In 1906, Reginald Fessenden contracted with General Electric to build the first alternator transmitter. G.E. continued to perfect alternator transmitter design, and at the time of this report, the Navy was operating one of G.E.'s 200 kilowatt alternators at New Brunswick, New Jersey. However, alternator transmitters were too bulky to place on shipboard, so G.E. built a vacuum-tube transmitter for the U.S.S. George Washington to use for the two-way tests. These tests took place in April through July of 1919, and, as reviewed in the article, interspersed with the two-way tests were occasional concert broadcasts. There was even an unsuccessful attempt to broadcast an Independence Day speech by President Woodrow Wilson, who was a passenger aboard the George Washington--however, when Wilson stood too far from the microphone to be heard, the naval personnel were too intimindated to ask him to move closer.

General Electric Review, October, 1920, pages 804-806:

Radiophone  Transmitter  on  the  U.S.S.  George  Washington



      This article and the one following describe the radio equipment installed on the U.S.S. George Washington to enable President Wilson to engage in direct telephonic communication with his officials in Washington while on his homeward trip from the Peace Conference in Paris. The makeup of the transmitting apparatus is described below and interesting details of its operation are furnished.--E

    During the first part of March, 1919, the Navy Department asked the Research Laboratory of the General Electric Company to install a radio telephone transmitter on the U.S.S. George Washington, to work in connection with the New Brunswick station, so that the President would be able to get into telephonic communication with Washington while still on the high seas. It will be remembered that at that time President Wilson was in Paris, attending the Peace Conference.
    The General Electric Company had then no finished equipment that would be suitable for such purpose and it was necessary to design and build a special set for this particular work. Fig. 1 Transmitter schematic
    It was decided to build a set employing a number of large pliotrons as generators of the high-frequency current required. There was in the Laboratory at that time a panel arranged to hold twelve of these large tubes and to this was added another section containing the necessary modulating apparatus.
    These tubes required a source of 1500 to 2000 volts direct current and a separate source of 20 volts direct current. Special motor-generators designed to operate from the ship's mains and to supply these voltages, together with their control and starting panels, were hurriedly put together and the whole apparatus shipped to Hoboken by auto truck. There the apparatus was assembled and installed on the boat, and although the time was so short that it could not be tested until after leaving the dock, it performed remarkably well and gave practically no trouble during the three months it remained abroad.
    The connections of the radiophone transmitter are shown in Fig. 1. In this diagram the actual number of tubes used in each stage are not shown ; also the details of the control circuit, by which the operator of the transmitter was able to supervise conversations and to connect the set with the receiving apparatus in the receiving room and to the President's suite, are omitted for the sake of simplicity.
    The action was as follows: The microphone transmitter M, a standard telephone desk set, was used and the currents generated by the voice were stepped up in voltage by the transformer T, and amplified by the pliotron tube A1. In this set this was a single small tube having a rated output of about 50 watts. The output circuit of this tube was connected at B through a capacity to the grids of two larger tubes of 200-watt capacity each, where the voice currents were still further amplified.
    Two similar tubes, C, were connected as oscillators to generate an alternating current of about 170,000 cycles (1800 meters). The high-frequency output (in watts) of these tubes is proportional to the voltage supplied to their plates. The plates of the tubes C and B are connected to the direct-current source through the high impedance I, and therefore any change in the current flowing through the tubes B will cause a voltage to be set up across this impedance and a resulting change in the output of C. In this way the high-frequency output of C is made to correspond with the voice currents supplied by the microphone M.
    R is a bank of twelve 200-watt tubes having their grids connected in multiple and coupled inductively to the oscillating circuit of C. The plates of these tubes were also connected in multiple and inductively coupled to the antennæ. The tubes R therefore acted simply as an amplifier for the fluctuating output of C.
    With an input to the plate circuits of all the tubes of 1600 volts and between two and three amperes, an antenna current having a steady value of from 30 to 33 amperes was obtained. An oscillogram taken when speaking into the microphone showed that the current in the antenna then fluctuated from 3 to 35 or more amperes. The wave length was at all times maintained at 1800 meters.
    After the installation of the set was completed on April 12th, a great many interesting tests were made before the President boarded the ship almost three months later.
    On April 14th, we talked to the U.S.S. Frederick, at that time about 150 miles ahead of us, and they reported: "Phone loud and strong, easily understood." On April 16th, the log reads: "Before beginning the 3:00 p.m. schedule a broadcast message was sent on the George Washington's spark transmitter at 600 meters and at 952 meters asking all ships to listen for our radiophone on 1800 meters and report how they received us and giving their position." About a dozen ships sent in reports. The ship farthest away that reported was about 320 miles from us. They reported "Phone fine on crystal with Marconi type receiver." The U.S.S. President Grant, about 150 miles from us, reported hearing our radiophone 75 feet from the head phones using a four stage amplifier. Fig. 2 Ship reports
    Fig. 2 is a chart showing the names and positions of the ships which reported that they had received on the test. This was one of the first tests of this sort and from then on we often entertained the operators of other ships by phonograph concerts transmitted via our radiophone. The log of the Navy receiving station at Otter Cliffs, Maine, shows that at one time, when we were 1000 miles away, the music of one of these concerts came in so loud that the sailors there danced to the tunes we played.
    On April 17th, when we were 2184 miles from Ambrose Light, Otter Cliffs reported hearing our tests but that the speech at that time was not clear.
    Up to this time the full output of the set had never been obtained because of some trouble with the power circuit supplying our motor generator sets. In Brest this trouble was remedied and thereafter all our tests were made at about full output.
    On April 27th, we began our first return trip, Secretary of War Baker and several thousand soldiers being on board. The static most of the time was very bad and we did not get into touch with Otter Cliffs until May 4th, when they reported: "Your telegraph signals excellent ; speech at first half of schedule loud but not clear and on the last half very loud and clear." We had made an adjustment in the middle of the schedule when we had discovered that the quality of the speech was poor. Four hours later they reported that they were copying our speech on a typewriter. Later in the day a number of commercial messages were transmitted by the radiophone set. The first one read as follows:
    "From U.S.S. George Washington via Otter Cliffs, Maine, to Perkins Street, New York City. 'Expect to see you Monday night. Love. (Signed) Ted. 3:30 p.m.' " As far as the writer is aware this was the first actual paid commercial message ever transmitted by radiophone from ship to shore. Later that day Secretary Baker spoke a few sentences over the phone which were received at Otter Cliffs.
    On the following day the set was used to transmit to New Brunswick, where the speech was automatically relayed over the wires to Washington. Several persons talked over the phone on the ship to people in Washington, Secretary of War Baker talking for some time with Assistant Secretary of Navy Roosevelt, and making arrangements to meet some relatives in New York upon his arrival there.
    On the occasion of one of the concerts which we gave on this trip the radio men on the U.S.S Pastores, then 600 miles distant, connected the loud speaking telephone on the bridge of the ship with the receiver and tuned in the George Washington's wave. One of the radio men, telling the writer of the incident, said that some of the officers and men were so surprised at the loudness and clearness of the music and voices that they at first refused to believe that the sounds were actually coming from the George Washington, and it was only with difficulty that they were convinced that the radio men were not up to some trickery.
    On May 10th the George Washington again started for France and on this trip numerous tests were made. We were in communication with New Brunswick until we were some 800 miles away, when the interference and static became so heavy that we gave up and all tests were discontinued until our arrival at Brest.
    While lying in Brest harbor we tested out the set and representatives from Admiral Sims' office in London listened for us there. We received the following report: "Your schedule received, signal strength ten, modulation good . . . ." London is about 300 miles from Brest and "strength ten" seems very loud.
    While we were in Brest the NC-4, enroute from Portugal to England, passed over and circled the ship. An effort was made to talk with her by the radiophone. The operator of the NC-4 in conversation with the writer later said that he heard our signals but that they were so loud with the amplifier that he was using as to be uncomfortable and scarcely understandable. If more time had been available it is probable that good communication could have been established with the NC-4 before she landed in England.
    On June 29th, we started on our return trip, President Wilson and party being aboard. When we had approached to within 1300 miles from Ambrose Light we picked up a message from Otter Cliffs to New Brunswick saying : "For information, can hear George Washington's wireless telephone fine; can copy solid but at present Glace Bay causing interference."
    On July 4th an effort was made to transmit the President's speech to the troops by radiophone. A telephone microphone was concealed on the stand where he was scheduled to speak, but due to a misunderstanding the President spoke on a lower deck some 20 feet from the microphone. All ships had been notified to listen for the President's speech, but only an occasional word could be heard. This was very much to be regretted, as the atmospheric conditions were splendid at the time. The writer read the President's speech in the phone a few hours afterward. Colonel Carr, Department Signal Officer of the Southwestern Department of the Signal Corps, has since told the writer that he heard portions of the speech on a small antenna in San Antonio, Texas. This distance was roughly 3000 miles and almost entirely over land.
    On July 5th and 6th, the static conditions were so bad that we had difficulty in getting into good communication with either New Brunswick or Otter Cliffs, but on the 7th we got several messages through, though it was not at all satisfactory. The ship was then only about 375 miles from Ambrose Light.
    Later in the day the conditions grew worse and it was not until the following morning that really satisfactory two-way communication was obtained and the President was able to send a message over the radiophone to Secretary Roosevelt in Washington.
The Hatchet, 1919, page 210:
George Washington wireless telephone microphone
The Easy Course in Home Radio (Lesson 5), 1922, page 47:
George Washington transmitter

North Platt (Nebraska) Semi Weekly Tribune, March 14, 1919, page 2:
Daniels on shore