunswick and the George Washington.

Duplex Arrangements at New Brunswick
    At New Brunswick, Mr. Burke Bradbury made arrangements for duplex operation by setting up the receiving apparatus at a point about four miles from the radio station, and sending the amplified received currents back to the radio station over an existing telephone line connecting New Brunswick with the Marconi receiving station at Belar.
Fig. 2: Duplex configuration

    Fig. 2 shows the connections used for duplex operation at New Brunswick. It will be noted that the received currents are introduced in series with the toll line leading from New Brunswick, enabling the party talking from Washington or any other point to hear the incoming radiophone speech over the same wires which transmit his own speech to New Brunswick, as in an ordinary land wire connection. It is also evident that, with this connection, the speech and signals picked up by the receiving apparatus at New Brunswick will modulate the alternator output and be re-radiated again at New Brunswick's wave length. Anyone listening on New Brunswick's wave length would, therefore, hear both sides of the conversation. This explains a point which puzzled many amateur operators, who reported that they heard the George Washington radiophone on 8000 or 13,600 meters, whereas the wave length of the radiophone on the George Washington was 1800 meters. The wave lengths of 8000 and 13,600 meters were both used at New Brunswick for radiophone tests at various times. The writer often heard short wave spark signals while listening to the New Brunswick radiophone on the George Washington, the signals being picked up by the receiving apparatus at New Brunswick and being re-radiated in the manner described.

Duplex Arrangements on U.S.S. George Washington
    The solution for duplex operation on the George Washington was necessarily different than for New Brunswick, as the receiving and transmitting apparatus could not be located at different points as at New Brunswick. The problem which presented itself, therefore, was to provide receiving apparatus sensitive enough to respond loudly to received currents of a few millionths of an ampere on 8000 meters, and yet be practically unresponsive to a radiation of thirty or more amperes at 1800 meters, radiating on an antenna stretched from the same masts as the receiving antenna.
Fig. 4: Antenna arrangements
Fig. 3: Duplex receiver connections
    As a solution for this problem, Mr. E. F. W. Alexanderson, Chief Engineer of the Radio Engineering Department, suggested the circuit shown in Fig. 3. This circuit was first tried out in Schenectady, using the same antenna for both receiving and sending. It was found possible to receive signals from Europe on long wave lengths and at the same time radiate ten amperes at 4000 meters on the same antenna, using either an Alexanderson alternator or a pliotron oscillator as the source of energy.
    On the George Washington, however, separate receiving and transmitting antennae were used, arranged as shown in Fig. 4.
    Fig. 3 shows the connections used in the duplex receiver on the George Washington. The current from the receiving antenna divides through two parallel branches; C1 being one branch, and FC2, the second branch. The "frequency trap" F is tuned to the transmitting wave length of 1800 meters, and offers an impedance of about 250,000 ohms at 1800 meters, but a very low impedance to long wave lengths. The condenser C1 has a capacity of 0.006 microfarads, and offers an impedance of about 160 ohms at 1800 meters. As the primary L1 of the receiving set is in series with the frequency trap, the interference from the local transmitter is reduced in the ratio of the impedances of the two branches to 1800 meters, or 160/250,000, so that the interference is reduced to 0.6 of one per cent of the interference that would be experienced without the filter circuit. The remaining interference is so small that it is easily taken care of by tuning alone.
    The variable condenser C2 tunes the primary of the receiving set to the long wave length which it is desired to receive. The inductance of the frequency trap F enters into the tuning of the primary circuit, and therefore the frequency trap offers practically no impedance to the long waves. The branch FC2L1 offers only a few ohms effective resistance to the long wave length to which it is tuned, while the branch C1 offers a comparatively high impedance, about 1200 ohms for 13,600 meters. It is, therefore, possible to receive the long waves at practically full intensity, and yet render the receiver very insensitive to the effects of the powerful radiation from the local transmitter.
    The remainder of the apparatus is the same as is used for receiving telegraph signals, excepting that it is adjusted to receive a wider band of frequencies than a telegraph receiver, in order to receive all components of the telephone wave necessary for clear speech. If the receiver is too sharply tuned, the quality of the telephone speech on long waves is very poor, because only one frequency is received strongly and other frequencies are suppressed. For long wave telephony, it is very essential to tune the receiver in such a manner that it is capable of receiving a band of frequencies within about 1000 cycles on either side of the carrier wave. On short wave lengths, 1000 cycles is a very small per cent of the carrier frequency and an ordinary sharply tuned telegraph receiver will receive telephone speech clearly. On wave lengths above 10,000 meters 1000 cycles is several per cent of the carrier frequency, and the average telegraph receiver is tuned too sharply to receive all of the frequencies necessary for clear speech, the speech sounding muffled and being very difficult to understand. With broad tuning, however, the long wave length speech may be received clearly with good quality.
    The adjustment of the duplex feature of this receiver is very simple. First, the receiving set is tuned to the wave length it is desired to receive. Then, the frequency trap F is adjusted until the interference and noise produced by the local transmitter disappear and the distant signals are heard clearly. The frequency trap adjustment is very sharp, and in the case of the George Washington the radiation from the transmitter was so powerful that the detector bulb was rendered inoperative until the frequency trap condenser was within a very few degrees of the correct position.
    Some idea of the effectiveness of the duplex feature of the receiving set may be gained from the following demonstrations: When the pliotron transmitter on the ship was in operation, radiating about 30 amp. on the main antenna at 1800 meters, the receiving antenna was sufficiently exposed to light a 40-watt lamp to full brilliancy when the lamp was placed in series with the receiving antenna. The intensity of the interfering current reaching the receiving set secondary was so small, however, that no interference whatever was experienced above 6000 meters when the pliotron transmitter was being used for continuous wave telegraphy with 150 amp. radiation. When the pliotron transmitter was modulated by voice or buzzer, the modulation could be heard weakly on 13,600 meters, and a little stronger on 8000 meters. The interference from the modulation was never strong enough to interfere with telegraph reception from New Brunswick on either 13,600 meters or 8000 meters, even in Brest Harbor, but it was strong enough to interfere slightly with the New Brunswick radiophone when the ship was several hundred miles away from New Brunswick, as the radiophone intensity was very much less than the intensity when New Brunswick was telegraphing. The interference or "side tone," however, was not loud enough or clear enough to understand; and it was necessary for intercommunication purposes to introduce an artificial side tone enabling the transmitter operator to speak to the receiving operator for changes in control, etc. For duplex telephony the side tone is not only unobjectionable, but is more or less desirable, as it more nearly approximates the conditions in the ordinary land wire telephone.
    It was noted that in very wet weather the side tone was appreciably stronger than in dry weather, probably due to leakage currents between the transmitting and receiving antennæ over the wet insulators. However, duplex operation was satisfactory in very wet weather, with a few exceptions. On one or two occasions, it was found that wet halyards swinging against the receiving antenna produced more or less inductive disturbance when the transmitter was working. During rough weather, when the ship was rolling and pitching badly, it was sometimes quite difficult to keep the receiver quiet when the transmitter was in operation. Loud inductive disturbances, synchronous with the rolling or pitching of the ship, were observed several times. In each case, the source of disturbance was found to be an antenna lead-in grounding somewhere on a metal stay or some other grounded metal object. As there were about twelve antennas on the ship, it was sometimes difficult to locate the source of disturbance immediately. It was found necessary to insulate each antenna lead-in carefully in such a manner that it could not swing against a grounded object when the ship was rolling and pitching in a storm.

Extracts from Log
    Mr. John Payne and the writer made two trips to France on the George Washington. On the first trip, the ship sailed from Hoboken on April 11th. Due to the short time available for installing the radiophone, the ship was out of range before arrangements were completed for duplex operation and no duplex conversations were tried.
    During this trip, the New Brunswick radiophone was operated at a wave length of 13,600 meters with an average antenna current of 120 amps. The New Brunswick radiophone was operated on definite schedules during the day and evening, and was received consistently up to about 1200 miles, with the exception of a few schedules when the ship was in the gulf stream, and heavy static interfered with reception. The speech was partially understood up to 2500 miles when the static conditions were favorable, and was heard but not understood at still greater distances.
    On the return trip from Brest, the New Brunswick radiophone was heard, but not understood, as soon as the ship left Brest Harbor. At a distance of 2000 miles from New Brunswick, complete sentences and orchestra selections were recognized on the ship. The orchestra music was obtained by placing a telephone near the orchestra at the New Brunswick Opera House, and also at the Hotel Klein. At a distance of 1200 miles practically all of New Brunswick's speech was understandable under normal static conditions.
    Due to very unfavorable static conditions, duplex conversations were not satisfactory until the ship was about 200 miles from New York, on the morning of May 5th. After establishing satisfactory duplex conversation with the engineers at New Brunswick and Navy officials in Washington, Secretary of War Baker on the George Washington held a conversation with Assistant Secretary of Navy Roosevelt in Washington. Secretary Baker remarked that "the connection was as good as over an ordinary toll line."
    The George Washington sailed from Hoboken again on May 10th. The first duplex conversations were held while a heavy sea was running and in a driving rain, and considerable difficulty was experienced from inductive disturbances caused by the lead-in on unused antenna; swinging against grounded objects. Later, fairly satisfactory conversations were held with New Brunswick up to a distance of about 400 miles from New York. At that distance, very bad static was experienced which made the reception unsatisfactory on both ends, particularly on the George Washington. After the ship was about 800 miles out the static conditions on the ship end were much improved, so that New Brunswick was easily understood again. New Brunswick could hear the George Washington at this distance, but could not understand the speech well enough to work duplex satisfactorily.
    While the George Washington was lying at anchor in Brest Harbor, some radiophone tests were made at New Brunswick, using a wave length of 8000 meters. During these tests, the average antenna current at New Brunswick was about 50 amp., as compared with an antenna current of 120 amp. at 13,600 meters. At night, the 8000-meter wave length radiophone was received much clearer and stronger than the 13,600-meter wave length, but very little was understood due to interference and static. However, it was possible to recognize selections sung at New Brunswick. On one occasion, Mr. W. W. Brown's voice was clearly recognized singing "America." Sometimes, at night, the 8000-meter wave with 50 amp. average antenna current was received by heterodyne note as strong or stronger than the 13,600-meter telegraph wave with an antenna current of 350 amp. In the daytime, the conditions were reversed, and the 13,600-meter wave length was received much stronger than the 8000-meter wave length. In fact, it was often impossible even to pick up the 8000-meter wave length in Brest Harbor in the daytime, although at night the same radiation was very strong.
    The George Washington sailed from Brest again on June 29th with President Wilson and party aboard. When the ship was 2400 miles from New York, New Brunswick began to make tests comparing the 8000-meter and the 13,600-meter wave lengths with 50 and 120 amp. average antenna current respectively. The first test was run at night, and the 8000-meter radiophone was understood fairly well at this distance of 2400 miles, while the 13,600-meter wave length could not be understood due to static, although the radiation in amperes was over twice as great. The next test was made on the following day, with daylight. The distance between New Brunswick and the George Washington was about 2000 miles. Both wave lengths were received about the same on this occasion. From this point on the 8000-meter wave length was received better than the 13,600-meter wave length, day or night. From a distance of 1600 miles until the ship docked, the 8000-meter radiophone was understandable at most schedules excepting when there was interference from Glace Bay.
    The first duplex conversation on this trip was held on July 7th, when the George Washington was about 375 miles from New York. The first conversation was not very satisfactory, due to static and interference at New Brunswick. Later in the day, fairly satisfactory duplex conversations were held between New Brunswick and the George Washington, but it was necessary to repeat some of the sentences several times before they could be understood. It was decided that the conditions were not quite favorable enough for President Wilson to talk, and it was hoped that the static conditions would improve later in the day. The static conditions became worse, however, so the presidential conversation was deferred until the following day. On the following morning, satisfactory communication was established, and the President was able to send a message over the radiophone to Assistant Secretary Roosevelt via radiophone to New Brunswick and thence to Washington over the toll line.

    On all trips, New Brunswick was received well up to about 300 or 400 miles. Between 300 and 500 miles from New York, the reception of New Brunswick radiophone was generally comparatively poor on every trip, due to very strong static. Beyond 600 miles from New York, the reception on the George Washington was generally good again on all trips. The static conditions seemed to be most unfavorable on the western edge of the Gulf Stream, and all signals appeared to be comparatively weak there.
    During all of these tests, the New Brunswick radiophone was received with the detector oscillating at zero beat. Non-oscillating detector reception was also tried, using a multistage radio frequency amplifier. Good reception was obtained, but it was found more practical for duplex operation to use the oscillating detector, as it was not as subject to interference from the transmitter as the radio frequency amplifier system, due to the fact that the radio frequency amplifier tended to amplify the high-frequency harmonics from the transmitter to a great extent.
    The duplex reception with the zero beat method of reception with oscillating detector was very satisfactory. Under normal conditions, the static produced a much louder sound than the side tone from the local transmitter, so that the distance that it was possible to receive New Brunswick's radiophone successfully was limited by the static rather than the side tone from the local radiophone transmitter. The master oscillator method of generating the radio frequency energy used by Mr. Payne on the George Washington radiophone produces a wave which is remarkably free from harmonics.
    When the pliotron set was being used for continuous wave or buzzer modulation telegraphy, sending commercial messages to Otter Cliffs, Maine, it was found possible to stand watch on New Brunswick without the slightest interference from the pliotron transmitter. This feature may be found useful in the future on large ships with heavy traffic, as it enables them to send and receive simultaneously. With this type of duplex, it is possible to work down to within about twice the wave length of the transmitter when the transmitter is being used for continuous wave telegraphy.
Fig. 5: Relaying messages