Quad ESL Design

[ak_47_boy]

I am realy interested in the point source technology used in quad esl's. Is point source the reason quads sound good? Link explaining delay rings:
http://www.euronet.nl/users/temagm/audio/esl63.htm

I have a few questions about them.

1. Are the bass panels last in the delay line?
2. Where could i get delay coils?
3. What is the delay between rings?
4. Won't some frequancys sound delayed? Or is the delay so short that there is not enoungh time to notice?
5. Is the different ring thicknesses for different frequencys? If they are why are they aranged in that order?

How could this idea be inproved upon?


thanks

[maudio]

As it is christmas holiday anyway, I suggest you do some reading Most answers to your questions can be found in these articles:

-Peter Baxandall wrote an excellent article on esl theory which also contains a wealth of detailed information on the Quad 57/63 models.

-You may also want to read the original papers from Peter Walker. Here is one to start with, a zillion others can be found all over the internet. Quadesl.org is a good starting point.

Btw I don't agree with you that the Quads are very good sounding esl's But that's a matter of personal taste I guess..

[I_Forgot]

I have been trying to simulate the delay line to get some idea of how effective it is at producing the "pulsating sphere" that Quad talked about in their marketing literature.

The problem I am running into is that I don't have any data on the coils. I'd bust open one of my speakers, but since I rebuilt them, getting at the boards with the coils is a major PITA. Does anyone have a loose coil sitting around? I have a quality LCR bridge that I can use to do some tests so I can develop a decent model.

In my initial playing around I tried simulating the coils (actually air core transformers with a shorted secondary, per the schematic). The results are not encouraging. The shorted coil makes the whole thing look like a smaller inductance. That is the opposite of what is needed based on my simulations of the entire delay line circuit, where I can't seem to get more than about 20 us delay between adjacent segments of the speaker, which seems too small to me.

I'll post some of the circuits and results when I reboot to windoze. I sure wish LT would come up with a linux version of SwitcherCAD! Maybe I'll try running it in wine...

I_F

[maudio]

I measured them long ago. As far as I know, the coils are around 4H.
You also need the capacity of the segments:
Capacity of the inner 6 rings is around 11 pF. The next one is 45pf, the most outer one is 72pf.
Bass pannels are 360pf total.

The shorted turn is to dampen the line, which is calculated exactly so the signal won't reflect back at the end. You can simulate that by placing a resistor in series with the shortened turn.

There is some more info about the coils in the Baxandall article if I remeber well.

[I_Forgot]

Wow! 4H seems awfully high considering their size, but I think that is the sort of number that starts to work in the simulations. I'll try it out.

I_F

[maudio]

Btw, the capacitances mentioned are measured directly on teh panels, also without wiring capacitance.
Also there is signifant capacitance in between the rings. The small 10p caps are there to compensate for them. Better omit them in simu.

The source the line is driven from (stepup trannie) is important as well, the transformers have a remarkable high dc resistance (several kohm, don't have the accurate number where I can find them...)

[moray james]

Remember that with a point source you have a panel that radiates hemispherical wave fronts. That means that you are exciting all the room modes especially consider the floor, ceiling as well as wall bounce. With a line source you no longer have to worry about floor and ceiling bounce and your wall bounce can be controlled. Just some food for thought.
Some years ago I co designed a point sourch panel speaker that was magnetically driven. All of our first designs were line source but we went to market with a point source version. For what it is worth we never had a point source panel that could touch our line source versions in terms of 3-D holographic image quality, not even close. Long story as to why we went with the point but it was not because it was better. Regards Moray James.

[I_Forgot]

I put them here so they'll have a permanent home:

http://mark.rehorst.com/Quad_ESL-63/index.htm

I_F

[Paul Ranson]

In the Baxandall document linked above (I think copied from 'The Loudspeaker and Headphone Handbook' edited by John Borwick.) he presents a simplified single ended schematic just containing the panel capacitances and with inductor values of 18H. Figure 3.44 in my copy of the Third Edition. I would expect this to be a good basis for simulation.

The actual inductors are apparently 3H, the delay per section is 20uS. The DC resistance of the output transformer secondaries is about 3.1K (each).

Paul

[Calvin]

Hi,

as Moray pointed out, one might have better results with a line source than with a point source.
Besides featuring very clever ideas, the Quads are still production cost optimized products. As a result I see some compromises that lead imo to the fact, that the ESL63 doesn´t really work as it is intended to!
Every ring of the diaphragm should be mechanically decoupled from it´s neighbouring rings. Otherwise there will be resonances and bending waves with increased distortion and narrower distribution than predicted!!?? There is a Patent dating from 1930 (!!) which describes exactly what the ESL63 does, or is intended to do.

And I´m with maudio: What is all the fuss about such a dull sounding speaker...no bass, no highs, no dynamics...every 63 I heard sounded as if a blanket were thrown over the speaker.

jauu
Calvin