Class D amplifiers and ESL's

[jmateus]

I built my electrostatics three years ago and I've been estatic with
their sound, either driven by my tube amplifiers or a JLH class A which
got to be my pride and joy.
But class D amplifiers came up on the scene some time ago and just
for kicks I bought one of those Sonic Impact 20 some dollar jobs which
I've been playing with good dynamic speakers. Good but not out of
this world.
Yesterday came the time for an experience, to connect the SI to the ESL's. Of course now I will say to myself "you should have never done that", and in fact I shouldn't, but the fact is the little thing called Sonic
Impact blew right before my eyes, it played for 2 or 3 seconds and then
quit working.
Can anyone help me to understand why the heck did my class T SI
blew that way? I've heard about amplifiers that are not very suitable
for the ESL, including regular class AB solid state amplifiers but I never
understood why...Is that because of the internal capacitance of the
ESL? What is wrong with 1,200 pf that the ESL show to the amplifier?
Is that capacitance determinant to the amplifier?
I really would like somebody to explain to me the true reason why certain amplifiers don't live well with electrostatics.
Thanks.

[SY]

Your ESL doesn't show 1200pF to the amp- there is the small matter of the step-up transformer.

[I_Forgot]

Take a look at this for a good explanation of amplifier stability (not just op-amps) and capacitive loads: Stability

You can put an Ohm or two resistor in series with your amp and it might help, but that's a less than ideal solution. The real solution is to design the amp with sufficient phase margin to prevent instability.

I used to have a Carver amp that was one of their real small, high powered things (M200t if I recall correctly). It didn't like my ESLs at all. Piece of junk... Eventually cooked the power transformer which started buzzing loudly after that (not while driving ESLs- it sounded like **** with the ESLs because it was oscillating at ultrasonic frequency). I was glad to throw it away.

I_F

[phase_accurate]

Class-d amps with pre-filter NFB (like this one) are usually quite robust regarding load impedance, apart from the subsequent response deviations.
There may be the odd case when the combined impedance of the filter and the load may cause trouble (like in your case). Another danger are the very LOW impedances of ESLs at high frequencies though I doubt that this was what caused your troubles since there is not much audio power needed up there usually.
Unfortunately many speaker manufacturers don't specify the impedance vs frequency properly for their ESLs. They just make statements like x Ohms @ y Hz.

IMO a switching amp with an output filter that is dimensioned exactly for that purpose might perform very well driving an ESL.

Regards

Charles

[jmateus]

Thanks for your answer phase accurate.
In fact this was my first opinion, that the odd impedance of the filter
and the capacitance of my ESL would have colide somehow...
As I said before I made the ESL, they were not any brand name
ready made, I made them from scratch so It's really difficult to say
what kind of impedance they have, the only thing I know is the output capacitance which is based on the physical dimensions of the panels.
Of course we can calculate de impedance (nominal, of course) as well
by other formulas taking other factors into consideration. But I never
did that, I was very happy they could play with no problems whatsoever
with my tube amplifiers or the JLH, class A so I never bother to do it.
Now that this happens with the SI, I'm debating very hard if I should
connect the module class D I'm working on now. This one is not that
cheap and I hate to have it damaged.
Oh well, thanks for your input.

[phase_accurate]

What type of module is it then ?

Regards

Charles

[jmateus]

It is a module that 41Hz.com sells on their web site and is based on
Tripath chipset 2350, it´s supposed to be 2x300W nominally and is
supplied by +- 45 volts.
Before I bought it I adressed the problem of the ESL and was told,
after the guy consulted Tripath, that there would be no problem in
driving the ESL.
However after this stupid experience with the SI, I don't know if
I should take that kind of risk.

[djmiddelkoop]

The problem with ESL is that they look like a resonable size capacitor to the amplifier, say a couple of uF depending on type and stepup transformer.
Most commonamps don't like this, because they are designed to drive resistive loads.
The better amps can take a phase shifting load without problems.

The class D amps work very different. To filter out the PWM high frequency there is a special filter at the output.
Guess what is doing this capacitive load, read ESL, to this filter ?

That is what happend with your amp and it obvious couldn't take it.
I have heard of more problems with class D and ESL, even from the better brands so be warned.

Dick.

[jmateus]

What you said makes a lot of sense although I already had a pretty
good ideia that the filter at the output of the amplifier would somehow
crash (for better term) with the combination ESL and transformer capacity.
Instabiity at high frequencies in regular push pull amplifiers would
have a similar efect.
Anyhow I'll be aware of that in my next endeavour...
Thanks a lot for your enlightment.

[samborambo]

I'd imagine the output filter (inductor) in a d-class would cause a "tank" effect with the load capacitance.

I've been giving ESL serious consideration in my next project. I've done some searching but can't seem to find anyone who has tried direct driving ESLs from a d-class by putting a step-up transformer before the output filter.....similar to many SMPS designs. Is there something fundamentally wrong with this approach?

Sam.