Oaktsh,
I too have had reliability problems with my Quads through the years. I only hope that the newer successors to the 63 are more reliable. In the 20+ years I’ve owned them, I have accumulated a stockpile of damaged but repairable panels sufficient to build a new pair of speakers (I’m not exaggerating, sadly). To date I have repaired my Quads only with new panels, since I am paranoid about being able to exactly match the diaphragm characteristics with my own repairs (tension, conductance, etc.). Someday, I had hoped to repair these panels and create a second pair, maybe even trying to make my own delays lines, or by just ordering new Quad front end cards.
I live in a high humidity and high salt environment here on the Atlantic coast in Florida. Although I always keep the Quads in an air-conditioned space, this must still be a factor. Also, for a long time I had inadvertently positioned my naked Quads in such a place that sunlight could stream through nearby windows and strike the backs of some of the panels at certain times of the day. The UV seemed to have hastened the failure of the adhesive that holds the PC-based perforated stators to the plastic egg crate structure. The thin PC board then falls in toward the diaphragm causing arcing, especially in the corners. So, with Quads running naked, one must either cover them up, as you have, or cover nearby windows with drapes or blinds.
I haven’t tried stacking my Quads. This idea seems to run counter to the point-source raison d’être of the Quads. You’d have to be clear about what are you trying to accomplish by doing that. The sweet spot would shrink even more, and there would be lobing in the vertical dimension due to wavelength-dependent constructive and destructive interference. If the speakers were positioned properly and very precisely, pointed directly to the listener's head, and if the listener’s head were positioned exactly equidistant from each speaker center, then it might sing. Stacking older Quads (the so-called 57 model) makes more sense because the tweeter is a quasi-line source and stacking merely extends the line height. With 63s, you’d have two bull’s eyes, separated by roughly 3 feet. That’s a long distance in terms of the wavelength of treble frequencies. If it could be tried without investing in lots of fixtures and stands, it might be worth trying, but it never appealed to me.
I do think a subwoofer is a good improvement, if only to reduce the excursion burden on the ESLs, increasing maximum SPL capability. I use my own twin subs with 4 Focal 10 inches drivers. I cross these over (passively to the ESLs) at around 70 or 80Hz if memory serves. That's pretty low. A higher cross-over frequency would further ease the burden of the 63s, but I am happy with my compromise. I would consider this option before stacking, if you haven’t already.
I too have had reliability problems with my Quads through the years. I only hope that the newer successors to the 63 are more reliable. In the 20+ years I’ve owned them, I have accumulated a stockpile of damaged but repairable panels sufficient to build a new pair of speakers (I’m not exaggerating, sadly). To date I have repaired my Quads only with new panels, since I am paranoid about being able to exactly match the diaphragm characteristics with my own repairs (tension, conductance, etc.). Someday, I had hoped to repair these panels and create a second pair, maybe even trying to make my own delays lines, or by just ordering new Quad front end cards.
I live in a high humidity and high salt environment here on the Atlantic coast in Florida. Although I always keep the Quads in an air-conditioned space, this must still be a factor. Also, for a long time I had inadvertently positioned my naked Quads in such a place that sunlight could stream through nearby windows and strike the backs of some of the panels at certain times of the day. The UV seemed to have hastened the failure of the adhesive that holds the PC-based perforated stators to the plastic egg crate structure. The thin PC board then falls in toward the diaphragm causing arcing, especially in the corners. So, with Quads running naked, one must either cover them up, as you have, or cover nearby windows with drapes or blinds.
I haven’t tried stacking my Quads. This idea seems to run counter to the point-source raison d’être of the Quads. You’d have to be clear about what are you trying to accomplish by doing that. The sweet spot would shrink even more, and there would be lobing in the vertical dimension due to wavelength-dependent constructive and destructive interference. If the speakers were positioned properly and very precisely, pointed directly to the listener's head, and if the listener’s head were positioned exactly equidistant from each speaker center, then it might sing. Stacking older Quads (the so-called 57 model) makes more sense because the tweeter is a quasi-line source and stacking merely extends the line height. With 63s, you’d have two bull’s eyes, separated by roughly 3 feet. That’s a long distance in terms of the wavelength of treble frequencies. If it could be tried without investing in lots of fixtures and stands, it might be worth trying, but it never appealed to me.
I do think a subwoofer is a good improvement, if only to reduce the excursion burden on the ESLs, increasing maximum SPL capability. I use my own twin subs with 4 Focal 10 inches drivers. I cross these over (passively to the ESLs) at around 70 or 80Hz if memory serves. That's pretty low. A higher cross-over frequency would further ease the burden of the 63s, but I am happy with my compromise. I would consider this option before stacking, if you haven’t already.
Quads and subs
Nice to hear about you using Quads with other subs than the obvious Gradient alternative.
Anyone out there with more experience with sub woofers to 63s?
I use a 10" Tannoy TS-10 for both channels, separating actively at 80 Hz. Would it be better to use two subs for each channels, or is that overkill?
My Quads are also "stripped" of the dust cover and grilles. Does anyone have any good suggestions for cloth to cover the speakers are not in use, to protect them from sunlight and dust? I assume natural materials aren't suitable?
Cheers
Nice to hear about you using Quads with other subs than the obvious Gradient alternative.
Anyone out there with more experience with sub woofers to 63s?
I use a 10" Tannoy TS-10 for both channels, separating actively at 80 Hz. Would it be better to use two subs for each channels, or is that overkill?
My Quads are also "stripped" of the dust cover and grilles. Does anyone have any good suggestions for cloth to cover the speakers are not in use, to protect them from sunlight and dust? I assume natural materials aren't suitable?
Cheers
Hi Brian,
Thank you for your kind and informative reply.
My goal was to add a little more bottom end to the native ESL63 without loosing its magic.
I have read many diffirent approaches to achieve this one being the stacking the 63 turning them on the sides and using the bottom 4 panels connected to the most outer ring of the top ESL63 - only ineffect using them as bass panel only - like this:
http://www.integracoustics.com/MUG/MUG/tweaks/quad/hey_you/stacking.html
This is in a sense trying to achieve newer quad 988/998 by adding more bass panel
The other options would be to using a dipole- something like this:
http://www.euronet.nl/users/temagm/audio/dipolesub.htm
Brain, you seem to favor the later approach.
Could you give more detail of your sub and cross over? Do you attenuate the bass on the ESL63?
Thanks.
Thank you for your kind and informative reply.
My goal was to add a little more bottom end to the native ESL63 without loosing its magic.
I have read many diffirent approaches to achieve this one being the stacking the 63 turning them on the sides and using the bottom 4 panels connected to the most outer ring of the top ESL63 - only ineffect using them as bass panel only - like this:
http://www.integracoustics.com/MUG/MUG/tweaks/quad/hey_you/stacking.html
This is in a sense trying to achieve newer quad 988/998 by adding more bass panel
The other options would be to using a dipole- something like this:
http://www.euronet.nl/users/temagm/audio/dipolesub.htm
Brain, you seem to favor the later approach.
Could you give more detail of your sub and cross over? Do you attenuate the bass on the ESL63?
Thanks.
This is my solution.
Two Peerless 831727 10" woofers. Under the grill at the left there is the amplifier. Linear fase active filter crossing at about 95 Hz with 12 dB/oct.. One for each side. They couple perfectly together. All in all there is good powerfull bass and all the tones are played.
There are many other tips here:
http://user.tninet.se/~vhw129w/mt_audio_design/
Two Peerless 831727 10" woofers. Under the grill at the left there is the amplifier. Linear fase active filter crossing at about 95 Hz with 12 dB/oct.. One for each side. They couple perfectly together. All in all there is good powerfull bass and all the tones are played.
There are many other tips here:
http://user.tninet.se/~vhw129w/mt_audio_design/
Hi,
adding a dipole sub could be the best solution. We had always success using such subs with ESLs, even when any other sub failed in making the ESL sound better. The Quad is an prime example. Not only does such a sub extend the frequency range, but it provides the ESL with a rocksolid bottom with similar distribution character, raised dynamic headroom and -when the ESL is crossed over actively too- more resolution and less distortion. As a positive sideeffect the Quads rise in height. The sonic character of dipole subs -very clean, less booming and high resolution- fits nicely to the ESL´s sonic charcater.
I strongly recommend to give it a try.
@Brian
Yes, You´re right about the frequency distribution of ESL´s -planar or curved in special- but what would a Quad be without ts massive EQing?
You can´t compare some simple flat metal with such a -admittedly very elaborated- electrostat (and thats the reason what I marvel them for) that the Quad is. As far as I know each and every commercially successful ESL provides for some EQing and controlled dispersion. Known examples: MartinLogan does this in the crossover network and by curving, Audiostatic uses electrical segmentation.
The questions for me still remaining are:
-a) Is such a high grade of effort needed in first place
-b) Does the product do what it is supposed to in theory
-c) Do the design parameters lead to an efficient design
a) Basically every practical ESL doesn´t give a linear frequency response, so some measurements are needed to linearize this parameter. Since mechanical measurements work nonlinear over amplitude and might alter the distribution character too, it mostly better to linearize electrically. Luckily these circuits can be easy to implement actively or passively. As P. Baxandall explained in his afforementioned article, one can reach excellent results with (curved) line sources with much less network complexity.
b) The design claims to generate a point source or at least dipolar lobes with wider distribution than a single membrane of such size could give. In practise the ´63 is very directional..a true one man-speaker as it was said elsewhere.
Walker´s as well as Baxandall´s explanation look at each of the membranes ring-elements as if they were independent. In such a case a broadened distribution could be expected yes, but the ´63s membrane works as one big piece mechanically. I assume that the bending waves that are generated by each (with dicrete timesteps-drive and non-linear frequency response resulting from the electrical network) membrane segment interfere with the electrically generated motion of each segment. In other words. Does the bending wave that the inner segment generates travel with the same velocity -hence delay- over the membrane towards the outer rim as the electric network´s delay?? And what happens to the bending wave that is generated by the outmost segment travelling towards the centre of the membrane? Imo this leads to phasing problems that prevent the speaker from reaching the claimed dispersion. The earlier mentioned patent of 1930 already describes a speaker that is very similar to the Quad 63. The main difference is that the patent postulates, that each segment of the membrane has to be mechanically independent from every other segment! Now Quad could have used some damping rings between each segment, but that probabely would have spoiled the efficiency considerably, because of the small ´width´ of the rings.
So, is the dispersion in the horizontal plane wider or better or is it significantly different to the segmented or curved line source? I don´t think so. Since a linesource has to be narrow and tall You can´t use the easy solutions for this kind of speaker in smaller less high speakers. In case of the Quad it leads (amongst others) to more network complexity.
c) High efficiency is another claim in Walker´s papers, but to be honest..a speaker generating less than 90dB@2.83V/1m can hardly be described as efficient, especially not when the transformer ratio is well above 1:200(!!). Typically such SPLs are reached nowadays with ratios of 1:50 to 1:100. Besides creating problems winding transformers with such high ratios, trannies with lower ratio are easier to wind, will give better specs as well as they usually sound better. The much lower voltages with lower ratios are less dangerous, are easier to insulate and insulative materials will be more stable (less aging).
A conclusion:
I´m of the strong belief, that the optimum design of an ESL is that of a tall strip-like Linesource. Ease of build, low complexity of EQing, easy controlable dispersion and good efficiency will lead to excellent sonic results that can´t be reached, let alone beaten by other designs.
But if You only can or want to build a smaller ESL with considerable less heigth the Quad-way is probabely the best to go.
jauu
Calvin
adding a dipole sub could be the best solution. We had always success using such subs with ESLs, even when any other sub failed in making the ESL sound better. The Quad is an prime example. Not only does such a sub extend the frequency range, but it provides the ESL with a rocksolid bottom with similar distribution character, raised dynamic headroom and -when the ESL is crossed over actively too- more resolution and less distortion. As a positive sideeffect the Quads rise in height. The sonic character of dipole subs -very clean, less booming and high resolution- fits nicely to the ESL´s sonic charcater.
I strongly recommend to give it a try.
@Brian
Yes, You´re right about the frequency distribution of ESL´s -planar or curved in special- but what would a Quad be without ts massive EQing?
You can´t compare some simple flat metal with such a -admittedly very elaborated- electrostat (and thats the reason what I marvel them for) that the Quad is. As far as I know each and every commercially successful ESL provides for some EQing and controlled dispersion. Known examples: MartinLogan does this in the crossover network and by curving, Audiostatic uses electrical segmentation.
The questions for me still remaining are:
-a) Is such a high grade of effort needed in first place
-b) Does the product do what it is supposed to in theory
-c) Do the design parameters lead to an efficient design
a) Basically every practical ESL doesn´t give a linear frequency response, so some measurements are needed to linearize this parameter. Since mechanical measurements work nonlinear over amplitude and might alter the distribution character too, it mostly better to linearize electrically. Luckily these circuits can be easy to implement actively or passively. As P. Baxandall explained in his afforementioned article, one can reach excellent results with (curved) line sources with much less network complexity.
b) The design claims to generate a point source or at least dipolar lobes with wider distribution than a single membrane of such size could give. In practise the ´63 is very directional..a true one man-speaker as it was said elsewhere.
Walker´s as well as Baxandall´s explanation look at each of the membranes ring-elements as if they were independent. In such a case a broadened distribution could be expected yes, but the ´63s membrane works as one big piece mechanically. I assume that the bending waves that are generated by each (with dicrete timesteps-drive and non-linear frequency response resulting from the electrical network) membrane segment interfere with the electrically generated motion of each segment. In other words. Does the bending wave that the inner segment generates travel with the same velocity -hence delay- over the membrane towards the outer rim as the electric network´s delay?? And what happens to the bending wave that is generated by the outmost segment travelling towards the centre of the membrane? Imo this leads to phasing problems that prevent the speaker from reaching the claimed dispersion. The earlier mentioned patent of 1930 already describes a speaker that is very similar to the Quad 63. The main difference is that the patent postulates, that each segment of the membrane has to be mechanically independent from every other segment! Now Quad could have used some damping rings between each segment, but that probabely would have spoiled the efficiency considerably, because of the small ´width´ of the rings.
So, is the dispersion in the horizontal plane wider or better or is it significantly different to the segmented or curved line source? I don´t think so. Since a linesource has to be narrow and tall You can´t use the easy solutions for this kind of speaker in smaller less high speakers. In case of the Quad it leads (amongst others) to more network complexity.
c) High efficiency is another claim in Walker´s papers, but to be honest..a speaker generating less than 90dB@2.83V/1m can hardly be described as efficient, especially not when the transformer ratio is well above 1:200(!!). Typically such SPLs are reached nowadays with ratios of 1:50 to 1:100. Besides creating problems winding transformers with such high ratios, trannies with lower ratio are easier to wind, will give better specs as well as they usually sound better. The much lower voltages with lower ratios are less dangerous, are easier to insulate and insulative materials will be more stable (less aging).
A conclusion:
I´m of the strong belief, that the optimum design of an ESL is that of a tall strip-like Linesource. Ease of build, low complexity of EQing, easy controlable dispersion and good efficiency will lead to excellent sonic results that can´t be reached, let alone beaten by other designs.
But if You only can or want to build a smaller ESL with considerable less heigth the Quad-way is probabely the best to go.
jauu
Calvin
In the book of E. Fikier there are measurements shown of the ESL63. The dispersion charasteristic (page 83) looks fairly good with 8 kHz having at least + and - 20 degrees dispersion. So why should it be a 'one-man'-speaker? Without doubt at may have the best stereo image if you sit exactly in the middle because of better phase.
MJ Dijkstra said:
Every speaker design ever made or even makeable has best performance in only one spot. That is the nature of stereo reproduction. Two people between the speakers will get the best result if they are sitting single file. If they are next to each other, each listener will be closer to one speaker than the other, throwing off the amplitude balance. Each listener will be listening to the two speakers at different angles throwing off the phase balance. Accurate stereo reproduction depends on both amplitude and phase being matched. Listening off center (as in two or more people sitting side by side) guarantees that amplitude and phase will be mismatched. There might be something you could do with a lot of DSP hardware and listener head position tracking to modify the amplitude and phase in real time and maintain the stereo image for that person, but it's still a one man show.
The reason one might try to broaden high frequency dispersion is so that the speakers don't sound too dull when listening off axis- like when one is doing other things around the house/room while the music is playing.
I_F
Calvin said:
Calvin,
You raise an interesting question that I have pondered in the past. To what degree does the Quad speaker act as one panel mechanically rather than as independent rings? In other words, to what degree does the motion caused by the forces from one driven delay ring propagate to other regions outside or inside this ring? If there is ripple propagation, what is the propagation velocity compared to the sequential delay line timing? Of course, the Quad design complicates this question because it actually uses four panels cutting across these rings, breaking up modes.
What follows are my thoughts based on intuition rather than a solid mathematical proof, so I reserve the right to be wrong. I hope others will add to this discussion.
I think what you're asking about is longitudinal versus transverse modes of vibration, and the coupling between those two modes. As a related case in point, the old Suma Aria (now sold by the Museatex/Meitner/Melior folks) was a single large membrane excited in the middle by a voice coil. (Moray James, didn’t you say that you co-designed that speaker?) The center excitation propagated outward longitudinally as well as transversely, as would ripples in a pond. Edge terminations provided mechanical impedance that approximated the impedance of the membrane, absorbing most of the energy to keep edge reflections from returning toward the middle again. This ripples-in-the-pond approach is analogous to what the Quad does, although by a very different technique. An ESL uses quite different membrane material, much lower in mass, as I understand it, compared to the Aria. Furthermore, any ESL, even the Quad with its various rings, exerts electrostatic influence on all parts of the diaphragm, even where there is no driving audio voltage. If you tried to move the diaphragm, it would have to work against the forces provided by the stators and the output Z of the driving transformer or delay line (think of it as a reciprocal microphone in that case. Yes, the Quad, or any ESL, can be used as a large microphone.).
At most audio frequencies, the stiffness (spring constant) of the ESL diaphragm is entirely negligible compared to the air’s radiation resistance (see the figure from a Walker paper below). Only at very high frequencies does the diaphragm’s mass reactance start to intrude, rolling off highs (we strive to make it start above the audio band). In other words, over the audio band, the air’s radiation resistance has more control over motion than will the diaphragm’s own material properties, a very important point to this discussion. This situation would not have been the case with the Sum Aria, I think.
By similar thinking, we have the fact that over most of the audio range, a large uniformly-driven ESL diaphragm will move as a flat plane, not as a curved drum-head-like surface. Almost all bending will take place at the outer edges only. At low frequencies, there will be transition to a curved surface displacement, as the relevant impedances trade in relative magnitude, and stiffness becomes a factor. The tighter the diaphragm is stretched, the greater its stiffness, and the higher the frequency becomes where this transition takes place.
So, I believe that the longitudinal ripples would be suppressed in an ESL to a great degree, although not entirely, of course. In other words, the membrane’s motion is under the tight control of the nearby local electrostatic field and of the air. A longitudinal ripple component, coming in sideways from an adjacent ring would be attenuated by the other “stronger” controlling forces. I expect that there is still some mutual mechanical leakage between rings, and that Peter Walker, being no dummy, must have accounted for that nit in the tweaking of his finished design. At worst, this would have had a smoothing or blending effect at the edges of the various rings forming a more continuously varying spherical approximation, perhaps even for the better. But I would not expect any major deviation from expected “theoretical” performance due to longitudinal ripple propagation.
Actually, I agree with you. This has been the focus of my DIY work (slowed by lack of time). Although mine is not dipolar...
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.