Advice wanted

I am planning on building my first ESL's. I have eventually sourced perforated mild steel sheets and 12 micron Mylar (the thinnest currently being imported into South Africa where I live). I think I will use 3m double-sided tape for spacers and liquid soap (at least initially) for coating the diaphragm. Ideally I would like to have an ESL panel/panels covering mid/high and a dipole woofer for the base (crossed as low as feasible given the other constraints).

I only know what I have read on the net. I am sure that the books on the subject (e.g. by Roger Sanders) would be very useful. But if I do order one, it would be weeks before it arrives.

Deciding what design to go with is not an easy matter. I would like a sufficiently large sweet spot for at leat two listeners at a distance of about 4m. For this reason, it seems to me that I should be making the panels tall and narrow, not wider than a foot (maybe as narrow as 3 inches) and about five feet high. Is is better to make one panel or stack two smaller ones?

As regards the stators, what hole size is optimum? I can choose almost anything. I was thinking around 3mm, but this is completely arbitrary. Also what thickness is best? 1mm? 1.6mm?

Almost all of the ESLs I have looked at have no added baffle width. Would this not be useful to cancel dipole cancellation at low frequencies, particularly for narrow panels? I haven't done the math, but maybe 6 to 12 inches added on either side. Does this give unwanted baffle step or diffraction effects? This also raises the question of how thick the baffle can/should be.

My biggest obstacle at this stage is the audio transformers. I have not managed to source any locally and I think it would probably be significantly cheaper to get them custom wound than importing them. The local import duties and shipment costs are pricey. Are technical details available anywhere or can somebody help me on this? I would like to stick to one audio transformer per side.

Any advice the list could offer would be much appreciated. I wish I had endless supplies of money and time to spend on research and experimentation...

As an aside, has anybody ever tried using a "digital" amplifier with an internal step-up transformer? It seems to me that such a beast could be directly coupled. The negative feedback would come from after the transformer (via a resistor divider?). Wouldn't this give much better control of the speaker? Presumably a lot of the LC effects could be greatly reduced. Switching noise at these voltages may be an issue though.

ESL advice

I can see that you are in the same boat as I was 3 years ago,
not knowing too much about building ESL panels and reading all
I could find about it. Another issue I had in common with you was
the "budget" wich is to be taken in consideration in an endeavour
like this.
My spacers were done in PVC also 3 mm thickness. PCV is an
excellent material to use for spacers, they are practically immune to temperature
changes and humidity. They are not cheap though... However they
can be very forgiven if you are to disassemble the panels, because
you can always to get rid of the glue used to glue the diaphragms.
By the way you should use Polyerethene glue which is powerful and
also independent of humidity. Actually it works better with a certain
degree of moisture...
I made my panels 12 inches wide and around 5 feet tall, that worked
very well for its sonic qualities, however I segmented the panels
puting two 2,5 feet panels on top of each other. I really don't
recommend to make them narrower than 12 inches unless you
want to make a tweeter element which in this case should be 2 to
2.5 inches.
This big issue and where the beef is, the audio transformer. The
prices are astronomical if you are to consider a very good quality transformer. My option was to use two transformers from a
delapidated tube amplifier and even thouhg is not the very best
as the turns ratio are considered, it suits the purpose and I
got good results with them. So, if you have access to an old and
beaten tube amplifier this is the way to go, the transformer is
wired backwards the secondary to the output of the amplifier,
the primary with the center tap to the minus side of the HV
These is my modest contribuition, I hope you are to succed as
I was even though with a lot of frustation and difficulties, as
you can imagine.
More questions re ESLs

Thanks for your input John. What performance do you get from your ESLs in terms of frequency response and sensitivity? Do you cross to a conventional woofer? 3mm spacing implies that you have some fairly low frequencies in your panels.

As regards the last comment in my post re switching amplifiers, I went and actually looked at the Application/Test Circuit for the TP2150B. The feedback comes before the reactive components, not after. Nevertheless, it still seems to me that a Class D amplifier would be far more suited to ESLs than a class AB. To use a term not normally associated with audio, ESLs have a very low power factor, i.e. a mostly reactive load. When driven by a class AB amplifier, peak currents occur when almost the full rail voltage is across the output devices and hence far more power is dissapated in the amp than the panels. The efficiency improvements of a class D amplifier would be even more pronounced than with conventional cone drivers.

Presumably the inductor in a class D amp would not be necessary as the inductance could be provided by the audio transformer.
Also, the power supply would need to be a lot less beefy. I am not sure if I am missing anything here.

I still have questions I would like advice on from my first post, including:
- optimum stator hole size
- increasing the baffle width
- audio transformer winding

ESL questions

Sorry for the delay, things have been quite hectic around here.

Yes, I cross over the ESL to a subwoofer around 200 HZ and this is
acomplished by an active 24 db crossover even though I really don´t
quite like it, when I take the crossover out of the circuit the sound
is much more transparent and the frequency response seems to
be better. Even though the crossover has top quality components
(high quality IC´s and polypropylene caps) the sound is quite muffled
compared to running the ESL directly from the amplifier. One thing
that I have to try is to use a passive line crossover but that´s another
story all together.
The hole diameter on the stators is 3,5 mm which at the time was
what I could find, however even 4 mm is alright.
Of course you can always increase the width size but from the pratical
point it´s something that only you could determine, I guess that
increase could lead to a lower response with the same 3 mm spacer
Transformer winding is something that I don´t have any experience
with and I think it would be a difficult endeavour if you are not
really experienced dealing with transformer windings. We are
talking here about a LOT of turns believe me... Like I said before
your best bet would be to buying already made trafos even
if you consider its relatively high price. Gary Waldron had trafos
for sale sometime ago for $48 each, designed specifically for
ESL however this has been a while ago and when I tried to get
a hold of him to buy those his e-mail adress wasn´t operative
anymore and his web site was also outdated. I don´t know
what happened and I´m convinced that it is not up and running
again (has been a while...)
Of course the other alternative is a tube amplifier transformer...

Now the class D amplifiers.
Sometime ago I bought one of those class T amplifiers, the
Sonic Impact job that used to sell for around 20 bucks. The
amplifier sounded really well with dynamic speakers even though it didn´t have too much power. I thought I would give it a try with
my ESL, well to make the story short, it blew right before my
eyes! I guess it didn´t like the ESL, since then I´ve tried to get
some input from this forum as far as other amplifiers including
the UCD modules and others. Nobody is very positive about it
and I don´t think I´m going to risk a few hundred dollars just
on a experience in which I don´t have the minimum assurance
that is going to be succeful.
This is the whole story about the class D amplifiers if I had a
lot of money to waste I would do it but unfortunately I can´t.
The same thing applies to Class AB amplifiers, some (the majority)
cannot drive the ESL others may be and I don´t want to risk
anymore of my stuff, I rather stick with the ones I know very
capable of driving them, the tube amplifiers and class A amps.
I have a JLH 40 watt amplifier that drives the ESL beautifully.
If I could give you anymore details of what I know about ESL
please feel free to say.
much power
ESL questions

I forgot to tell you that the way they are the response of my ESL go
down to about 55 to 60 Hz running them without the sub you can
ear low frequencies around those but of course they don´t have the
same level as the other, they are still useful but not enough. This
is the reason why I´m using the sub but of course I already knew that before I built the ESL. I ran a test (several tests) with different confi-
gurations and my conclusion was that I would have to use the sub.
I hope this helps.
I was originally considering crossing at around 200Hz as well. But I am sure it would improve things if I could cross to 100Hz or lower. Obviously this all depends on the ES panels and what kind of performance I can get.

By my reckoning, dipole cancellation (from omni-directional drivers) should start at a much higher frequency than the 200Hz you are crossing at. It is obviously negated by the directionality of the panels.

I am not sure what kind of excursion you get at 100Hz, but would it not be possible to add some equalization to extend the low frequency response?

As for transformers, I had the same attitude as you until I started reading dhenryp's thread on has ribbon array. He would he own transformer from apparently very little experience. Granted, it was very simple in comparison but I did a little more reading and it definitely doesn't seem to be an impossibility to wind my own, although it might be very time consuming.

I was concerned about the quality of transformers in the $50 price range. From what I have read, the transformer is likely to be at least as great as source of distortion as the panel. The transformer seems (I may be way off the mark here) to be the critical item in the whole ESL. I could potentially hand wind a toroid with a few (hopefully not too many) hours of labour. I would presume the performance would be better than an E-I core, which seems to be the standard. I need more data on the subject and it is not very easy to come across.

Class D amplifiers still seem like a good solution to me. Why yours fried is a bit of a mystery. The ESL is a far more complex load than a 4/8 ohm cone driver. Maybe some additional compensation is needed to prevent instability. I may ask this question elsewhere.

ESL build up

Hi Bret

Yes, it´s definitely possible to add equalization to the low frequencies
but be careful with that, you don´t want too much boost otherwise
is going to affect the performance of the ESL exactly on the excursion
of the panels, too much is too much for the panels. There are schemas
of that on the Web (the audio circuit is the place to check). I really do
not like this type of boost I always see this arrangement as an unatural
way of modifying what seems to be natural coming out of the panels,
but that´s me talking.
Yes, you´re right, the cheap transformers have always the drawback
as However the ones I cited from Barry had a very
good report quality/price and that was mentioned by various people
especially on that site Actually if you
have never visited this site it's a very worthwhile experience to do
is a wealth good source of information on ESL. When you open it
click on Loudspeakers and again on Electrostatics, you're going to
find lots of information.
On the class D subject I sugest you ask in this forum again even though
I couldn't find definitive answers, some people responded but with
no assertive information that I could use. May be you get lucky...
Last but not the least I've read somewhere that toroids weren't very well suited for electrostatics don't know why, but don't quote
me on this, it's something in the back of my mind that I can't
Hammond P-T1650F

I think you should choose your transformer (since you decide to go
for a tube transformer) with a bit less primary impedance because
this will translate in more turns ratio.
The P-T1650K has an impedance of 3,400 CT ohms, is rated for 50 W
and even though it´s about $20 more expensive it represents a
better value over the F version which is only for 25 W.
I don´t know exactly the prices of these transformers in South Africa
but I figured these will be similar to the ones in the States.
Hi John,

I am not sure what you are getting at in terms of turns ratios. Remember that you have to use the audio transformer "backwards" and so the higher the "primary" impedance the better.

1650F impedance ratio 7600:4 turns ratio ~44:1
1650K impedance ratio 3400:4 turns ratio ~29:1

i.e. the 1650F has the higher turns ratio, which is why I suggested it. How much power rating do I need? Traditional numbers don't really seem to apply.

I can understand that you would want a high power amplifier for two reasons:
1) Bigger output voltages require lower turns ratio transformers for the same stator voltages (or conversely, higher stator voltages for a given transformer)
2) The load is mostly reactive and hence the amp has to dissipate a lot of power internally (for a class AB)

But what rating transformer is actually required? Mark Rehorst suggests 15W to 20W for 30Hz output on his web page, which is what I based the choice on. Presumably larger is better, but how large?

From my understanding of the physics, ESLs should actually be quite efficient in terms of electrical to acoustical power conversion. (They may have low sensitivity, however).

Audio transformers are somewhat unusual in that they give ratings in Watts rather than VA (which is normally the important value). ESLs may have low W but significantly higher VA requirements due to their low power factors.

Any input I can get on this subject would help tremendously.


The power which a step-up transformer can handle depends on the frequency. You can have a rated 50 Watt transformer at 50 Hz, but at 20 Hz it can handle only around 10 Watt. Putting more power in it results in (core) saturation; distortion!
Is this 10 watt at 20 Hz enough? It ususally is because most music has little bass content below 50 Hz. If your ESL has low efficiency and your playing organ music on high levels, than you have a problem.
My advice is to make the spacing rather small for high efficiency.
Specially in case of limited step-up ratio which the old tube-amp transformers usually have.
Best regards
Audio transformers

Hi Bret

You are absolutely right, somehow I got confused about the two
transformers in calculating the turns ratio.
Np/Ns = Square root of Zp/Zs gives exactly the numbers you mentioned
43.5 for the F and 29 for the K, I must have swapped the numbers
somehow. This way the F sufix is the suitable one for the ESL as you
stated provided the impedance numbers provided by the manufacturer
are correct, even though the turns ratio is not the best suitable for
this purpose, it must be a little higher according to what I've read
since a few years ago. But anyway it's acceptable as mine proved to be.

I've read yesterday an interesting elaboration on the reason why
certain amplifiers - class AB - are not suitable for driving ESL speakers
this was the first time I've read something convincing about the
subject. In another posting on this forum I asked why the amplifiers
especially class AB couldn't drive the ESL without being destroyed
and I never got a convincing reason. Now, by accident, I came across
something that makes some sense to me. And I quote an answer
that was given by Hugh Dean of Aksa amplifiers to the question of
a member of his forum.
"The PP SS amplifier, both class AB and A, has one mortal enemy.
Amp Kriptonite is called cross conduction (CC). CC puts one rail
in direct communication with the other - we can forget about the
speaker here - via the output transistors. A horrific current flows,
aided and abetted by the largish capacitance at the rails, and the output devices lay down their lives in the electronic equivalent of a massive auto burnout. So, what causes CC? Oscillation; uncontrolled switching of the output devices precipitated by phase shifts across the global feedback network. Negative feedback
MUST remain NFB, and when it strays into the PFB region, the amp oscillates in short order."

I really must confess that class A amplifier do not suffer from this
prolem, my JLH is an example of this, it drives my ESL beautifully
without any strain whatsoever, by what concers class AB this
explanation is more than sufficient to me.
If you want to read more about the subject check this site,
Sorry about this long dissertation but I figured that you would
be interested to know about this.
ESL and transformers

MJ Dijkstra

You've got a point here, the power of a transformer is directly dependent on the frequency. But the same reasoning applies to a less power trafo so I think that the more power a trafo has the better it is.
Of course we should always consider a small spacing in all circumstances and that's where the tube-amp trafos come into place
with reasonable results albeith its less than ideal turns ratio.

Conclusion? A higher primary impedance and high power rating will
do the purpose, right?
Re: Audio transformers

jmateus said:

I've read yesterday an interesting elaboration on the reason why
certain amplifiers - class AB - are not suitable for driving ESL speakers
this was the first time I've read something convincing about the
subject. In another posting on this forum I asked why the amplifiers
especially class AB couldn't drive the ESL without being destroyed
and I never got a convincing reason. Now, by accident, I came across
something that makes some sense to me. And I quote an answer
that was given by Hugh Dean of Aksa amplifiers to the question of
a member of his forum.

That's very interesting. So the class AB amps I've had driving my ESLs all these years are actually blown? I never would have suspected- they sound great! Hmmmmmm.

ANY class of amp can safely drive ESLs, but it must be designed to remain stable with capacitive loads. Some commercial amps are and some are not. I have found Carver's cheap solid state amps to be unstable with ESLs, for example. I have also found cheap Soundcraftsmen amps to be perfectly stable with ESLs (and ribbons, and every other type of speaker I have ever connected). A cheap, 80s vintage Technics amp I have is also completely stable with my ESLs. None of these are class A amps.

Don't let talk of amplifier instability stop you from making or buying ESLs. There are plenty of options ranging from cheap to expensive to drive ESLs.