• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

DIY Direct-Drive amp for Estats

RetconCat

New Member
Joined
Feb 16, 2021
Messages
4
Likes
1
And so begins yet another needlessly complex project. I should disclaim before the talk of how dangerous dealing with multi-kV systems is and how I'll hurt myself that I'm in my last year of an electrical engineering undergrad, so I know how to work with high voltage safely.

I'm in the design stage for a set of DIY ESLs (a jazzman-inspired segmented wire stator design), and I'm thinking about my amplification options. As a starting point, I'll probably just use some MiniDSP plate amps and a transformer, but as a fun electrical engineering project, I want to build a direct drive amp system and completely remove the transformer from the equation. I've looked into a few topologies but I figured I'd ask the room here and see if anyone has any particularly good leads on designs or high voltage rated components.

Option 1 is a class AB MOSFET design based on this design. The schematic on that website is known to be inaccurate (connections where there shouldn't be, the random pointless diode), but I've got a working version simulated, and it shouldn't be too hard to expand that from a +/- 1kV design to +/- 3kV to match my stator bias voltage. The main reason why I'm not particularly enthused with this design is it would mean chaining a stupid number of fets to get the voltage rating I need using easily available 500V FETs, or shelling out for some even higher voltage units that I have a suspicion might not sound the greatest as they're designed for power regulation.

Option 2 is to go a little crazy with a class D topology using a similar output transistor stack for voltage requirements. Only reason this is really on the table over the class AB is it gets rid of some analog design screwyness and lets me easily do an optically isolated connection to a digital driver and DAC, preventing me from accidentally zapping all my other gear with a few kV. (plus bonus power efficiency)

Option 3 is to go full insanity and go for a tube amp. The thought here is that there should be tubes available that can directly handle my whole 3kV target voltage and then some without having to resort to stacking them like the FETs in the other design. The problem with this is that I have very little experience selecting tubes for a design. FETs and BJTs I can just slog through Digikey's parametric filters, find things that match my specs and pretty much call it a day at that, but AFAIK there isn't really a good parametric database of tubes out there that would let me find appropriate tubes for my output section.

If y'all have any wisdom on useful designs, good FETs or tubes to use for these, anything like that, I'd be interested to see it, or if you think that it's stupid overkill and I should just put my time and money into some good quality transformers and normal amps, all input is welcome.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
I've owned various ESLs for most of my life.

I'd throw out option 1.

Option 2 looks like the way to go if you can make it work. I've knowledge of circuits and can do some simple design, but off the top of my head don't have useful suggestions other than it looks like the way to go. Since there are solid state radio transmitters of high voltage I'd think those parts can be used in a class D design for the output stage. But that is just a half informed guess.

Option 3 is very viable. I think you'll end up with mostly transmitter tube for the voltage involved. I assume you know that at one time Acoustat offered a direct drive tube amp for their speakers. Which are of course pretty much made like your planned panel design. Some info on it here:
http://www.beercityaudio.com/index.php/acoustat-x-10-electrostatic-electrostat-tube-amplifier.html

Here is one you can buy ready made:
http://tubeaudiostore.com/eslelodidram.html
$6500 for 4000 volt output which is a bit low. I'd like 5 or 6 kvolts myself.

Rodger Modjeski designed one you can see some posts about it here. Unfortunately he passed away in 2019.
https://www.audiocircle.com/index.php?topic=58015.0

Magnetron tubes at high voltage have been used in microwave ovens. I believe they have been or are being phased out for GaN devices that are able to operate at those voltages. You just have to run your switching amp at 2.45 ghz I guess.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
I've also thought a sort of reverse version of the David Berning amp could work for solid state high voltage drive of an ESL. It is a variation on switching amps anyway.

http://davidberning.com/products/zh270

He used tubes running at a high frequency that went thru a transformer designed for that high frequency. And then modulated that result. Seems you could do something similar in reverse. Run transistors at a high frequency via transfomers designed to step up the voltage at the high frequency, and modulate the result. This is all just spit-balling without thinking thru problems. I'd think such a switched via transformer amps might get you the results of direct drive better than direct drive itself. And better than the conventional wide-band transformers because the switching transformers don't need to be wide bandwidth just high frequency.
 

Maki

Active Member
Joined
Oct 24, 2018
Messages
252
Likes
478
Maybe look into the Cree C2M1000170D. They're used for electrostatic headphone amps and are rated 1700V drain to source and have very low output capacitance. The problem is the P-channel counterpart - doesn't seem like that exists.
 
Top Bottom