Having hit a dead end with the corona wind loudspeaker I went back to plasma. Luckily all the amplifier
and power supply kit was only too willing to drive anything I could throw at it. These experiments provided
some of the prettiest effects.
| 40 plasmas using brass screws. Worked, difficult to align and too much heat
produced so melts the plastics.
|| A ceramic tile with 15 plasmas running on the surface. Not long lasting
but much louder as the tile acts as a reflector.
| 40 plasmas using titanium nitride plated sewing needles. Not bad, needles melt
and wood burns!
|| A very many needle array - plastic melts and catches fire. Shortest lived and
quite time consuming to build.
| A row of plasmas on a granite tiles. Glowing orange as they slowly melt
|| A more successful test using M12 dome head nuts as electrodes.
The last of these was the start of the next large design phase. Having determined the basic requirements
for driving many plasmas at once I could tidy it up and arrange for the largest and hottest design yet.
In order to increase the effective frequency response of the plasmas you simply need more of them - low notes
just need more mass to create them effectively. Trying to drive several in parallel was always troublesome
due to the properties of plasmas - if one strikes it tends to get all the current and prevents the others striking
unless they had individual supplies. I solved this by putting the arcs in series, initially using a
screwdriver to manually strike them.
Once I found that the M12 dome nuts made a good cathode I designed a special
series electrode that effectively looked like a rivet. It had a domed end and a pointed end in one fat
mushroom shape. I tried a number of materials but good old stainless steel worked best.
To the right you can see them arranged on ceramic sticks. You can see the white plasma between each
electrode. I arranged the sticks on the crossing of several scissor tongs. This allowed me to push them
close together to strike then move them apart in one movement to create the plasma. I eventually motorised
this process on the final vertical arrangement shown left.
As far as I'm aware this a unique arrangement. 14 series plasmas, arranged vertically on a moving striking
system around 1.2m in height. The process was to simply lower the scissor so the electrodes were almost
touching then fire up the PSU. Small plasmas would strike then the motor at the top would wind the scissors
open until each plasma was 1 to 2cm in length. Too far and they went out. Too much audio bass and they
went out. Too much breeze and they went out. But when they were running it is the most impressive and
dangerous thing I think I've created. The plasmas glowed bright violet and licked and flickered around the
electrodes. Once it had been on for a few minutes the electrodes themselves glowed cherry red. This
used all the voltage and current I could muster and everything ran at its full 1.5kW capacity.
It sounded good. Even with 14 fat plasmas there was little bass but I had got the response down to 100Hz
measurable and flat from around 300Hz. Even with all that power being used it was still not loud
enough. The loudest I recorded was 101dB - but that was with noise and distortion. It would comfortably
do 85dB without distortion. It was also reasonably reliable once I had upgraded the ballast resistors.
This speaker didn't produce ozone. This is one major difference between corona wind and plasma side
effects. Corona wind speakers like to combine oxygen molecules to produce unstable ozone but there is no heat
present to do what a plasma does. A plasma runs around 6500 degrees centigrade and combine oxygen and
nitrogen quite effectively to produce various nitrogen oxides, mainly nitrogen dioxide. This high power
arrangement produced NO2 in large amounts. I didn't realise initially as NO2 doesn't smell strongly like
ozone. Once I obtained a meter the truth was revealed. A stream of hot NO2 came off the top of it and
filled the room very quickly to beyond toxic levels. Unfortunately NO2 is difficult to remove. I ran
the speaker with extraction which was less than satisfactory.
To increase sound levels I tried two things. A rear parabolic reflector and a tractrix horn.
| A rear parabola reflected the sound forward.
|| This unusually shaped tractrix horn had the plasmas running at the back.
Both methods worked. They increased the sound level and for all of the effort of the tractrix the simpler
parabola proved more effective and practical.
With an improved speaker I could start to tell a limitation of the 12kV amplifier I designed. Due to the
simple stacked arrangement the upper frequency response of the amplifier was limited and started to roll off after
10kHz. As plasma speakers are capable of much more this seemd a shame so I altered the amplifier
design. I built a pre-amplifier whose main job was to drive 12 IR LEDs. These were individually coupled
to each FET via fibre optic and a photodiode allowing each to now run with local feedback. This is difficult
to do electrically as the top FET is at a 12kV potential, but the fibre optic completely isolated each FET from the
low voltage pre amp circuit. This proved very successful and procided a lower distortion and wider bandwidth
amp capable of far better control of the plasma but still not compromising the single ended class a design.
However even with the the parabola and improved amp the sound levels were still not there. Over time
the electrodes' life was questionable and other materials didn't cure the issue. Most importantly I
couldn't cure the toxic gas issue as there was too much to filter or adsorb. I was hitting a dead end with
the limits of my design so took the only option left.