Technical electro-stat. questions.

MartinLogan Audio Owners Forum

Help Support MartinLogan Audio Owners Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Beat_Dominator

Well-known member
Joined
Aug 23, 2006
Messages
646
Reaction score
0
Location
Phoenix, AZ
I am bored at work right now and I was pondering the inner workings of ESLs.

In a conventional cone driver the current and voltage (power) flow through the voice coil and interact with the motor structure's magnet/s to produce the field that moves the coil in and out of the gap. From what I've seen, the average cone driver has an efficiency of a few percent.

So here are my questions:

1) How exactly does an electro-stat. "use" amplifier power?
2) How efficient can an ESL panel be?
3) How does the charge generated by the wall plug interact with the amplifier's signal, if at all.

I know how an ESL works for the most part, but I can't quite wrap my mind around the use of the analog signal out of the amp. It almost seems to me that there would be no difference between giving the speaker 800 watts and giving it 5 watts. In a cone driver that difference makes for stronger magnetic fields..... is it the same for an ESL?
 
Generally the membrane is given as large a negative charge as possible by the high voltage charging circuit. The audio signal is attached to the stators and causes the charge on the stators to vary in response to the music signal. Typically the signal to one of the stators is out of phase with the other. As one stator becomes more negatively charged it repels the membrane. At the same time the other stator is becoming less negatively charged and is therefore attracting the membrane giving ESL their symemetrical push/pull nature. Thus amplifier power is used to vary the charge on the stators causing the membrane to vibrate in a push/pull balanced drive motion.

The efficiency of an ESL is a function of the voltages applied to the membrane, the membranes resistance to breakdown (the point at which the voltages arc across the stators and through the membrane) and undoubtedly many other factors. The higher the breakdown voltage the greater the efficiency. One of the important factor is the breakdown voltage of air. The voltage of the panels can never be higher than this. One company, Dayton-Wright, enclosed their ESL's in a bag filled with a gas (sulphur hexaflouride) IIRC that more than doubled the breakdown voltage of air. They would play really loudly. However they needed to be refilled from time to time and SF6 ain't cheap.

Current ESL technology seems to be limited to efficiencies around 90-91db/1w/1m due to improvements in the coatings on the stators and the membranes.

Anyone else please feel free to correct or add to any errors in the above general description.
 
Last edited:
Beat_Dominator said:
It almost seems to me that there would be no difference between giving the speaker 800 watts and giving it 5 watts. In a cone driver that difference makes for stronger magnetic fields..... is it the same for an ESL?
I've been experimenting with various coatings on the membrane of ESL speakers and can see the logic you describe. One point I should make is that when the diaphragm is fully charged or maybe even over charged with static current it takes almost nothing to drive them. They are extremely sensitive to small amounts of current (amplified sound) in which case I suspect a 10 watt amplifier could easily max out the travelling distance of the diaphragm (membrane). Which by the way is very loud.

It would appear then because an electrostatic speaker is voltage driven and has less mass to move if would give reason to conclude that they can in theory make better use of the energy from an amplifier. The down side to this is of course is how well can the diaphragm be controlled hence the reason to limit the amount negative charge to the diaphragm.

Beat_Dominator said:
How does the charge generated by the wall plug interact with the amplifier's signal, if at all.
The other interesting issue you pointed out is the current from the wall plug. That could just have well been substituted by a battery (as a source of current) or even a common high voltage source delivering a static charge to several speakers. I have also experimented with this where one speaker is unplugged from the wall and the static charge is borrowed from a alternate charged speaker and successfully allowed the unplugged speaker work with this borrow source of static current.

The logic here is that the negative static charge for the diaphragm doesn't have to be integrated with the step-up transformer for this to work although quite often this is how these speakers are designed.
 
Very interesting stuff! I was afraid I hadn't explaind myself very well but you both have touched upon the very questions I had.

enilsen: Would I be correct to assume then, from what you have said, that you could attain a maximum output from the ESL panel with an amplifier that produced around 10 watts and would not clip at low (<1ohm) loads?

risabet: Do you know how close ML runs their panels to that breakdown voltage? Would anything be gained by increasing the voltage run through the panels if there is a little bit of headroom?
 
Beat_Dominator said:
Very interesting stuff! I was afraid I hadn't explaind myself very well but you both have touched upon the very questions I had.

risabet: Do you know how close ML runs their panels to that breakdown voltage? Would anything be gained by increasing the voltage run through the panels if there is a little bit of headroom?

Interesting you asked. A while back I had a problem with a voltage regulator on my Clarity (promptly replaced by ML) and the voltage on the membrane increased an unknown amount. The change was that that speaker played about 3 to 5 dB louder than its partner and periodically made a crackling noise.
In speaking to JP at ML about it he stated that one of the effects of a higher voltage is higher efficiency at the risk of arcing. Personally I wouldn't risk arcing the panels with higher static loads for a relatively small increase in output.
 
Beat_Dominator said:
enilsen: Would I be correct to assume then, from what you have said, that you could attain a maximum output from the ESL panel with an amplifier that produced around 10 watts and would not clip at low (<1ohm) loads?
Yes you could, but what I found was just because you can make the speakers more efficient doesn't mean they will sound better. A great idea would be if you have a specific use for a super sensitive electrostatic speaker like for the electric solo guitar only. On-stage it would be crisp, clear, fast and dynamic. 10 watts would sound like a 1000 watts.

In due time I believe they could produce a diaphragm that would have active properties and could dynamically regulate both surface resistance and tension. A form of feedback logic would reasonable to incorporate so it could compare real-time diaphragm movement with signal input. Hence air humidity and temperature can be factored in yielding optimal performance. Alternatively the use of sealed units using selective gasses would produce the most stabile operating environment.
 
Beat_Dominator said:
We can dream can't we? ;)
I dream all the time. I think the way forward is to build a new type of High Voltage amplifier thus eliminating the need for step-up transformers to drive the stators.

Most of us keep thinking in the box when in reality we should be thinking in simpler terms. One of the biggest problems we face is constantly trying to adapt other standards to our needs and in our efforts have to sacrifice in many ways.

The fancy electronics required to drive electrostatic speakers is a consequence of a global standard used for coil speakers. Coil speakers need to create a strong magnetic field and rely on the amperage in the current to make this work. It is a very inefficient way of using energy to move a diaphragm and consequently suffers dearly because of this.

If we turn the logic around and use the voltage to create the static pulling force we will soon find out that the amperage is almost irrelevant. We are suddenly dealing with less moving parts and the energy needed to produce same amount of sound is a lot less.

From this I would conclude that matching the correct output voltage from the amplifier to the stators directly would in reality be the most efficient way to produce sound. The less electronics between the source the better.
 
Yes, an amplifier designed around an electro-stat..... not the other way around. Seems like parts made for those high voltages would be costly though.
 
I think this has been done! Acoustat made a speaker (model unknown) that included a Servo-Charged amp that drove the panels directly, at least I think it was direct drive.
 
Beat_Dominator said:
Pfft Acoustat. I want Krell!

I can see the sparks now as the massive Krell current actually melts the stators and the membrane burst into flames :)
 
Beat_Dominator said:
Yes, an amplifier designed around an electro-stat..... not the other way around. Seems like parts made for those high voltages would be costly though.
Anything built in small numbers will always be costly. The first television was costly and yet today you can pick up a CRT monitor for next to nothing. This (the TV) is a high voltage device that in theory would contain many similar parts needed for a high voltage amplifier.

We are surrounded by many devices that need high voltages to work. Photo copiers, laser printers, air purifiers, camera flash, strobe lights and those illuminating fly zappers not to mention all those gizmos that waste time emulating lightning storms. Oh year and Martin Logans of course.

As you can see from the advancements in all those technologies the manufacturing costs are relevant to market needs. Build amplifiers using of-the-shelf components and I suspect they can come out favourable in price and performance. The fact that no one does it doesn’t mean it can’t be done.

BTW. I don't think high voltage amplifiers will be large in size like the Krell. I also suspect operating temperatures will have less of impact on design. Heat in terms of lost energy is due to the way we need to control the current needed to drive coil engines (amperage dependent). Hmm! Welding machines come to mind when you describe the Krell.
 
You're right, welding machinary does fit that mold. I guess I need to pick up a TIG welder and hook it up to my MLs :p
 
enilsen said:
Hmm! Welding machines come to mind when you describe the Krell.
A friend of mine restored a Harman/Kardon Citation II amplifier like the one I have, and he welded two razor blades together with it, by running it mono-bridged, with a 1khz input @ 3v, and a TiG welding head ( I think ) jury-rigged to the speaker terminals. Most incredible thing I ever saw. I'm too chicken to try that on my own amp though. Sorry for the hijack, but I had to share that!
 
Innersound, another ESL manufacturer, makes an amp dedicated to drive its ESL panels. It has been well-reviewed.
 
aliveatfive said:
Innersound, another ESL manufacturer, makes an amp dedicated to drive its ESL panels. It has been well-reviewed.
True, but is still compromised to work with coil speakers and can probably handle the welding test quite comfortably.

I think we have to let go of the 4 -> 8 ohm concept in order to handle the huge variation of resistance the electrostatic element presents.

I read about someone who built his own electrostatic amplifier using tubes to power the electrostatic headphones (no step-up transformer needed). Grant you its output was only 600 volts but it work very well indeed. :)
 
Enilsen,

Do you know how much drive voltage (kV RMS) would be required to direct drive a ML Sequel II? THe HV bias is about 1.1kV, does that mean the max signal level is limited to that too?

Jan Didden
 
From what I have read on this subject I believe you would need between 2 kv to 3.5 kv make this work and up to 7 kv for peak output on the stators. Step-up transformers simulate this process by induction. Hence a 1 to 50 -> 100 ratio is used and 70 volts (peak amplifier output) becomes 3.5 kv -> 7 kv (stator input).

The bias voltage on the other hand is subject to how good the diaphragm can retain its charge. An estimated 1.1 kv could quite well be enough. If the diaphragm coating has too little resistance then arcing (sparks) can occur and the voltage would have to be reduced. If the diaphragm coating has too much resistance then voltage would have to be increased (within reason).

The answer to your question is probably that the bias voltage has a constant value (like the magnet on a dynamic speaker) and the stator voltage is dynamic in relation to the amplifier output (0 to 7 kv).
 

Latest posts

Back
Top