Megatron Electrostatic Headphone Amplifier

[JoaMat]

Schematics auto switching. Not sure it's accurate, but that's what I've done.

New Grid Voltage Control boards.

One trimmer for offset and the other for balance so, both sides are controlled simultaneously. Two jumpers. For 300B/2A3, EL34 and EML 20B-V4. Works al right, quite easy to use.

The -510V is achieved with help from GRHV 110V supply. Here is one with Hammond 2VA trafo on board - 2.75in x 2in. 
 

New idea - servo controlled grid voltage.
Okey, I've tried servo controlled Megatron (with solid state CCS) before and that failed...

[kevin gilmore]

better put a 300v zener across the output of the h11d

 

[JoaMat]

Thanks a lot.

I've been thinking of a 300K resistor cross the h11d to limit the upper grid voltage.

[simmconn]

Looking at the CTR curve I would choose the operating range between 1 to 2mA of If. Since the nominal CTR is 20%, the phototransistor side would be designed around 200uA to 400uA. Adjust the resistors such that the desired normal bias voltage is centered around this range. The phototransistor could operate off the 110V supply if you don’t need the full 510V range for the bias adjustment (why would you?).

[Shawn]

On 10/10/2025 at 5:31 AM, JoaMat said:

Schematics auto switching. Not sure it's accurate, but that's what I've done.

I checked your schematic and compared it with your previous hardware photos, then corrected a few small errors(From both sides of us).

Here’s the updated layout. It’s basically reverse-engineered from the pictures you posted earlier.

I also derived the schematic from it, but it’s quite messy at the moment. If anyone’s interested, I can post that version as well.

 

Edited October 12 by Shawn
Fixed the layout

[JoaMat]

 

I believe R41, FB2, C16 and C17 are connected as red line shows.

[Shawn]

1 hour ago, JoaMat said:

I believe R41, FB2, C16 and C17 are connected as red line shows.

Eagle eye!

Good catch. I’ve updated the image and added the missing R41–FB2–C16–C17 connections as you indicated. Thanks!

[JoaMat]

Made a control board for filament smps today.

 Some pictures from my kitchen.

 
Idle voltage – no tube

 
2A3

 
300B

 
and EML 20B-V4

 

 

[simmconn]

It’s a pretty ingenious design. It takes advantage of the fact that a cold filament draws more current than when it’s hot and thermally stable. The circuit outputs 2.5V by default at no load in case someone does a hot-plugging. An 2A3 draws more current than 2.5A when cold, and gradually drops down to 2.5A. The output is maintained at 2.5V. If it’s a 300B instead, the current would drop further, say, below 1.6A. The circuit would then switch to 5V, and the threshold also doubles to 3.2A. There could be some brief back and forth, but as the filament gets hotter, the current will drop further towards 1.2A while the circuit stays at 5V output.

It should be okay for a DIYer to build a clone for his/her own use, but integrating the same in a “DIY” unit that sells for an atrocious price at “the marketplace” would probably be a different matter.

[JoaMat]

Three kinds of switching mode filament supplies made in my kitchen.

 

Right: Traco brick, only 5V with some fine adjustment, designed by Kevin Gilmore and slightly modified by me.

Left: DC-DC Step Down Converter XL4015, adjustable voltage and current limiter, have used it for 2.5V to 6.3V.

Middle: TPS62913 Low Ripple Buck Converter (green board) My own design but done with help of TI datasheet (all info needed is there). Copper color board is an Elekit regulator with 2A3 and 300B auto detection. I’ve added a trimmer for fine adjustment of voltage.

All three seem to work all right with my electrostatic amplifiers.

I also have Pete Millett, Tentlabs and Coleman filament regulators. They are all bulky with heat sinks and produces a lot of heat.

On 10/14/2025 at 7:57 PM, simmconn said:

It’s a pretty ingenious design.

Yes, I do agree.

On 10/14/2025 at 7:57 PM, simmconn said:

It should be okay for a DIYer to build a clone for his/her own use, but integrating the same in a “DIY” unit that sells for an atrocious price at “the marketplace” would probably be a different matter.

Yes, at some point it (probably) will be a different matter.

[simmconn]

A lot of those filament regulators focus on reducing the differential mode noise. I think the DHT would be more sensitive to common mode noise, because the filament itself is the cathode in the signal path. Depending on the circuit topology, you may or may not be able to bypass the noise to ground. The common mode noise starts from the secondary windings on the power transformer, and gets more complicated with the switching regulators noise riding on top of it. It depends  more on the implementation than the pretty numbers in the datasheet.