I am planning a pair of ESLs using round (coin shaped) drivers with stators built of 'egg crate' lighting framing and 10 per inch aluminum or copper screening. The stator to diaphram gap would be 1/8". I am using round drivers to reduce the beaming associated with planar speakers.
Update April 2000
I have just fininshed a pair of the Waldron/Sanders Compact ESL /TL speakers and have been pleased with the results. I had originally planned to build a pair of full range ESLs but the more I researched the more it became apparent that a hybrid system was the way to go.
The stators are constructed of "egg crate" lighting frame and 10 per inch aluminum screen. My first attempt to attach the screen with 3M Super 77 spray adhesive didn't work. The screen came off the frame when I heat stretched the mylar. However, I was successful attaching the screen using a steam iron. If you attempt to go this route use lots of heat, pressure and patience.
My amplifiers have limited wattage so I have played the speakers only a little but I have been impressed. I will keep you posted on further progress as I get my amplifiers and crossovers set up.
I would also like to thank Barry Waldron for all of his advice. He was able to answer many of my questions on ESLs and thus made my project possible.
Update August 2000
I am not ready to call my speaker a success just yet. My fifty watt amp worked hard just trying to produce a medium sound level. I skimped on some parts and will need to replace them for the speakers to work properly.I used a bias supply/transformer assembly from a Janszen Z210ah to run my ESLs and it just did not work well. As I later determined, this assembly was probably intended for a small tweeter array, not a large, almost full range ESL speaker. I did the math and determined the bias supply was only good for about 850-900v. No wonder I couldn't hear anything! I don't know what the transformer turns ratio and frequency range performance is but I suspect it isn't right for this application either.
With a more powerful bias (3-5kv) and better step up transformers I expect the speakers to play loud and clear.
Along the way I became further convinced that other things I planned to add later on in the project are absolutely necessary from the begining for the speaker to sound correct. These aren't optional items - they are necessities:
- A good active electronic crossover with a steep slope (24db) is a must. I was using a simple first order (6db) passive biamp crossover between the preamp and amplifier until I could build or buy a quality electronic unit. It is easy to make one of these as it only takes a few parts and a little math. My completed version worked and seemed to crossover at the desired frequency however the bass seemed mushy and both speakers seemed to lose volume around the crossover frequency. I moved the crossover frequency to a few settings and got similar results. I wouldn't even think about using a conventional passive crossover network between the power amp and speaker. The loads are just too different between a ESL and dynamic speaker for good results.
- Frequency correction is also a must. The ESLs themselves lacked full midrange and bass and sounded bright as ESL's naturally do. However adding the dynamic speakers added more bass but still left the midrange and low-midrange lacking. There was hole right in the middle of the frequency range. I am sure some of this would improve with better bias and transformers to drive the ESLs but it would still not be a level response. With a hybrid speaker as I am builiding, one might be tempted to downplay the importance of adding frequency correction to boost lower octaves after all there is a dynamic speaker making up the difference, however don't skip this important addition. The frequency response of an ESL is naturaly not level and it must be flattened to sound correct whether it plays with another speaker or not.
Barry Waldron provide me a lot of valuable advice on the speaker project including the above items. Believe him-he knows what he's talking about.
I need to amend the above description of the problems regarding frequency response. I did a litttle work and discovered the transformers I was using are not suitable for the frequency range I was trying to use them for.
I spent an evening and made a pair of simple ESL headphone drivers, partially out of curiousity and partially to find out if the bias supply/transformer set I had would be good for another application like this one. The drivers were made of printed circuit board material 2.75" (apx 9cm) in diameter with a 1mm diaphram to stator gap. Total cost for this project was about $5.
The drivers worked fine, had nice detail and had plenty of volume. The stators could have used a few more holes as the lack of open space seemed to cause a little coloration and a bite to high tones. I don't remember what this effect is called but I have heard others with stators made of Lincaine describe it too.
Anyway the biggest problem was, you guessed it, was a lack of bass and low midrange. I have read several descriptions of others who built headphone drivers and none of them described a lack of bass. In fact there was usually a small rise around a few hundred hz. This leads me to believe the transformers I was using were not suitable for full range.
I still think that ESLs need frequency correction to sound correct however my problems with frequency response and a hole in the low midrange might have been exagerated by transformers that were not delivering full range performance.
I will get back to this project sometime with all new bias, transformers, and active crossovers with frequency correction. When I do I will let you know my results.