Current vs voltage drive ESL?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Unless Im way off - I believe that using voltage drive with an electrostat is to bypass the step up transformer in the signal path. Typically, an amplifier has way too much current and far too low voltage to drive the panel - so there is a step up transformer in the speaker. Tube amps suffer the opposite problem - they produce too much voltage and not enough current - so they use a step down transformer on the output. So with a tube amp running an electrostat, you have a step down transformer, followed by a step up transformer. So you can do away with both, running the bare signal from the amp right to the panels. Don't try it without consulting the proper authorities.
I'm sure you're right but I was wondering more along the line why a current amplifier can't be used to drive esl's?
U=RI and all...

Since I'm a no0b with the electronics bit there's a lot for me to learn.
Since the stators aren't touching can there really be a current running through the circuit?

The definition for an electric field found in my physics book reads
E being the electric field [V/m], V2 and V1 the electric potential [V] of the stators, a the distance [m] between the stators and x, well I'm not really sure what the x is doing there?
Please correct me if I make an error. It would appear the electricfield can exist without a current as long as there's a difference in electric potential.

If I remember things correctly it is the change in the electric field that's causing the charged diaphragm to move?

P being the power [W], U the Voltage [V] and R the resistance [omh]. In our case R isn't really R but Z. This would explain power needed by the esl.

Still I feel like I'm missing something really essential.
How could there not be a current? It really makes my head hurt. :dead:

Please, can someone explain this to me?
An electrostatic speaker can be modelled in the first approximation as a capacitor.
Therefore there is no continuous current through it but there can be alternative current whenever the voltage on the capacitor changes.
You can review the chapters about RLC circuits in the physics textbooks to understand the power needed to charge and discharge a capacitor at a certain voltage and a certain frequency.

An ideal capacitor doesn't dissipate power. If you want to have a better model, you need to add a dissipative subcircuit. A very simple one would be a resistor in series with the capacitor.
Maybe someone with more knowledge about the ESLs can talk about better models for them.
Joined 2008
Paid Member
The place to start thinking is this: a monumental acoustic sound of 130dB is one electric watt (please correct me if I am wrong). ESLs make lots of sound without needing much power to speak. Cone speakers are very inefficient (that's a cue that cone speakers are not a smart way to make sound).

Great problems driving an ESL with a transformer and only because they are so wonderful even when poorly driven do they sound so much better than anything else. The transformer output looks screwy to the ESL panel and the transformer makes a screwy load to a conventional hifi amp.

So, you can make a high voltage direct drive amp with a resistor as the load (and by the way, an ESL strapped across that resistor) or you can hand-tinker with transformers.

Correction for earlier poster: ordinary tube amps make only a small fraction of the drive voltage needed by an ESL. You can ditch their regular 10k ohm to 8 ohm matching output transformer and substitute a mild step-up transformer. Not a bad way to power ESLs.
Yes, now we're getting somewhere.
I had missed the obvious, esl's are capacitors and should be treated that way. :)
A capacitor can never hold a higher charge than the voltage across it can supply. The current flowing only flows while charging the capacitor. This way it won't matter if the amp can put out 1 or 1000 amps as long as the voltage is to low.

Bypassing the panel with a resistor sounds like a interesting idea but I wonder if it'll really work? Current will flow continously and there will be a voltage across the resistor. Put enough current through it and the voltage will be high enough to charge the panel.
However the panel is frequency dependant and the reactance will vary. Current division between resistor and panel will cause voltage fluctuations, possibly not so great for driving the screens? I don't know? I'm just thinking out loud here?
Bringing up the voltage across a resistor would also waste a lot of energy in the form of heat, not a very efficient way of doing it?
Still I don't know the inner mechanics at work here and it might be a viable of doing it?

as with all audio gear, compromises dictate the end-result.
what current drive promises is:
- a linear frequency response compared to a 6dB-rise with voltage drive
What is usually and too easily forgotten and often not told are the premises and disadvantages of current drive which are:
- cd (current-drive) is not applicable to the entire audio frequency range
- it applies only to the far field range. Under near field conditions, Headphones e.g, linear response is achieved with voltage sources
- the membrane size must be small compared to the radiated wavelength, so it doesn´t apply for example to line-sources
- small membrane areas mean high impedance values, mean high voltages, mean high transformation factors of the audio tranny
- if implemented with a voltage source feeding the panel via resistors (as in the Quads) the voltage demands are seriously increased, hence transformation factors of ridicolously high values required
- in praxis cd doesn´t make equing obsolete as the schematics of all Quads proove.
- since the load defines the current and voltage demand from the amp, a current amp would require the same supply voltage and current range as a voltage amp

So current drive is not per se advantageous in driving speakers, neither dynamic nor electrostatic. One must closely look if and how a claimed Pro applies in praxis and if the unavoidably occurring Cons don´t eat up any possible improvement.

Current vs. voltage drive


try to dig up the old articles from Peter Walken... He shows that the frequency response is flat when the supplied current is proportional to the audio signal. Sadly, the ESL is reactive, so when you drive it with a voltage source instead of a current source, you get a rising response for higher frequencies.

There is a simulator to play around with here: Electrostatic Loudspeaker (ESL) Simulator

It is possible to make a special amp with current-sensing feedback to accomodate the ESL, or you can use a line-level equalizer.

But when you use electrical segmentation (RC sectionizing), then if you do it right you can restore the flat frequency response for voltage drive. This is because the higher the frequency, the smaller the radiating area then becomes. Using this system, you wouldn't need the above tweaks.

[...]The voltage driven sim always ending around ~120dB@20kHz. [...]

Try varying the D/S spacing...

Sure, as with any sim, one needs to realize that it is just a graphical representation of an equation which was derived as a simplified model of reality with a number of assumptions taken. At least the website explains in detail what the simplifications and assumptions are, so people can judge for themselves...
Like Calvin said, interesting sim but I've never seen such curves irl?
It's hard to get anything usefull from it when you don't know what to trust.
It would be very nice indeed to get a guesstimate for the sensitivity based on the most common esl parameters but that sim's very odd.

Since starting this thread I've dusted of my old physics books and actually found some interesting stuff. Equations I've seen in audio contexts suddenly make more sense.
There's still a lot of stuff that's unacounted for but I'm getting there. :)

To reconnect somewhat to my original question.
Current- vs voltage drive.

Is this statement true and a good summation?
ESL's are capacitors and as such they can never hold a greater voltage than the peak voltage across them.

And maybe...
The amount of current pumped into them is irrelevant as long as you fill the minimum requirements and you can't raise voltage with current capacity. (Except by using other means like transformers etc)
Is this statement true and a good summation?
ESL's are capacitors and as such they can never hold a greater voltage than the peak voltage across them.

As a very, very, very simplified model you could say that an ESL panel is a capacitor, yes. Of course such a model is even worse than that sim, because a capacitor will never produce any power, acoustic or other.

I think your statement is circular... Basically you're saying "a capacitor can never have more voltage across it than the peak voltage across it." That seems obvious to me.

Remark that you can say the same thing about an inductive load driven by a perfect voltage-drive amplifier. The amps job is to shunt any back-EMF produced by the inductor. Just as it is the amps job to source and sink any reactive current flowing through a capacitive load. In both cases, if the amp cannot meet that goal, the result is compromised.

So perhaps you mean to say "a capacitor will never generate a back-EMF voltage when driven by an imperfect voltage source"?

The amount of current pumped into them is irrelevant as long as you fill the minimum requirements and you can't raise voltage with current capacity. (Except by using other means like transformers etc)

If voltage drive is your target, yes you could say that. Actually what the amp is doing is pumping a certain amount of charge Q (depends on the signal voltage) from one stator to the other and back every period. The peak value of current is Q times frequency, which explains why ESL impedance is so low at high frequencies.

Many ESL systems will fail to meet this reactive current requirement due to a combination of inadequate current sourcing capability in the amp and bandwidth limitations in the step-up transformer.

What we ultimately want is high sound pressure, so we need a large electrostatic force acting on the charges of the charged diaphragm. The force acting on a unit area of diaphragm is the product of the charge density Q on the diaphragm and the (AC) electric field strength E between the stators. To maximize Q, we must maximize the polarization voltage. To maximize E, we can either increase the signal drive voltage V or decrease the D/S spacing d, since E=V/d. But decreasing d reduces the maximum mechanical excursion of the diaphragm, so it limits max LF sound pressure. Therefore, we still must maximize V.

So I guess, the ultimate reason why ESLs are inherently high voltage devices, is because of the high compressibility of air... If air were less compressible, it would take much less force, and hence less diaphragm excursion, to get the same pressure.

Concerning the current vs. voltage issue:
In theory a capacitor could be charged up to an arbitrary voltage by a current source. But one needs a current source that just "keeps on going", even in the face of the growing voltage on the capacitor. Unfortunately, this is exactly where electronic current sources fail. They cannot handle a load whose voltage exceeds their supply voltage.

A mechanical current source (charge conveyor belt) could do it (as in a Van de Graaff generator) but unfortunately, those work only for DC.

Thanks, that's pretty much what I had gathered but I was having difficulties putting it into words. :)

It would explain why a high voltage source is needed.

I suppose the next step involves estimating how high a voltage we need? Maybe the topic for a new thread? Involving all the scientific stuff of course. ;)

ESLs are high voltage devices because they are high impedance devices ( several kOhms to several 100kOhms, at least with typical dimensions and in the audio frequency range, compared to the typical 2-20Ohms of dynamic speakers). Similarities occur to tubes and FETs vers. bipolar transistors.

Yes that´s a Man´s amp..... a powerman´s, like Tim Allen is.....isn´t that frightening? :D

The regrettable part is that 90% of the time, only 10% of the voltage is needed. This is because of the large crest factor (ratio between peak voltage and rms voltage) of the typical music signal.

When I first tested my big full range ESLs (~3.5mm D/S spacing!) I used a solid state amp and some 50Hz toroidal transformers, which generated only about 500Vrms or so. I had to turn the amp all the way up, but it sounded really loud! Unexpectedly loud.

Of course the catch is that I didn't have any headroom left for peaks, so on everything except vocals or flute, the speakers sounded like ****. :(

Morale of the story: beware of this trap, the system might sound good at low levels, but be sure to test with extremely dynamic music too, and verify the peak reproduction.

For a hybrid ESL+dynamic woofer, a DD amplifier which can reproduce 100% of the peaks is certainly a possibility. For full-range ESL you'll find you always end up with a rail voltage requirement which is... unpractical. Unless you can live with a low acoustic output power, that is...

The choice of rail voltage for a full-range DD amplifier is thus really a compromise between feasibility on the one hand and how much of those 10% peaks you want to accommodate on the other hand.

ESL design is one long chain of trade-offs ;)

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.