Arg! forgot to include comment about the contribution of the resistive component of the airload to the total load presented to the DD amp.
As an example, if we hooked up all the ESL63 panels and drove them uniformally, the total electrical load from roughly 300Hz upward(ie where the airload is nearly purely resistive) can be approximated by a parallel RC circuit with:
C = 900pf (total stator to stator capacitance)
R = 25Meg (from total resistive airload)
There is a resistive component, but phase would still be -85deg or more.
As an example, if we hooked up all the ESL63 panels and drove them uniformally, the total electrical load from roughly 300Hz upward(ie where the airload is nearly purely resistive) can be approximated by a parallel RC circuit with:
C = 900pf (total stator to stator capacitance)
R = 25Meg (from total resistive airload)
There is a resistive component, but phase would still be -85deg or more.
So the '7' areas are driven from R5/R7 then?
When I then look at the ESL 989, are the 'extra' bass panels driven from the same '8' tap?
Jan
I wonder how Acoustat panels would respond to DD amp drive if they are unsegmented and driven full range? There dont seem to be any impedance plots of raw Acoustat panels - can that actually be done?
Published impedance curves from Acoustat includes the transformer while the Stereophile has transformer and a woofer crossing over at 300Hz:
It should be possible to do freq response and impedance measurements on the panels, as long as you take care of safety.
What I did with the Quads was to measure level vs. freq at the xformer secondary, which is the panel voltage vs. freq.
I also used a small resistor in series with the panel and measured the voltage vs. freq across it, which gives me the current vs. freq.
Then V/I of course gives me impedance vs. freq. The phase angle will give you an idea of how 'capacitive' it is.
Of course you can also measure the pure capacitance part directly.
Jan
What I did with the Quads was to measure level vs. freq at the xformer secondary, which is the panel voltage vs. freq.
I also used a small resistor in series with the panel and measured the voltage vs. freq across it, which gives me the current vs. freq.
Then V/I of course gives me impedance vs. freq. The phase angle will give you an idea of how 'capacitive' it is.
Of course you can also measure the pure capacitance part directly.
Jan
Yes. Areas marked "7" are driven from R5/R7.
I have not personally inspected any 6 panel Quads, but I know stokessd owned a pair at some point and mentioned that the additional bass panels were just connected in parallel with the other bass panels driven from the "8" tap with no other changes to the circuit. I can't seem to locate the post at the moment though.
OK, just for my understanding. Let us assume that the added panels have the same characteristics as the existing ones.
If the drive signal to those panels remains the same*, then, by putting them in parallel, they get the same drive as before so the output power doubles (double the # of panels).
This would give +3dB SPL in the low register. Correct?
Jan
* I think that is a reasonable assumption as all those delay line sections predominantly work in the mid- and high frequency bands.
If the drive signal to those panels remains the same*, then, by putting them in parallel, they get the same drive as before so the output power doubles (double the # of panels).
This would give +3dB SPL in the low register. Correct?
Jan
* I think that is a reasonable assumption as all those delay line sections predominantly work in the mid- and high frequency bands.
The short answer is that you would get roughly 1-1.5dB lift in the bass below about 150Hz and the bass is extended by about 1/4 octave.
The details require understanding a few principles from the Baxandal ESL chapter.
1) The farfield SPL response is directly proportional to the summation of the stator-to-stator currents through all ESL panels. (ie Walker Equation)
2) For panels driven directly from the transformer, the SPL increases +6dB/Oct with increasing frequency. Double the panel area and the total current doubles so SPL increases by +6dB
3) If you put a high value resistance in series with a panel, the current would be constant vs frequency, so the farfield response would be flat. Double the panel area and the current is still the same so SPL stays the same. But with twice the area, the diaphragm excursion is half what it was for a single panel.
4) With moderate resistance in series with a panel, there will be a transition from flat response at higher frequencies to +6dB/oct slope at lower frequencies. Double the panel area and the transition frequency falls by an octave.
5) For circular or low aspect ratio rectangular panels, the mass of the airload on the diaphragm at resonance is roughly proportional to Area^(3/2) when operating in free-air with no baffling. So if you have a panel with resonance Fs when played by itself, it will be roughly Fs/1.6 if you place two panels adjacent to each other.
Although the top and bottom panels of the ESL63 are often called the bass panels, all 4 panels contribute to the bass response. The top and bottom panels do have roughly 1.5Meg in series with them, so they are driven at constant current above about 300Hz. Adding 2 more “bass” panels increases the total radiating area by a factor of 6/4 = 1.5. So we might expect to see an increase in bass response by 20*LOG(1.5) = +3.5dB. But, since the panels are added in parallel with those already driven through the 1.5Meg, above 300Hz there is essentially no change. Below 300Hz there is a progressive increase that reaches the expected +3.5dB by about 60Hz.
However, since the area was increased by a factor of 1.5 the mass of the airload is increased by roughly 80% so resonance frequency would be reduced by about 35%. Having baffling around the panels and having the panels set on the floor also affects the mass of the airload so the theoretical reduction in resonance frequency of 35% for panels in free air is not seen. With a bit more math, I’d estimate the value at about 20-25%. When resonance is reduced, the equalization peak from the underdamped resonance (Q = 2.5) is shifted down as well. The end result is an overall lift of the response below 150Hz by about 1.5dB and an extension of the bass by about 1/4 octave.
The details require understanding a few principles from the Baxandal ESL chapter.
1) The farfield SPL response is directly proportional to the summation of the stator-to-stator currents through all ESL panels. (ie Walker Equation)
2) For panels driven directly from the transformer, the SPL increases +6dB/Oct with increasing frequency. Double the panel area and the total current doubles so SPL increases by +6dB
3) If you put a high value resistance in series with a panel, the current would be constant vs frequency, so the farfield response would be flat. Double the panel area and the current is still the same so SPL stays the same. But with twice the area, the diaphragm excursion is half what it was for a single panel.
4) With moderate resistance in series with a panel, there will be a transition from flat response at higher frequencies to +6dB/oct slope at lower frequencies. Double the panel area and the transition frequency falls by an octave.
5) For circular or low aspect ratio rectangular panels, the mass of the airload on the diaphragm at resonance is roughly proportional to Area^(3/2) when operating in free-air with no baffling. So if you have a panel with resonance Fs when played by itself, it will be roughly Fs/1.6 if you place two panels adjacent to each other.
Although the top and bottom panels of the ESL63 are often called the bass panels, all 4 panels contribute to the bass response. The top and bottom panels do have roughly 1.5Meg in series with them, so they are driven at constant current above about 300Hz. Adding 2 more “bass” panels increases the total radiating area by a factor of 6/4 = 1.5. So we might expect to see an increase in bass response by 20*LOG(1.5) = +3.5dB. But, since the panels are added in parallel with those already driven through the 1.5Meg, above 300Hz there is essentially no change. Below 300Hz there is a progressive increase that reaches the expected +3.5dB by about 60Hz.
However, since the area was increased by a factor of 1.5 the mass of the airload is increased by roughly 80% so resonance frequency would be reduced by about 35%. Having baffling around the panels and having the panels set on the floor also affects the mass of the airload so the theoretical reduction in resonance frequency of 35% for panels in free air is not seen. With a bit more math, I’d estimate the value at about 20-25%. When resonance is reduced, the equalization peak from the underdamped resonance (Q = 2.5) is shifted down as well. The end result is an overall lift of the response below 150Hz by about 1.5dB and an extension of the bass by about 1/4 octave.
Last edited:
Stephen, thanks, you can explain it very clearly.
Yesterday I had a slow day so I started to speculate. What if I added more (bass) panels to my 989's.
Say 2 or three on each side, in portrait orientation. Doubling the # of 'bass' panels.
It would also provide some baffling for the existing panels, sort of like adding wings, but then active wings (how's that for a sales pitch ;-).
With 2 panels at each side the total radiation area would increase by a factor 10/6 or roughly 1.7 times.
It would extend the bass a bit further, but more important (for me) it would allow higher playing levels.
I guess you can't continue to add panels with no other changes so the drive level/impedance must probably be adapted.
Does that sound like crazy to you?
Jan
Yesterday I had a slow day so I started to speculate. What if I added more (bass) panels to my 989's.
Say 2 or three on each side, in portrait orientation. Doubling the # of 'bass' panels.
It would also provide some baffling for the existing panels, sort of like adding wings, but then active wings (how's that for a sales pitch ;-).
With 2 panels at each side the total radiation area would increase by a factor 10/6 or roughly 1.7 times.
It would extend the bass a bit further, but more important (for me) it would allow higher playing levels.
I guess you can't continue to add panels with no other changes so the drive level/impedance must probably be adapted.
Does that sound like crazy to you?
Jan
Hi Jan,
I'm in the process of getting my ESL-63 up and running again after being stored away for many years. My thoughts on adding bass extension to the quads has always been adding either Gradient SW63 or John Krekovsky cardioid style bass system. The rationale for this is to relieve the quad of the low frequency content since no amount of bass panels will change the quads relative high panel resonances and I therefor want to relieve the quads of these frequencies as much as possible.
I realize putting the 989's on top of an pair of Gradient SW-63 is not feasible
Just my thoughts.
Mogens (who is following your DD project with great interest)
I'm in the process of getting my ESL-63 up and running again after being stored away for many years. My thoughts on adding bass extension to the quads has always been adding either Gradient SW63 or John Krekovsky cardioid style bass system. The rationale for this is to relieve the quad of the low frequency content since no amount of bass panels will change the quads relative high panel resonances and I therefor want to relieve the quads of these frequencies as much as possible.
I realize putting the 989's on top of an pair of Gradient SW-63 is not feasible
Just my thoughts.
Mogens (who is following your DD project with great interest)
Hi Mogens, yes a separate sub is probably the sensible thing to do.
Although I often hear people complain about the difficulty to 'match' an electromagnetic driver to an electrostatic one.
I don't know how much of that is popular lore - I have owned two different pairs of Martin Logans and they sounded OK to me. in the lower department.
Jan
Although I often hear people complain about the difficulty to 'match' an electromagnetic driver to an electrostatic one.
I don't know how much of that is popular lore - I have owned two different pairs of Martin Logans and they sounded OK to me. in the lower department.
Jan
Hi Jan,
I think much of "difficulty to match" might be caused using the wrong type of bass system. Also, I have never heard an conventional sub-woofer sound great with a pair of quads. Not that I have heard that many.
I might also add that I'm as such not necessarily interested in an sub-woofer, but more an bass extension to relieve the quads of the low-frequency content in order to have more head-room for the quads. From what I have read about adding the Gradient SW-63 it allows for nearly 10 dB higher volume. That's more than enough for my 35 m2 room.
Don't know if you know this Dutch site https://www.meddens.eu/audio/dipolesub.htm
Btw. I have never been a fan of Martin Logan. Friends of mine has had different types going back to the original CLSII. I have always thought they sounded weird compared to the quad. That's of course my opinion and taste.
Mogens
I think much of "difficulty to match" might be caused using the wrong type of bass system. Also, I have never heard an conventional sub-woofer sound great with a pair of quads. Not that I have heard that many.
I might also add that I'm as such not necessarily interested in an sub-woofer, but more an bass extension to relieve the quads of the low-frequency content in order to have more head-room for the quads. From what I have read about adding the Gradient SW-63 it allows for nearly 10 dB higher volume. That's more than enough for my 35 m2 room.
Don't know if you know this Dutch site https://www.meddens.eu/audio/dipolesub.htm
Btw. I have never been a fan of Martin Logan. Friends of mine has had different types going back to the original CLSII. I have always thought they sounded weird compared to the quad. That's of course my opinion and taste.
Mogens
Hi Jan,
Sure. https://drive.google.com/file/d/1tArMcDPhRkNdUp8XKOG9TWhwkSD9E07S/view?usp=sharing Please try this link.
I can't say if 1K Euro is reasonably or not. Big woofers and cabinet does cost some money either way.
Mogens
Sure. https://drive.google.com/file/d/1tArMcDPhRkNdUp8XKOG9TWhwkSD9E07S/view?usp=sharing Please try this link.
I can't say if 1K Euro is reasonably or not. Big woofers and cabinet does cost some money either way.
Mogens
Adding that many panels to an ESL63 type design does start to sound a bit crazy to me...definitely diminishing returns.
Remember that no matter how many additional bass panels you add in parallel with the existing one, the middle two panels are still being driven directly from the transformer at low frequencies so they will still overload at the same input signal level. I guess if you added enough panels such that you had a rising bass response, you could then use a line-level EQ to flatten the rise and thus achieve some headroom for those panels.
Find attached GB815978 and the mostly equivalent US patents.
You may find it interesting that the patents discuss the concept of “active” baffling for low frequencies that you mentioned. See Fig 12 in GB patent or Fig 10 in US patents. If only driven at low frequencies, they would be mostly acoustically transparent at high frequencies so would avoid the usual cavity effects of U-frame baffles, or mid-range diffraction effects of large flat baffles.
Thanks for the patents Stephen. Fascinating reading.
On the bass panels, I understand why they also want to center panels to contribute to the lower octaves, everything helps. Smart move.
And you can't continue to double the number of panels, definitely diminishing returns.
It was more a thought experiment for me.
Another slant: what is your opinion on adding a subwoofer à la Gradient?
It would relief the ESL of the lowest notes, probably lowering IMD although the audibility of that is questionable.
But would help the system to play louder, no?
Jan
On the bass panels, I understand why they also want to center panels to contribute to the lower octaves, everything helps. Smart move.
And you can't continue to double the number of panels, definitely diminishing returns.
It was more a thought experiment for me.
Another slant: what is your opinion on adding a subwoofer à la Gradient?
It would relief the ESL of the lowest notes, probably lowering IMD although the audibility of that is questionable.
But would help the system to play louder, no?
Jan
You are correct, the top and bottom panels in the stack of 6 are just attached to the 2nd and 5th panels in the stack with the same little curly jumpers as the inner bass panels are attached to the mid panels (with the concentric rings). So the additional panels are indeed simply in parallel with the inner bass panels. There are no changes to the delay line or anything else in the electrical drive side of the speakers.
I have a set of 2912's in my main system, and I've owned 989's as well. (BTW, the 989's need the panel rails attached to the outer frame extrusions because they are so floppy that the panels can hit and damage the dust covers in shipping. Nice work from the hacks that took over the quad name). I don't find that the 6 panel quads are really worth the extra size and cost. They look ridiculous in my room, and the little performance gain isn't worth it. IF you want more bass extension a sub crossed low (100hz or slightly below) with a steep active crossover is the best way IMHO.
I use a pair of 2805's in my shop (driven by the remarkable Benchmark AHB2 amp and their accompanying DAC3), the 4 panel quads are really the way to go. And frankly the 63's are the nicest looking of all the modern quads; in my opinion of course.
Sheldon