My ESL attempt

[SM7UYJ]

I’ve had plans to try to build a pair of ESL speakers for some years now.
Finaly I have got all the material I need and have done some experiments with different coatings, spacings etc.

The plan for the speakers are as follows:

Use two ESL-panels/speaker and 2 (or 4) 12” woofers for the low end.

I have built the ”mid”-panels from 1000x400mm alu.sheets, 2mm spacing and 12µm Hostaphan film- polymercoated.

The ”tweeter”-panels are 1000x200mm, 4.8µm Hostaphan – but not shure what spacing to use and if i should divide the panel into 2 or 4 cells?

Using two toroids(80VA) as audiotransformer (ratio 1:80) and have abt. 4kV on the stators.

If done right, what SPL can one expect from a panel of this size and how low should it be able to play?


/R

[SM7UYJ]

One test panel playing.



Changed the diapragm support to 5 "dots" in the 2:nd version!

/R

[jjbunn]

Nice! The panel looks great.

What HV are you using, and most importantly: what do they sound like?

[arend-jan]

It's hard to predict actual SPL levels with any accuracy. Theoretical SPL levels can be calculated (see for example my simulator at quadesl.nl).

How low it can play depends on the size and tension of the diaphragm. It's a balancing act to get good SPL, sensitivity and low frequencies. Not much options but to go trial and error on this one.

For the tweeter panel 1mm spacing (stator to stator) would be enough. IMO 200mm is too wide to get good horizontal dispersion at high frequencies. You'd rather have something 1/10 of that.

I wonder what kind of polymer coating you are using. Can you tell us a little more about that?

[SM7UYJ]

I taught I had abt. 4kV on the stators, but practical tests point to a much lower bias!

Have a HV-probe on its way, so will measure it shortly.

I'm not looking for an absolute SPL, but rather in what region DIY panels tend to land!?

The coating used on these panels are from MT Audio.

I have experimented with graphite, soap, fabric softner, glue etc.

Next test will probably be Elvamide doped.

The "tweeter panel" was planed to be divided into 2 or 4 sections verticaly.

/R

[arend-jan]

Do you have 4KV of audio signal on the stators, or do you have 4KV of bias voltage on the diaphragm?

If you hook up two toroids of 1:80 in series (is that what you are doing?) then you can get about 3KV of audio drive (depends on the amplifier of course, I assumed 50W in 8 Ohms)

So how high is the bias voltage?


You can devide the tweeter panel into multiple cells, but if you are going to drive them with the same signal, then it won't help you to improve dispersion. I think a strip of 3cm wide along the side of the panel in the picture would work nicely as a tweeter.

[Calvin]

Hi,

if done correctly, You could expect SPLs of >100dB@4m with the smaller panel in a freq-range >300Hz.
You should get lower SPLs with the larger panel in the same frequency range, because of the larger d/s. If done properly the frequency range won´t extend to lower frequencies anyway! The thicker membrane will in fact lead to a higher fo because of its superior tensile strength and therefore much higher posible mechanical tension! So the larger d/s is the wrong way to go, since You can´t use it for good! This is a very typical failour nearly every beginner makes. Smallest possible d/s needed for the lowest sensible fo is the way to success

jauu
Calvin

[SM7UYJ]

The voltage was meant as 4kV on the diaphragm!
In reality i suspect it only to be ~2kV, but I'll find out when I get my HV-probe.

The choice of spacing was due to avalible double sided tape with good glue.
I now also have 1,5mm PVC and 2mm old Bakelite strips avalible.
I'm tempted to try the Bakelite as it would make a real "retro" ESL - I hope the smell after cutting the strips fade with time!

Is there a good formula for the tension off the diaphragm or is it a trade secret and eaquals trial and error for me?

The info I have found point to everything from 0.8kg/5cm to "tight as " ?

On my present panel i tensioned it to ~3.5kg/10cm.
With 2mm spacing I can get it to touch the stators on some bass notes at medium SPL levels, but with i HP-filter I can crank my GainClone almost to max (~50W).

Right now I'm using two 1:40 toroids - so the 1:80 is for a pair!

/R

[arend-jan]

There is no 'best' tension, it depends on what you want. There are two opposing factors involved:

1. the higher the tension, the higher you can make your bias voltage (membrane stability), the higher the sensitivity of the panel.

2. the higher the tension, the higher the resonance frequency, so you lose low frequency output.

Stator spacing also effects sensitivity a lot.

So you have to balance these things to get what you want from your speaker.

[arend-jan]

Another tip: get a cheap microphone like the Behringer ECM8000 and hook it up to your PC. Run ARTA to make measurements. This will give you a lot of valuable info!