ESL amp - switching amp?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hey, I've been looking for an amp setup which would drive an ESL without the need for a step-up transformer or high voltage transistors. I know the general consensus regarding switching amps is that they suck, but I've heard they can be done right and made to sound good. To my understanding, they're just like switching power supplies, but the signal/feedback is used to control the voltage, rather than a fixed voltage.

So I've been looking for schematics of switching amps made from discrete components, perhaps using a 555 timer ic. Anyone have any schematics to offer or any other useful information?
The only way to drive an ESL is with a high voltage. That means you need either a step up transformer or high voltage transistors/tubes, or both.

Switching amps are another subject entirely. There is no reason why you couldn't use high voltage transistors at the output of a switching amp...

There are some good switching amp controller chips available now (check TI, Tripath, and others) that might actually let you get a high fidelity output.

The one and only
Joined 2001
Paid Member
I think you're gonna need either a step up transformer
or high voltage outputs.

You are probably better off building an ordinary switching
amp and stepping up the audio, as the speed required to
get very fast rise times on these high voltages is probably
beyond the capacity of the parts.
Ah, well the reason I'd go with a switching amp for driving an ESL is that I can get whatever voltage I want, and power mosfets are the tried and tested method. I can get kV mosfets fairly easily, too. But, yeah, I guess I would have to use HV mosfets for output. I guess designing a switching amp is very different than designing a linear amp. Well, I'll post a schematic later, when I get something preliminary set up. I'm planning on using a 555 and a couple HV mosfets.
Since you want to drive an ESL I assume you are after a result that is pretty good quality. I'd ditch the 555 and go for one of the more modern, application-specific parts from TI or Tripath or one of the other companies that make switching amp controller ICs.

Hi Altaic

Driving an ESL by a switching amp is in fact very tempting and would have great advantages. But you will encounter problems.

First of all you need some experience in EMC since a switching amplifier is a precision RF device used for audio amplification.

Apart from that two things might be causing troubles in your case:

The ESL is a mainly capacitive load that would of course simplify the output filter a bit (you need one capacitor less). But most LC filters don't like complex or imaginary loads (except for frequencies that are very much below cutoff, thats why they can excel in the bass region) so you would have to add some resistivity as well (I don't know if this is proper English for what I mean) which will decrease your efficiency.

As Nelson Pass already pointed out: While the slew-rate of the output stage of a switching amp of moderate output voltage capability would already be some kilovolts/microsecond it would get even higher for an output stage with +-500 volts (as an example) drive capability.

I once developed a class-d amplifier (if you like I can fax you the schematic diagram) that was intended to be used as a lab power source.
Because of this I was contacted by somebody who wanted to develop a class-d amplifier for driving a piezo actuator (output voltage 0 to 1000 volts). While the snubber networks of my small amplifier dissipated 2 watts approx, dissipation would have risen to 100 Watts for his amp simply by the fact that these losses increase with the square of the voltage in question (and proportional to switching frequency and capacitance used).

As mentioned by MRehorst, decent switching amps don't use NE555s for the modulator, they either use some sort of delta-sigma modulator (i.e. Tripath and Sharp) or a precision Triangle generator and comparator (I intentionally omitted digital PWM generation).


Thanks for all the advise, guys. I really appreciate it.

I'm heading off to Canada for a few days, unfortunately for this project, but when I get back I'm going to give the art of electronics a skim. There's got to be a way to cancel out the distortion, or at least minimize it. Perhaps use some sort of hybrid tube-mosfet design?
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.