When I add silicone dots to an ESL panel, the main purpose is to add stability and adjust the resonant frequency upward. I have not noticed any increase in resonance modes with two rows of dots, but I usually apply a tightly woven damping cloth to the back stator similar to what QUAD did in there ESL63. This dampens all of the diaphragm resonance modes except the fundamental which remains with a Q between 2 and 3.
Viewing pictures of the beautifully constructed Capaciti panels, I have often wondered why so many rows of silicone dots are used when Audiostatic uses only one row. Whatever the reasons, it obviously works well from what I have seen in posted response measurements.
Concerning sectioning of the diaphragm into strips of different widths. I have never understood the benefit for extending LF response. If anything, it seems like you are loosing radiating area for LF. With 3 strips as you mentioned with resonances at 60, 80, 100 Hz. The 100Hz section rolls off at 12dB / octave below 100Hz, so by 60 Hz it is adding nearly nothing to the LF output. You have effectively thrown away that area in terms of LF output. Perhaps I am thinking of it wrong, but I figured you want as much area as possible pushing out the LF sound. Maybe it works OK for response equalization if no resonance damping is used on the strips leaving them with hi-Q fundamental resonance. But, this compromises spectral decay as you mentioned. Personally, I much prefer the sound once some damping is added.
However, if the diaphragm is sectioned in to 3 equal widths with resonance of, say, 70Hz, all the diaphragm area would be moving together equally for the LF output. This should give similar results to a silicone dot damped diaphragm with the same total area and same resonant frequency, shouldn't it?
bolserst
Hi Bolsert,
if you damp the rear your will loose transparency of sound. I tried this this with different materials and the result is that you loose breathing of voices and that the sound seems to stick more to the panels.
My measurements prooved that by showing increased spectral decay in the midrange.
I do not know waht others think about the quads, but imo they do not sound very open, but sound like wrapped into something. And this something are the dust cover, the streamin resistance of the PCB stator and the rear damping.
But you are right, dampening makes resonance control much easier and in addition the rear damping makes anESL less critical to rear wall interaction.
Splitting in different sections as ment for the finals do not mean that you gain sound pressure in the lowest frequencies (e.g. 40 Hz) but in the upper low range (60-120HZ), without electrical equalization. If you listen to a Final ESL you will recognize that there is no deep bass.
Imo some Fullrange ESL do not suffer from the aspect of missing deep bass, but from a sound pressure dull in the range between 80 and 200 Hz, where the open baffle effect kills efficiency. There is no worse sound for an fullrange ESL havin a dull from 80-200 Hz and a extended pressure at 40-50 Hz. It sounds boomy at low and anemic at upper bass frequencies. This is why most people call ESL-Fullrangers a one note bass transducer.
As a result dont focus on best performance at 40 Hz, squeezing out the last decibels in that range, but take care that your frequency range is linear from 80 to 500 Hz, which is tricky or even impossible enough for a small ESL with limited area.
Capaciti
if you damp the rear your will loose transparency of sound. I tried this this with different materials and the result is that you loose breathing of voices and that the sound seems to stick more to the panels.
My measurements prooved that by showing increased spectral decay in the midrange.
I do not know waht others think about the quads, but imo they do not sound very open, but sound like wrapped into something. And this something are the dust cover, the streamin resistance of the PCB stator and the rear damping.
But you are right, dampening makes resonance control much easier and in addition the rear damping makes anESL less critical to rear wall interaction.
Splitting in different sections as ment for the finals do not mean that you gain sound pressure in the lowest frequencies (e.g. 40 Hz) but in the upper low range (60-120HZ), without electrical equalization. If you listen to a Final ESL you will recognize that there is no deep bass.
Imo some Fullrange ESL do not suffer from the aspect of missing deep bass, but from a sound pressure dull in the range between 80 and 200 Hz, where the open baffle effect kills efficiency. There is no worse sound for an fullrange ESL havin a dull from 80-200 Hz and a extended pressure at 40-50 Hz. It sounds boomy at low and anemic at upper bass frequencies. This is why most people call ESL-Fullrangers a one note bass transducer.
As a result dont focus on best performance at 40 Hz, squeezing out the last decibels in that range, but take care that your frequency range is linear from 80 to 500 Hz, which is tricky or even impossible enough for a small ESL with limited area.
Capaciti
Hi,
damping mats are quite effective in damping resonances as well as sucking life out of the ESL
Their effect is frequency and amplitude dependent, which is not what one wants to achieve in first place. And besides it spoils the optics, which is what we all like most about ESLs...don´t we?
jauu
Calvin
damping mats are quite effective in damping resonances as well as sucking life out of the ESL
Their effect is frequency and amplitude dependent, which is not what one wants to achieve in first place. And besides it spoils the optics, which is what we all like most about ESLs...don´t we? jauu
Calvin
Visual appeal of ESLs was NOT the strong point of mine (rebuilt and modified Acoustats)! I tried them with and without damping mats. I'm not sure that I'd call the effect of the damping bad, just different. It definitely helped to tighten the mid bass and focus the images better, at least in my room. I had about 50% of the active area in thin felt, spaced about a centimeter from the diaphragm.
I too have found that damping material can have a negative effect in the midrange, but improved transient response in the mid-bass. As usual with speaker design...there are trade-offs. The trick is to find a cloth that will provide enough acoustic resistance to damp the diaphragm resonance adequately without reflecting back the midrange sound and creating a resonant cavity between the damping material and the diaphragm. Very thin synthetic felt (1/32”) works quite well. The more common 1/16” felt also works reasonably well, but the deficiencies in the midrange start to show up. I have also found some tight weave polyester cloth that work…damping is not quite as good in the bass, but midrange effects are nearly unmeasurable. Still searching
Ah! This makes sense to me now. By splitting up the panel in to different width sections they are NOT trying to extend the LF response. Rather, they are trying to fill in the valley between where the dipole baffle effect starts rolling off the bass and the large resonance peak from a single diaphragm at 40-50Hz. But, they use up panel area doing this and are not left with enough for adequate deep bass output.
The Audiostatic ES100 is a perfect example of a one note bass transducer with anemic mid-bass. With the right kind of music, it could sound wonderful…very transparent midrange with amazing imaging capability. But with a lot of pop music or large orchestral type music it fell short sounding thin yet boomy at the same time.
I am not trying to create a full range ESL. Rather, I am trying to extend the LF of the ESL from the 200-300Hz used by the hybrids I’ve built down to 80Hz. Besides the obvious benefit of extending the ESL “magical” sonic qualities down another octave or two, my thoughts are that the crossover will be more seamless here since the wavelengths are much longer. However, as you mentioned, this area(between 80-200Hz) is where most ESLs perform with the poorest efficiency as the dipole baffle roll off is in full effect. My thought was to set the diaphragm resonance near the desired crossover point and damp to Q of 2 – 3. This way I will effectively gain 6dB efficiency to counteract one octave of the dipole roll off. Any remaining deficiency in response in this range would need to be made up by 1) increased area, 2) adding some small wings on the edges of the ESL to delay the dipole baffle roll off, or 3) increased signal drive in this frequency range(Acoustat uses separate LF transformers, Quad ESL57 adds windings for the LF, active EQ) 4) drop the efficiency of the midrange and highs to match the lows.
1) & 2) are easy enough...hoping I don't have to resort to complications like 3)
4) is also easy, but throwing away ESL efficiency never feels like the right solution
The efficiency of a damping cloth can be increased by glueing the cloth on the perforated stator (quad63) or in case of a wire stator, using a seperate perforated sheet with the cloth glued on (DIY netherlands). A fixed cloth won't move and is purely resistive. I found that a thin speaker grill cloth just 'wrapped around' the speaker worked but only very moderate compared to silicone dots in taming the fundamental resonance. I think this is because the wrapped on cloth will move along thus being less effective.
I do not know how silicone dots compare to the fixed cloth.
I do not know how silicone dots compare to the fixed cloth.
I agree, a fixed cloth does improve the damping. Also the closer you can get the cloth to the diaphragm the better the damping. If you attach the cloth on the inside of the back stator(like QUAD 63) you can use a cloth with a more open weave and still damp the resonance. The open weave then has less effect on the midrange. ALso the open weave does leave the ESL with a more transparent appearance which most people like.
Are you saying that you have found adding silicone dots to reduce the Q of the fundamental resonance? In my experiments, the Q and height of the response peak stayed pretty constant (+15db to +20dB) as I added more and more silicone dots. The only change was the frequency of the resonance increased as the number of dots increased.
Hmmmmm....perhaps I need to try some different brands of silicone. Maybe some have better damping properties than others.
Hi bolsert,
a peak up to 20db at resonance indicates a mismatch between membran width and tensioning of the mebran.
I guess you tension the membran mechanically. You should not. imo a heavily streched membran do not sound best. The more important is, that you need months or even years break in time to reach the point , where the tesion reduced itself to a constant and long lasting value. E.g your resonance of 60 Hz wuill decrease to possibly 40 Hz and Q decreases as well.
Better just to make thermal tensioning. glue the membran with lowest tension to the frame. The tension should bejust as high that major wrinkles are reduced.
Take a heat gun, best using one with a temperature indicator and set it to 250°C. Move it like a spray gun over the entire area. Keep a distance of about 5cm to the membran. Repeat it 5 times, let the membran cool down for 20 minutes and repeat the treatment 5 times again.
This wont end up in the best tension you might achieve figuring out a combination of mechanical and thermal treatment, but for the beginning it provides much better results than mechanical tensioning only.
I am sure that you end up with completely different results, which might surprise you
Capaciti
a peak up to 20db at resonance indicates a mismatch between membran width and tensioning of the mebran.
I guess you tension the membran mechanically. You should not. imo a heavily streched membran do not sound best. The more important is, that you need months or even years break in time to reach the point , where the tesion reduced itself to a constant and long lasting value. E.g your resonance of 60 Hz wuill decrease to possibly 40 Hz and Q decreases as well.
Better just to make thermal tensioning. glue the membran with lowest tension to the frame. The tension should bejust as high that major wrinkles are reduced.
Take a heat gun, best using one with a temperature indicator and set it to 250°C. Move it like a spray gun over the entire area. Keep a distance of about 5cm to the membran. Repeat it 5 times, let the membran cool down for 20 minutes and repeat the treatment 5 times again.
This wont end up in the best tension you might achieve figuring out a combination of mechanical and thermal treatment, but for the beginning it provides much better results than mechanical tensioning only.
I am sure that you end up with completely different results, which might surprise you
Capaciti
Hi
jauu
Calvin
This can be done quite successfully without damping but with usage of a dampened highpass of 2nd order. See this thread: http://www.diyaudio.com/forums/showthread.php?threadid=64620 especially the pdfs Multidoc3 and 4.
jauu
Calvin
Hello Capaciti,
Thanks for your suggestions. I should have mentioned that the +20dB peak is withought any damping at all. Even adding thin grill cloth to front and back to help keep dust out reduces this by 3-6dB.
I had tried mechanical tensioning, but as you mention the results always seem to change with time. I have been using the thermal tensioning method lately, but use a model airplane film shrink iron which is temperature controlled. It is nice to have a visual LED light indicactor to let you know if you are at the right temperature, or too hot/cold.
http://www.coverite.com/accys/covr2700.html
I usually set it to 325 degF (162 degC) and go over the film several times. Matching of tension from panel to panel is quite good. I'm not sure if this method tensions the film in the same way that a heat gun does or not, but it is easy to do.
Once the diaphragm has been tensioned, is it your experience that adding silicone dots reduces the Q of the resonance along with moving it up in frequency? This is the impression I have gotten from several comments on this forum, but personally I have not noticed the silicone dots providing much, if any damping.
Your comment about there being a mismatch between panel width and tension got me to thinking. Perhaps it is the diaphragm material. I am currently using 1/4mil (6 micron) Hostaphan. Maybe some thinner 3 micron film will provide a better match resulting in improved damping.
bolserst
Attachments
QUAD ESL 57 Woofer Panel Damping
Just slightly off topic for this thread, but....
Does anybody know if any damping material was used in the woofer panels of the original Quad ESL 57? The measurements I have seen showing a LF Q of about 2.5 sure makes me think there is, but I have not found any pictures showing the addition of any woven cloth mesh(like the ESL63) or felt(like the tweeter section of the ESL 57).
Perhaps the use of Saran instead of Mylar provided the needed damping.
Just slightly off topic for this thread, but....
Does anybody know if any damping material was used in the woofer panels of the original Quad ESL 57? The measurements I have seen showing a LF Q of about 2.5 sure makes me think there is, but I have not found any pictures showing the addition of any woven cloth mesh(like the ESL63) or felt(like the tweeter section of the ESL 57).
Perhaps the use of Saran instead of Mylar provided the needed damping.
@ Harry,
no i mean 250 °C since the distance of 5 cm reduces the effective heat to the film surface.
A temperature of 150-160°C would be right for an closed oven, where the heated air molecules are close to the surface.
Check out a heat gun by setting it to 250°C and using a temperature probe, which is included in some affordable multi testers. You will see that a distnace of about 5cm to the probe will show readings about 150°C
@ bolsert,
6µm is fine, smaller like 3µm is not as forgiving wrong tensioning and or heat treatment. The more layers of molecule chains (means thicker film), the more heat and tension will be dissipated. On the other hand the mass of 12µm is causing bandwith limitations at highest frequencies in the audio range.
basically you are right, the silicone dots wont lower Q, since those dots do not add damping, but a strategic distribution of the dots results in a strategic distribution of resonance modes of the membran. If you set them right, you will e.g. create two resonaces 30 and 50 hz with limited Q out of one big 20 Hz resonance with extended Q whithout dots.
As i mentioned before the silicone dot method can optimize a given panel geometrie to best compromise regarding low frequency performance, but bad dots create disturbing sound and to find out the right positions of the dots is a hell of trial and error.
If one is capable of FEM calculations it might assist, but i guess that it is nearly impossible to setup the correct model for the film behaviour. E.g. the polyester film isnt a simple elastic system, but different behaviour depending on conditions for stress and elongation. And Computers are stupid. if you feed them wrong the output is wrong.
OOOOOh, i later on recognized your heat treatment tool, makes a lot of sense to me. The hot surface is in direct contact to the film, thus showing best and reproducible heat transfer to the film. You know, there is always something to learn or at least to think about.....
Capaciti
no i mean 250 °C since the distance of 5 cm reduces the effective heat to the film surface.
A temperature of 150-160°C would be right for an closed oven, where the heated air molecules are close to the surface.
Check out a heat gun by setting it to 250°C and using a temperature probe, which is included in some affordable multi testers. You will see that a distnace of about 5cm to the probe will show readings about 150°C
@ bolsert,
6µm is fine, smaller like 3µm is not as forgiving wrong tensioning and or heat treatment. The more layers of molecule chains (means thicker film), the more heat and tension will be dissipated. On the other hand the mass of 12µm is causing bandwith limitations at highest frequencies in the audio range.
basically you are right, the silicone dots wont lower Q, since those dots do not add damping, but a strategic distribution of the dots results in a strategic distribution of resonance modes of the membran. If you set them right, you will e.g. create two resonaces 30 and 50 hz with limited Q out of one big 20 Hz resonance with extended Q whithout dots.
As i mentioned before the silicone dot method can optimize a given panel geometrie to best compromise regarding low frequency performance, but bad dots create disturbing sound and to find out the right positions of the dots is a hell of trial and error.
If one is capable of FEM calculations it might assist, but i guess that it is nearly impossible to setup the correct model for the film behaviour. E.g. the polyester film isnt a simple elastic system, but different behaviour depending on conditions for stress and elongation. And Computers are stupid. if you feed them wrong the output is wrong.
OOOOOh, i later on recognized your heat treatment tool, makes a lot of sense to me. The hot surface is in direct contact to the film, thus showing best and reproducible heat transfer to the film. You know, there is always something to learn or at least to think about.....
Capaciti
Last edited:
Yes there is, it's a some sort of natural fibre cloth on the inside of the back grill.
Some people like to take it out because of poor bass performance, but that is because the speakers are old and the bass gets weaker (lowered sensitivity) and the resonance frequency shifts up in frequency (sometimes over 100Hz!). The cure is to repair the panels and leave the damping mats in, that way it goes all the way down to 40 Hz, and the bass can sound surprisingly well.
This is one speaker that does not sound better without damping material. In fact it will lose it's wonderful natural midrange if you do.
Thank you.
Yes, with my closed oven a temperature of 150-160° C is right. I have never used a heat gun for tensioning in the past. So I have checked this today with a Leister heat gun set to 250° C. In the first minutes the temperature at 5-6 cm is around 150° C as you have mentioned. After full warm up of the gun I had to increase the distance to 9-10 cm. Only want to mention this if someone has bigger panels where considerably longer working time is needed.
I had similar experience with 3µm film. This thin one is not forgiving a wrong treatment. The sound is 'hard and edgy' then right from the beginning. And this type of sound will not go away with time. It's ruined forever.
As a starting guide, passive resonance tests are very helpful for this. It's easier then to play around with different places for the dots. The exact middle of the panel is not always the best place when you want to split the resonance peak into two smaller ones.
For visualizing the temperature of the film (or hot spots), Arend-Jan mentioned a coating with thermochromic paint. It changes color with temperature. But I have never used it. 'Cool' thing.
Temperature sensitive paint
Harry
I really appreciate you taking the time to answer the many questions my experiments have raised.
My hat is off to you for researching a distribution of silcone dots to optimize low frequency performance. I can only imagine the amount of time and frustration involved in the trial and error process. I may mention this problem to a few guys at work that are fluent in dynamic FEM analysis to see if they would be interested in seeing what solutions NASTRAN might be able to come up with.
Glad to see that I provided some information in return for you to think about.
Heating the 6µm film with a heat gun always made me nervouse. Too far away, and uniform heating did not take place. To close, or hovering in one place too long and you risk burning a hole in the diaphragm. With the Coverite iron, once the temperature is set correctly, I can glide it over the film as needed for uniform treatment without any concern for damaging the film.
Hi SY,
The few Acoustat panels I have seen used some fairly thick felt damping pads in the middle 1/3 area of the panel, and nothing on the outter area. Did you remove these pads before applying your thin felt? You mention only applying the damping to 50% of the active area. Can you describe where/why you applied the damping? Any Pics?
I have not experimented yet with partial panel damping...maybe this a better solution to avoiding contamination of the midrange resposne.
bolserst
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.