Background to Plasma Sound

So we wanted to make some kind of audio component, with a plasma to let you control sound with a magnet...
Turns out this works pretty well for noise in fluorescent tubes, and this noise can also be made to interact with input signals, to make some pretty wild output sounds.

This page has got technical details, prototype demos and other background stuff.

 

Plasma

A plasma is an electric flame - glowing gas energetic enough to be ionised, which makes it fluid-like and electrically conductive. It exists in many forms, from inside a neon indicator bulb to the heart of a star or lightning bolt.
The passage of current through a plasma is often unstable - current usually causes ionisation, which makes the plasma more conductive, and so allows more current to flow, etc... known as negative differential resistance, this makes the voltage required to sustain the plasma lower than the striking voltage, and is able to sustain relaxation oscillations. 

Making noise with plasma

We found passing a low DC current through a small fluorescent tube would make noise all right, with a frequency content that broadly varied with tube voltage. Interestingly, applying an increasing amount of radio-frequency (RF) energy to the plasma could then change the noise, often through a range of harsh shrieking and purer tones, until the current ultimately became stabilised - meaning the plasma's behaviour in a circuit could be varied from smoothly conductive to noisily unstable. Separately modulating the DC and RF supplies will cause a tube to make a wide range of vocal noises, that can often sound really funny - shrieks, whistles, buzzes, squeaks, farts, groans... 

The short film below shows our first dual-modulated prototype:

Different fluorescent tubes can produce quite different sounds. This video has clips of six types of tube making a range of noises:

The plasma noise is sensitive to magnetic fields and a strong magnet can affect the sound in different ways at the two electrodes. Electron paths curve in magnetic fields, and the effect is assumed to be from the extending of active electron paths.

A fluorescent tube is sealed and contains only a tiny amount of gas, so to hear really acoustic and animalistic noises was at first very surprising. In more familiar circumstances, these kinds of noise are often made as gas flows through some kind of constriction - like stretching the neck of a balloon, or squeezing air past vocal cords - and the dynamics of current flow in a plasma like this are clearly analogous enough for them to sound similar. Populations of positive ions and negative electrons determine the potential profile of a plasma, and they must be delicately balanced for it to exist at all. Both of these species have to pass this 'constriction' for current to flow, which occurs through 'sheath' regions at the electrodes - much of the sound we are using seems to be generated in the sheath region at the anode, although different audible effects are also associated with the cathode sheath.

Plasma in a synth

Glowing gas that can make weird noises is obviously pretty special stuff, but on its own doesn't really make a musical instrument. Plasma's ability to fluctuate and generate noise has been used before to make tone oscillators, in early electronic instruments like the Trautonium. These however needed various types of external circuit to keep them in tune, and we've instead experimented with applying tuned signals to the tube control circuit, which then interact with the plasma's own dynamic behaviour in different ways depending on how the tube is driven. Using digital control the machine can build a rough profile of the plasma sound from a given tube, which then lets it 'play' the tube to be in some kind of tune with the applied signal.
This arrangement is essentially a valve pre-amp that sources its current through the plasma tube, with an internal oscillator to generate the 'input' tuned signal. Strong input signals with lots of plasma noise generally make an overdriven sound, and at different conditions the output can vary to sound a bit like brass, reed or string instruments.

Tubes and electrodes

Almost all experiments so far have been done with lighting fluorescents, though we are now starting to prototype custom tubes with Julia at www.neonunity.co.uk, who will be making tubes for the synths. We still have lots to find out about what noises can be made with different types of plasma tube.  
Sounds produced by fluorescents can vary a lot, even between tubes of identical brand/spec. Many do not have a stable region within the operating range, and some go through a periodic cycle of noise/brightness. These instabilities likely occur due to small variations in the coating applied to the electrodes to lower the work function (potential barrier electrons have to overcome to leave the surface), which erodes from particle bombardment, and the distribution of the small amount of mercury in the tube. Much of the variation is reduced after a few hours' use, and a small magnet near to the anode helps to stabilise excessive noise.

 
Our own tubes will not contain mercury or have electrode coatings, and we hope they will be more stable/consistent. Until we get a bit further with them we won't know quite how this machine will be able to sound! In the meantime, we have enough fluorescents that test well to release a limited number of instruments. Details of ordering and availability can be found here.

Schematic

This is the main circuit for driving the tube and producing the audio signal. The RF generator uses ECC81 to drive two 50T coils at 6-7 MHz, regulated by one ECC82 triode. This has to float at high voltage, so the heater circuit is elevated to contrain the voltage between heater and cathode, and the RF control signal is applied using an optocoupler. The synth has additional valve circuitry for audio tone and EQ stages, and low-voltage audio out and control signal generation.

And why telmatronics?
Telmatron means something like 'swamp tube' - chosen to describe the effect of the RF plasma excitation, which does progressively swamp the audio signal. It also seems appropriate, given the wet noises some tubes make!

Links

These sites have been interesting and/or helpful for this project. Much information about plasmas and synthesisers is online, wikipedia is brilliant.

Fascinating history of electronic instruments
https://120years.net/wordpress/

Julia Bickerstaff's awesome neon creations
www.neonunity.co.uk 

Great collection of unusual instruments and resources
http://www.oddmusic.com/

Crazy synthesisers!
http://www.deviantsynth.com/

Fantastic resource for valve circuit design
http://www.valvewizard.co.uk/

Another great site for valve electronics
https://www.angelfire.com/electronic/funwithtubes/

GBVP - a project to restart valve manufacture in the UK
https://brimaruk.com/menugbvp/great-british-valve-project-2/

Synth news sites
https://www.matrixsynth.com/
https://www.synthtopia.com/
https://synthanatomy.com

 

This article was updated on August 17, 2025