It's not the coating works better at high frequencies, it works worse at low frequencies. The goal of an ESL panel is to have the charge fixed and uniform on the diaphragm. Because the charge can escape from the diaphragm, we need to replace that charge that is lost (with the bias supply). In order for it to be distributed on the diaphragm, there has to be a conductive coating on the diaphragm. But that coating should have a very high resistance so the charge can't move around on the diaphragm. If the charge moves around, the diaphragm motion will not be uniform and will be non-linear. So the charge should stay fixed, but over what timescale? The charge should stay fixed over the period of motion of the lowest frequency you care about that the diaphragm produces.
So if you have a low resistance coating, the charge can move more quickly, but if you are only using the panel as a tweeter, you might not care, but even though the charge can more faster (than a high resistance coating), your music signal is even faster still.
As a general rule, you want as high a resistance coating as you can which allows your diaphragm to fully charge up.
Low resistance coatings can also become a fire hazard because an arc can move laterally across the panel setting it on fire. This is really only an issue with metalized mylar (based on my experiments), but it was enough of a worry for Quad to put in a clamp circuit to shut down the 63's after an arc was detected.
Sheldon
quadesl.com
All Sheldon says seems correct to my knowledge. But the problem DIYers face is the difficulty of assessing or measuring almost any parameter of the coating.
Anybody have a handle on measuring a coating? Light passing through??? Weight???
With Dayton-Wright cells with big spacing in my open-air frames, I am able to use very high bias voltages, like 10kV. But for safety, I have series resistance in my home-brew bias supply on the order of 200 megOhms. It is easy to measure speaker loudness (or by simply listening). So my speakers take like 12 hours to get near to full loudness. That's a kind of measure of something in the bias/coating system.
Ben
ill be given it a try soon, got my airbrush in got all the air hoses and solenoids. so hopefully i can rig it up to my cnc machine soon, so i got a reliable way of spraying , then i can Fidel with the mixture of acrylic, carbon black and alcohol. to get the resistance im looking for. witch is the one of quad. i cant say what the resistance is but i can compare the 2
I no longer try for a low crossover point.
Midrange in my room turns out better in sealed cabinets, 2 push pull Dynavox 6.5" drivers, 1st order filtered at 1100 hertz, the mini panels sitting atop the cabinets, with 3.5" deep wings going about 1/3rd of the way up.
They're digitally highpassed at 70 hertz to a pair of open baffle, push pull subs.
The preamp allows digital delay to each channel for time alignment; in real time, it's easy to hear the effect of even 2 or 3 milliseconds on bass integration.
This actually gives more headroom, as it allowed me to lower the plate amp volume by half, to maintain the same volume at the listening position.
6db slope let's plenty of 'stat ambience come through, vocal integrity is preserved, and dynamics are the best I've heard at home.
Very hard to design for a low crossover freq. But besides the overt advantages of ESLs, having a crossover point outside the range of humanly sensitive hearing addresses the challenges of integrating cones and ESLs.
I'd guess a crossover below maybe 250 would be a nice goal. Then you could nave a single cone handle everything below if the cabinet sits between the ESL arrays. No kidding, that simple design might be wonderful.
I am fortunate in having panels that can be crossed over around 130 Hz. That means I just need the usual kind of sub-woofering and can place my subs anywhere they work best (and my wife tolerates). Once you get down to maybe 140 Hz (and have clean drivers and 24 dB/8ave crossover slopes), there is undetectable localization on music, despite what you may have heard about lab experiments with sine waves.
Ben
Last edited:
I did try several drivers, Alpair 12P and 12PW, were the next best for integration, but a little weak in bass.
Some old Danish Peerless 6 1/2" in ported cabinets were as good as the Dynavox pairs, but I need them in a 2nd system.
For some reason, I've never gotten open baffle voice to sound right in my room, but they work fine in other speakers I've sold, and helped setup.
In my case, two per channel Dynavox in sealed cabinets, wired in series, are a good match in efficiency, nearly ruler flat frequency response, and using the voice coil inductance, I get by with a smaller air core inductor, which appeals to my sense of cheapness.
For music, no sub required. For movies, open baffle bass now centred under the TV, crossed at 70 hertz, is awesomely loud, without disturbing the neighbors.
I've always liked the Dynavox, no notch filters required, like some of the Scanspeak drivers clients request, and no zobels either, like the Peerless others request.
And I never feel I'm risking an uber expensive driver, like the lowish power handling Alpairs, which need to be active, or use an L Pad.
Last edited:
Just got around to trying two 18" woofers on an open baffle, between the left and right mini panels.
I'm using a vintage recapped Miller Kreisel LP1S crossover.
With its shallow slope at 120 hertz, and 12mH inductor on one woofer, and a 4.7mH on the other, this is wonderful sound.
No pant flapping deep bass, but a hundred watts to each woofer, and 60 watts to the panels, is enough for great dynamics at my listening levels.
To be safe, I've changed out the supplied 22uF cap for a 15uF.
First order slopes are pretty easy going.
Thanks for the idea.
Yes indeed. If you can locate the woofer(s) near the L and R, no need to fret over higher freq leaking from the woofers and spoiling the localization.
But I have found substantial changes in woofer freq response depending on room location. So then you would need sharp slopes.
Ben
The panels are first order filtered at 525 hertz, and while the woofers are summed to mono, one is first order filtered at 275 hertz, and the other at 60 hertz.
They are side by side in push pull , no more than 2.5 feet from each panel .
The LP1S starts at 2nd order at 80 hertz, running up to 4th order at 120 hertz, if I understood the review properly.
I'm fortunate to have many different values of inductors at hand, these seem to work well. I think front wall reflections plus the dipole nature must be "fluffing up" the output of the mini panels, otherwise the large dip in the crossover makes no sense at all.
If I use the calculator recommendations, there's so much midbass the vocals become muddy.
At 98db/watt, with a qts of.69 at @40 hertz, these go deep enough for me, and are just loafing along.
The ancient SAE 2200 now powering them eliminated a hum problem from using a pretty new plate amp, frankly sounds better, while its meter leds never peak at more than .75 watts.
Totally arrived at the right combination of inductors plus the line level 2nd order filtering by trial and error; I don't have the math skills for more than rudimentary guesses at compound filter slopes, combining with room and speaker location variables.
This doesn't have same kind of punch that sealed cabinets provide, but on pop and acoustic music it sounds natural to me, and I just finished rewatching King Kong,Godzilla, and Rogue One.
I never had the sense I was missing anything, and preferred it to 5.1 surround with Tannoy fronts,and Neo 10 surrounds.
Last edited:
MIT,
is this the 6 1/2 " you speakof?
Dynavox LW6004PMR-N 6-1/2" Poly Cone Woofer
Xmax looks quite limited BUT FR is serious
is this the 6 1/2 " you speakof?
Dynavox LW6004PMR-N 6-1/2" Poly Cone Woofer
Xmax looks quite limited BUT FR is serious
You can't put ESL panels in sealed cabinets. I think the reason is that the ultralight membrane is too affected by resonances within the cabinet, unlike dynamic driver cones which are extremely stiff and heavy in comparison.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.