Jazzman's ESL/TL speakers

[Toppsy]

steve s wrote

Scott thats a challenge for your next design

Certainly a challenge that Scott would likely take up, but then I'd have to build the prototypes, lucky me hey .

I have to say that these speakers are a truly wonderful build. Great work there Charlie, I like the looks and build quality very much. I still have a fondness for ESL's but I gave up on trying to integrate a cone sub-woofer with mid/treble electrostatic panels and get the slower reacting cone to fully match the speed of the ESL panels. At one time with my ER Audio ESLIII's I thought I had cracked it but then a fellow mate of mine came over for a listen and brought along his classical music trained son with him and he instantly picked up on the speed mismatch. Me I could not discern any speed mismatch (perhaps something to do with my age?) but had to bow to his superior trained ear. It would be interesting to find out if the TL approach would resolve this issue. Something to ponder on.

I know Martin Logan sell a lot of their ESL's with an integrated sub, but having heard a couple pairs of their speakers I have not been convinced they work that well together despite all the excellent media reviews.

[Jazzman53]

steve s wrote

Scott thats a challenge for your next design

Certainly a challenge that Scott would likely take up, but then I'd have to build the prototypes, lucky me hey .

I have to say that these speakers are a truly wonderful build. Great work there Charlie, I like the looks and build quality very much. I still have a fondness for ESL's but I gave up on trying to integrate a cone sub-woofer with mid/treble electrostatic panels and get the slower reacting cone to fully match the speed of the ESL panels. At one time with my ER Audio ESLIII's I thought I had cracked it but then a fellow mate of mine came over for a listen and brought along his classical music trained son with him and he instantly picked up on the speed mismatch. Me I could not discern any speed mismatch (perhaps something to do with my age?) but had to bow to his superior trained ear. It would be interesting to find out if the TL approach would resolve this issue. Something to ponder on.

I know Martin Logan sell a lot of their ESL's with an integrated sub, but having heard a couple pairs of their speakers I have not been convinced they work that well together despite all the excellent media reviews.

I agree that ML's don't blend as well as reviewed.

Roger Sanders contends that low system Q is essential and only in a horn or TL can a woofer hope to blend well with an ultra-fast ESL. A horn for bass frequencies would be truly huge so that leaves a TL. His recommendations for the woofer are low-Q, low moving mass and a strong motor (magnet) coupled with a weak (low inductance) voice coil-- and low inductance trumps all other parameters. This is the approach I took with my hybrid ESL.

[Scottmoose]

An interesting field. I can't design electrostatics themselves since that is driver design rather than system, but assuming a quality existing panel of limited BW, integrating into a complete speaker would certainly be an interesting challenge.

A ~aperiodic TL has a very good impulse response with rapid decay / little ringing. Only a critically damped sealed box is likely to improve on it, although a very low tuned EBS reflex will not be shabby through the main passband. Low system Q and TL are a problematic configuration though, unless you can get a driver with a very low F0. A low Q driver with a moderate F0 will see mass-corner rise, and in a heavily damped TL, an over-damped LF response, so it would seem the above has achieved this balance well. Front horns certainly work, but as noted, need to be very large and / or corner loaded for this kind of use. There are ways & means, but none small.

As for the drivers, low VC inductance is certainly valuable inasmuch as that is what determines the ultimate upper rolloff of the driver. Usually it will be rolling off before that though through other mechanical effects, but it's still worth having to ensure the electrical low pass effect of the VC inductance is at a reasonably high frequency.

Big magnet = no bad thing, but Q is the guide re driver behaviour at F0 & the ultimate motor strength through the rising response ('acceleration') BW, since that shows the relationship between the electrical damping and the driver's mass / compliance.

[Jazzman53]

I believe Sanders' recommendation for the strong magnet motor is to offset the driving force/efficiency lost by choosing a weak voice coil.

And I'm starting to realize how little I know about TL's

[rowuk]

I think the BIGGEST problem has more to due with the entire system rather than the match at crossover. We have to match UHF extension with LF extension. If we are using a large ESL panel and no dedicated UHF driver, this could make the bottom sound completely wrong - and the woofer is not even at fault. Actually I think a sealed box would be easiest to integrate - it can have the softest slope at its bottom knee. Perhaps a heavily damped TL could get close.

If we do not crossover near the panels resonance, rather at least an octave higher, most of the problems should be solved

[Scottmoose]

I believe Sanders' recommendation for the strong magnet motor is to offset the driving force/efficiency lost by choosing a weak voice coil.

And I'm starting to realize how little I know about TL's

A weak voice coil? A VC can't be 'weak' or 'strong' except in terms of its power-handling capacity, and low inductance is not automatically indicative of this. I'm sorry, but Saunders appear to be talking absolute nonsense. Where does this come from -is it some kind of article?

[Scottmoose]

I think the BIGGEST problem has more to due with the entire system rather than the match at crossover. We have to match UHF extension with LF extension. If we are using a large ESL panel and no dedicated UHF driver, this could make the bottom sound completely wrong - and the woofer is not even at fault. Actually I think a sealed box would be easiest to integrate - it can have the softest slope at its bottom knee. Perhaps a heavily damped TL could get close.

If we do not crossover near the panels resonance, rather at least an octave higher, most of the problems should be solved

To an extent I'd agree, although I suspect the major problem is poor system matching between MC bass units & the panel, and inadequate XO design, e.g. running the LF driver up higher than it's capable of doing, not adequately suppressing breakup modes in its stopband, & insufficient attention to the polar & power responses.

A well designed aperiodic TL will equal or improve upon most sealed boxes other than a critically damped alignment. A very low tuned EBS reflex can also get close. In the case of the system here, the TL & panel have clearly been well integrated.

[Jazzman53]

[Scottmoose]

[Jazzman53]

The XO points seem to be pulled out of a hat somewhat; were we taking about a high pass filter above the driver's natural resonance, it would seem more reasonable. As far as a low pass is concerned, it doesn't make a heck of a lot of difference how far above F0 it is. However, higher order filters typically exhibit greater GD around the crossover frequency. This can be over-stated and depends on the quality of the filter design, how steep the slopes are and where they are located.

Not sure we're on the same page here... the "diaphragm resonance" refers to the ESL diaphram, not the woofer diaphragm. Sanders sets the crossover to work around the ESL's drum-head resonance, and chooses a woofer with compatible range.

[IslandPink]

This kind of integration problem crops up with every really great driver unit. If you're running a ribbon, or a compression driver/horn for upper mids ; or ESL panels ; the big task is how to get the next driver down ( in freq ) to match up . It's the tonal/transient qualities that are really hard to get right .
I think a lot has to do with driver cone mass . You're going from a diaphragm that's really light to something a lot heavier . I've only just managed to get a decent blend to the Radian745+horn by using a low cone-mass FE208Ez on OB . I also think low Rms ( which these have) has a lot to do with getting good tonal integration at low drive levels .
For the ESL / bass issue it's almost the same . Martin Logan's solution is almost as bad as you could choose , I know a mate of mine in Essex talks regularly about how he needs to get down to work and make a better DIY bass option for those, as it's quite poor . I guess it's just a compact commercial solution as usual . What Charlie has here is a serious and much less compromised approach to the problem .
I can't get my head around all of the trade-offs well enough to know exactly what Rms, Qts or BL combination is the best , but this unit is one good data point .

[Jazzman53]

Here's an interesting white paper (not from Sanders) on woofer speed:
http://www.adireaudio.com/Files/WooferSpeed.pdf

And here is a compilation of forum posts by Roger Sanders. It's a rather lengthy read and does not contain the particular post I was looking for but it does explain his design philosophy in depth:
http://www.whatsbestforum.com/showthrea ... er-Sanders

And here is a Youtube video explaining his preference for flat panels:
https://www.youtube.com/watch?feature=p ... df3VA06iSA

[steve s]

Here's an interesting white paper (not from Sanders) on woofer speed:
http://www.adireaudio.com/Files/WooferSpeed.pdf

]

I have this one printed out from quite a few years ago, to me it provides more questions than answers

Is adding an inductor in series the same as a 1st order low pass, surely it would affect the phase and add an impedance? Would that not have the same effect on the driver response.

If bl is calculated as a constant, the substituting of high bl driver will change the effect ?

The test rig uses what i would guess is a high current 3400 watt amp, not my 4 watt valve amp ? That a rather different situation to most peoples set ups,
And a miss match as in my view... How many 6.5 " woofers do you see connected to pa amps?

I find briggs books from the 50's much more relevant as when i try a range of varying bl drivers i've come up with a similar end results to what he describes.

Im certainly no expert in this field by the way..?

[Jazzman53]

Here's an interesting white paper (not from Sanders) on woofer speed:
http://www.adireaudio.com/Files/WooferSpeed.pdf

]

I have this one printed out from quite a few years ago, to me it provides more questions than answers

Is adding an inductor in series the same as a 1st order low pass, surely it would affect the phase and add an impedance? Would that not have the same effect on the driver response.

If bl is calculated as a constant, the substituting of high bl driver will change the effect ?

The test rig uses what i would guess is a high current 3400 watt amp, not my 4 watt valve amp ? That a rather different situation to most peoples set ups,
And a miss match as in my view... How many 6.5 " woofers do you see connected to pa amps?

I find briggs books from the 50's much more relevant as when i try a range of varying bl drivers i've come up with a similar end results to what he describes.

Im certainly no expert in this field by the way..?

I'm sure that I have less expertise than you. My interest is more focused on whether voice coil inductance impedes a woofer's transient speed. If it impedes changes in current thru the coil, then it seems logical to me that it would and/or produce a phase shift between the woofer and stat panel. Does that make sense?

[rowuk]

Instead of looking for a fictive reason for concern, would it not make more sense to simply listen to the woofer by itself, full range for a considerable amount of time to HEAR what it does to music in various frequency ranges?

An accellerometer could quickly prove speed but not SOUND. Even if a speed difference were measurable, that fact is insignificant if those speed differences only result in suporting our conjecture.

There is little or no correlation between specs and sound.

If the woofer sounds great into the midrange, then it is great for that use case - regardless if our measurement shows a speed difference.

The ESL does not sound the way it does because of speed. The sound has more to do with the amount of air being moved, with that size, there is phase cancellation (directivity), which creates a completely different interaction with the room. There is also a great influence on the sound made by the plastic foil and its tension.

Speed is a convenient term that has no correlation to sound. We know how much speed is necessary for 20 khz, 1k or 20 hz. Audiophiles still talk about "fast bass" which does not exist.