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.
Dear all,

I am currently building a driver for a Stax Lambda Pro before I start work on my own electrostatic headphone. the first version will be built to low cost, a later version may be more Super HiFi. The headphones come from ebay are in beutiful condition but I assume they are working fine, I dont know yet and getting something working (and cheaply) tested is the first objective.

This is the first installment. It would have the title "Voltage doublers, how many stages?"

I am basing the voltage doubler / "ladder", on the stax circuit. ""high-quality-stax-headphone-adapter""

Now I just happen to have lots (8 unused) torroidal 230 V -> 2X22 V transformers from a bargin ebay buy, but they do look rather old and the sticky tape sealing then is loosing its sticky but they are not so bad as to need electrical tape. I thought I might play with these to drive the headphones in the short term. I suspect I will apresiate the knowledge gained.

So I thought to save my unused transformers I would build more stages on the voltage doubler and run the system off 44 V AC, well the doubler is less and less a doubler with distance from the supply. Tomorrow I will set up a second transformer effectively making a over built 1:1 transformer so will have 240V, and in series making 88V driving it though I suspect if I was buying transformers for the job I might want 100V like the original circuit as my estimates suggest the numbers would be right on the money for voltages. Some times it good to simulate in reality, much as I should learn how to do this on the computer.

Like the Circuit I attached, I used 0.1 uF capacitors (650 and 1000V) and 1000V diodes, but I guess each stage only sees the magnitude of the AC signal sent to the ladder so lower ratings woudl have been fine? I found little benefit after the first 6 diodes. The 7th doubler brings me to about 276V and the 8th makes no benefit at all.

I assume the reduction in effectiveness of the voltage doublers is due to the capacitors smoothing the AC more and more over time, does this mean the frequency of the voltage doubler / "ladder" would be very limited in maximum frequency of operation making frequencies like 1 Mhz out of practicality? Or possibly due to leakage in my breadboard (at such low voltages I hope not, but it was super cheap)

I see Quad uses 8 diodes in its power supply and a 10th of the size of the capacitors, used by Stax, so knowing Quad amplifiers and how they optimise for performance/cost I assume all 8 provide a benefit, so I doubt my experience with 6 stages being the limit to voltage doublers being practical / cost effective. I assume the same frequancy of supply (50Hz) with much lower capacitance's and closer to ideal diodes would improve the stages efficiency in driving an ideal unleaking speaker.

All thoughts apresiated as I am nearly completely self taught in electronics and have a reasonable skill in maths.


  • 5c0055-71242.pdf
    211.8 KB · Views: 103
  • 2f989048_staxsrd7mk2schematic.png
    276.3 KB · Views: 772
Are the diminishing returns on the V multiplier a result of the meter (loading) you are using to measure the voltage?

Why multiply from 44 or 88?

See jpg

Even if you start and end with 230VAC it will be less multiplication than with 44 or 88.


  • transformers.jpg
    6.7 KB · Views: 700
I was using a Multimeter, thankyou, I should have thought of this. I also like your suggestion of using 240V as a good voltage to multiply. Is thier any reason to use much higher transformer outputs, I would imagine beyond 250V the costs just increase?

I have an old osciliscope (again from ebay) that I will have played with not enough and will report back. I read in many places that voltage doublers dont deliver currrent so I was prepared, just did not engage the brain.
Ok so my osciliscope gives aproximately the same answer DC wave with 320 V to about 270 V peaking so I guess its got the same order of magnitude of input resistance. the 7th is noticbly smoothing so maybe a bad connection rather than load resistance.

I guess I need to make an extremly large voltage devider with a lot of resistance for meeasurement or is thier a better solution?
Thanks Doug,

With a transformer driving a trasnsformer in reverse I can get all the voltage I need. It seems I have over 2000 V at the end of my ladder. Just two diodes and I reach the desired voltage. I have 4 X 2.2 Mega Ohm that I guess a simple rcrcrcr network might help with smoothing if it prooves an issue.
So now thats resolved I guess its time to work with the lethal step up transformer parts.

As I stated before I have a few 30 VA 230 VAC to 2 X 22V AC transformers in my bits box.

I am guessing I can just use the two, 22 Volt secondaries in parallel and connect these to the amplifier and use the mains output (230 V) to drive the stators?

Time to chop up some old molex plugs (from broken atx power supplies) until the Stax sockets come in the post for the next stage. This is a little scary as the voltages + power I guess could be lethal on the end of a 50 W power amp. Im going to use the Quad 306 power amp as it has some protection against overload and driving a short circuit that I know works.

I do now wish I had bought some designed for function insulation tape, something to buy on Monday.

I have some candles about the house, can I use any wax as an insulator or is only some wax (bee's wax for example) as a high voltage insulator?
It may work,But you may find you will need a higher step-up ratio.
You can add a custom primary to do this.
Head phones typically need about 1:20 to 1:50 transformation ratio.
Also you may need to use 2 or 4 cores to get a good low frequency response without distortions.

Keep in mind what frequency the transformer was designed for (50hz or 60Hz) and to get to half the frequency input 25Hz/30Hz with a 22v winding then 11v is the most you can apply to it or you will get core saturation distortions.
This will only get you 110v maximum with your transformer.

If you were to add a custom primary to give you a 1:40 ratio then your maximum output would still be 110V at 25Hz/30Hz but with only a 2.75Vrms drive signal and you would need 4 cores to get 440Vrms, the range needed to fully drive a headphone.
Although 100V to 220v may be enough for to listen to and much of this will be determined by the D/S and your bias voltage.

The Micro driver I made was the size of a credit card,

A Micro ESL Made From Old Credit Card's

It had a D/S of about .030" and used bias of 500V.
The transformer I used was from a EL34 P-P tube amp.
I didn't measure it but from the requirements of the tubes I am guessing that the ratio was about 1:25 to 1:35 or so.
It got to a comfortably loud level and was a pleasure to listen to.
Even the bass sounded great with no distortions.
Although that transformer was deisgned to go that low as well.

I have gotten similar results with my Mini Desktop ESL's at low levels and listen to them fullrange with no issues providing I don't exceed the voltage/frequency saturation requirements of the 60Hz designed transformers.

Yes, These voltage can be lethal coming from an amp that can deliver the current!!
So do be very careful!!!!!!!!!!!!!
I have been nailed by them before and they pack a Walop.!!!!
More so than a bias supply that has a current limiting resistor on it!!

jer :)

P.S. A good conformal coating to use is some Plain ole' Clear Acrylic Enamel spray paint.
I have tested this stuff to be good up to 1500V per mil and even as high as 2100v in some cases but definitely good for at least 1000v per mil (.001").
I used this on my HV supply and have had no issues yet at 13.6KV.
I also use Clear silicone rubber as well, it is messy to use and hard to get off should you need to repair something, but it is one of the best thing out there as well.
Last edited:
Some success but not HiFi yet.

Thankyou Geraldfryjr.

Your quiet right it does work, and yes I do think I need a higher step up transformer. 1:10 is quiet low volume.

When driving at distortion generating levels on the Quad 306 its an OK volume for listening in a quite house, so I assume a second transformer per channel may give me enough step up volume, definitely further experiments to come.

At first I rigged up the plug in the wrong way. A mirror image of what I want as the articles I read did not make it clear if they where viewing from front or back. The sound was fine on one channel but mains hmm was on the other channel. I doubt I could have broken anything as its not going to make the flash point no mater how you wire it I assume. After wiring it the correct way (see attachment I found on the web) the sound is stereo and I think louder and better.

The Bias is set to 620 V (a little over the targeted 580 V) but I think I will add a few filter stages to reduce the modulation on the Bias and they might bring down the voltage a small amount. At the moment I don't have any filtering on the bias supply.

Since I don't yet have a Stax socket to (yet as three are due in the post), I cut a small terminal block into 5 parts and attached them to the headphone plugs pins. In fact this construction is only terminal blocks to connect the components, and dropped solder on any multi core wire to avoid loose strands.

The treble is already good, and I do have a big smile on my face, only slightly effected by the bass distortion. Expect an update soon with the filtered bias supply.


  • 001nq.jpg
    92.3 KB · Views: 627
Very Good I am glad that you have them running!!

I was quite amazed the first time I powered up the Micro ESL and you have them in stereo!!!
I have yet to listen to any of my builds in stereo since 2003 when I first made them,He,he,he,he !!!

I was certainly impressed with how much clean Bass I was getting with only a .030" D/S.
I thought the mine had no treble at first maybe because the holes were to small and it turned out that it was the sample that I was listening to and what was lacking and not my system.

I made the Micro ESL in 2010 and only listened for few days and then I tore it apart to make the pictures.
I can reassemble it at any time though.
I have everything to make a second one as that was the plan too.
Since then I started a new one using some clear CD Blanks.

I am now back working on an affordable DIY transformer, as that is what is so needed right now for all!!


Dear all,

I put in some filtering on the Bias (rcrcr) and I have a big smile on my face. This is now sounding good the bass issues where mostly due to Bias modulation.

I can now play quiet loud for about 2 mins until the Quad 306 gets upset I then have to reset its protection circuit. Keeping the volume down to low levels give just ace sound.

So it seems to me I should start to look into drive circuits and a nice box.
I also have a Stax Labda Pro and I had before de SRD-7 unit. Today i use a home made transformer with a amorphous c-core.

Here some measurements i made from the original transformer and my homemade transformer.
An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Dear esltransformer,

That is very interesting data though I am not sure I understand it all yet. It seems you have made an impressive transformer though, and I would be interested in what you thought of off the shelf alternatives as winding a transformer seems to require tools and I already have a lot of tools when I move house next (which might be quiet soon)

The original transformer is 1:50 has a -3db point at about 10 Hz under the test graph 1. It would be interesting to know the power you where using, are situation Blue and Red with and without a 1 ohm resistor in series like graph 2? It is also interesting to see the 1 Ohm series resistor and how this improves things, but again would be interesting to see the power involved. The final graph I think explains that the capacity for bass and how the transformers can saturate and how yours passes more VA before saturation at 10Hz. Please correct me if I misunderstand as I am new to all the details.

Since your graph states clearly that the transformer is a 1:50 ratio I can now look at transformers to buy in an informed way? Do I still need the input peak current and peak voltage delivered to your headphones at a little too loud volume, or does the Bass Responce graph tell me the VA I need?


Driving options and thoughts with opinions appreciated.

I have just bought at bargain price (52:90 Euro) a probably broken Stax SRD-5 and SR-3 so I could experiment with rebuilding the headphone and look and measure the transformers as esltransfomer has done above. I dont think I would have saved my self 50 Euro had I seen your notes even if the most important measurement I needed was the transformer. This said having a driver to refurbish will make the process of making a DIY driver for the newer Stax easier.

I assume that others have more data than I here, so I will state what I think I know and what might be valid conclusions.

0) Stax amplifiers headroom is highly variable, increasingly reading on internet suggests people are planning peak output should be from +-1000V down to as low as +- 100V. No clear concerns value is available, so maybe I should assume around 100V RMS and about 500V peaks.

1) High step up ratio's introduce lots of extra costs to transformer manufacture when passing the full audio range.

2) Direct drive amplifiers with Valves or transistors are scary high voltages that unless used by class A amplifiers will need large amounts of capacitance to cope with instantaneous power demands, the same sort that kill people who make hobby electronics with a beer in the hand.

3) Power transformers with greater than 240 V secondary cost a lot new.


(a) It seems to me that unless you have a well developed product or intend it to sit in a wax filled box direct drive amplifiers are very lethal to make mostly due to the power supplies running at +-500V regulated which are not things I want to go wrong.

(b) Transformers lend them selves to electrostatic amplifier circuits as they block DC and provide a simple way for push pull stators to behave as you should like. This said as the step up ratio or power or bandwidth increases so does the cost dramatically.

(c) So the conclusion for a little safety and for ultimate HiFi would be to use a high voltage amplifier and then boost the voltage using a transformer using a small step up ratio. So a superfi solution would be to drive the transformer with +-250 V and use a 1:4 step up transformer. I suggest it might be possible to use a simple circuit like Hiagra, Sugden or JLH but with much lower Bias and about 3 W output. Much like this valve amplifer..

For a price efficient solution some of the budget valve to speaker output transformers available on ebay / Farnell might do and cost a lot less time and money to put in a box. A lot of single ended unspecified (apart from what tube drove them) exist on ebay from old Valve gear, probably most have not enough capacity to play lower than 50 Hz.

What factors should I look for, I am guessing 4 W handling down to 20 Hz, I assume 1:50 step up is required and at least 1000V rated or is the high voltage on the output too high a specification, as I cant see anything that matches this spec?

Why do you think a transformer with lower turns ratio is beneficial? It will need to have far more primary turns; as a consequence more difficult to manufacture. And come with the same problems as a low-ratio transformer : core saturation, parasitic inductance & capacitance, etc. Furthermore a common amplifier can't be used.
For an electrostatic headphone a couple of good quality mains toroids(that we use for ESLs) of lets say 230:20V should provide excellent results IMO, especially because a lot of problems are solved by much lower turns ratio needed than for ESLs.
I(and some others) have done measurements of those in the past; my conclusion was that when driven from a low impedance source a quality main's toroid can provide harmonic distortion in the same order or even less than electronics, and with bandwidth flat to 20 kHz.

Hi,Why do you think a transformer with lower turns ratio is beneficial?

Ok I am either confused, or misusing terminology I or both so thanks for helping.

It will need to have far more primary turns; as a consequence more difficult to manufacture. And come with the same problems as a low-ratio transformer : core saturation, parasitic inductance & capacitance, etc.

Thank you for taking the time to challenge these wrong assumptions, it drastically changes my conclusions. This is interesting :)

So since theirs less primary turns in a reversed mains 240:20V torrid, a quality conventional mains transformer reversed is likely to to be better at keeping resonance due to inductance and capacitance under control to well beyond the audio range of say 50 Khz or higher that was always a problem for valve amplifier designers in the 1960's.

For an electrostatic headphone a couple of good quality mains toroids(that we use for ESLs) of lets say 230:20V should provide excellent results IMO.
They seemed to but mine dont, lets explore why they dont, later so as not to get distracted here.
especially because a lot of problems are solved by much lower turns ratio needed than for ESLs.
OK now you seem to me to be contradicting your self. maybe some clarification would help as its not clear to me now what was said before. Are you saying because I don't need 1:125 step up ration its not an issue, but the difficulties are not hard to resolve with 1:50 and so no harder to resolve than 1:10 or 1:4. I think I read these Electrostatics have 0.5 mm DS spacing and are Bias around 580V

I(and some others) have done measurements of those in the past; my conclusion was that when driven from a low impedance source a quality main's toroid can provide harmonic distortion in the same order or even less than electronics, and with bandwidth flat to 20 kHz.

I honestly believe you are correct. and would love suggestions available in Europe?

How about this?
How about this? data sheet attached. I cant help thinking that a step up ratio of 1:10 (22 : 230) is not high enough particularly in light of the measurements saying 1:50 and how loud the Quad 306 would be playing with normal speakers with my 22V:230V experiment.

At this moment as the amplifier is unhappy, is this because:

0) How stupid to run both 22 V secondaries in parallel? (I imagine this is not the issue)

1) The 30 VA transformer is saturating on the bass? If so what should I expect to use 50 VA?

2) I have no series resistor and we hit a resonance issue and current draw on the amp is high enough to click in protection.

3) The load presented by the speaker is too low when stepped up just by 1:10 and I need a higher step up ratio, so the amplifier is very inefficient is struggling to deliver the required current for such high voltages and I am listening to the linearity of the protection circuitry?

I suspect it might be all the above.

I think its time to put a resistor in series with the primary and watch the amp output. And look for a suitable result form your testing thread :)


  • 2n310.pdf
    27.2 KB · Views: 50

1)Internal and load capacitance seen at primary is increased at turns ratio squared. So, assuming the same amplifier is used, one with higher ratio is more difficult to drive. 1:50 transformer is a lot easier to drive than 1:100. But 1:100 transformer will provide 6db more output! So no surprise it's requesting 4 times the power.

2)It's very likely that saturation of the core will be similar , but varying to a degree between 30, 50 and lets say 100VA mains toroids(assuming the same voltages for sure). A smaller transformer has simply much more turns so the difference in core saturation point will not be large.

3)I do not think that reaching 50kHz is feasible or even needed. Yes, I a bit overstated "flat to 20 kHz". I have measured several toroids and they had
1-3 db rise @20 kHz.Resonance was > 20 kHz for all samples. But that's easily corrected with a resistor in series with primary or secondary and IMO is not of too big of a concern.

4) In general, the fact you are running into problems with 1:10 turns ratio suggests there is something else going wrong. For example, I am running ESLs with 6 toroids per channel for a total turns ratio 1:115. At 20 kHz the impedance is quite low, about 1 Ohm or even less. Despite the fact these transformers are rated at 230:12V, I can even run them full range to quite loud levels before cores start to saturate. And the amp is nothing special.

5) You can connect secondaries in parallel of the same transformer, but only if they have exactly the same amount of turns and voltages are very closely matched. Otherwise it will draw large amounts of current.

Some hints :
Have you measured DC voltage at output of your amplifier ? A DC voltage of few tens of millivolts could cause serious problems with core saturation. Toroids are especially vulnerable to this.
Is the amplifier stable ?
Basically a 22:230 volt toroid should start saturating at above 22Vrms@50Hz. Maybe@25V or so seriously. If that won't work then there is something else going wrong.
One of my old measurements is preserved in this thread :
Later I found substantially better transformers.
I am using ones from a firm called INDEL, it's from Poland so might not be widely available.
basically when hunting for a toroid you should be looking for units that look quality built(wound nicely, etc). I have found this can have a positive correlation with measured performance.

Very good description Lukas,I have typed that stuff out so many times My fingers hurt!!! He,he,he,he

Two things that could happening probably already mentioned is that the amp may not like the low DC resistance of the primary winding.
I have to use at least a .5 to 1 ohm resistor in series to keep my Crown DC300AII stable at idle.
Right now I am using 1.1 ohm resistor plus the primary resistance of .4 ohm.
With 25milivolts of idle voltage this means there should be about 16.6 Miliamps of current flowing through the winding (confirmed 13ma with my meter).

This is not the original winding,
The original winding had much much less resistance lets say .01 ohms and if my offset voltage was still 25mv then the current would be closer to 2.5 amps of current running through the primary at an idle state!!!

This may be why your amp is self protecting.

Second, The self capacitance of the transformer may be quite high.
Typically a smaller transformer will have a higher capacitance due to having more total turns than a larger one will, up to a point due to physical reasons.

I have been testing a Antek AS-1206 lately and because of its high capacitance it will shut down my amp due to overheating at only 14Vrms (20V peak) at 20Khz and much quicker at 40Vrms ( 56.5Vpeak) while it is drawing about 3 or 4 amps with no extra capacitance load on it.
At 40V peak (28.Vrms) I measured something like 2 to 3 amps at 20Khz with no extra load capacitance and this was about 80 to 120 of power that was being wasted in just the transformer causing the amp to heat up rather quickly.
This is just using its own self capacitance as a load.
When the HV windings where in series it was double this and was when I noticed how quickly the amp was overheating and shutdown.
The amp has no fan on it so it gives me a good indication of how much of a load is on it before I get a chance to measure it.

Stay away from the Shielded types as I have found that the shield adds to the total capacitance and can arc to it due to inadequate insulation and short some windings, as I have just experienced.
Luckily I was able to repair it this time and continue and I will present all of my data on this core soon!

jer :)

P.S. I got very similar THD measurements as you did Lukas with my two samples of cores that I am using,some were even lower depending on what I used for a signal generator.
My computer has about .005% THD and my signal generator is about .05% THD.
And I got only slightly higher (10% more) than either of the sources at full output as measured from the transformers output.
More on this later!!
Last edited:
Dear Lukas and geraldfryjr,

Thanks for the feedback, I will look into this soon, and try a series resistor with the primary and a second transformer but not tonight simply because I went drinking with friends tonight for a couple of hours, so no games tonight with high voltage as I want the morning to happen for me too :)

Thanks again

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.