what amplifier for driving esl

[kavermei]

I'm currently building a fairly high-voltage amp (B+ will be 1500V) with low-rp triodes in PP configuration (GM70's, to be precise) feeding a low-ratio (about 1:3) step-up transformer.

Think of it as an "almost-DD" amp; the transformers are just there to provide safety and to boost the voltage a little (my speakers are full-range esl's with pretty high D/S spacing). The pretty low rp of the GM70 tubes will hopefully reduce the transformer's imperfections to a minimum, even without feedback.

I expect it to work very well, but it's not quite finished yet...

Kenneth

 

[bentoronto]

I'm currently building a fairly high-voltage amp (B+ will be 1500V) with low-rp triodes in PP configuration (GM70's, to be precise) feeding a low-ratio (about 1:3) step-up transformer.

Think of it as an "almost-DD" amp; the transformers are just there to provide safety and to boost the voltage a little (my speakers are full-range esl's with pretty high D/S spacing). The pretty low rp of the GM70 tubes will hopefully reduce the transformer's imperfections to a minimum, even without feedback.

I expect it to work very well, but it's not quite finished yet...

Kenneth

Smart in many ways.

An alternative is to use higher B+ and DC-blocking capacitors.

I couldn't say where the "quality bottleneck" is or which approach comes closer to 100% DD.

My resistor-load amp had feedback right from the stators. Again, I don't know if an ESL provides useful negative feedback the way motional impedance can be used in a cone driver (I think speaker feedback is the last frontier of feedback). Anybody can say?

 

[kavermei]

RC coupling with a 1500V B+ would probably work well with a hybrid system with low D/S spacing.

If the voltage needs to be higher (as in my case), there is a problem with raising the B+: it's hard to find fairly linear, low rp tubes in the >2kV voltage range. Many tubes in that range are made for pulse operation and therefore have terrible linearity and insane heater requirements. And they usually require gobs of grid current.

So if there were a hypothetical tube identical to, say, the GM70, but with a 5kV B+ rating then most probably, the best sound would be obtained by RC coupling this tube to the speaker. But between transformer-coupled GM70 at 1500V and RC coupled "unsuited" tube at 5kV, I'm not sure which would give the best sound.

When the amp is finished, I'll certainly post pictures, so keep an eye out here!

BTW Ben, can you say some more about your RC-coupled amp and speakers?

Kenneth

 

[markusA]

An OTL + Step down transformer is a bit silly, because the entire point of an OTL is to NOT have an output transformer. I'd say just build an amp with a low output Z transformer would be my advice.
Probably less expensive

Edit: once you step down using a transformer with that OTL, what has happened to the voltage swing?

Looks like I managed to be unclear anyway.
I was thinking one could use an OTL amp and a lower step-up ratio instead of the usual tubeamp with a step-down tranny + a high ratio step-up?

 

[spocintosh]

Why not just use any Bongiorno designed amp ?
Will drive ESL's without any problem as these are among his personal favorites.

 

[kavermei]

Looks like I managed to be unclear anyway.
I was thinking one could use an OTL amp and a lower step-up ratio instead of the usual tubeamp with a step-down tranny + a high ratio step-up?

Of course this is possible, but the transformer will need to be custom designed, and thus probably more expensive.

Another difficulty is when you use a tube stage which would normally require feedback from the speaker output (as is usual with pentode output stages). With two cascaded transformers, it can be taken from the low-voltage intermediary. When using only one transformer, the feedback needs to be taken from the high voltage output, and would require expensive high-voltage resistors.

So there are benefits to each of the alternatives.

Kenneth

 

[bentoronto]

From what reading and experimenting I did long ago, at the very most ideal, a 1:100 transformer is a very weak link in an ESL system. That is true both as to the kind of load is presents to the amp (weird impedances), frequency range it can handle in light of its windings and inductances and reactances, as well as "hysterical" gremlins. Therefore, anything you can do to curb it, is very beneficial. Best is obviously DD. Next best is an amp needing a lower ratio.

I know that little mic and stage coupling transformers can be made acceptably low in distortion and without being inside a feedback loop. (They are coming back into style to control ground loops.)

But, I'd like to ask, are a lot of DIYers mistakenly taking any old toroid transformer (or a couple in series) from their spares box and thinking it will add little distortion or frequency discriminations? Or is this application not prone to transformer distortions (because it doesn't carry large B+ currents like tube-amp output transformers)?


BTW, my impression is that few people have feedback from the speaker side of their 1:100 transformer. But it seems to me you really ought to.

 

[kavermei]

A lot of DIYers are just operating a toroid in open loop between their amp and panels, since it is so easy to do -- no need to modify the amp! -- and the result is not terribly bad.

Only a part of those move on to more elaborate drive methods...

 

[bentoronto]

As I've said many times, us ESL devotees overlook many shortcomings of ESLs in order to gain the special good qualities.

I'd like to see data about the distortions introduced by those randomly selected toroids.

 

[tiefbassuebertr]

Driving ESL without Transformer (OTL)

I am looking for diy projects and commercial products concerning amplifiers with very high supply and output voltage for direct-driving of ESL Panels. Thanks for advices

 

[405man]

I found this

High-voltage amplifier uses simplified circuit | EDN

And figure 1

I wondered if this could be extended for higher voltage operation, it may be suitable for direct drive electrostatic headphones. It should be possible to use high voltage BJTs. There is some discussion here about high voltage transistors

http://www.diyaudio.com/forums/power-supplies/170678-looking-high-voltage-1kv-transistor.html

Stuart

 

[bear]

The EDN circuit (and you need to look at the comments for a version with fewer errors) has super high values for the gate resistors. I'm not certain, but it seems to me that given the gate capacitance, this will make for an amp that my be somewhat slow, or have issues in terms of transients. The author does not show any square waves for reference.

One way to overcome this problem might be on the driver side, but this defeats the low parts count aim of the design. Multiple drivers, each stacked down in voltage, probably bipolars (to eliminate the input C issues) possibly would address this issue.

 

[geraldfryjr]

The EDN circuit has many obvious mistakes, I couldn't get it to work even in a simulator!!
I had extensively searched out the author and circuit and found no rebuttal or corrections to that article or circuit quite some time ago ( 7 years to be exact)!!

Here are just a few threads that I have already started such research on this subject,

http://www.diyaudio.com/forums/planars-exotics/221050-looking-hv-fets-direct-drive.html#post3192392

http://www.diyaudio.com/forums/head...discrete-class-headphone-amp.html#post2596223

http://www.diyaudio.com/forums/planars-exotics/80714-another-direct-drive-thread.html#post930459

http://www.diyaudio.com/forums/planars-exotics/221050-looking-hv-fets-direct-drive.html#post3240065

Do It Yourself - Electrostatic Speakers - Project: ESL H.V. Amp by Neil S. Mckean

http://www.diyaudio.com/forums/solid-state/170945-8-watt-hifi-few-parts.html#post2255998

Here is a schematic of the Stax version,

http://www.diyaudio.com/forums/plan...trostatic-headphone-driver-2.html#post3371857

http://www.diyaudio.com/forums/pass...-headphone-amplifier-concept.html#post3753546

http://www.diyaudio.com/forums/planars-exotics/248210-hv-transconductance-amp-idea.html#post3755723


I have finally got myself a small stack of HVFET's from STM ranging from 900v to 1500v and I just need to get working on a power supply for such an amp.

I haven't done much with my ESL projects lately because I have been working with Microcontrollers and IC's at the moment for the Preamp/crossover unit I am building for my desktop system.

Cheers!!

jer

The Head-Wize site had a bunch of low current designs by ChuMoy and Gilmore that are geared for headphone and could be adapted for an ESL as well.
They were rebuilding the site and all of the pages are scattered and hard to find last time i checked, Head-Fi has alot of the same data and projects there too.
Search for the Blue Hawaii amplifier it is a tube hybrid that uses some EL34's, I would love to build that one as well.

Here are the links!!! He,he,he,he,he

http://www.diyaudio.com/forums/planars-exotics/226208-esl-headphone-driving-options.html#post3294890