High Voltage Direct Drive amplifier for Plasma loudspeaker

[thadman]

I'd like to build a simple prototype plasma tweeter in the next few weeks to demonstrate the concept.

A design that pulses a high frequency resonance (>100khz) greatly reduces the voltage demands on the amplifier and will maximize energy transfer. However, such a circuit would require complex analysis.

I believe a simple, yet elegant design would result from directly coupling the output of the amplifier to the electrodes.

For the tweeter prototype, I'd like to use a separation distance of 1cm between the electrodes and operate them in anti-phase (push-pull).

Dielectric breakdown of the air occurs at a potential of 3kV/cm. Since the electrodes are defined by a separation distance of 1cm, we will require 1.5kV on the positive electrode and -1.5kV on the negative electrode. Once this potential is reached, the air will ionize and transition from a gas (~excellent insulator) to a plasma (~excellent conductor). Resistance will be VERY low.

For the amplifier prototype, I'd like use a push-pull voltage source. If the amplifier operates in push-pull mode, we can operate the electrodes in anti-phase. If the amplifier is a voltage source, we can control the potential between the electrodes.

I'd like to build two different amplifiers and contrast their performance.


1) Class A amplifier utilizing Vacuum tubes


Goals:

-maximum linearity
-minimum 100W


Advantages:

Vacuum tubes are characterized by a linear response and the ability to provide extremely high voltages. In addition, Vacuum tubes would be able to provide a continuous waveform to the electrodes. As a result, once the plasma was formed (transient state), it would be inherently stable.


Disadvantages:

Vacuum tubes are very inefficient and quite expensive. They will require significant cooling. However, I am interested in heat transport so I do not consider this a significant disadvantage.

A digital to analog converter will have to be used. This will require additional components and will introduce an error (however small).


Notes:

A stable plasma requires a specific "holding current". As a result, a solution must be provided which maintains the "holding current" through the electrodes while not interfering with the response of the tweeter. Since an AC waveform is being passed between the electrodes, there will be a point at which current is zero. I'm not sure if this is simply a "theoretical" problem since the waveform will not spend any finite amount of time at a position since it is continuous.

I believe an AC voltage source power supply will be required. How would you approach designing the power supply?

What Vacuum tubes would you recommend?

Do Ultralinear (high fidelity) Vacuum tubes exist which can provide 1.5kV on the electrodes?


2) Class D amplifier utilizing Insulated Gate Bipolar Transistors


Goals:

-Maximum linearity
-minimum 1kW
-easily scalable to 10kW


Advantages:

IGBTs are characterized by their high efficiency, fast switching, and the ability to handle extremely high voltages and large currents. I believe IGBTs would easily be capable of providing 1.5kV to the electrodes and can switch over 100,000 times per second. Additional IGBTs could be connected in parallel to provide any desired amount of current.

IGBTs would not require a digital to analog converter. They can be switched at the discrete steps corresponding to the exact representation of the digital wave from the source.


Disadvantages:

IGBTs could not provide a continuous waveform to the electrodes. In addition, semiconductors are inherently non-linear (can this be corrected with feedback?).


Notes:

A stable plasma requires a specific "holding current". As a result, a solution must be provided which maintains the "holding current" through the electrodes while not interfering with the response of the tweeter. The IGBTs are characterized by a stabilization time and will exist in the "OFF" position for a finite amount of time. Since the IGBTs are able to switch very quickly (>100,000 times per second), the time the IGBTs spend in the "OFF" position will be very small. I do not assume any current will be passing through the electrodes when the IGBTs are in the "OFF" position. How can I relate this to the holding current? Will the plasma still be destabilized even if this time is made very small?

An H-bridge utilizing IGBTs could provide current in either direction with a DC voltage source power supply. How would you approach designing the power supply?

What IGBTs would you recommend?

What should I use to control the IGBTs?

Thanks,
Thadman

[Arnulf]

Originally Posted by thadman

A design that modulates a high frequency resonance (>100khz) greatly reduces the voltage demands on the amplifier and will maximize energy transfer. However, such a circuit would require complex analysis.

I'm not sure what this "complex analysis" is but it sounds like what you want is an AM LF transmitter.

You remarked that tubes are inefficient further down your post. Heating current aside, they will be inefficient in your design by the virtue of the choice you've made (to go class A & to operate without carrier modulation).

For the tweeter prototype, I'd like to use a separation distance of 1cm between the electrodes and operate them in anti-phase (push-pull).

Why restrict yourself to a fixed gap ? My suggestion (based on plasma speaker some kid here did last year as his school project, he used one of those SMPS controller chips as modulator and step-up transformer) would be to take tungsten welding rods to serve as your electrodes and put them into electrode clamps (the things that grip the electrodes when welding, I'm not sure what the correct English term is for those). You can adjust the spacing to get optimum operation.

A digital to analog converter will have to be used. This will require additional components and will introduce an error (however small).

These aren't hard to come by - every sound source has one (TV, computer, portable MP3 player, etc.)

Since an AC waveform is being passed between the electrodes, there will be a point at which current is zero. I'm not sure if this is simply a "theoretical" problem since the waveform will not spend any finite amount of time at a position since it is continuous.

What happens during quiet passages or silent portions of the recording ? I think this is just one more reason not to try to reinvent the wheel - use suitable carrier (you say 100 KHz is optimal, I suggest adjustable oscillator frequency) and modulate it with your audio input. As long as you don't get anywhere near full modulation, the plasma arc should be uninterrupted.

I believe an AC voltage source power supply will be required. How would you approach designing the power supply?

Tubes operate on DC.

What Vacuum tubes would you recommend?

The kind that might kill you, considering your choices of desirable output power and voltages involved - radio transmitter tubes.

Do Ultralinear (high fidelity) Vacuum tubes exist which can provide 1.5kV on the electrodes?

Tubes don't provide voltage, they "burn it up" so to speak (they act as input-dependent resistors).

"Ultralinear" is a term that describes a specific feedback arrangement from output transformer for pentodes (and tetrodes).

I strongly suggest you invest some more time to familiarize yourself with what you're planning to do. Voltages of a couple of kV are extremely dangerous.

[thadman]

Originally Posted by Arnulf

I'm not sure what this "complex analysis" is but it sounds like what you want is an AM LF transmitter.

If the design requires complex analysis, it will most likely require a significant number of components and funds invested. I'm not arrogant enough to assume my first design will be perfect or I won't make any mistakes. It's a learning process. Assuming I'm going to make mistakes, I'd prefer if it didn't involve a significant number of components and funds invested. Once I've reached the point where I feel I understand the problem at a sufficiently high level, I will attempt the pulsed HF resonance design.

Originally Posted by Arnulf

Tubes operate on DC.

Thanks for the correction. This is the kind of feedback I'm looking for

Originally Posted by Arnulf

The kind that might kill you, considering your choices of desirable output power and voltages involved - radio transmitter tubes.

Could you recommend any resources where I can learn more Vacuum tubes? Could you recommend any Vacuum tube distributors or manufacturers? I'd like to be able to determine and source the optimal Vacuum tube for my design.

Originally Posted by Arnulf

I strongly suggest you invest some more time to familiarize yourself with what you're planning to do. Voltages of a couple of kV are extremely dangerous.

This post simply represents a few hours of personal research on amplifier design. I'm not going to physically build anything until I'm confident in the mechanisms involved and have familiarized myself with the required safety precautions.

BTW, thanks for taking the time to respond to my post. I really appreciate it.

Best Regards,
Thadman

[Arnulf]

Originally Posted by thadman

If the design requires complex analysis, it will most likely require a significant number of components and funds invested. I'm not arrogant enough to assume my first design will be perfect or I won't make any mistakes. It's a learning process. Assuming I'm going to make mistakes, I'd prefer if it didn't involve a significant number of components and funds invested. Once I've reached the point where I feel I understand the problem at a sufficiently high level, I will attempt the pulsed HF resonance design.

Well, it's your choice but if I was only beginning my venture into an unknown area I'd try to follow and established path, not pioneer my own. Copying existing (proven) RF designs seems like the most logical course of action to me.

Could you recommend any resources where I can learn more Vacuum tubes?

Radiotron Designer's Handbook, 4th edition is avaliable in PDF format online (I don't have the URL handy but it should be easy to find) and should IMO be your starting point. A bible of tube design, I believe it covers all the areas of interest to you, that is audo and RF amplification, oscillators, supplies, etc.

Could you recommend any Vacuum tube distributors or manufacturers? I'd like to be able to determine and source the optimal Vacuum tube for my design.

I am not familair with RF tubes but it is hard to beat the prices for Soviet NOS gear. There is plenty of military grade stuff floating around ePay. I use Soviet tubes for audio as do many others, they work fine (just as fine as new production stuff, but at fraction of the price of even cheapest lowest quality Chinese substitutes).

This post simply represents a few hours of personal research on amplifier design. I'm not going to physically build anything until I'm confident in the mechanisms involved and have familiarized myself with the required safety precautions.

Very high voltages are very lethal and expensive to play with (supply filtering would be a serious issue with 3+ kV supply). 100W audio amplifier is doable at at significantly lower voltages, but I wonder if those are enough to create the plasma arc. On top of that I cannot possibly imagine why the need for 100W tweeter, scientific curiosity aside. Why not start with something less demanding and work your way up only if necessary ? This would cut down on start-up costs and safety hazards.

Again, my suggestion would be to:

1: Have a look at existing plasma speaker designs (all I've seen use modulated carrier, step up transformer and adjustable gap width; none use AF AC or fixed gap) and try not to reinvent anything when existing solution works just fine and doesn't add new variables into the equation. Perhaps try to build one as a reference from scrap parts from an old CRT TV ?

2: Have a look at tube AM transmitter designs because that's what a tube equivalent of existing (solid state powered) plasma speaker design would be.

3: Have a look at local laws and regulations regarding RF transmission in the band of your interest - your project might cause serious interference to LW/MW communications and/or public radio.

[Arnulf]

Here's a link I just found, you might find it very informative:

Ulrich Haumann's DIY PLASMA TWEETER

[kevinkr]

One other thing is do your experiments in a very well ventilated space, plasma tweeters can produce massive quantities of ozone. My own exposure to operating plasma tweeters resulted in a severe asthma attack - I have avoided them since.

[thadman]

Originally Posted by kevinkr

One other thing is do your experiments in a very well ventilated space, plasma tweeters can produce massive quantities of ozone. My own exposure to operating plasma tweeters resulted in a severe asthma attack - I have avoided them since.

Do solutions exist which can attract and neutralize the produced ozone?

[Michael Koster]

Originally Posted by thadman

Do solutions exist which can attract and neutralize the produced ozone?

Hmm Plasma consists of ionized gas... ionized oxygen can easily form ozone... the ozone molecule is not electrically charged... nothing jumps out at me but a little research might yield something. There may be ways to operate as to minimize ozone production.

I agree with previous posters that you will need something to strike and maintain the arc in the absence of audio signal. There is a significant amount of work needed to ionize the air which would create a lot of distortion and probably dropouts. It's hard enough as it is without introducing new variables. An AM system is simple and relatively easy to get working.

There is a ~27MHz frequency allowed for industrial use that you might look into.

I've become interested in a different arc-less design that would use modulated DC but it's purely theoretical at this point. Voltages would be in the 50KV+ range at low current making implementation a big question mark.

Michael

[godfrey]

Originally Posted by thadman

Dielectric breakdown of the air occurs at a potential of 3kV/cm.

I think it's more like 3kV/mm, which means you'd need about 30kV to initiate the breakdown.

OTOH, as you pointed out, resistance of the plasma will be low. If the arc is maintained with a DC current, it should be possible to modulate it with an audio signal of reasonable voltage.

[Arnulf]

Originally Posted by thadman

Do solutions exist which can attract and neutralize the produced ozone?

Ozone (O3) will break down (into oxigen, O2, in presence of nascent oxigen atoms, O) and react with materials around it; i.e. it will oxidize most metals (copper, iron, etc.) and react with other substances to form more or less toxic gasses (various nitrous oxides etc.). It will react more willingly in the presence of CFCs and some other compounds(remember ozone depletion in upper atmosphere ? you could actually benefit from it in this case).

My suggestion: ventilate your premises well and you'll be just fine, unless you otherwise have problems around photocopiers, laser printers and similar ozone generating devices.