Fresh ESL Research

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Have mailed Matthew Lattis to see if he is interested in getting envolved in this discussion. Have also spread the word as far and wide as I know how to. If everybody interested here sends out an invitation to as many buds and sites as they know of we may well get a real deal discussion going here. What say you? Your interest may well yield you enough information to build a damn good set of panels on the very cheap and have them up and running in time for the holidays this fall and the praise of all you know.
As an aside someone somewhere said that ESL's were not tube friendly. Well I have run Acoustat 0ne plus 0nes with a set of Antique Sound Labs Wave 8 mono blocks very nicely with good play back level. Acoustats have six wires per inch. The panels I am building have just as much stator to diaphragm spaceing as Acoustats but much thinner insulation so in reality the stators are closer (more efficient) and they have twenty wires per inch. You pick up six db of output every time you double the stator wires. Do the math for twenty wires per inch. These panels will play very loud on not a lot of watts. Best regards moray James.
Moray James, I like your enthusiasm. Lattis’ article in AudioXpress is indeed excellent and will save me some additional work that I had been planning to do down the road. Moray, just a slight correction on your point about achieving a 6 dB doubling in sensitivity for each halving of stator wire-to-wire spacing, IMO: As the wire-to-wire spacing becomes much less than the stator-diaphragm spacing, there are diminishing returns in sensitivity, since the electric field becomes essentially uniform across the diaphragm. Further reductions in spacing will yield negligible improvements when going in this direction. See Figure 17 in Lattis’ article (upper right area where curves crowd together). The 6 dB rule does apply over wide ranges for changes in bias voltage and stator-diaphragm spacing, both of which translate directly into electric field potential (volts/meter) as seen by the diaphragm.

I hope Lattis joins this discussion. I want to commend him for his excellent work and I look forward to his next article. I would also ask him if he understands the cause of frequency response anomalies in the 1 to 2 KHz range which must be related to the larger frame dimensions and not directly to the various stator designs, because the wavelength is too long. But this anomalous region does vary with changes in stator design. What is the panel’s primary resonance frequency? Also would like to know his testing set up. The sloping response indicates that the panel was probably not in an infinite baffle. How far away was the mike? What kind of mike, etc.?
Wire spaceing

Hi Brian Beck: I believe that you are right about very close spaceing and gains in output. That said I think that a lot of people look to commercial products to see what works. Perhaps people do not take the time to put themselves into the manufacturers shoes though. When compeating in the commercial market place there are very many more important factors to be concerned with other than how well a product performs. Acoustat panels are a good case in point. With only six wires per inch (resulting in a good sounding panel) there is very much room for significant improvement by the DIY builder. With fine guage double build magnet wire or Kynar wire wrap wire the 40% open area would be easy to achieve. Fine pitch threaded rod can be used to easily build a spaceing jig for a very consistant panel which should far exceed the factory product.
I hope that my enthusiasm pays off and we can generate some great new ideas. There is so much knowledge here to be found and so many bright minds sitting behind the keyboards. Thanks for joining in. Hope that I can count on you to help beat the bushes. Best regards Moray James.
Hello Moray

I read the article to and was under the impression that the perforated metal stators were on par and even a little better than the wire stators. Are you going with the wires because you allready are in the building phase or do you think that there is a benefit?

I am seriously considering a copy of the fullrange found here:

This is pretty much what I have been leaning towards for about a year and a half. Doing all the research that I can. With all the books and net trawling ( should add sorting and rejecting to ) I'm leaning to the foam tape spacers and the rubbed on graphite conductive coating. What are you doing in your panels?

I'll look forward to the discussion in this thread and the coming article in AudioXpress.

Hey Mark: the panels that I am building are to replace some old Acoustat panels. Acoustat uses 6 wires per inch with a PVC dielectric. I am useing 20 wires per inch with a Heavy build magnet wire good for 12 Kv which is far beyound the PVC used in the Acoustats. Yes you are right that in Matthew's article the perf metal had the highest output. The closest spaceing Matthew tested was 16 wires per inch. A 32 wire per inch panel could be made easily but with a lot more winding (which would add 6 db to Matthew's 16/inch panels). I believe that wire stators represent the best possible compromise for DIY with known and consistant dielectric strength. If you have a perf stator and it gets screwed up at the powder coaters well you have to start from scratch which is a lot more expensive than magnet wire. Magnet wire also allows for stator segmentation as well. You can wind as much or as little wire as you like where you like. You can make a panel as efficient as a perf metal stator I believe.
Foam tapes are great however the technical advice I have received from 3M says that they will not hold a high stretch rate diaphragm but are fine for shrink tensions. Another factor in high output is diaphragm tension. The current batch of panels being built will be useing Licron spray for the diaphragm high resistance coating. Hope this helps. please join into the discussion on the DIY ESL project. Look forward to seeing you there. Best regards Moray James.
I urge some caution with the "better insulation".

Strickland found that when he built panels with "better" teflon insulation that they worked fine in Florida, but when they moved the operation to Arizona, not so fine. The lack of humidity changed things.

So, apparently the PVC is *good* because it leaks!

If you've looked at the Acoustat wire, it is rather fat in insulation... so it will handle a high kv rating, but still manages to leak the required amount. Or so it seems.

_-_-bear :Pawprint:
Check out the link I posted

In the link I posted they used polyethylene as a coating and were able to really jack up the polarising voltage. But I may checkout the wire to. THere may actually be a cost benefit. Thin guage magnet wire is not that expensive. 84 square feet of perferated metal will not be cheap! The only problem I can see is the amount of tension that will be developed with the close spacing. THe frames needed to support them will not be trivial. Probably Steel tubing would do the trick.

Related to ESL but if not Please excuse me

I have been in contact with Arto Kolinummi and asked him if he would mind an attempt at reverse engineering the full range transformer that he and a friend created. HE said go for it!

Check out the transformer section.

A step up trannie does effect the sound of an ESL and they to require some design interest to many designs stand on the shoulders of others. What they did was actually legitamate design and engineering. THe actual design and engineering work is above my current knowlege of transformers. But the actual winding is probably within my capabilities. But I'm pretty sure that there are members that could take a stab at it.

Bear: while you are right about Jim Strickland's findings I never heard tell of Jim mentioning the thickness of dielectric he tried out. suppose that I could ask Jim. I live in Calgary Alberta and while not as dry as Arizona ii's a far cry from Florida. I have built excellent panels with Kynar wire wrap wire as well as double build (Heavy build as it is known now) magnet wire. They far exceed the output of a standard Acoustat panel. Esl's can produce small amounts of ozone and over the years (they are all old now) the dielectric gets spongy and voltage can "punch" through on big peaks. I am comfortable with useing newer stronger dielectrics for my panels. Your point is worth investigation for those liveing in super dry areas . I will see if I can get Jim Strickland to comment on this. Wish me luck.

Mark: I have been useing the same grid frame idea as used by Acoustat (Jim Strickland) which is egg crate light louvre material. This is available in Styrene and Acrylic. The latter is much more expensive but far superior. Both materials are available in half inch thickness as well as three eights thickness. I prefer the 3/8's inch version. Both materials are available in white and black. The magnet wire is strung by hand useing a simple and inexpensive jig made with threaded rod either 1/4-20 or 1/4-32 will do fine. I would suggest that if you want to match or better the perf metal stator material then 1/4-32 and 30 guage heavy build magnet wire should get you there. No matter how you build you are going to get structural resonances and I think that they will be higher "Q" and lower frequency with steel perf plate than with light louvre and fine wire. The plastic louvre is plenty strong enough to do this job. Further the louvre based structure is far lighter than the perf steel stator design.
Another advantage of insulated wire is that you can pick your dielectric and it will be for all intent and purposes perfect in thickness and consistancy further all wire is fully tested for quality of dielectric strength etc. as a matter of course. When manufacturers choose a powder coat material for perf metal stators their number one concern is NOT how good it sounds! Thier number one concern is how well the powder coating builds up at the sharp edges of the punched holes. Poor build up at the edges or inconsistant build up at the edges and you simply throw that stator panel in the garbage as it will arc over there first, so now you have an unhappy customer and a damaged reputation. That is the reason that Nylon powder coat is the number one choice material for perf stators.
A standard Acoustat panel is approx. 9 inches wide and about 46 inches long. Acoustat did not drive the outside 1/2 inch of thier diaphragms. If you drive the whole surface of the diaphragm with stator wires you can pick up almost an extra square foot of driven radiating surface area so in a 0ne plus 0ne configuration that is almost an extra 4 square feet of driven radiating area. Not an insubstantial area to be sure. Combine that with (just a guess here) about 15 more db of efficiency and I would say you have a pretty fine speaker in your listening room!

For those with Acoustat's try reduceing the load resistor value of the High Voltage supply from the stock 500M ohms down to about 10-50 M ohms for some badly needed extra volts on your diaphragms. Well worth the effort. Nice improvement all around and it is cheap and easy to do. Best regards Moray James.
I really hope you guys come up with a great DIY ESL spks :D ESL speakers are my favouite spks for mids I still can remember the first time I heard a pair of stacked Quad ESL 57. Have a look at Beveridge website some interesting articles there as well.

Btw do also consider Beveridge's approach as well. The ESL diaphragm need not to be so big to generate the necessary SPL. I've heard the Beveridge ESL used in conjunction with Entec subs quite nice sounding.....
BTW: not sure if I know what you mean regarding Harold Beveridge's aproach. I will guess given your text that you are referring to building a hybrid speaker. Well I will go on the record here to say that I agree with Harold there as well as Roger Sanders and others. Bass is best produced via a dynamic driver. Smaller cheaper and easier to impliment. Just incase you did not know a fully modified Acoustat can dance circles around many of today's high end big dollar speakers. They can be found very inexpensively. A buddy just found a set of model 3's in Vancouver BC on eBay for the grand total of $53.00 Canadian. That's like $20.00 US each!
Without sounding smug I think that what is being talked about here is a great sounding speaker which is easy to build and inexpensive as audio projects go. For that matter I think that if it was just the equal of an Acustat 0ne plus 0ne I think that most DIY audio buffs would feel that they had died and gone to heaven(as I said a fully upgraded interface and panel with top notch electronics). However I know that the designs being discussed here are superior to the Acoustat by a wide margin. I am interested in takeing these ideas further. To build them cheaper and easier with higher quality and consistancy. To cross them over either passive or actively to quality compact subs. Anybody here know how to passively cross over a stat? I hope that there will be enough interest to get this rolling. I have tried to stoke a good fire to see if we can as a collective make some steam.
For an example I know that Jack Sutton makes a fine wide range ESL line source only four inches wide. That is a pretty domestic foot print. Most folks these days have a sub or two anyway so no extra space required there. I know that while I_Forgott will cringe stat panels have been horn loaded too. Dosen't get much more sexy than that! Point is that there should be something for everybody here. I am not so interested in a how to developing here as I am of developing concepts with working building blocks that can be used to build what ever it is you want.
How about a little brain storming throw out some ideas. Simple, complicated or crazy. Nothing gets laughed at. Sometimes the ideas you initally think of as krazy or just plain stupid can turn out really cool or spawn terrific ideas as a result of a fanciful thought. At the very least the odd ones out can keep you entertained. I am sure that Wright had his share of laughs and snickers when he decided to buuild an ESL inside a plastic film bladder full of sulferhexafluride. Maybe but Dayton-Wright's were pretty cool sounding speakers. Have some fun that's why we are doing this. Best regards Moray James.
As Bear alludes, Strickland used mil-spec PVC insulated wire in the Acoustat design, finding that “better” insulation robbed too much of the electric field within the insulation.

The volume resistivities for some common plastic insulators are as follows (all in ohm-cm at 25 deg C):

Polystyrene 10^18
Teflon 10^17
Polyethylene 10^17
Polypropylene 6x10^16
Polyvinyl-chloride (PVC) 10^14
Enamel ???

As you can see, the PVC is 1000 times leakier than Teflon or polyethylene. With PVC the bias voltage stays across the higher resistance of the air gap (think of it as just a simple voltage divider circuit) where it can do desired work. This applies until the bias voltage is raised so high that air ionization begins to occur and then relatively more voltage is dropped in the insulation than in the lower R air.

With polyethylene or Teflon, yes, you can really jack up the bias voltage, but it’s not doing you much good because much of the V is across the insulation and not in the air gap. Humidity matters both in determining air ionization levels, but also, depending on how hygroscopic the insulation material is, in how much insulation resistance drops with moisture.

In a private correspondence with Jim Strickland, he told me that they found that PE acted like a capacitor and held on to surface charges even after the HV bias was removed. This would result in continued playing with phase reversals. His bottom line: use PVC.
Hi Moray James,

Actually I'm refering to dipole vs monopole type of radiation pattern. Btw I've heard Acoustat 2,3 and 2+2.... before very nice speakers in the stock form I'm sure the upgraded version should be better. The Beveridge ESL does project a nice and huge soundstage quite different from the dipole ie Acoustat type of course I do admit the Diying a Beveridge cabinet won't be that easy but I really like the Beveridge sound. Incidently I live in warm and humid Singapore , many ESL speakers like the early Quad, Soundlabs do not survive long in our climate. All the best in your Diy project i'll be following it closely.. :D
That would be a copyright violation, and besides Ed Dell's little magazine needs all the support it can get from us DIYers. To subscribe:

The article you mentioned is in the latest (June 2005) issue. Outside the US or Canada, the international rate for three years of subscription is $152.00US, hardly criminal in this day of $100+ designer capacitors. That’s only about $4.22US per issue. I'm not affiliated with AudioXpress, just a subscriber.
Coupla comments...

I don't think you can make a transformer that actually works full range and results in a flat response without other stuff that reduces the sensitivity and/or flattens the response.

Suggest looking at Strickland's patent and what Audiostatic proposed (but oddly never seemed to use in production?) as alternatives.

You can get magnet wire with a "Nylon" type insulation , iirc, dunno what the specs on "nylon" might be compared to PVC.

I've used the 3/8 styrene light louvres for some experiments... wondered why Acoustat went with the taller stuff - seems to me less is more when it comes to HF response...

The Beveridge design requires that you build some fairly complex "waveguides" and mate them with the cells so that they don't have any transition issues between the cell and the waveguide... but interesting stuff that is...

If I were to build up cells based on the Acoustat method, I think I'd want to find a way to make two "light louvres" stick together so I could make a single diaphragm without having that [gap] at the ~4ft mark that happens to coincide with where your ear level is when sitting down!! Maybe someone makes them in 8ft lengths for the long commercial bulbs?? Hmmmm...

I suspect (for no evidentiary reason) that the purpose of not running the Acoustat diaphragm doping out to the edges was:
A) There's no real output to be had at the edges
B) Applying the stuff *after* the diaphragm has been stretched on one side kept it neat and from having leakage issues to the plastic grids.

I don't *know* that they applied the dope to the diaphragm after it was stretched on one half, but looking at the constuction, that's what I think since imho it's not really heat shrink mylar on there. It's my guess on this...

Crossing over to an ESL?

I use nothing or a first order on the ESL, depending on where it likes to roll off on its own, and if it can handle the LF energy in the first place (talking stuff like an Acoustat panel now), and a 4th order to the subs @ ~60 Hz. This gives a very nice relationship between the dynamic driver below 100 Hz. and the ESL which with the first order is rolling off below ~125 or so... the main thing is to get the woofer's response to nil above 100 Hz. Getting an exact and perfect ly flat response in the area where they still interact is fairly non critical unless you happen to find a *huge* bump or suck out in your room at your listening position.

I also like to position the ESLs with respect to the subs for best impulse response when measured, this always seems to sound better than a random relationship between the two.

In my case, I designed the subs so that I could stand my ESLs or anything else I like on top of the subs, making this relationship fairly straightforward to achieve.

The downside of putting speakers on top of subs is floor bounce - which likes to "suck out" things... so the ~100 Hz pass off works that way too... the floor bounce creating a nice notch just at the right frequency.

In a room with a wider dimension or with smaller subs, putting them side-by side would work too... fyi. And, no floor bounce issue.

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