The TRANSMITTER W/T No.1 is also known as the STERLING transmitter, and also as the No.1 AIRCRAFT TRANSMITTER SPARK. It is a spark transmitter used by British aircraft in World War 1. It is a small wooden box with pancake coil on the front, and uses a long trailing aerial. The intended use was to report the fall of artillery shells, and to report the position of enemy submarines. It was originally used as a one way system, but later on, a receiver was also fitted to the aircraft. It was usually mounted on a tray on the side of the aircraft fuselage, and the trailing aerial spool along side it. It measures 7.5” x 6.5” x 5” and weighs 9 pounds 12 onces.

Figure: Front

“The aerial was 120 feet of trailing wire which gave it an operating range of 8-10 miles. It was stranded copper wire with a 3 pound weight on the end was wound out through an insulated gland in the floor.” [Ref 2]


At the outbreak of the Great War in August 1914, the RFC moved to France taking all their aircraft that were fitted with wireless equipment. There was Lefroy, Leslie Miller and Rouzet transmitters. They were first used in the Battle of Marne in September 1914. In March 1915, the first of the new Sterling transmitters was used in the Battle of Neuve Chapelle. The Sterling set was then improved, and designated Wireless Transmitter Type 1. In 1917 this transmitter was replaced by the Type 52. In England, the Number 9 squadron was formed, and valves were experimented with, to replace spark gap technology. Radio Telephony experiments were also undertaken at the Wireless Testing Park, at Biggin Hill in Kent, under Major C.E. Prince a Marconi engineer. This resulted in the Mark 1, which used a V24 High Frequency amplifier valve. [Ref 8]


The Sterling Spark transmitter comes in several different forms. It appears that the original 1914 version was in a square wooden box with the coil on the front face. It used a 6 volt accumulator 30 watt input, and covered the frequency range 100 to 260 meters. Louis Meulstee states that 1331 units were made by the Sterling Tel. Co. and the W.D. Factory. His Compendium 1 [Ref 3], shows a photograph of a version in a rectangular wooden box, and he calls this No. 1 Aircraft Transmitter Spark. He states that 3958 units (and variations) were made by Marconi W.T.Co.Ltd. and the W.D. Factory. My example is probably one of these, as it is called a TRANSMITTER W/T No.1, made by the W/T Factory WD SOUTHGATE, although it is in a square box, has a serial number of 4144, and tunes 90 to 301 meters. He also states that there was a No.2 model for use in hot climates, where the ebonite panels were replaced with ebonite-bushed wood.

The RNAS/RAF had a Type 52 which was similar, and used an 8 volt accumulator 40 watt input, and had a frequency range of 100 to 335 meters. The Type 52M had a frequency range of 150 to 410 meters. The Type 52A had a frequency range of 150 to 410 meters, but used a wind driven alternator which provided 150 watt input. Photograph in Compendium 1 [Ref 4]. It was available as a Two Unit Type, which had the batteries in a separate box. It was also available as a Single Unit type with batteries in the same box [Ref 1].

The set was modified in 1915 to become the W/T Trench Set 50 watt D.C. (BF set). The photograph in Compendium 1 [Ref 5], shows that it used the same spark gap, induction coil and buzzer.

The photograph in Compendium 1 [Ref 6], shows the W/T Trench Set 130 watt Wilson also used the same spark gap.

The photograph in Compendium 1 [Ref 7], shows that Unit HT W/T used a similar induction coil and buzzer.

In 1918 the Connecticut Telephone and Electric Company made a similar transmitter called the Airplane Radio Telegraph Transmitting Set Box Type BC-15A which is part of the SCR 65-A. This is very similar but with the coil on the top of the box, tunes 90 to 340 meters and has all the terminals on the panel [Ref 2]. The frequency is marked by rubber blocks on the coil turns, rather than a scale on the spokes.

Figure: BC-15A (SCR 65-A)


The design is very simple. It consists of a tuned circuit, which is excited by a spark gap. The tuned circuit is very robust, as it has a high voltage applied to it and the exciter consumes about 30 watts. The coil consists of 12 turns of edge wound copper strip, air cored and mounted on ebonite spokes. There is a tap to select the desired frequency, and the spokes have a calibrated scale on them to assist in tuning. There is another tap to select the aerial connection. An ammeter measures the aerial current, but is only used when tuning. This can be mounted on the top of the transmitter, or removed and mounted where it can be easily seen. The 0.5 amp hot wire meter can be in the aerial lead or in the earth lead, but it is shorted out when defective, or not required.

The spark gap is excited by an induction coil with a built in buzzer. The Type 52 calls this a “high speed hammer make and break” and calls the reed a “trembler blade”. The similar SCR 65-A calls this a “vibrator adjustment” and calls the moving reed a “hammer”. A 6 volt battery powers the induction coil.


It is a small wooden box with an ebonite front panel that holds the coil. The top is an ebonite panel that holds the aerial terminal and ammeter, and has a window to view the spark gap and spark. On the right hand end, is a door, which opened to reveal the spark gap and induction coil buzzer. This door is sealed with an “india rubber washer”. This would be to provide some safety, as the sparks from the buzzer and spark gap, could ignite any gas or petrol fumes. A morse key was used by the operator which simply connected the battery to the induction coil.

The terminals are a mixture of the usual brass fittings, but with aluminium nuts. This is probably to save weight, but it must have been expensive as brass was more common than aluminium.


The side door opens to reveal three adjustments. One is for the spark gap, and has a lockable adjuster on the front panel. It should be set, so that a reliable crisp spark occurs. This is when the gap is not too close, and not too wide. The other two adjustments are on the induction coil. The buzzer should be adjusted for reliable operation. It has two screws. One sets the closeness of the contact, which at one extreme adjustment has the contacts open, and prevents the buzzer operating at all. At the other extreme, the contacts are closed and the buzzer, draws current without moving. The correct adjustment is between these two positions. The other adjustment sets the hammer travel, and allows some characteristic to be added to personalize the spark. This is to help distinguish between two or more spark transmitters when operating simultaneously. All these adjustments interact.

Figure: Coil Taps


The transmitter was in good condition, but it came with no instructions or documentation. The wooden box and internal wood was in very good condition. The insides were clean, and only required some dust to be removed. The outside was cleaned. The cork gasket on the top was hard and missing in parts. The rubber gasket on the spark chamber door was hard, and the door would not close. The aerial ammeter was missing. The tuning tap wander lead was intact but very frayed, and the aerial tap wander lead was missing altogether. The spark gap was clean and showed very little pitting or burning, perhaps it had not been used much.

It was easily disassembled, as it was held together with wood screws. The Induction coil tested as being continuous with no obvious short circuits. It was powered up but would not buzz or vibrate, but just make a single spark when adjusted. The “contact screw and trembler blade” were examined, and no contacts were found, absolutely none, they were missing, just a hole in each piece where they should be. A heavy duty relay with large contacts, was dismantled, and the contacts removed. One contact was riveted on to the blade. The adjuster screw, was drilled and tapped where the old contact had been, as there was a small hole evident. The new contact had a small protruding shaft that was used to hold it on when it was riveted. This was gently threaded and then screwed into the adjuster screw. It was reassembled and tested. It buzzed properly.

Figure: Old Contacts (bottom) and relay contacts (top and middle)

Figure: Repaired Contacts and “hammer”

Figure: Induction Coil

The spark gap was cleaned and connected to an LM ERICCSON PTY. LTD. TRIMAX DIVISION IONISTAION TESTER TYPE G1B. This is not the best instrument for testing a spark gap, but it has an easily controllable high voltage. This instrument is used for non-destructive testing of insulation, or capacitors, or anything of high voltage. The voltage is adjustable from zero to 10 kV, but it is DC and not a spikey waveform like the induction coil will produce. Also the voltage is a high impedance source, so that it does not destroy the device being tested. A built in audio amplifier and a speaker give an indication when leakage occurs. The sound rises from silence, to a hiss as it leaks, then a loud “crack” and then breaks down.

Figure: Ionisation Tester

The gap was adjusted very closely. The voltage was slowly wound up to 6 kilovolts (kV) before a small blue spark occurred. The gap was widened progressively, and the breakdown voltage measured. At about 10 kV, the spark changed to a wider, bluer, noiser spark, and this appeared “healthier” (whatever that means).

GAP WIDTH (mm)       VOLTAGE (kV)
0.5                                         6
1.0                                         8
1.5                                        10
2.0                                        12

The manual talks about adjusting “…. by ear, care being taken to get a good crisp spark and a clear even note.” Adjusting an active spark, is tricky, and I suggest adjusting the gap, un-energised, then testing the spark. I attempted to adjust the gap, while it was sparking, and got a shock twice, before abandoning that method! The insulation is not adequate to protect your fingers on the front panel when using the adjusting knob.

Figure: Spark

Some connections inside were raw copper and some were plated brass strips. Other connections were heavy wires and these were frayed, the cotton insulation was worn, broken or missing, and the rubber insulation underneath was brittle or missing. The old insulation and cotton was removed and replaced, and the ends were whipped, using black twine. Two of the lugs on the battery wires were loose and the solder joints were dry. A new wander lead and a coil tap were made.

Figure: Whipped Lead

The coil was intact, but with a few wobbles in it. These were straightend so as to restore the concentric arrangement. In two places, part of the edge appeared melted. This may have been due to a loose contact of the wander lead during use.

Figure: Coil

The condenser was a sealed wooden box, with an ebonite lid. It was stamped 002 and measured near this value (0.002 uF).

Figure: Condenser

The coil and condenser were measured with the grid dip meter, and gave a fairly sharp dip. Various taps were tried and the frequency was measured.

TAP              FREQUENCY (mHz)
100                        3.0
200                        1.5
300                        1.0

The case was in good condition and need nothing but a clean.

Figure: Case


The transmitter was connected to a 6 volt DC power supply that could provide 10 amps. A large morse key was connected and the buzzer was adjusted, even tough it had been adjusted before. The spark gap was adjusted too, but the setting from previous testing proved to be the best. It gave a lovely blue to white spark, and sounded “good”. It was not very noisy (acoustically). The buzzer contacts gave a small green spark.

An ammeter was connected to a dummy aerial. Testes were tried, but the regulated power supply behaved erratically, doing a current shutdown, then coming back up to 6 volts. It sounded like a “burp ….. burp …. burp”. A friend who was watching speculated that RF might be getting into the power supply. I fitted 2 ferrite rings to the power leads, and suddenly the power supply was stable, showing 6 volts at about 8 amps. It was difficult to tunr it to a 50 ohm load.

Figure: Testing

There is a video of the spark gap working here…..

There is a video of the spectrum here…..

Figure: Spectrum normal, 0 hertz marker at left, 2 FM stations at right (100 mhz)

Figure: Spectrum (showing output)


To tune the Type 52, the aerial and earth terminals are not connected when tuning the frequency. The earth is normally connected to the aircraft frame. The aerial is normally connected to the trailing aerial. A wave meter is used to tune the primary coil. The aerial tap is then adjusted, and the coupling should be “more than 2 turns from the primary tap”, more turns for longer wavelengths. The Type 52 aerial can be up to 1100 feet long and the frequency range can be 100 to 335 meters, or 410 meters or 600 meters, depending on the model [Ref 1].

The No.1 AIRCRAFT TRANSMITTER SPARK can have an aerial 100 to 200 feet long for a frequency range of 100 to 260 meters [Ref 3].

To tune the SCR 65-A, a phantom antenna is used, so that the transmit frequency can be set with the main tap, while the aircraft is on the ground. The antenna and counterpoise terminals are not connected when tuning the frequency. The counterpoise is normally bonded to the aircraft wire stays. The antenna is normally connected to the trailing aerial. The phantom antenna is adjusted for the antenna and frequency to be used. The antenna is 150 to 250 feet long for a frequency range of 90 to 340 meters [Ref 9]. Then the antenna tap is adjusted for “maximum radiation”, which is probably measured with a wavemeter.

Figure: Instructions SCR 65-A


It is difficult to criticise a device that was built 100 years ago, as modern techniques have changed so much. However, an easier method to tune the transmit frequency would be an advantage. When tuning the 80 meter band, there is only about 2 turns of the coil used, so with the fixed condenser the L/C ratio would be poor, and the Q would be less than ideal, and probably broad. Perhaps a variable condenser would be an advantage. An internal aerial ammeter would help. A different method of coupling the aerial would be an improvement, as a tap on the same coil, is probably too tight for the aerial coupling. The aircraft contained little metal, so a short circuit from the aerial was unlikely, and crew contact and shock probably minimal. Enclosing the coil would protect it, and reduce the possibility of contact or sparks, but it seems a common mounting procedure was to attach the transmitter to the outside of the cockpit, and the aerial spool was similarly mounted. It is unknown how it was fitted to flying boats, but it was probably inside the aircraft.


This is a very unusual transmitter, and it is amazing that it needed so little restoration. It is also gratifying that it still works. It is extremely simple so that is in its favor.


[1] Royal Naval Air Service W/T Apparatus. Transmitter TYPE 52, September 1917
[2] SCR 65-A auction site photographs
[3] Compendium 1, page 97, Louis Meulstee.
[4] Compendium 1, page 98, Louis Meulstee.
[5] Compendium 1, page 21, Louis Meulstee.
[6] Compendium 1, page 26, Louis Meulstee.
[7] Compendium 1, page 94, Louis Meulstee.
[8] Aircraft Radio, a brief History, Part 1: 1899-1939. Stephen Pope. Radio Bygones No.61 October/November 1999
[9] Air Corps Communications: World War 1, Radio Age Vol. 13, No. 11, Nov 1987

Ray Robinson