Hello All,
I'm building some 12x48 perf metal esl panels assembled with 3/4 wide 1/16 thick double-sided foam tape. I'm trying to decide whether to use 1/4" width foam strips or 1/2" diameter foam dots for the inner spacers and would like to hear recommendations from the esl builders here. Bear in mind that my stators are less than perfectly flat (tack-weld one together from two pieces) so I will need to be conservative with the distance between spacers in order to keep the diaphram centered.
The attached sketch shows (4) dot layouts on a 12x48 panel-- should I use 1/4 strip spacers or one of these dot layouts? Please let me hear your opinions!
Chaz
I'm building some 12x48 perf metal esl panels assembled with 3/4 wide 1/16 thick double-sided foam tape. I'm trying to decide whether to use 1/4" width foam strips or 1/2" diameter foam dots for the inner spacers and would like to hear recommendations from the esl builders here. Bear in mind that my stators are less than perfectly flat (tack-weld one together from two pieces) so I will need to be conservative with the distance between spacers in order to keep the diaphram centered.
The attached sketch shows (4) dot layouts on a 12x48 panel-- should I use 1/4 strip spacers or one of these dot layouts? Please let me hear your opinions!
Chaz
An externally hosted image should be here but it was not working when we last tested it.
it's a done-deal now
OK...
No responses (yet) to the dots versus strip spacer query. Oh well, it's a moot issue for this project now, since I've already assembled one of the esl panels this evening using foam tape strip-spacers. Still, I would like to hear any opinions on sonic differences between dots and strips (resonant modes, resonant frequency, character of the sound).
I was alarmed and astonished to find that the 3-M foam tape didn't stick to the stators very well. I concluded the culprit was the satin clear poly topcoats on the stators rather than the tape. The stators were first spray-coated with about 7 or 8 mils of latex house primer, then black Krylon from an aerosol can, then three sprayed coats of some Minwax clear satin polyurethane left-over from another project. I'm thinking the satin filler in the poly (which dulls the shine) inhibited adhesion to the tape so I advise DON'T USE IT!
Anyway, (due to the poor adhesion) I was afraid the dots were not going to provide enough adhering surface to hold the stators together so I opted for two 3/8" wide vertical strip spacers 3.5" apart and I sanded and solvent cleaned the stator topcoat at those locations; which seemed to improve the adhesion somewhat and the finished panel looks ok.
LESSON LEARNED: Spring for powder coating
Chaz
OK...
No responses (yet) to the dots versus strip spacer query. Oh well, it's a moot issue for this project now, since I've already assembled one of the esl panels this evening using foam tape strip-spacers. Still, I would like to hear any opinions on sonic differences between dots and strips (resonant modes, resonant frequency, character of the sound).
I was alarmed and astonished to find that the 3-M foam tape didn't stick to the stators very well. I concluded the culprit was the satin clear poly topcoats on the stators rather than the tape. The stators were first spray-coated with about 7 or 8 mils of latex house primer, then black Krylon from an aerosol can, then three sprayed coats of some Minwax clear satin polyurethane left-over from another project. I'm thinking the satin filler in the poly (which dulls the shine) inhibited adhesion to the tape so I advise DON'T USE IT!
Anyway, (due to the poor adhesion) I was afraid the dots were not going to provide enough adhering surface to hold the stators together so I opted for two 3/8" wide vertical strip spacers 3.5" apart and I sanded and solvent cleaned the stator topcoat at those locations; which seemed to improve the adhesion somewhat and the finished panel looks ok.
LESSON LEARNED: Spring for powder coating
Chaz
so you know...
high gloss coatings have the most solid content in them so will provide greater dielectric insulatiom. 3M acrylic tapes require a couple of things to work well and to achieve full strength. They are heat and pressure. Get and keep your panels warm and get some weight on the joints. Does not matter how you go about doing either just so you do.
high gloss coatings have the most solid content in them so will provide greater dielectric insulatiom. 3M acrylic tapes require a couple of things to work well and to achieve full strength. They are heat and pressure. Get and keep your panels warm and get some weight on the joints. Does not matter how you go about doing either just so you do.
I went and bought 300 of these things: 3M™ Bumpon™ Protective Products-Shop 3M
Will be trying them on curved stators ("here goes nothing"). Anybody tried dot spacers with curved, is it doable? I planned also to put a M3 nylon screw through each dot so I can compress the stators together (reduce d/s from 3,6mm to 3mm) and make a very nonresonant sandwitch out of them.
Added capacitance should be very small as the area the dots take up is only 10% of the horizontal spacer strips. I also wonder if I get more sensitivity and one very low fundamental frequency as the membrane moves much like one big area.
Will be trying them on curved stators ("here goes nothing"). Anybody tried dot spacers with curved, is it doable? I planned also to put a M3 nylon screw through each dot so I can compress the stators together (reduce d/s from 3,6mm to 3mm) and make a very nonresonant sandwitch out of them.
Added capacitance should be very small as the area the dots take up is only 10% of the horizontal spacer strips. I also wonder if I get more sensitivity and one very low fundamental frequency as the membrane moves much like one big area.
First step, making holes (for the nylon screws) to the polyurethane dots with leather hole punch tool:
I might get to membraining tomorrow or the day after that. I will practice first with 12µm mylar to see if I can make them work in the first place.
An externally hosted image should be here but it was not working when we last tested it.
I might get to membraining tomorrow or the day after that. I will practice first with 12µm mylar to see if I can make them work in the first place.
Hi,
The thing I liked very much about silicone dots is:
- you can always add some additional dots if the effect is not enough
- you don't have to worry about the thickness of the dots very much as it settles between the mylar and stator on itself
- you can put them on one side of the mylar only (at the back-side of the esl usually)
-silicone dots have good insulating properties
-they stick well to (clean) mylar
-they can easily be removed (you need to replace the mylar though)
About your curved esl; If the dots are replacing the horizontal curved spacers, than I doubt if these dots are good enough to support the curved shape of the mylar.
The thing I liked very much about silicone dots is:
- you can always add some additional dots if the effect is not enough
- you don't have to worry about the thickness of the dots very much as it settles between the mylar and stator on itself
- you can put them on one side of the mylar only (at the back-side of the esl usually)
-silicone dots have good insulating properties
-they stick well to (clean) mylar
-they can easily be removed (you need to replace the mylar though)
About your curved esl; If the dots are replacing the horizontal curved spacers, than I doubt if these dots are good enough to support the curved shape of the mylar.
Dot system did not work. The membrane got ~1,5mm away from the rear staor between each square. At least I tried. Will be going for horizontal spacers, they worked nice, i already made a test hack.
(Aluminum membrane is just for testing, I have it plenty around.)
. Will be going for horizontal spacers, they worked nice, i already made a test hack.An externally hosted image should be here but it was not working when we last tested it.
(Aluminum membrane is just for testing, I have it plenty around.)
An externally hosted image should be here but it was not working when we last tested it.
Some food for thought regarding vertical acrylic spacers, where the width of each section follows an exponential relationship, in this (long) building thread.
Also with measurements here and here.
Harry
Also with measurements here and here.
Harry
Last edited:
Hi, I actually own those very speakers (is it not obvious from the pictures
). I will be changin to horizontal spacers because vertical spacing is also problematic due the same reason that the membrane gets closer to back stator in center of each segment. The widest segment is/was 11-12cm wide and the membrane got only ~1-1,5mm away from rear stator. Because of this I got air conduction in that spot when I boosted the bias to >4kV. It would not pay off increasing the number of vertical segments, the radiating area gets smaller, SPL is lost and capacitance increases.
Horizontal spacers keep the distance better equal between the stators and they take less area compared to vertical. I also think that the segment size with horizontal spacers is better regarding the standing waves of the membrane and the fundamental resonance - narrow and long segments resonate longer than more square-like...
I drew this comparison couple of days ago which I ment to post here:
I will be using actually only 10 horizontal strips that reduces their area to ~460cm2, which is still ~41% smaller than the vertical spacing. The layout is also incorrect, I will place the biggest segments on the bottom to get the floor boost to the lowest fundamentals.
Last edited:
Hi, I actually own those very speakers (is it not obvious from the pictures
Ahhh, thanks for the info.
What I would change first in this speaker construction, is to change the metallized membrane to one with a high resistance coating. Also in my Esl-63 I got better results (compared with the original graphite coating).
Harry
What MartinLogan did to trun Down the out put Of the panels....Not that i think you wount to do this nor do i....Thay ran more horz bars..
The SL3 panels have the most output an sound the best to me..
The SL3 panels 12"X48" there are 11 hroz bars then the top an bottem so it 13 in all...an the 5 in the middel of the panels are 5" a part....an at the top an the bottem thay are 2"...an the others are 3"...this gives Max output of the panels..An best hi-end....
The more horz bars an closer togather... turns the panels output down....Just my finding..thanks for the input...goodluck
The SL3 panels have the most output an sound the best to me..
The SL3 panels 12"X48" there are 11 hroz bars then the top an bottem so it 13 in all...an the 5 in the middel of the panels are 5" a part....an at the top an the bottem thay are 2"...an the others are 3"...this gives Max output of the panels..An best hi-end....
The more horz bars an closer togather... turns the panels output down....Just my finding..thanks for the input...goodluck
Last edited:
Hi, already did that
. New membrane is 6µm Hostaphan polished with an antistatic vinyl polish spray.Here are some pictures of the horizontal spacers:
Back stator:
An externally hosted image should be here but it was not working when we last tested it.
Front stator:
An externally hosted image should be here but it was not working when we last tested it.
Close-up:
An externally hosted image should be here but it was not working when we last tested it.
I will be putting five nylon bolts through each spacer (50 bolts per speaker) and squeeze the hell out of the stators and make them very non-resonant. I have already tried it with 16 bolts with acryl spacers, and it made a big difference. The stators were not completely damped with this setup however. With PVC spacers, which are very damped by their nature, and 50 bolts, the stator will be dead as a rock.
Lots of work but it does pay off.
Hi,
the speakers Harry linked to, are really a fine piece of woodwork. For the panel, well, there are several points in the design and building that could be improved.
The metallized membrane beeing one, the way of tensioning the diaphragm another, vertical spacers probabely the worst. The 1:75 Plitron audio tranny will offer the poor amplifiers a tortureous impedance, with close to shortcut values at the upper bandwidth limit and very large phase angles between 2kHz-10kHz. Such a large panel shouldn´t be matched to the amp with more than 1:50, or You can count the number of amplifiers able-not-only-to-drive-the-thingies-stable-but-still-sounding-well by the five fingers of Your hand.
The distance of the spacers should follow the rule of shortest distance that still allows for the full d/s stroke. The rule of thumb of 1:70-1:100 (under the premise that the HV-bias is optimally high) still applies. Larger distances cost on efficienc and dynamic range (because of lower allowable HV-Bias) as well as shortening the distances (d/s getting larger than required). The spreading of resonances only works with considerable differences in segment area and dimensions. Care should be taken, that the varying dimensions dont cost too much on efficiency. If the panel is of a hybrid-type, the resonance peak probabely won´t matter anyway and the spacer distances can be designed for highest efficiency. Only FR-esls might profit from resonance distribution, but then, those are the inferior systems anyway.
The positioning of the smallest segments towards top and bottom of the panel helps to prevent the diaphragm to develop wrinkles at the edges.
To put it the other way round will probabely result in warps and wrinkles in the edges of the panel.
The soft double sided tape used as spacer introduces alot of damping to the stator. But both stator sheets must match precisely in curvature (or the front stator slightly, very slightly less curvature, if the rim is clamped well together), otherwise they will separate slightly in distance over time, with a following decrease in output level. The small soft tape strips are not able to cope with a constant pull over time and will give in. Also the soft strips won´t be perfectly flat but follow the hole contour of the stator sheet --> tiny warps. Besides the obviously faster, easier and as such cheaper mounting of the strips, I assume this to be one reason for ML to change from the soft 3M tape spacers to the clear hard spacers.
Since the hard cored spacers also feature thin glue layers the panels are still well dampened if glued together (MLs former panels were just glued at top and bottom, leaving the vast majority pieces of rattling metal). Then there´s no need to screw the stators together, which btw, is a true PITA, besides the inferior optics.
jauu
Calvin
the speakers Harry linked to, are really a fine piece of woodwork. For the panel, well, there are several points in the design and building that could be improved.
The metallized membrane beeing one, the way of tensioning the diaphragm another, vertical spacers probabely the worst. The 1:75 Plitron audio tranny will offer the poor amplifiers a tortureous impedance, with close to shortcut values at the upper bandwidth limit and very large phase angles between 2kHz-10kHz. Such a large panel shouldn´t be matched to the amp with more than 1:50, or You can count the number of amplifiers able-not-only-to-drive-the-thingies-stable-but-still-sounding-well by the five fingers of Your hand.
The distance of the spacers should follow the rule of shortest distance that still allows for the full d/s stroke. The rule of thumb of 1:70-1:100 (under the premise that the HV-bias is optimally high) still applies. Larger distances cost on efficienc and dynamic range (because of lower allowable HV-Bias) as well as shortening the distances (d/s getting larger than required). The spreading of resonances only works with considerable differences in segment area and dimensions. Care should be taken, that the varying dimensions dont cost too much on efficiency. If the panel is of a hybrid-type, the resonance peak probabely won´t matter anyway and the spacer distances can be designed for highest efficiency. Only FR-esls might profit from resonance distribution, but then, those are the inferior systems anyway.
The positioning of the smallest segments towards top and bottom of the panel helps to prevent the diaphragm to develop wrinkles at the edges.
To put it the other way round will probabely result in warps and wrinkles in the edges of the panel.
The soft double sided tape used as spacer introduces alot of damping to the stator. But both stator sheets must match precisely in curvature (or the front stator slightly, very slightly less curvature, if the rim is clamped well together), otherwise they will separate slightly in distance over time, with a following decrease in output level. The small soft tape strips are not able to cope with a constant pull over time and will give in. Also the soft strips won´t be perfectly flat but follow the hole contour of the stator sheet --> tiny warps. Besides the obviously faster, easier and as such cheaper mounting of the strips, I assume this to be one reason for ML to change from the soft 3M tape spacers to the clear hard spacers.
Since the hard cored spacers also feature thin glue layers the panels are still well dampened if glued together (MLs former panels were just glued at top and bottom, leaving the vast majority pieces of rattling metal). Then there´s no need to screw the stators together, which btw, is a true PITA, besides the inferior optics.
jauu
Calvin
Hi Calvin, thanks for the answer. I agree with aluminum membrane and the vertical spacers. What do you mean with that the way of tensioning the membrane is something one could improve upon?
Here are some pictures I took (pardon for messy room):
Tension tapes:
50mm wide tate that locks and secures the membrane from all over (underneath is also the charge ring):
Black linen based tape that will define the looks that will show under the acryl side strips of the front stataor (I wanted black looks all the way.)
Small pieces of transparent tape is put on the membrane before I made the holes with soldering iron for the nylon screws that go through the spacers. In picrture the holes are already done:
And here is one ready speaker with all the 50pcs bylon bolts going through the stators squeezing them completely together:
The sensitivity increased by 2-3dB compared to the vertical spacers. If I wanted I can also lure in some more sensitivity by squeezing the ylon bolts hard because that will squeeze the PVC spacers a bit to maybe 2,5mm d/s. The stator sandwitch is very sturdy when knocked!
The capacitance dropped by 100pF to ~2700pF.
The speakers will have line level crossover. Maybe 1st order around 200Hz. I have to try different approaches, right now I'm testing them with modded DCX2496 but will go for passive crossover and leave the DCX for bass. Dipole line array woofers from floor to the ceiling with 5pcs of 18" drivers per side will take care of the low registers.
Here are some pictures I took (pardon for messy room
):Tension tapes:
50mm wide tate that locks and secures the membrane from all over (underneath is also the charge ring):
Black linen based tape that will define the looks that will show under the acryl side strips of the front stataor (I wanted black looks all the way.)
Small pieces of transparent tape is put on the membrane before I made the holes with soldering iron for the nylon screws that go through the spacers. In picrture the holes are already done:
And here is one ready speaker with all the 50pcs bylon bolts going through the stators squeezing them completely together:
The sensitivity increased by 2-3dB compared to the vertical spacers. If I wanted I can also lure in some more sensitivity by squeezing the ylon bolts hard because that will squeeze the PVC spacers a bit to maybe 2,5mm d/s. The stator sandwitch is very sturdy when knocked!
The capacitance dropped by 100pF to ~2700pF.
The speakers will have line level crossover. Maybe 1st order around 200Hz. I have to try different approaches, right now I'm testing them with modded DCX2496 but will go for passive crossover and leave the DCX for bass. Dipole line array woofers from floor to the ceiling with 5pcs of 18" drivers per side will take care of the low registers.
Hi,
the tensioning as done in the thread by hand will hardly give an even distribution of mechanical tension and is prone to panel-2-panel variations, i.e inconsistancies.
With the vertical spacers the diaphragm could probabely be heat treated after mechanical tensioning, but not so with horizontal spacers.
Here a dedicated tensioning frame is required to achieve high precision and consistancy.
Too I´d suggest to pull the diaphragm around the stator rims to the backside and to fix it there also (at least at top and botttom rim).
This keeps the tension up over a longer period.
Simple taping as shown in the thread will hardly produce longtime stable glue joints.
With apropriate procedures and tooling one could achieve Fs variations < +-1Hz and amplitude variances < +-0.5dB.
2.5mm of d/s for a panel crossed at ~200Hz is like an huge empty airplane Hangar.
Every mil of unused d/s costs on efficiency, since the efficiency drops quadratically with distance.
Even if the HV-Bias is raised to counter the drop, the needed wattage raises.
The loss is so severe that the dynamic range shrinks which even the increased possible excusrsion can´t completely compensate for.
The THD usually rises sharply below 200Hz, which also speaks against large d/s values.
Regarding the large size of the panel I´d opt for a lower crossover of maybe 120-150Hz (or even 100Hz if the resonance Fs allows for) to put the d/s to a good use.
If mated with bass drivers as large as 18" the crossover freq needs to be low.
Keep the working range of such large basses possibly below 100Hz.
They don´t play and integrate well at 200Hz, especially not in a dipole arrangement.
Crossing over at 100Hz or below You might think of using the Basses (and a lower number of them) in CBs.
One driver might already be enough then.
At such a low xover the panels distribution has taken on a dipolar global shape (at >150Hz it transits smoothly into a dipolar cylindrical shape), that may probless be mated with the global distribution shape of the woofer.
jauu
Calvin
the tensioning as done in the thread by hand will hardly give an even distribution of mechanical tension and is prone to panel-2-panel variations, i.e inconsistancies.
With the vertical spacers the diaphragm could probabely be heat treated after mechanical tensioning, but not so with horizontal spacers.
Here a dedicated tensioning frame is required to achieve high precision and consistancy.
Too I´d suggest to pull the diaphragm around the stator rims to the backside and to fix it there also (at least at top and botttom rim).
This keeps the tension up over a longer period.
Simple taping as shown in the thread will hardly produce longtime stable glue joints.
With apropriate procedures and tooling one could achieve Fs variations < +-1Hz and amplitude variances < +-0.5dB.
2.5mm of d/s for a panel crossed at ~200Hz is like an huge empty airplane Hangar.
Every mil of unused d/s costs on efficiency, since the efficiency drops quadratically with distance.
Even if the HV-Bias is raised to counter the drop, the needed wattage raises.
The loss is so severe that the dynamic range shrinks which even the increased possible excusrsion can´t completely compensate for.
The THD usually rises sharply below 200Hz, which also speaks against large d/s values.
Regarding the large size of the panel I´d opt for a lower crossover of maybe 120-150Hz (or even 100Hz if the resonance Fs allows for) to put the d/s to a good use.
If mated with bass drivers as large as 18" the crossover freq needs to be low.
Keep the working range of such large basses possibly below 100Hz.
They don´t play and integrate well at 200Hz, especially not in a dipole arrangement.
Crossing over at 100Hz or below You might think of using the Basses (and a lower number of them) in CBs.
One driver might already be enough then.
At such a low xover the panels distribution has taken on a dipolar global shape (at >150Hz it transits smoothly into a dipolar cylindrical shape), that may probless be mated with the global distribution shape of the woofer.
jauu
Calvin
Tensioning tapes are really sturdy s*it, it does not worry me the least. 3M makes very good tapes these days. The glue joints are actually compressed agains the stators/acrylic side strips with 22pcs M6 nylon bolts that can be tightened hard.
I can live with +/- 10Hz differences, although they are not going to be that big (I know from last time I membraned them, that was with the vertical spacers) and they get closer each other in burn in.
That number I said is just electrical crossoverpoint, acoustical crossover point will be totally another story as we all know. I manly make the high pass in order to protect the Plitron from saturation. I usually use only 1st and 2nd order filters, they sound the best.
Here are plots for full range, 100Hz, 150hz, 200hz and 250Hz 1st order filters, measured in the room, mic in the listening postion.
I can easily integrate woofers there with every plot because woofers are actively crossed and equalized. I would try an acoustic crossover of ~70-100Hz regardless, I'm not going to fill the room mess at the upper bass with woofers. The excess group delay is always dependant on the actual room response, even sharp parametric EQs don't mess things if the response stays smooth. DCX adds maybe 1-2ms of delay, which is quite annoying but haver to live with it.
I think 150Hz or above that would protect the Plitron with 50Vrms max drive voltage quite good.
Woofers have <1mH inductance and can be used as high as 200Hz with qood guality (first minor THD peak at 350Hz and cone resonance at 1,2khz). No CBs for me with dipoles (been there). Dipole matches with dipole and dipole bass sounds best to my ears, and I don't mind the huge sizer of the line arrays.
I can live with +/- 10Hz differences, although they are not going to be that big (I know from last time I membraned them, that was with the vertical spacers) and they get closer each other in burn in.
That number I said is just electrical crossoverpoint, acoustical crossover point will be totally another story as we all know. I manly make the high pass in order to protect the Plitron from saturation. I usually use only 1st and 2nd order filters, they sound the best.
Here are plots for full range, 100Hz, 150hz, 200hz and 250Hz 1st order filters, measured in the room, mic in the listening postion.
An externally hosted image should be here but it was not working when we last tested it.
I can easily integrate woofers there with every plot because woofers are actively crossed and equalized. I would try an acoustic crossover of ~70-100Hz regardless, I'm not going to fill the room mess at the upper bass with woofers. The excess group delay is always dependant on the actual room response, even sharp parametric EQs don't mess things if the response stays smooth. DCX adds maybe 1-2ms of delay, which is quite annoying but haver to live with it.
I think 150Hz or above that would protect the Plitron with 50Vrms max drive voltage quite good.
Woofers have <1mH inductance and can be used as high as 200Hz with qood guality (first minor THD peak at 350Hz and cone resonance at 1,2khz). No CBs for me with dipoles (been there)
. Dipole matches with dipole and dipole bass sounds best to my ears, and I don't mind the huge sizer of the line arrays.
Last edited:
Yes, maybe on some ESL's but not in these with the Plitrons. They make normal cone floorstanders green from the low THD numbers at the bass
.I tightened the nylon bolts and the sensitivity increased like hell. Here's a 105dB sweep with 250Hz 1st order filter, bias 4,7kV and the drive voltage might be ~25Vrms. The THD is 0.2-0.3% @ 105dB (where it peaks so high) and H2 rules with quite a margin (cannot measure higher harmonics, they are at the noise floor of the measuring equipment
). Very clean low end also.An externally hosted image should be here but it was not working when we last tested it.
I have to say that after this sweep I have not stopped smiling. Although my ears ring a bit from 30min sweeping
. It's time to start on the other one so I get them to equal condition.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.