Making an electrostatic headphone driver.

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The next stage of my tweaking the home made driver will start soon, thank you for giving me some advice on the resistance range.

The current circuit on the SRD-5 has 4 Ohm in series and 30 Ohms in parallel with the low tension in the transformer. But I guess their designs had to work with Amplifiers designed for 16 Ohms, and compromise on sound for universality and not breaking amplifiers?

The capacitor, in parallel with the headphone is not understood my me? Why is it like that? AM I right with my first guess that it was making the speakers have a low pass filter, the more I think about it this explanation it seems wrong, and since the headphones behave like a very non ideal capacitor why add more capacitance to the high tension?



The srd-5 has a different transformer i think then the srd-7. At least the srd-7 doesn't use those ridiculously high serieresistors and parallel capacitors.
Dear all.

As some of you may remeber earlier in this thread I stated that I had baught an SRD-4. This is designed for electret headphones and I plan to convert it to support my Stax Lambda Pro.

So looking inside the first thing I notice is stax has been nasty and cut the center tap lead on the high tension.

To work around this I am thinking that I should connect 10 M Ohm resistors to each of the transformer high tension outputs of the transformer and use this for the bias, I guess it will be fine and just take a little longer to charge the headphones than a lower resisitance source. If thier is soem probelm with this idea, I could always just use one outp[ut to set the bias, but I suspect this will reduce the maximum output level by 3dB.

I would rather do this than risk the fiddle with the very short lead on the transformer, as I fear solder getting on the windings.

What do you guys think?


Yes, that should work fine just try to match the two resistor as closely as you can.
I have also powered my ESL's with one connection to the transformer on either side with the bias.
If the DC resistance of the winding is low enough it will make little difference in the balance.

Besides it is just a static charge with little to no DC flowing.
If your winding is on the order 50 to 100 ohms or more then It may make a difference at only 500V or so.
Try it both ways, if the winding resistance is high enough then it may cause an imbalance on one of the stator's during very high voltage swings.

jer :)
Yes, that should work fine.
jer :)

Thanks for the reassurance, I'm working on it now.

The inside of the SRD-4 is just enough bigger to be much much nicer to work on than the SRD-3. Furthermore it looks like the same box used by multiple models, and the base has holes for potentially bolting down a bigger transformer, both transformers are at 90' to each other, which I imagine reduces cross talk.

All the people who have made transformers and a few like me who have just destroyed them (don't worry not this time), know that transformer wire is insulated and rather fragile when the diameter is thin. Maybe for "HiFi" or marketing reasons, but I suspect cost, the SRD-3 and SRD-4 use the wire direct out of the transformer which is almost impossible to work with and very fragile. The solution is don't. Having done this twice I can be sure, first thing you do is hot glue a terminal block to the top of the transformer. The "tin" the end of the transformer wire, as if you just place it in the terminal blocks it will not conduct. The wire is covered in a very very thin highly insulating film, I imagine making the wire more prone to snapping so do this first :)

Time to finish this job :)

Well I have done it and it sounds better than the experiments with the SRD-5 SR3.

Not in a box yet, and running a little low in Bias, 130 V AC input and a 4 diode multiplier so maybe 520 V.

I haven't quiet finished the power supply as I intend to put a voltage divider to make measurements as suggested by "jer"

I have to say that the sound is definitely the best yet.

I have been thinking the box the energiser is in is rather attractive, and maybe I want the bias in a different box, with the mains isolation transformer. This should require a nice plug and socket. (For now I will use the two bolt holes in the bass that could allow cable in.

Also looking closer in the the possible plugs I found the plugs and sockets used in the SRD-5 and SR-3 they are using the IP40 / IEC 60130-9 which is rated 250 VAC so on the edge of its rating. IP68 seems to be a better plug to use for electrostatics. Is there a non proprietary electrostatic headphone plug socket combination?

I will now have to do comparisons between the driver transformers especially considering how easy it is to do such tests now. Measurements to come in time.
Dear all,

Some of my Software Engineering work has been taking some of my time.

So what have I been doing relevant, on the headphones. Well not as much as I was. I have just stripped down the SR3's to the driver which I did not dissemble. The construction of the driver is it seams a set of moulded plastic bolted together. The stators are thin perforated sheets that look like they have been stamp pressed in a simple mold, a bit like a crown top bottle. These are suspended on the perspex like plastic that surrounds the assembly.

The world facing side of the SR3 is what looks like high grade injection molded plastic shell with metal grill with plastic coating that I think looked modern in the 1960's. Behind this metal grill is a layer of what looks like cotton. and then a layer of what looks like wool. This is then a covered with a course material holding back the long haired wool and then you can see the metal stator. On the ear facing side there is a protective plastic membrane I guess to stop sweat getting near the stators. Then there is the angled plastic shell with a series of holes you have undoubtedly seen on pictures of SR3's and finally the plastic imitation ear pads that mount on your ears. If these where my main headphoens they would be replaced with AKG ear pads as I much prefer my ear not to be touched by the headphones.

Now the head band on the SR3 is wonderful and can be pop-ed out to spread the weight of the phones across your head very lightly. The metal headband on the SR3 is impressive and is much better than the Lambda Pro's. The spring pulling the headphones toward your ears is not adjustable in force, but feels fine to me, I do think my AKG 500 headphones are more comfortable and its only due to the ear pads, in every other way I think the SR3 outclasses my other headphones in looks and comfort.

The electrostatic drivers, are attached to the wire nicely using brass contacts, in a way no one would bother today for consumer electronics, it feels more like something that comes from a yacht chandlers. (Shop for buying bits for leisure boats)

Its amasing how much grime these 1960's headphones had on them. I should have cleaned them before electrically fixing the SRD-5 I thought as I started to clean them. I then washed everything in grappa. Cut Cleaning Alcohol is not so cheap in Germany as bad grappa. Then feeling there was some grime still on it I put the plastic parts that could go in for a cycle of the dish washer clean them nice and new, and it worked very well, and but some bits will wait until I rewire them.

I have 2 incompatible systems that cant talk to each other so I cant separate out the sound of the headphone and that of the driver this is due to someone wireing the SRD5 + SR3 with incompatible plugs and sockets. My overall impression of the two electrostatic headphones is that the Lambda Pro + SRD4 + Home made bias, sounds more authoritative in the bass and is just more real.

It cant be denied that the SR3/SRD-5 is one amasing headphone, It sort of sounds a bit like my factory reconditioned Quad ESL 57's, great what Bass we have but less than you want. Bass light but you mostly forgive, for the lovely mid range. In comparison he Lambda Pro with my customised SD4 seems to have lots more bass, but the differences reduce as the mid range is reached, with the Lambda Pros sounding more natural.

I notice that the SRD-4 has bigger transformers than the SRD-5. They both look too small. This is especially true when I compare to my 8 transformers all 8 30VA bolted to a board I made for testing, which I must do. I would be surprised if the Stax transformers are 10 VA in power per channel, they are smaller from memmory than a 1960's 8W PP tube amp I had.

The SRD-5 is clearly 1960's in design while the SRD-4 is clearly 1980's in design. Odd they made such an inconsistent numbering.
Over all I am very impressed with the SR3/SRD-5, and might experiment with some bass boost but I suspect the SRD-5 transformer is limiting performance more than the SR3 headphones.


Dear all,

I have not yet gone the perfectionist route and started on direct drive but I am seriously thinking about it now, but since all suitable NPN transistors seem to be going end of life I think the output might have to be a tube/step up transformer or both :(

But lets not get distracted from my current mission which is to make my 2 SRD-4 drivers have a charge Bias. In the end these are relatively easy to connect the center tap to a resistor. and that to a terminal block.

At first I was shocked by the SRD-5 from the mid 60's taking the AC direct from the mains. Then I thought about it and wondered if their really is any risk?

Is this circuit safe?

The values are
R1,2,8 = 200K ,
All other resistors 10 M,
All capacitors 0.1uF.

R8 makes a potential divider, which indicates the output is 580V across R8-R10 so I assume on the output also.

The advantage is almost no stray magnetic fields for the power supply, and I don't need to order any transformers.

If the circuit is not safe, and I need to order transformers, I had thought:

Transformer 2,0VA, 230V/6V/220mA, ta40/B, 36,0x28,5x30,0mm, 100g

Transformer 2,0VA, 230V/12V/170mA, ta40/B, 36,0x28,5x30,0mm, 100g

To step the voltage down to 110V, before the voltage allowing more fine control of the voltage.



  • volageBiasStax.png
    40.9 KB · Views: 203
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.