DIY super high efficiency/frequency ESL line array

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I have been intrigued with the idea of building a personal line array for quite some time, and after recently purchasing a book on DIY ESLs i've decided they are the ticket.

Im going to be running them on an open baffle crossed over to a stack of dynamic cone drivers (probably 8-12 Dayton RS180s) around 1500hz. I have a Behringer DCX2496 available for all crossover duties, and will be purchasing a DEQ2496 for equalization and measurements.

I'd like the line to match the sensitivity and height of the RS180 array, so im aiming for a 7' by 1" line for maximum displacement. After researching ESLs, ive come to the conclusion that stator to stator spacing will determine the efficiency vs displacement (wider spacing=bad efficiency, but better displacement and likewise)

How wide should my stator to stator spacing be to maximize efficiency and dynamics/output in the >1000hz range? I'd like to match the distortion and dynamic performance of the RS180 stack if at all possible.

I'd like to avoid beaming if at all possible and have concluded that 1" wide diaphragm would disperse quite evenly. Is that a good width to aim for?

How much $$$ in materials should these cost to construct?

Where do I purchase a transformer? What do I look for in them?

All comments/questions are welcome
 
Try Russ at Just Real Music - I've been thinking of trying to get him to help me do almost the exact thing your looking to do for a while now. His power supplies and step up transformers are very nice. His panels are a bit shy on efficiency though - he wants to produce bass

I'm game if you want to start a project - 1K up and 100 db/w/m is what I'm looking for. Russ lives a few miles from me. It seems he doesn't see a 'market' for such a driver - I DO!!! All we need to do is build the panels themselves and use his electronics and trannies. :)

Russ- Just Real Music Stats

I just sent him an email and invited hiim into this discussion
 
I think that 1-2 mm stator to diaphragm spacing would allow you to make a pretty sensitive driver for that frequency range. It's probably best to make a few test drivers to be sure. Since your final drivers will be only 1" wide, you can make full sized test drivers. The expense will be very low.

I_F
 
Yeah, I'd definitely be down for a project!

The efficiency of lower midrange/fullrange ESLs isnt...THAT bad (80-90dB im guessing) and efficiency is determined by the square root of the distance of the stator to stator spacing so small distance differences can equate to much higher efficiency. With this in mind, I believe 100dB+ is quite achievable im just a noob at ESLs at the moment.
 
Hi Folks,

a ESL should be run in open baffle mode to sound best, but since You need a certain width of the speaker anyway, it´s good -with regard to nearly every aspect of the design- to size up the diaphragm area (width) as much as possible and to size down the ´inactive baffle´ area. You can roughly say, that a width of 100mm (4") is good for frequencies >800Hz with the use of a baffle and 150mm (6") without -the baffle or a simple bracing should give a width of 8" to 10" to cope for the phase cancellation. For frequencies above 1kHz a diaphragm travel of <0,5mm (~0,02") is sufficient. Add some safety margin against building tolerances and acoustic coupling and You get a d/s distance of max. 1mm (0.04"). I´d opt for just 0.5mm as a starting point, if You´re able to build with small tolerances in such small dimensions.
Forget the idea of high efficiency and small width! You can´t have both! If You need good dispersion character with an ESL but don´t want to sacrifice completely on efficiency, You should think of electrically segmented ESLs (wire or rod or PCB-stators come to mind). Simple metal sheet stators won´t work for this. Efficiency or good dispersion...not both!

But still You will probabely not get even close to 100dB/W/m!!!!!
As a line source of such great height (7´) the SPL will rise as distance increases and will reach its maximum at 3-4m (10´-13´) distance. Above this range SPL will fall, but with less than the 6dB/doubling of distance of an monopolic working driver.
When You get to collect experience with ESLs You´ll learn that measuring on a distance of 1m will give rather useless results with a very high degree of possibility. Nearfield measurements (1-4" distance) and ´in-room-at hearing-distance´ are a way to go. And since the impedance of most ESLs varies greatly it´s common use to measure SPL normalized to 2,83V.

I agree with Russ that such a thin strip ESL has no real market chance, since this is the territory where magnetostatic drivers shine.
Those can be build cheap but with excellent results and high efficiency and are usually a better choice!
If You allow for a greater bandwidth (lets say from ~300Hz on) You come to the field where electrostats reign. With an panel of 8" to 10" or even more width You may start from 300-400Hz with good efficiency and high SPLmax and You make use of the superior sound quality and the extremely low distortion of the ESL against dynamic drivers in the mid and upper frequency range.

So my recommendation is:
If You stay with Your concept of a crossover freq of ~1kHz, You will be better and easier off with a megnetostatic line source like Bohlender et al.

If You´re open to a lower crossover point, You could think of an ESL- panel of ML Sequel size (10"x50") crossed over >300Hz.
The dimensions of both speaker solutions would be quite similar.

jauuu
Calvin

ps. Solution nr. 2 could look like this.
An externally hosted image should be here but it was not working when we last tested it.
 
High Efficiency and small width...I dont think id be sacrificing diaphragm area at all? The line should have a diaphragm area that is several orders of magnitude larger than a conventional dome tweeter or those BG Neo drivers, how would I be sacrificing anything? A 7' line that is 1" wide would have a surface area of 542cm^2 (thats more than most 12s)

Im open to any options really, as long as I meet the criteria of low distortion, high sensitivity, and cheap and easy to construct. How would I go about building a magnetostatic driver? (ESLs sound like a piece of cake to construct compared to other transducer types)
 
I share your enthusiasm for the basic idea of a woofer array coupled to an ESL, but I also second Calvin's comments. I've been designing something very similar to the project you described, but based on experience with previous ESL construction projects, I'm aiming for a lower crossover point than you're suggesting. I intend to use a vertical array of 8" Dayton woofers crossed over to an ESL panel at 300-400 Hz. As I see it, the lower crossover point has significant advantages.

The woofer array and ESL panel will always have some spatial separation between them. As a result, near the crossover frequency there will be two spatially separated sources radiating the same sound. This leads to interference---non-smooth variations in frequency response as a function of listening angle. This problem can be addressed by reducing the distance between the two sound sources, or by reducing the crossover frequency (increasing the wavelength at the crossover). Since there's a lower limit to how closely you can position the woofer array and ESL panel, reducing the crossover frequency can be quite helpful. At 1 kHz sound has a wavelength of a foot or so. A typical woofer array-to-tweeter distance is therefore going to be at least half a wavelength. If the crossover were at 300-400 Hz, then the woofers and tweeter would behave as if they were almost coincident (they'd be something like one seventh or one eighth of a wavelength apart instead of one half).

Also, if you rely on very small diaphragm-to-stator spacing for high efficiency, you start to run into difficulties keeping the diaphragm from touching the stators. It's impossible to make things perfectly flat, and if you rely on something like a 1/32" spacing---or even smaller, given your sensitivity target---you're going to have a tough time avoiding sparks. I tried making some prototypes with 1/32" spacing and it was a real drag. I'm confident that with more careful construction a high efficiency panel with 1/32" spacing could be made, but you mentioned that you were looking for something easy to build. Panels with very small diaphragm-to-stator distances don't fit into that category. The construction becomes much more finicky as the spacing drops below 1 mm.

In addition, you'll find you need some sort of frame to hold the ESL stators, and it needs to have some thickness in order to be sufficiently stiff. If you use, for example, 3/4" thick MDF and cut a 1" slot in it to accommodate your ESL tweeter, you introduce a nasty cavity resonance. Even if you bevel the edges it's still going to have a major influence on the panel's response. If the panel were six or ten inches wide, the same 3/4" frame thickness would cause far fewer problems. You can still, as Calvin pointed out, use a segmented design to maintain good directivity characteristics. Just put a very narrow tweeter strip down the middle of the panel so that it's several inches from the nearest obstruction.

Finally, one of the advantages of the ESL panels is that they exhibit very low distortion over much of their operating range. Why not take advantage of that over a wider frequency range, as long as you're building one? The woofer will act pistonically below 1 KHz or so, but if you look at the frequency response of almost any 7"-8" woofer you'll see the first glitches start to appear at 1-2 KHz. Those glitches are not going to be supressed by much if you use a 1 KHz crossover frequency. If you cross over an octave or more below that point (300-500 Hz) you end up with a dynamic woofer operating at the frequencies it handles best, and the ESL operating at the frequencies that it handles best. Granted, virtually all dynamic speakers cross over higher than 1 KHz and get away with it. But with an ESL covering the upper octaves you have the freedom to use the woofers only in the range in which they behave nearly ideally. It seems a shame to throw away that freedom. Besides, increasing the radiating area of an ESL panel has very little impact on the cost or complexity of the construction. The same can't be said of most other speaker technologies.

Any way you cut it (or cross it over) I agree that an array of OB woofers mated with an ESL panel has lots of potential. I'll be very interested to see how your design progresses. Best of luck!

Few
 
Why cross so high? You'll be mostly listening to the woofers, not the hard to build expensive, and good performing, ESL portion.

Think about the cavity and diaphragm resonance issues with a 7' x 1 inch cavity. the 1 inch part will make the resonance pretty high, even crossing it over like it's a tweeter, you may still have to notch that resonance out.

Based on my experiments, I didn't like crossing my panels over as high as 300 Hz, and by 500Hz it wasn't sounding very good at all. IT was exhibiting that weird phasey sound that ML speakers tend to have. I prefer to push the crossover as low as I can. I use a 132 Hz crossover point, but my panel is 20" + frame wide.


Sheldon
 
RTR Panels Wanted

stokessd said:
Why cross so high? You'll be mostly listening to the woofers, not the hard to build expensive, and good performing, ESL portion.

Sheldon

Because they will be be (in my case) mated to 100 db plus sensitive line arrays and I don't really want to use some power sucking lifeless HF drivers (like in my Acoustats).

I have used the little RTR panels in arrays with large horns with success in the past. Nine of the RTRs in a vertical array biamped will do the job. I'm a horn guy by nature....

Anyone have any?

I sold mine -

:xeye:
An externally hosted image should be here but it was not working when we last tested it.
 
Hi,

the problem was, is and will be that You won´t be able to reach an efficiency of 100dB/2.83V/at whatever distance with a classical flat ESL!
One interesting idea -and the only one I could think of at the moment- could be to fold the diaphragm in a way similar to the AMTs. This could lead to a greater diaphragm size with same panel width or to a smaller width with the same membrane area. The stator would not be a single pair but it would be a array of small rectangular and massive rods with always changing polarity and the membrane would be zigzagged between those rods.
An externally hosted image should be here but it was not working when we last tested it.

For a good HF-response the depth of the driver would probabely have to be less than 5mm. To build such a thing would of course be a pain in the ....... ;)
As an theoretical advantage, besides that the smaller width supplies for a better dispersion characteristic, I see that there´s no stator in front or behind the diaphragm through which the sound has to propagate. The acoustical openness is therefore much higher than with a flat panel ESL. The stators on the other hand could and actually they must(!) be massive, which gives the highest field strength and most homogenous field distribution! Which is an advantage for small values of d/s.
Further do I assume that the membrane wouldn´t need very much mechanical tension since of the small dimensions of the stator rods. Maybe no mechanical tension in vertical direction would be needed anyway!??!
To keep stray capacitances (losses) at a reasonable value the diaphragm coating should only be right opposite the stator rods which leaves us with the nice problem to solve how to do this and how to connect this multitude of strips to a bias supply?

..and so on...and so on :D

jauu
Calvin
 
Calvin,
Great minds think alike! I have a quickly assembled prototype of a folded ESL sitting next to me as I write this. I put it together last summer, but assembled it in a rush and introduced a short circuit somewhere and haven't had a chance to go back and do it right. I used copper clad fiberglass boards (an unusually thick version of the boards usually used for making printed circuit boards) for the stators. I etched away the copper where it would have just increased stray capacitance, and used 1/16" thick 3M foam douoble-stick tape for spacers.

I'm not quite sure what the relationship between depth and high frequency response will be. I made my stators much deeper than you suggested (several cm instead of 5 mm) in order to get some feel for the way the drivers behave. It's not clear to me that they must be as shallow as 5 mm to achieve good high frequency response, but I'd be very interested in hearing about your reasoning. I know the Heil drivers were quite shallow, but they also had magnetic field constraints to contend with.

By the way, Peter Walker of Quad fame included a device of this type in one of his ESL patents, so we're a few decades too late if we want to claim to be pioneers. I'd love to get one up and working nonetheless!

Few
 
Hi,

my assumption for a small depth of the driver is, that the depth of the cavity that is formed by the membrane and the stators might lead to phase probs with higher freqencies. Too the membrane won´t work in a pistonic way and probabely not even in an a curved way, but the pressure building up in the cavity will rather send kind of a shockwave running over the membrane. This would introduce distortion as well as a reduction in efficiency. The membrane is getting ´stiffer´ the smaller the depth becomes.

jauu
Calvin
 
ever heard of the solosound panels , there 10cm w - 15 cm high.
4 of them make a nice 90 db already. they where used from 350 hz, so maybe you can decrease the 1,8 mm stator disctance to 1,5 or maybe even 1 and get a real nice SPL, only possible if you gona cross them over higer up in like 750 or whatever, just trial and error. also could use metalized mylar also huge SPL boost, but nasty load resistance.
 
Hi,

with such a small panel (width, btw. which Panel anyway?) You hardly get the claimed efficiency below ~800Hz. Using the panel just above so high frequencies a reduction in d/s will be useful and necessary and will rise efficiency clearly.
The choice of diaphragm material is of low importance regarding efficiency, as well as the question of the conductivity values. But high conductivity has severe disadvantages with regard to other design aspects of a panel. Those disadvantages ragard safety aspects (arcing) and linearity, but definitely not SPL or load resistance! You´ve got something seriously wrong.

jauu
Calvin
 
its pretty clear theres is a connection between conductivity and SPL, the other coating does not play that loud. i look and listen to 2 right now , the same panels only differents in the coating. and the one with alu does sound louder.


and for the crossover frequenty of those panels. search solostatic
they only use 4 of those panels and they claim 90Db

and they sound great so i dont think they have a gab between 350 - 800
 
Hi,

if there´s a clear difference in SPL noticeable than thats a sign for a defect or a wrong coating material! As long as the coating doesn´t reach a too high resistivity, the choice of material doesn´t influence the SPL noticeably (btw. just one invisible fingerprint can spoil the hole thing!). So rather have a look for leakages in the softer panel or bad connection to the membrane!

I suggest rather than me searching around on a slowly loading website, You provide for a precise address!

What someone claims and what reality says, that can be quite differen matters. :cool:
First of all the SPL statements are incomplete since the data is not precise. No measuiring voltage and no impedance plot is given, as the supply with precise measurements over all is terribly low.

And again, panels of such small width will suffer from phase cancellation (depending on the baffle size) below ~1kHz. As any passive correction for this effect cuts on efficiency it´s quite unbelievable that they reach an effective SPL of 90dB@2.83V(8Ohms)/1m from 350Hz on. Measuring an unequalized panel at ~1-2kHz (where SPL is highest), recalculating for a lower impedance and voila we have the desired high number, while in effect @350Hz -10dB would be probable.

My critics is not about the resultant sound quality but solely about the belief in claimed specs that seem unrealistic and can´t be proofed because of missing information

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