Hello,
I need a little help getting my segmented ESLs up and running.
1-A slight background noise when you put your ear close to the panel.
2- When I turn off the panels, a crumpled mylar noise that lasts until the voltage drops to about 30 volts.
Power supply: 5300 volts (tested OK)
D/S :1.5 mm
Transformer 4x 230/6Volts 30 VA (RS-Component).
Resistor: PR03 Vishay 750 Volts ( 2 resistors per segment)
Mylar: 3 microns.
Otherwise works well with a low-power amp.
Thanks for your help.
Serge B.
I need a little help getting my segmented ESLs up and running.
1-A slight background noise when you put your ear close to the panel.
2- When I turn off the panels, a crumpled mylar noise that lasts until the voltage drops to about 30 volts.
Power supply: 5300 volts (tested OK)
D/S :1.5 mm
Transformer 4x 230/6Volts 30 VA (RS-Component).
Resistor: PR03 Vishay 750 Volts ( 2 resistors per segment)
Mylar: 3 microns.
Otherwise works well with a low-power amp.
Thanks for your help.
Serge B.
From your description it sounds like your mylar has collapsed to the stator because the bias voltage is too high for the diaphragm size, D/S, and tension you have.
Usually for 1.5mm spacing, bias voltage in the 2.5kV - 3kV range is used.
What is the size of your diaphragm? and how did you tension the mylar?
Usually for 1.5mm spacing, bias voltage in the 2.5kV - 3kV range is used.
What is the size of your diaphragm? and how did you tension the mylar?
Hello,
the panel measures 1.35x 25.4cm, and the mylar is stretched using the JazzMan inner tube method.
When I remove the connections to the secondary of the Transformers no more noise.
The coating is antistatic Static 100 equivalent to Licron on resistance.
1.5 mm PVC x 82 wire (20-12-6-8)
I'll try lowering the voltage to 3Kv to test.
When I listen to music, I get saturation in the treble.
Thank you for your contribution.
Serge B.
the panel measures 1.35x 25.4cm, and the mylar is stretched using the JazzMan inner tube method.
When I remove the connections to the secondary of the Transformers no more noise.
The coating is antistatic Static 100 equivalent to Licron on resistance.
1.5 mm PVC x 82 wire (20-12-6-8)
I'll try lowering the voltage to 3Kv to test.
When I listen to music, I get saturation in the treble.
Thank you for your contribution.
Serge B.
Are you using a large value resistor between your HV supply and the diaphragm? And by large, I mean hundreds of megohms. And I agree with bolserst, your bias voltage is likely too high for your design.
Hey, Charlie (a.k.a. Jazzman) here,
A good rule of thumb for the bias voltage is to target for 1/2 the breakdown voltage of the air gap. The breakdown voltage of air is about 100 volts/mil, and 1.5mm is about 60 mils, so the breakdown voltage would be about 6KV for your gap dimension. Dividing by 2 = 3KV for the recommended bias voltage.
Does your 25cm wide panel have any center diaphragm supports?
The bike tube jig tensioning method shown on Jazzman's ESL Page stretched [thicker] 6-micron film to 1% - 1.25% elongation, on a panel about the same width as yours. The Jazzman panel also has two center support spacers, so the diaphragm spans only 1/3 the panel width unsupported.
3-micron film is less strong/more elastic than 6-micron film, so stretching it to the same elongation would not produce the same tension (significantly less, I would think).
The Jazzman panel is configured so conservatively with respect to diaphragm support spans, that I'm thinking you're OK with less tension-- if the diaphragm isn't spanning the entire 25cm width unsupported. That's why I asked if you used any center diaphragm supports.
A good rule of thumb for the bias voltage is to target for 1/2 the breakdown voltage of the air gap. The breakdown voltage of air is about 100 volts/mil, and 1.5mm is about 60 mils, so the breakdown voltage would be about 6KV for your gap dimension. Dividing by 2 = 3KV for the recommended bias voltage.
Does your 25cm wide panel have any center diaphragm supports?
The bike tube jig tensioning method shown on Jazzman's ESL Page stretched [thicker] 6-micron film to 1% - 1.25% elongation, on a panel about the same width as yours. The Jazzman panel also has two center support spacers, so the diaphragm spans only 1/3 the panel width unsupported.
3-micron film is less strong/more elastic than 6-micron film, so stretching it to the same elongation would not produce the same tension (significantly less, I would think).
The Jazzman panel is configured so conservatively with respect to diaphragm support spans, that I'm thinking you're OK with less tension-- if the diaphragm isn't spanning the entire 25cm width unsupported. That's why I asked if you used any center diaphragm supports.
Last edited:
Yes the bias voltage is too high.The rule of thumb here is about 500V/mm of the gap, but it is also related to the diameter of the mylar.Once it's over 100 square cm , an additional 0.25mm should be added to the gap for every doubling of the membrane surface.The saturation in the highs comes from a too high capacitance so you deffinitely need higher gap or lower bias.
Hello everyone,
I will test the polarization at 3000 Volts today, in fact I had not divided by 2 the voltage.
CharlieM, I was largely inspired by your work, except for the voltage, which was a mistake on my part.
The mylar is very stretched, maybe too much.
I don't have a vertical center support.
I have a 3x 10Mega resistor in series with the power supply.
I think you're all right, the voltage is too high.
I'll let you know the result.
Many thanks to all of you.
Serge B.
I will test the polarization at 3000 Volts today, in fact I had not divided by 2 the voltage.
CharlieM, I was largely inspired by your work, except for the voltage, which was a mistake on my part.
The mylar is very stretched, maybe too much.
I don't have a vertical center support.
I have a 3x 10Mega resistor in series with the power supply.
I think you're all right, the voltage is too high.
I'll let you know the result.
Many thanks to all of you.
Serge B.
I carried out the tests at 2.7KV, the noise persists but at a lower level.
When I switch off the mains supply, the crumpled mylar noise is still present but at a lower level for about 1-2 minutes, then I remove the tension wire feeding the mylar, the crumpled mylar noise is still present for about 1-2 minutes.
Can mylar film be too stretched?
I stretched a test mylar to breaking point and measured the elongation, then stretched the real mylar to just below breaking point.
Thanks for your help
Serge B.
When I switch off the mains supply, the crumpled mylar noise is still present but at a lower level for about 1-2 minutes, then I remove the tension wire feeding the mylar, the crumpled mylar noise is still present for about 1-2 minutes.
Can mylar film be too stretched?
I stretched a test mylar to breaking point and measured the elongation, then stretched the real mylar to just below breaking point.
Thanks for your help
Serge B.
I don't believe excessive tension in the Mylar could cause the problem you described.
Reducing the bias voltage to 2.7KV would have resolved the problem-- IF the diaphragm tension is not so low that it's still being pulled into a stator at the reduced voltage.
Assuming the diaphragm tension is sufficient for the span between supports, the problem must then lie elsewhere.
Is the static sound constant or intermittent?
Did you build a pair of speakers? If so, do both speakers exhibit the same problem?
Any small piece of debris between the stator and diaphragm could create a shorting path. Look for trash along the edge spacers and center support spacers. Debris shorting would create an intermittent popping sound [probably] emanating from a specific spot on the panel.
If the static sound is emanating from the entire panel, and the diaphragm isn't excessively loose/pulling into a stator, then I would suspect the problem is within the bias supply.
I once had a static issue which followed the bias supply when I swapped it out. Thinking the problem might be a cold solder joint, I reflowed all the solder joints but that didn't resolve it. I then cleaned all the flux residue off the back of the board and the problem went away. I then coated the backside of the board with BG Chemicals' silicone modified conformal coating, as a precaution to add a bit more insulation between the pads.
From that experience, I could only speculate that my board layout had certain solder pads (with high potential between them) in too-close proximity, and perhaps humidity absorption in the flux residue created a shorting path between those pads. All I knew for sure was that the problem went away after I cleaned the backside of the board.
Some years later I had a similar static issue which also followed the bias supply, but the board was perfectly clean. I probed components on the board but couldn't isolate the source of the problem, so I replaced the entire bias supply and the problem went away.
These kinds of problems can make you pull your hair out!
Reducing the bias voltage to 2.7KV would have resolved the problem-- IF the diaphragm tension is not so low that it's still being pulled into a stator at the reduced voltage.
Assuming the diaphragm tension is sufficient for the span between supports, the problem must then lie elsewhere.
Is the static sound constant or intermittent?
Did you build a pair of speakers? If so, do both speakers exhibit the same problem?
Any small piece of debris between the stator and diaphragm could create a shorting path. Look for trash along the edge spacers and center support spacers. Debris shorting would create an intermittent popping sound [probably] emanating from a specific spot on the panel.
If the static sound is emanating from the entire panel, and the diaphragm isn't excessively loose/pulling into a stator, then I would suspect the problem is within the bias supply.
I once had a static issue which followed the bias supply when I swapped it out. Thinking the problem might be a cold solder joint, I reflowed all the solder joints but that didn't resolve it. I then cleaned all the flux residue off the back of the board and the problem went away. I then coated the backside of the board with BG Chemicals' silicone modified conformal coating, as a precaution to add a bit more insulation between the pads.
From that experience, I could only speculate that my board layout had certain solder pads (with high potential between them) in too-close proximity, and perhaps humidity absorption in the flux residue created a shorting path between those pads. All I knew for sure was that the problem went away after I cleaned the backside of the board.
Some years later I had a similar static issue which also followed the bias supply, but the board was perfectly clean. I probed components on the board but couldn't isolate the source of the problem, so I replaced the entire bias supply and the problem went away.
These kinds of problems can make you pull your hair out!
Forgot to add:
A cold solder joint in any of the resistor network connections of your segmented panel could also create arcing sounds. This would likely manifest as as intermittent popping sound emanating from the entire panel but could also be localized to an area driven the wire group connected to the faulty solder joint.
A cold solder joint in any of the resistor network connections of your segmented panel could also create arcing sounds. This would likely manifest as as intermittent popping sound emanating from the entire panel but could also be localized to an area driven the wire group connected to the faulty solder joint.
Last edited:
When the membrame collapses into one of the stators, you should be able to see that by shining a flashlight on it at an angle. A collapsed membrame also tends to pop back after the bias is removed with quite an audible bang. But if build tolerances are too high, it can also occur only locally.
If it's not that, you likely have an insulation problem. That can be hard to trace. Did you coat the entire membrame area? What frame material did you use?
I always prefer to coat only the inner part of the membrame that is not in contact with the frame. Use a thin line of graphite paint around the coated area that at only one point extends to the frame for the bias connection.
Edit I once had a sizzling esl after recoating the membrame. For some strange reason the old and new layers of coating interacted in some weird manner. Had to replace the membrame to solve it in the end.
If it's not that, you likely have an insulation problem. That can be hard to trace. Did you coat the entire membrame area? What frame material did you use?
I always prefer to coat only the inner part of the membrame that is not in contact with the frame. Use a thin line of graphite paint around the coated area that at only one point extends to the frame for the bias connection.
Edit I once had a sizzling esl after recoating the membrame. For some strange reason the old and new layers of coating interacted in some weird manner. Had to replace the membrame to solve it in the end.
That could very well be your amplifier. ESL's require a lot of current at higher frequencies, if you amp has overcurrent protection it is not unlikely it kicks in at relatively low volume causing distortion. Or the amp becomes unstable due to the reactive load and is oscillating.
Try another amp to rule it out as the cause of the treble problems.
The second panel is being listened to, nothing to do with the first:
Just a tiny hiss when listening at 2 cm maybe normal?
CharlieM: could the product used to prevent the stator from sticking to the work board cause a problem if the product has gone onto the stator wires?
I take the 1st stator and look at the controls that CharlieM recommends.
Maudio: I put adhesive tape all around the panel, covering about 5mm of the membrane (CharlieM) before spraying the antistatic product. I also tested in the dark, no arcing.
Serge B.
Just a tiny hiss when listening at 2 cm maybe normal?
CharlieM: could the product used to prevent the stator from sticking to the work board cause a problem if the product has gone onto the stator wires?
I take the 1st stator and look at the controls that CharlieM recommends.
Maudio: I put adhesive tape all around the panel, covering about 5mm of the membrane (CharlieM) before spraying the antistatic product. I also tested in the dark, no arcing.
Serge B.
A good panel is dead quiet, so the hiss is definately not good.
But the fact that you sprayed the coating on is a red flag. With spraying, the stuff gets everywhere including all the places where it shouldn't. That might very well be the root of your issue.
With stuff in spray cans, I use a plastic cup to spray the product in, far away from the panel. The liquid collected in the cup then can be brushed on the membrame in a controlled manner.
But the fact that you sprayed the coating on is a red flag. With spraying, the stuff gets everywhere including all the places where it shouldn't. That might very well be the root of your issue.
With stuff in spray cans, I use a plastic cup to spray the product in, far away from the panel. The liquid collected in the cup then can be brushed on the membrame in a controlled manner.
Whatever product (wax?) you used to coat the work surface to prevent the wires from sticking to it would only be a problem if it were conductive.
I think it's very unlikely that any wax type product would be conductive.
If you masked off the edges of the panel sufficiently, and your wire support structure is non-metallic, there would not likely be a problem caused by you spraying on the coating. I've used sprayed-on Licron Crystal on at least 16 masked panels over the past 8 years and it's never been a problem.
Maudio's advice that your panel should be dead quiet is correct-- if it isn't, there's a problem somewhere.
Is the crackling sound emanating from the entire panel or is it localized to a specific area?
Can you post some pics of the panel and bias supply board and segmentation resistor connections?
Last edited:
Hello ChatlieM,
The files Panel+resistances + power_supply.
I recorded the noise but I can't put the file on diyaudio.
Thanks
The files Panel+resistances + power_supply.
I recorded the noise but I can't put the file on diyaudio.
Thanks
Attachments
Last edited:
Nice build!
You can zip the audio file, zips are allowed. Not sure about file size though.
I see you use wooden supports for the stator wires. Are these stiff enough? The panel is quite wide. If they flex too much, that could be another possible cause for the distortion you mentioned. The forces are quite substantial at higher volumes.
You can zip the audio file, zips are allowed. Not sure about file size though.
I see you use wooden supports for the stator wires. Are these stiff enough? The panel is quite wide. If they flex too much, that could be another possible cause for the distortion you mentioned. The forces are quite substantial at higher volumes.
Hello,
Thanks for the Zip tip.
The stator is made of "Iroko" wood, triple impregnated with epoxy resin diluted in : 20%, 10% 5%. The wood is very rigid and is used for boat building.
Noise occurs without a signal and with a signal there is no noise.
The stator is CNC machined and the horizontal bars are grooved to insert the stator wires 1.5mm deep.
Serge B.
Thanks for the Zip tip.
The stator is made of "Iroko" wood, triple impregnated with epoxy resin diluted in : 20%, 10% 5%. The wood is very rigid and is used for boat building.
Noise occurs without a signal and with a signal there is no noise.
The stator is CNC machined and the horizontal bars are grooved to insert the stator wires 1.5mm deep.
Serge B.
Last edited:
Didn't know this Iroko wood. Looks like a very good material for this application, the epoxy impregnating will prevent the wood from absorbing moisture. And it looks very nice indeed.
The noise is a bit difficult to judge, but it sure does sound like the sound of a membrane collapsed in a stator. Does it stop suddenly after some time or does it just fade away after power cut?
The membrane tension is there to keep the membrane centered between the stators. But with bias applied, the membrane will always be pulled more towards one stator as the tolerances will never be zero. So a membrane under bias without signal will be sort of dome shaped. Since there are limits on the achievable membrane tension, there is a rule of thumb in ESL construction that the ratio between panel width and DS should be under 100 for the membrane to be stable. Some sources mention even lower values. So for a 1,5 mm DS, 150mm would be the max allowable panel width. According to that rule, your panel is too wide for the DS spacing used. That makes it very difficult to prevent the membrane from touching the stators with bias applied. It is simply not possible to tension the membrane enough to prevent this, given the mechanical properties of mylar. So I expect this is what's happening here.
You can try to lower the bias even further. Use a variac if you have access to one. Start at zero, slowly turn up the bias until the noise begins. Then turn it down just a bit. That point is the max bias you can use. I expect it to be very low, probably too low to be useful. And even at this point, the possible excursion of the membrane will be almost zero. But it will at least confirm the theory.
The noise is a bit difficult to judge, but it sure does sound like the sound of a membrane collapsed in a stator. Does it stop suddenly after some time or does it just fade away after power cut?
The membrane tension is there to keep the membrane centered between the stators. But with bias applied, the membrane will always be pulled more towards one stator as the tolerances will never be zero. So a membrane under bias without signal will be sort of dome shaped. Since there are limits on the achievable membrane tension, there is a rule of thumb in ESL construction that the ratio between panel width and DS should be under 100 for the membrane to be stable. Some sources mention even lower values. So for a 1,5 mm DS, 150mm would be the max allowable panel width. According to that rule, your panel is too wide for the DS spacing used. That makes it very difficult to prevent the membrane from touching the stators with bias applied. It is simply not possible to tension the membrane enough to prevent this, given the mechanical properties of mylar. So I expect this is what's happening here.
You can try to lower the bias even further. Use a variac if you have access to one. Start at zero, slowly turn up the bias until the noise begins. Then turn it down just a bit. That point is the max bias you can use. I expect it to be very low, probably too low to be useful. And even at this point, the possible excursion of the membrane will be almost zero. But it will at least confirm the theory.
Last edited: