The Auxetophone & Other Compressed-Air Gramophones

10 Dec 2014

Yet more on modern
compressed-air loudspeakers

The Short but Loud Story of Compressed-Air Amplification

"Have you heard the auxetophone? It is to be hoped not. All Mr. Parsons' turbines will be wanted to take long-suffering humanity out of earshot of his diabolical invention"

Back to Home PageBack to The Museum

The Auxetophone Updated
The Pathe Gramophone
The Fortophon Starktonmaschine
The Stentorphone
The Gaumont system
The Fluidics Approach
165 dB SPL or your money back

Before electronic amplification became practical, the volume obtainable from gramophones was strictly limited. Here are some early efforts to do something about it.

Edison considers the use of compressed amplifiers to overcome the problem of lack of replay volume, and they apparently appear in a British patent he took out in 1878.

Two Englishmen, Horace Short and Sir Charles A Parsons (yes, the steam turbine man) introduced the compressed air amplifiers known as Auxetophones. Horace Short began the development of the idea and was granted a patent in 1898, and again in 1901. The patent rights were sold to Parsons in 1903. Parsons, who was noted for his skill as a craftsman, took on the development of the Auxetophone as a hobby when he was already financially secure from his steam turbine business, and applied it to musical instruments as well as gramophones.

Left: The dreaded Auxetophone ready to open fire on a defenceless audience.

This is a Victor machine, owned by Rene Rondeau.

The heart of the Auxetophone, housed in the "Parsons-Short soundbox" at the end of the tone arm, was the air-control valve. This was a sort of grid-iron valve consisting of a metal comb rigidly connected to the needle carrier, which opened and closed fine slots in the valve seating. The resulting pulsations of compressed air were passed to the horn.
Air pressure was provided by a 1/6 horsepower electric motor, and particles were removed from the air by what appears to have been a sort of oil-bath filter, as once used on cars. Clearly the air-control valve would have been susceptible to dirt and fluff.
One of the problems seems to have been the absence of any kind of volume control; probably some sort of air-pressure regulator would have worked. No technical details seem to be available on its distortion performance, but from the comment above it was probably pretty bad.

Note that although electric power drives the compressor, the turntable motor is still a clockwork three-spring motor wound by hand, which seems rather strange. Perhaps some sort of electrical remontoire (rewinder) for the driving spring would have worked? The small silver plate near the crank-handle is the on/off switch for the compressor.

You can now hear an Auxetophone for yourself; Rene Rondeau has put a video on Youtube.
Go to (external link) and select the Victor Auxetophone video.

The distortion, while bad, is not as awful as I had been led to believe by the various comments made on it. What is noticeable is a loud background hiss of compressed air.
The volume also does not seem as devastating as claimed, but this is hard to judge as the recording equipment used to make the video probably had some sort of automatic level control.

Horace Short was one of the brothers who founded the Short aeroplane manufacture business, which still exists in Belfast. See (external link)

I have found a brief reference that says: "1898: The compressed air Auxetophone is first used to broadcast records of operatic arias from the tops of the Blackpool Tower in England and the Eiffel Tower in France." This was one Short's first prototypes; it gives you an idea of just how loud this thing could be.

Left: One of Short's patents, this one dated 1901.

This version of the air valve appears to have a grid-type valve y hinged at the bottom that approaches and recedes from a series of slots to control the air supply, which enters at a'.
Fig 1 shows a voice-amplification system, with a microphone-type diaphragm at the right, which moves the valve via rod v.
Fig 6 appears to show a cylinder-phonograph version, with stylus m contacting the cylinder.

In 1903 Short sold the patent rights to Parsons, and that year another British patent (No 10468) was granted, following Sir Charles' improvements to the air valve, the critical part of the system. In a letter written in 1921 to Sir Ambrose Fleming (The inventor of the thermionic diode valve) Parsons says:

"I worked at this subject as a hobby in my workshop at home and tried many types of valve- double-beat, slide-valves with multiple openings, then a form of valve made of sheet metal on edge like a fireworks cracker, and lastly the "comb-valve"- much the best because it delivered a flat-faced sound wave into the trumpet and it is not liable to be impeded or struck by small particles of dirt. It is similar to Short's. I made valves of comb pitches from 1/50 of an inch for reproducing from faint phonograph records, up to some of 1/4 inch pitch for attachment to double-bass stringed instruments. The very fine ones were made of hard gold, the rest of magnalium.* The air-valve reproducer was shown at the Royal Society about 1904, on a gramophone. Professor Johnston Stoney, FRS, was much interested, suggested the name "Auxetophone", and treated the matter mathematically.
If the motion of the valve is expressed in a series of sine terms (Fourier) the the sound wave produced is the first differential, and consequently the harmonics are much increased in amplitude above the fundamental, and the tone much increased in richness. This was found to be the case when used on the gramophone or when actuated from the bridge of a stringed instrument.
It was shown soon afterwards in the Library of the Royal Institution and notices appeared in the papers, and then Short's letter reached me. Previously I had not made any patent search and was not aware of his patent. Edison's had either been cited by the Patent Office or Marks and Clerk had known of it.
It appeared that Short had played an instrument on the top of the Eiffel Tower some years before. He (Short) was at that time connected with Colonel Gouroagh of the Edison Company, but when I met him he had very little money and readily assented to sell his patent to me for 700 down, and an agreement for four years at 400 per annum. Soon afterwards, the Gramophone Company of 21 City Road bought mine and Short's rights for gramophones and phonographs for all countries for 5000. I retained the rights for musical instruments.
The valve you have (taken to pieces) was made by Short in our Shops (at Heaton) and was played on a double-bass at the Promenade Concerts at Queen's Hall all one winter about 1906.
We spent much time and money in endeavours to introduce it on violins, 'cellos, and double-bass instruments, but were virtually blocked or boycotted by the Musical Fraternity, because they found it would reduce the number of executants from one-fifth to one-tenth for the same volume of sound. I dropped the whole matter, and Short was employed on other work including experimental attempts to make diamonds. He left our service about 1914 to join his brothers making sea-planes, and he died in about 1917, leaving 70,000, having made large profits from the Short Folding Wing Sea Plane.
The limiting factor to greater magnification of sound by means of an air-valve seemed to be viscous resistance in passing through minute apertures. Experiments showed that this was very marked below 1/1000 inch aperture. Hence there results a limit to the fineness of the comb. I was never able to obtain an actual magnification of the voice by means of an air-valve, Your (Fleming's) ionic valve has solved this problem."

* An alloy of aluminium and magnesium.

This letter appears to indicate that Parsons was interested in compressed-air amplification and developed the comb-valve before he bought Short's patent, which shows a different kind of valve. Perhaps it was the patent on the general principle, rather than the details of the valve technology, which Parsons wished to acquire.
It also raises the question of whether Short's voice-amplifier actually worked; Parsons says he could not make one that worked. Note also that Parsons does not seem to appreciate that one amplified double-bass is not remotely the same thing as five basses playing in chorus. And one can hardly blame musicians for being unenthusiastic about an invention intended to put most of them out of work.

How exactly did the air-control valve work? Since sound consists of positive and negative pressure changes, you might imagine that while the machine was "quiescent" there was a steady flow of air which was increased for positive excursions and decreased for negative ones. If this was not the case- and this point currently remains obscure- then the sound waveform would have been half-wave rectified, which would explain the poor sound quality. The supply of air to both sides of the tonebox may be related to this, or it may be to do with balancing the valve to reduce the frictional forces on it.
Later versions of the air-valve introduced partial balancing so that the relatively small needle forces could better control the pressurised air. A spring-loaded piston powered by the air supply acted on a wire-spring lever attached to the valve cover.
Parsons' first experimental air-valves had combs made of boxwood, the slits being cut with a jeweller's saw. The air pressure used in early tests was 2 to 3 psi.

Despite Parsons' and Short's efforts, the subjective results were apparently still somewhat short of perfection. One reaction from an Edinburgh journal was:
"Have you heard the auxetophone? It is to be hoped not. All Mr. Parsons' turbines will be wanted to take long-suffering humanity out of earshot of his diabolical invention".

The following letter was sent on 18th May 1909 by Sir Henry Wood, who co-founded the Promenade concerts in 1895:

"Dear Mr Parsons,
I am making a new orchestral arrangement for next season's Promenade Concerts at Queen's Hall of Wagner's "Siegfried" and I particularly want to do the scene with the Dragon's voice.
On the stage, this is always sung into a very large megaphone, but in my arrangement I want to do it on a very big bass tuba. do you think it would be possible for me to utilise your auxetophone? Of course I have never heard a tuba reinforced by your splendid invention, but perhaps during June, when I shall be fairly free, you could arrange for me to bring my player down."

Unfortunately nothing came of this innovative proposal, for Sir Henry wrote in the following June:

"Very many thanks for your kind letter. I had hoped to be able to the Royal College on the day you suggested, but was prevented. I now find that a bass tuba played into an ordinary very large megaphone (which I am having made by Hawkes*) gives the desired effect. With kind regards and many thanks..."

* Who went on to become part of the famous musical company Boosey & Hawkes.

A Victor auxetophone-gramophone was demonstrated to the public in November 1906 at a Trade Exhibition at Earl's Court. Mr S H Sheard recalled that the correct operating pressure had not yet been determined:

"We used about six or seven pounds to the square inch, with the rather amusing result that those of the audience who took the front rows of chairs very quickly clapped their hands to their ears and made their way to the back of the building. It was subsequently found that a pressure of not more than two-and-a-half pounds to the square inch was adequate." Which if nothing else proves that varying the operating pressure was a viable way of altering the volume.

In 1922-23, when wireless broadcasting had become established, the long-neglected auxetophones were resurrected at Parsons' Heaton works by Mr A Q Carnegie, one of Parsons' colleagues. A gramophone-type valve was driven by a magnetic motor, producing a high-volume output at a time when the capabilities of electronic valve amplifiers were very limited. The idea was not however pursued, one reason being that the patents had expired and this would limit the profitability of future business. The loudspeaker was still in use at the Heaton works in 1933.

Parsons took out three patents in all:

No. 10,468 (1903) Improvements in Sound Reproducers or Intensifiers applicable to Phonographs, Gramophones, Telephones and the like.

No. 10,469 (1903) Improvements in and relating to musical instruments.

No. 10,892 (1904) Improvements in and relating to Reproducers or Resonators for Gramophones, Phonographs and the like.

Left: Close-up of the Victor machine pictured above, showing the tonearm, carrying the air-control valve.

One wonders about the sideforces on the needle generated by that armoured compressed-air pipe.

Note that air is fed to both sides of the valve box.

Left: A close-up of the Victor valve box, showing the air-control comb valve. The thumbscrew at the bottom is to secure the gramophone needle. The screw at the top presumably adjusts the air valve.

The valve itself consisted of a pair of combs fitted tightly together. When quiescent, the two combs overlapped their teeth to form a seal, stopping the flow of air. When the needle bar moved, it slid one comb sideways, opened the spaces between the teeth and allowed air to flow.
Air was supplied to the input side of the combs, but some of the pressurized air was also fed to the horn (or output) side. The force holding the comb valve down when the groove of the record had not moved the needlebar was supplied by the slight overpressure on the horn side.

Left: Inside the Victor.

The electric motor in the centre drives a small rotary-vane compressor, which feeds the sizable black air reservoir above, via the silver cylinder which is an oil separator.
No cutout switch to prevent over-pressurisation is visible, but it might have been housed in the plinth at the top, where the motor wires appear to go. One hopes there was at least a safety-valve. The working pressure was only 2.5 psi.

Note rubber mountings for the motor-compressor assembly.

This machine is owned by Rene Rondeau.

Left: The compressor of the Victor Auxetophone seen at the top of this page.

The compressor is at the left and the electric motor is on the right. Two nickel wick-type oilers can be seen just below the shaft bearing on the right of the compressor casing. These oilers have cotton wicks that stick up and contact the shaft; they are kept filled with light machine oil.

Photograph by kind permission of The Montana Phonograph Co.

More photographs of this Auxetophone can be seen at:

Left: The compressor of the Victor Auxetophone seen at the top of this page.

The end of the compressor has been removed to show the rotating-vane principle of operation. The brass wheel turns anticlockwise, and air is drawn in by the sliding vanes through the numerous small holes on the left of the compressor casing, and pushed out through similar holes (not visible) on the right, and into the nickel-plated vertical cylinder, which is an oil separator. The operating speed is 1100 rpm.

Photograph by kind permission of The Montana Phonograph Co.

Left: The Auxetophone was manufactured by Victor from 1906 ("the New Pneumatic Victor") until 1918.

Despite this long production run few were made. One reason was the price. 100 in 1906 was a year's wages for a clerk. The machine was not suitable for home use- but then it was not intended to be. The sound volume was reported as "extremely loud " and would have deafened the average home listener, not to mention his neighbours.

Auxetophone sales literature promoted its use in "large residences" but the real market appears to have been restricted to open-air cafes in parks and similar, despite the need for an electricity supply. Dance halls, theaters, and restaurants typically hired small or large bands, and there is not much doubt that their music sounded much better.

Left: This appears to be an early model Auxetophone, possibly a prototype.

A Subjective Evaluation of The Auxetophone:
The following listening report was kindly provided by Dan Gilmore.
"I have heard a functioning Auxetophone and the only way to describe it is to liken it to standing in front of an air-raid siren, or a firing squad (LOL!). They are loud. Probably about 120+ db at about 3 feet from the horn. I'm guessing at the output volume because it was beyond the threshold of pain at about 10 feet.

The sound quality is fairly good provided you are about 50 feet away. It sounds like the music has undergone pulse code modulation. It's a very harsh sounding phonograph unless you are at a substantial distance away from it. I would think that any frequency beyond about 2000 Hz and below about 500 Hz are stripped out by the process judging from the way it sounded. A good approximation of how an Auxetophone sounded would be like listening to the record played through a bullhorn. Very flat sounding, but loud. I can see why previous owners modified them to use an Exhibition sound box.

120 or so dB is indeed loud! If you put a dB meter (unweighted sensitivity) in front of a Victor with an Exhibition soundbox at about 3 feet from the horn, and use a loud tone steel needle, acoustic era recordings will punch out as much as 85-90 dB. Edison Diamond Disc phonographs can punch out a whopping 95-100 dB. An Auxetophone at the 'threshold of pain' packs energy like a locomotive."

I understand that a picture of the auxetophone connected to a cello during the performance of a concerto with full orchestra is shown in the C A Parsons Co. Heaton Works Journal for December 1934. (Volume 1, No 6) I have no idea what an Auxetophoned cello sounded like; we can only be confident that it was LOUD.

Left: Early model of the Auxetophone valve for gramophone use.

Left: Standard model Auxetophone with upper part removed to show the valve grating.

Left: A cello fitted with an Auxetophone.

Presumably there is a flexible connection between the air-valve on the cello and the fixed horn. Note the pressure gauge laying on the floor at bottom left. This appears to be the same horn shown connected to a gramophone in the picture above.

Note: Appleyard's biography of Parsons states positively it is a cello, but I have been told that, from the shape of the shoulders, this picture actually shows a double-bass. The horn must therefore be very large.

I received the following message from Mr Ted Bowman in December 2013:

"In the 1920's my late father constructed a loud-hailer for use at Woolsington airport (now Newcastle airport) which enabled voice communication from the ground to pilots flying overhead. This consisted of a Parsons Auxetophone in which the air valve was actuated by a "reed" earphone. At that time my father earned his living building custom-made radio receivers for wealthy Northumbrians."

On a website devoted to the London tube system, there is said: "More than one attempt has been made to encourage considerate behaviour on escalators; an attempt was made, for example, in 1920 using a kind of compressed air amplifier that took a signal from a gramophone recording to bawl out 'Stand on the Right'. I have not so far been able to find out any more about this.

Bibliography: "Charles Parsons: His Life and Work" by Rollo Appleyard. Constable & Co Ltd, 1933.

In 1903 in Germany Oscar Messter patented his new Auxtephone system which used compressed air amplifiers to feed special loudspeakers. No more details known at present.

Left: Pathe compressed-air gramophone.

This model has a hand-operated compressor. The length of the hose seems to indicate that the unlucky pump operative could be tucked away well out of sight for public-address applications.


Left: The Fortophon Starktonmaschine.

This German machine pays very close homage to the Victor design; it may have been made under licence or may have been straight plagarism. Note that the electrically-driven compressor is definitely a rotary model.

"Starkton" discs were cut with wider groove modulation to produce high volume.


Left: The Gaydon Stentorphone of 1914.

The Stentorphone was produced by H A Gaydon.

The Creed telegraph equipment company were active with the Stentorphone public address system in 1919-21. The valve power amplifier had not yet appeared and the Stentorphone allowed gramophone records to be amplified via a sound-box consisting of a comb-valve which sounds very much like Auxetophone technology. A number of Stentorphones were made and used at exhibitions, open-air concerts, and the like.

Obscure Fact: The name Stentorphone was used for an intercom system fitted to Cadillac cars around 1928, for communication with the chauffeur. How it worked is unknown- it was probably a simple voice-pipe.

In 1903 French engineer Leon Gaumont was granted patents for loudspeaker systems to go with his sound on disc talking films, which used one of Berliner's Gramophones.

In 1910 Gaumont demonstrated his Chronophone system, which synchronised sound and film, at the Gaumont Palace in Paris. The compressed-air amplifier, whiuch he called the Eglephone, was just a part of the whole system. The volume was enough for an audience of 4000. Initially the longest moving picture that could be made with synchronised sound was only 200ft, due to the limited playing time of the Gramophone record. (Projection was at 16 frames per second) Gaumont surmounted this problem by having two gramophone platters; a deft operator could switch between them to give a more or less continuous soundtrack.

Above: The Gaumont Chronophone: 1910.

Note the twin gramophones, driven from a common electric motor between them. An air hose goes to each valvebox from the control valve just under the air pressure gauge; I suspect that this control valve allowed the operator to crossfade between the two gramophones. DJ in the house!
Below this, there is a light-coloured metal manifold which connects the the gramophone output pipes to the two horns. Quite how the record arms are pivoted so they can follow the track on the disc is a bit unclear, but there appears to be some sort of ball joint where they enter the manifold.

Example in CNAM, the Conservatoire National des Arts et Metiers in Paris. Author's photograph.

Above: The Gaumont Chronophone: closeup of air valve and manifold. 1910.

For some reason that is so far quite obscure, the throat of each horn is studded with screw-heads.

Example in CNAM, the Conservatoire National des Arts et Metiers in Paris. Author's photograph.

Above: The Gaumont Chronophone: closeup of valvebox. 1910.

Compressed air enters the valvebox through the central pipe, and exits through the two pipes at each side. Note the diagonal wire and spring to reduced the otherwise enormous downward tracking force.

Example in CNAM, the Conservatoire National des Arts et Metiers in Paris. Author's photograph.

Above: The Gaumont Chronophone: rear view. 1910.

From this angle it can be seen that the wider parts of the horns are also studded with mysterious screw-heads.

To the right is the electrical control panel of the system.

Example in CNAM, the Conservatoire National des Arts et Metiers in Paris. Author's photograph.

Left: Electrically-driven air compressor for the Gaumont Chronophone

The brass label indicates that the motor ran off 110V. There is a rather dangerous-looking knife switch at the rear.

Example in CNAM, the Conservatoire National des Arts et Metiers in Paris. Author's photograph.

Gaumont was the first to suggest placing loudspeakers behind the screen, and carrying them about to follow the images on the screen! The human panpot was born.

Long after the machinery depicted here came fluidics, which is fluid control without mechanical parts, and there have been a few attempts in the past to use this technology for the direct amplification of sound. See The Fluidic Gramophone which appears to have got nowhere.
Is this an idea whose time has come? I suspect not, as reading a CD with jets of compressed air presents some interesting technological challenges...

Surprisingly, it's important. Very important, in its specialised field. We probably wouldn't have got to the moon without it. Compressed air modulated by valves is used to generate enormous sound levels in test chambers.

Left: Compressed-air noise generator.

The machine in the picture can generate a Sound Pressure Level of 165 dB from an air supply at 120 psi. The air is modulated by a servo-driven reciprocating poppet valve that sounds as though it is the direct descendant of Parson's comb-valve.

To put this in perspective, 130 dB SPL causes instant hearing damage. About the only thing in the world that generates such extreme levels as 165 dB is a rocket engine at close quarters- and that is exactly what these noise generators are used for- testing rockets and jet aircraft to make sure they can withstand the noise of their own engines.

For further details see: (external link)

Sometimes it is Rocket Science.

Research on these devices continues. To get up to date on air-valves and the like, see Experiments on A Compressed-air Loudspeaker (1988) by Glendinning, Nelson, and Elliott at Southampton University. (NB, PDF download) Their experiments used a sliding-plate valve driven by an electrodynamic vibrator, and levels of at least 150 dB SPL were achieved. The authors note that such loudspeakers have been available since the 1950's, and that units with acoustic power outputs from 4 kW to 20 kW are manufactured by Ling Electronics. Ling Electronics were bought out by Data Physics in 2009. It looks like they are still in the high intensity sound business.

Left: Cylindical compressed-air loudspeaker valve

According to Glendinning, Nelson, and Elliott, this is a typical contemporay design. There can be problems with wear of the cylindrical valve, and unbalanced axial forces due to air pressure imbalances. The bleed screw marked "bias valve" is one way of partly coping with the latter.

Left: Sliding plate compressed-air loudspeaker valve

The experimental design used by Glendinning, Nelson, and Elliott. The sliding plate is supported on aerostatic thrust bearings to eliminate friction and allow operation without lubrication at high temperatures.

Back to Home PageBack to The Museum EntranceTop of this page