Thanks for confirmation that I understood your setup correctly. I look forward to your measurements.
Theory would suggest that your measurements will show a very well damped bass(~Q=0.5) with roll-off starting about 1/3 octave above the natural resonance of the panel. At higher frequencies you will see ripples in the response that reduce in magnitude as you go higher and higher and the felt filters out more of the rear diaphragm response. I’d estimate at least +/-3dB ripples. This may also account for the increased depth to the soundstage you are experiencing. It could also just be that sandwiching two sets of panels together improved the reproduction of mids/highs due to increased mechanical stiffness and damping of the stators.
One other thing you might experiment with is adding a LP filter to the amplifier driving the rear ESLs, so they only contribute at low frequenies …maybe < 500Hz. This would remove the HF response ripples while retaining the improved LF output.
I agree with your assessment of the likely outcome of adding a barrier between the panels. My experience is that adding any acoustically opaque barrier like MDF in close proximity to the ESLs will results in diminshing the LF response. You will find that increasing the density or amount of felt will have an incrementally similar effect althought not as drastic.
Theory would suggest that your measurements will show a very well damped bass(~Q=0.5) with roll-off starting about 1/3 octave above the natural resonance of the panel. At higher frequencies you will see ripples in the response that reduce in magnitude as you go higher and higher and the felt filters out more of the rear diaphragm response. I’d estimate at least +/-3dB ripples. This may also account for the increased depth to the soundstage you are experiencing. It could also just be that sandwiching two sets of panels together improved the reproduction of mids/highs due to increased mechanical stiffness and damping of the stators.
One other thing you might experiment with is adding a LP filter to the amplifier driving the rear ESLs, so they only contribute at low frequenies …maybe < 500Hz. This would remove the HF response ripples while retaining the improved LF output.
I agree with your assessment of the likely outcome of adding a barrier between the panels. My experience is that adding any acoustically opaque barrier like MDF in close proximity to the ESLs will results in diminshing the LF response. You will find that increasing the density or amount of felt will have an incrementally similar effect althought not as drastic.
You may find it interesting to know that positioning two identical ESLs closely one behind the other results in the resonance frequency going up, not down. Seems counter-intuitive, thinking about it from a diaphragm mass perspective. But, with ESLs, at low frequencies nearly the entire moving mass is made up of the airload on the front and back of the diaphragm…the diaphragm mass itself is only a minor increment. With two ESLs positions one behind the other, the moving mass is essentially unchanged, with just the small increment due to the trapped air between them. But, the suspension stiffness has been doubled. So, resonance frequency increases. Theoretically with very close spacing the increase could be as much as a factor of 1.4. In practice, I usually see in the range of 1.2 to 1.3 with diaphragm-to-diaphragm spacing of about an inch.
Thanks for your response. I was wondering if adding more felt would do what you suggest. I think I will also try the same set up but with 3/8" felt as opposed to the 5/8".
ah I see and makes sense now . small mass increase but 2X spring rate.
Sooo why do "they" do it? what's the advantage of dual dia?
I was talking to a designer back in the 80s that told me the best bass he ever heard was a large dual stacked ESL and I believe M Logan does it as well on the bass section of the CLS
ESLs are limited in the force per unit area that can be generated before the breakdown of air in the gap. This is an electric field strength limitation. See plot and ionization pic here: First time ESL builder
One way to increase the force per unit area limitation is with dual diaphragms. This is what ML was after in their CLX, getting the bass ESL panels loud enough to match up with the high frequency panels.
I don't think you will be able to adjust damping by changing position of two silk screen mesh relative to each other. At least I did not find it possible when I was measuring meshes and experimented with stacks of 2 or 3. Position and/or orientation seemed to have little affect on the resulting acoustic resistance. It was always very close to the sum of the individual resistances.
Also, unlike thick felt, silk screen mesh will let mid/high frequencies from the rear panel pass right thru nearly unimpeded. In the configuration you are experimenting with this will result in increasing the amplitude of the ripples in the midrange and high frequencies.
Thanks for that insight. In that case, it seems a balance needs to be arrived at with the felt thickness. Right now, the bottom of the front side panels are vented. If you look closely at the pic you will see the frame struts extend right to the floor. So across the bottom of the panel, it's open to the cavity. The back side is closed all around. How much of an influence do you feel this is having on it's present behavior? How could I expect it to change if I close it? So right now it's not isobaric.
Thanks for the expert info. I went looking for some measurements to see how the CLX does with bass output with the dual diaphragm isobaric bass panels:
https://www.hi-fiworld.co.uk/images/stories/Loudspeakers/Reviews/martin-logan-clx-fr1.jpg
Nice looking mid-bass levels from the CLX but given the expense and size of the dual diaphragm CLX bass panels the low-bass output seems disappointing.
Was this a failure of translating theory into practice or do the physics of ESL's doom them in terms of pumping out low bass?
I would not expect any noticeable change if you closed it off. Consider that the diaphragms of the front panels “look” acoustically like huge open windows to the rear panels. So, adding a small leak path at the bottom won’t have much effect at all.
Essentially yes, it is a matter of tradeoffs in size, output, and potential for overload. The CLX woofer panel is only about 650 in^2 with roughly 1/8” stroke. For reference, this is similar output capability to a single 12” dipole woofer on a flat baffle. However, dynamic woofers with their progressive spiders and surrounds, handle overload much more gracefully than ESLs do. With dipoles, you need on the order of 8x the displacement to go another octave lower at the same output level. So, not too surprising that you don’t find many ESLs that can play loud down low. The SoundLab A1s with nearly 4x the area do a credible job of playing deep bass, but even those can be overdriven with many modern pop recordings.
ML could have set the panel resonance lower, but you wouldn’t be able to play the deep bass(<50hz) very loud before the diaphragm would hit the stator…not a pleasant sound as you may know. The Quad ESL63 has similar radiating area and D/S with a bit lower diaphragm resonance and it runs into this issue. The CLS (predecessor to the CLX) also was easily overloaded with deep bass. My guess is that resonance was set to avoid making ugly noises or triggering some sort of protect mode while playing loudly. If a version was made with built in amplifier and DSP, perhaps a dynamic limiter could be included to allow playing deep bass at lower levels, but progressively limit it as the wick is turned up. Sounds like a nice DIY project
Another trick, that you can do to extend the lower limit is to increase the pathway for the sound pressure between front and back.
So by adding some "vings" that extends backwards lets say the double distance, that will make the diaphragm to move 1/2 the distance.
It will also reduce distortion with MORE than a factor of 2!
It will increase the rigidity, as well, something that ESL 63 lacks...
That's why i like bigger baffles...
So by adding some "vings" that extends backwards lets say the double distance, that will make the diaphragm to move 1/2 the distance.
It will also reduce distortion with MORE than a factor of 2!
It will increase the rigidity, as well, something that ESL 63 lacks...
That's why i like bigger baffles...
FWIW, after living with this arrangement for a while I can unambiguously say transient performance is noticeably improved. So I've come to the conclusion that isobarically coupling panels together tightens response more than simply rendering them more sensitive which is what I initially thought. Probably why low level resolution is also improved. Sounds a lot like my Meridian 105 monos with the Kef 105s. Very revealingly interesting comparison in that the speed of the Meridians is quite striking in their ability to resolve. So a testament to their performance since it took this arrangement with ESLs to experience/duplicate that, a quality that I've found could very well be lacking in an esl even compared to a dynamic driver. However, this is remarkably more of everything that the Meridians and KEFs have to offer, a lot more. But there appears to be a threshold/tipping point by which overall performance is recognizable over and above the basic parameters we use to assess sq., ie: sound stage, FR, IMD, dynamics, transient response, etc. It sounds 'purer', more 'precise', more 'real'. Exactly what set the Meridians apart from the 'also rans' or it's contemporaries, or, the typical 'ho hum' sound we're used to, in a word- speed. It's just amazing that a dynamic system can outperform an esl by what defines it's very superior attributes. Even more amazing is that boosting it's performance is what it took to recognize that. Here, with the speakers...there, with the amp. Bottom line is, amplifiers 'absolutely' don't all sound the same.
um, or do speakers
um, or do speakers
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.