ESL woofer- anybody game?

[stokessd]

They guy to ask here is Sheldon Stokes- He repairs Quads on the side.

Did I hear my name?!


The original Quad ESL had the conductive coating on the outside of the bass panels. They were not made of circuit board material, but hard plastic. Including the stator thickness, the diaphragm to stator spacing is about 0.2"

The Quad ESL63's use a thin circuit board like material with the copper conductive regions on the outside. The D/S spacing is about 2.5".

I tend to agree with Calvin, an ESL woofer could be made to work, but it's far from ideal. There are two big issues with an ESL woofer (I should say subwoofer, because we are concerned with the bottom couple octaves here).

1) the ESL is a dipole, and it's Vas is enormous, so you can't really put it in a box. This means that the output is going to be rolling off at about 6 dB per octave starting when the wavelength reproduced is about 4X the shortest dimension. For my 20" wide ESLs, this equates to a low frequency rolloff starting about 800 Hz or so. I EQed the snot out of them and can get respectable performance to 130Hz, where I cross over to a very good moving coil driver in a transmission line enclosure (Seas Excel, in a 100 lb box) That combination can de-bone chickens with the dynamics. However, continuing to EQ lower and lower results in transformer saturation issues, and general loss of dynamics and impact because the driver is just beating the air around it more than doing anything overly productive.

2) ESL's tend to have a nasty resonant point (by moving coil standards), which you'll have to damp out. The cell geometry and construction tricks can move that peak around. For example, I use stabilizing "dots" rather than a support that runs the length of the panel. This gives me the mylar support I need for stability, but results in a much lower resonant point, than an unbroken support. You can damp this resonance electrically, but it's going to be a deep notch, and will take a lot of measuring and tuning. You could also damp it physically: The Quad 63's use a fine nylon screen over one of the stators (the back one), to damp out the resonant peak. This strikes me as a much better idea than trying to do it electrically.

All your musical energy is in the bass, so you will need some big and very good transformers to handle the power without saturating or getting generally nasty with your signals.

One of the things that my DIY ESL's have taught me is that my ear isn't overly sensitive to bass reinforcement type. My friends and I built a set of dipole woofers (using Lambda acoustics woofers), and it was very difficult to justify the increased size, weight, and presence of the dipoles vs. a good transmission line. But that's with a crossover point at 130 Hz (4th order). Martin logan has a habit (or did) of crossing over higher and more shallowly, so there is more musical information coming out of the woofer. I tried to minimize that at the expense of ultimate SPL (due to EQing the panel).


My view (take with NaCl):
If I were to build another set of ESLs (I've already got the parts, I'm just waiting for my motivation to arrive), I'd built a wire frame (like the audiostatic speakers) and segment it. This would kill two birds with one stone. First it would vastly improve the dispersion, second I would mechanically EQ it by electrically bringing in more and more of the panel width as frequency drops. This would allow my crossover to be much more simple, and I'd bet would get me into the 100Hz range or a little lower just with resistors on the stator wires. Then I'd cross over to a nice metal cone woofer in a resonant free enclosure. I guarantee that I can outperform a door sized ESL with a good 8" woofer in a TL enclosure.

 

[glorocks]

The guru has spoken. Thank you, Sheldon. I built your DIY 'stats a couple of years ago, and actually bought a pair of trafos from you. The result was very good- I'm still listening to them. How are you, by the way?
I know most folks like to cross into a dynamic LF at some point, and I did that, using T-Line enclosures as you did. The result was a real crowd pleaser, although the 3M double sided tape doesn't hold the membrane as long as we would like.
I am still, however, curious about what would happen if I could get those next 2 octaves from an electrostat. (I'm crossing at 200 Hz)
The whole idea of this forum is to experiment, for better or worse, so I am going to press on. We all certainly hope you will tune in here from time to time -

 

[glorocks]

Perhaps you will share your ideas for building a wire frame? I tried to build one many years ago ala Roger Sanders and it was a build nightmare, ultimately not ever working.
Thanks

 

[bear]

Acoustats measure to 35Hz.
I have measured them there... a long time ago, but they go to 35Hz.

Tyu, that is an "autoformer" and I would personally not use that method.

...the "other side" of the bias supply is at ground. Now one of the stators is at ground.
It can not move WRT ground. Only the other side of the xfmr develops voltage WRT ground. That is now 2x the voltage that was formerly present.

Perhaps I'm missing something here...

_-_-bear

 

[a.wayne]

Is that 3db down at 35 hz ..?

 

[john65b]

And which model Acoustat are we talking of?

Only the other side of the xfmr develops voltage WRT ground.

So the diaphragm is no longer push pull, "driven" only from one stator. Planars (maggies) work this way too - as the magnets are only on one side of the diaphragm.

I just don't get why output can be higher if one stator is not "driving" the diaphragm....seems like it should be 6db less...

Maybe your getting out of the LF phase cancellation issue with ESL by going single stator? Maybe only higher output just on the LF?

Sorry if this is an ESL 101 stuff...I am new to ESL's.

 

[Forman313]

Is it even possible to build an ESL subwoofer?

Im just thinking out loud here.. and I have no experience on the subject.
Sub bass is generated by moving massive amounts of air. And with the Xmax of an ESL panel, it just doesnt seem possible without an enormous diaphragm surface area. IE. enough to cover the etire wall and then some.

A woofer for moderate SPL levels sounds reasonable if enough money and brains are put into it, but sub bass? Really?

I love ESLs. I have a pair of stax and I absolutely love them. But I cant imagine the already huge stax panels scaled up for use in a whole room.
An H-shaped frame will help a lot, but to what extent are you planning on keeping the "ESL-sound" ?

Suddenly im not so certain anymore.. with some compromise and a lot of MBAT, it might work.
Im still not on board with the sub part, but a real nice woofer...


Good luck man

Now get a move on. Get it done .. I cant wait to read about your next undertaking

 

[tyu]

Less just say thay sold about 1000pr of the SL3 with this setup an thay sounded great...I have had one pr of SL3 with the SE setup.4 wire panels.....an a pr setup like there schematic....When thay went to the All pushpull setup thay had to put the limiter in....To stop panel noise from the tranfourmers......Bear i dont think you are missing anything...it is what it is..Most say it the same as pushpull with the back of the panel on the neg.input from the amp....i have as siad i have put this on all my ESL an it sounds better the any pushpull setup i have had....mybe if you were winding your own tranfourmers...you could get the winding off the centap the same but i have not seen a tube amps output tranfourmer...or any tranfourmer that had maching sides....
As Siad....You GET more output2-3db.
Less Fluxing of the panel better sound less smearing.
Less arking...i have fond this in some bad panels..Setup both ways..
I have found No use of the Earth ground on the setup tranfourmers ....just the neg.input from the amps in any of the ML...thay did the earth on the frist CLS an Seqels...then move away from this an i have all ways floted my stock Acoustats an Soundlabs an got much better sound.

On paper people say a 5watt SE class A tube amp wont have a nufe to drive a real speakers.. we no this is not the case...SE.5watt sounds more like a 30-50watt pushpull tube amp...
An look at the Pass Zen amps...one fet ran classA..sounds more like 50watts..

This has been working great ....from the 90s...

I am only after geting the best sound out of ESL....An this is a Diy form...No games No BS..But like we all no if you dont go you wont NO...
Thanks for your time

 

[hella356]

In my smallish room, I briefly set up my Acoustat 2+2 (modded Medallions with Counterpoint amp) alongside my own Acoustat 3+1 config (stock Servo amps), this combo having 8 panels per side. I don't have any measuring equipment, but playing test tones I was getting very strong response through the entire bass range down to 25Hz with diminished, but usable output at 20Hz.

 

[tyu]

Are thay anyone who think theM3 thats 3ea. 9"x48" Acoustat panels can,t give 35hz????

I have had a pr of M4 Acoustat..that 2ea.9"x48" an 2 ea 8"x48" in each speaker..All stock parts in the 121 interfaces....the highend was rooled down to i would say 16khz
BUt the low end was.....well look the speakers are big....i well say it as if you had 8ea. 12"drivers pr side ezey...so go fig...good 35hz...but i well be the frist too say the top is roll down....an runing them full rang FOR ME.... thay did not get it...today i have a pr of Acoustat M3 with a 12"x48"....the M3s With just the Stock Acoustat T1 base tranfourm..only get 200hz down an the 12"x48"creaved panel takes the 200hz up....Best base ever...... thanks for any info on ESL

 

[stokessd]

The guru has spoken. Thank you, Sheldon. I built your DIY 'stats a couple of years ago, and actually bought a pair of trafos from you. The result was very good- I'm still listening to them. How are you, by the way?-

Be careful, I'm no Guru, I'm just a guy who figured out a few things the hard way.

I'm doing well, fixing Quads from time to time. I'm enjoying small systems and trying to get good sound in small footprints.

Sheldon

 

[stokessd]

Perhaps you will share your ideas for building a wire frame? I tried to build one many years ago ala Roger Sanders and it was a build nightmare, ultimately not ever working.
Thanks

I'm planning to use 14 gauge solid core conduit wire that I straighten on a jig. I'm building the rest of the frames with lexan. Here's my first cut at CAD models.


Sheldon

 

[stokessd]

Are thay anyone who think theM3 thats 3ea. 9"x48" Acoustat panels can,t give 35hz????

I have had a pr of M4 Acoustat..that 2ea.9"x48" an 2 ea 8"x48" in each speaker..All stock parts in the 121 interfaces....the highend was rooled down to i would say 16khz
BUt the low end was.....well look the speakers are big....i well say it as if you had 8ea. 12"drivers pr side ezey...so go fig...good 35hz...but i well be the frist too say the top is roll down....an runing them full rang FOR ME.... thay did not get it...today i have a pr of Acoustat M3 with a 12"x48"....the M3s With just the Stock Acoustat T1 base tranfourm..only get 200hz down an the 12"x48"creaved panel takes the 200hz up....Best base ever...... thanks for any info on ESL

I can believe that you get output to 35Hz. Those are big panels, and they will have a resonant point down there in that range. That resonance combined with room reinforcement can produce usable output down there. The Quad63's, which many people use without subs, have four panels of about 6" x 24". They are reasonable down into the 40 Hz range depending on the room. The 63's have a dust cover resonance in the 50 Hz range, so that is complicating things from a response standpoint.

Sheldon

 

[tyu]

Sheldon
All you points are well taken.....I hope most here no that even thoe my room is 18'Wx25L'x14h..... that if you can get 35hz out of any speakers it would be almost at the end of 60'.....so in my room with the 35hz...i may get a goooooooooood45-50hz...maybe ...i have only had stock 57s...for what thay do... i never heard better..so your work must kik them up a big noch..I won,t to put a socaled dust cover..on other esls...the dustcover fits the ear like a glove..that one of the 57 clame to fame..
Thanks for your time an any info on ESL

 

[bolserst]

..Bear i dont think you are missing anything...it is what it is..Most say it the same as pushpull with the back of the panel on the neg.input from the amp....i have as siad i have put this on all my ESL an it sounds better the any pushpull setup i have had

Hello Tuy,
In your schematic you show the negative input from the amp connected to one side of the secondary.
Where is the other terminal of the bias supply connected to?

BTW, to a first order approximation it doesn't really matter whether the stators are driven pushpull, all pull, or all push. In a constant charge ESL, all the charged diaphragm sees is the voltage gradient of the electrostatic field between the two stators. It can't tell what voltages are being used to generate the electrostatic field...the charge just reacts to direction and magnitude of the field.
Constant voltage ESLs are obviously a different situation.

...You GET more output2-3db.

Have you measured this? at the same input voltage you get 2-3dB higher output when you change the location of the ground reference?
This has not been my experience.

 

[bolserst]

Perhaps you will share your ideas for building a wire frame? I tried to build one many years ago ala Roger Sanders and it was a build nightmare, ultimately not ever working.

Here is a link to a thread from last year that includes some good information on building wire stators.
It's fairly short, but has some attached project files, and several links to other postings with information on wire stretching, gluing, and the like.
http://www.diyaudio.com/forums/plan...-esl-project-file-translated.html#post2241752


stokessd pretty much covered all the basic concerns with trying to go low with ESLs. The only thing I might add is that extending your ESL completely floor to ceiling is advantageous if you are trying to go low because the dipole roll off slope is reduced from -6dB/oct to -3dB/oct. This improves your output capability below 100Hz.

 

[tyu]

Bolserst so you say

Where is the other terminal of the bias supply connected to?

Were it should be ...it a bias... it stans on it own....who that would like to get the best sound thay could ...would wont to put one side of a bias tranfourmer AC winding that in the Acoustats would be 750v AC....Now if the earth is there with all it noise... it would on paper be 0...not what i call hi end sound but will work....let the bias be on it own.... sounds best .....this is how i have setup any ESL i have owned ML ,Sound lab,Acoustats....thay work an sound great....i only post on thing that i done....that have worket for years...
As i sead ML stop earthing in 93...I have work on logans...Thanks to MR Powers who was at ML....allways help me.... from the Aerius i ,To 3pr Of SL3s,Ascent,Ascent i ,2pr Prodigy......an have redone all the panels bias feeds...that alone give 2db more output...an just the Neg from the amps input is on the center tap in there setups now from ML....

An i have found if the earth is put on the centap....i think it because the tranfourmers winding are not the same from the center tap...It like trying to dance with one foot nail down..... can cause more over load an Arking of the panels.
I well say agine the bias is the ESL if it right you got great sound....if not goodluck


An as for the more output in the SE setup...it like a tube amp....i have here ...it output tranfourmer has 0-4-8 ohm taps only....now i use 4ohm speakers so when i put them on the 4ohm tap... An then go to to the 8ohm tap.... thay play louder on the 8ohm tap like 2-3db
So i use the 8ohm tap it sounds better to me.....now if we put this on paper....its like the 8ohm is the hole winding...the 4ohm is the center tap...i say this because it sound the same when the Setup tranfourmer in the ESL is setup what i call SE...with no centap....but here the deal....just try it....you can always go back to what ever setup you have.....
Thanks for your time an any info on esls

 

[glorocks]

Sheldon-
How are you planning on cutting the guide channels for the stator wires on your spacers? I have an idea forming, but I need to try it in the shop. If I have any fingers left, I'll let you all know.
BTW, I'm not trying go get subwoofer response with my project. (elctrostatic walls anyone?)

 

[geraldfryjr]

Here are some TIG wire stators that I made but are unfinished as speakers.
I had burned up one soldering gun in the process,

http://www.diyaudio.com/forums/planars-exotics/198555-els-construction-questions.html#post2747919

and,

http://www.diyaudio.com/forums/plan...ectrostatic-speakers-dummies.html#post2547781

Just another idea !

jer

 

[stokessd]

I've got a little CNC milling machine, I was going to use that to cut the slots. I would use my Incra fence on my table saw, but the table saw blade leaves a bit too wide a kerf (even a narrow kerf one).

Sheldon