While I am well aware that push-pull ESL designs are the way to go for ultimate fidelity, I am currently experimenting with basic electrostatic concepts and I am baffled by the behavior of my singled-ended test circuit.
I have constructed the simplest type of ESL consisting of an aluminized mylar / perforated steel condenser sandwich in series with an audio transformer secondary and a 300v dc bias supply (voltage doubled AC with two .1 uf caps in series). A small all-in-one desk stereo readily drives the primary of the transformer.
When the power supply is initially plugged-in, the mylar adheres very tightly to the perforated plate - so much so that the indentations from the stator perforations are clearly visible in the mylar surface. Sensitivity is very high and the music plays cleanly.
After a few minutes (with or without music signal), the mylar starts to lose its attraction and eventually starts to peel away from the stator. If music is playing during this time, the volume can be heard to decrease significantly as the sensitivity is reduced.
Unplugging the power supply from the wall does not do anything, but placing a resistor across the two series .1 uf capacitors of the bias supply causes the mylar to snap back into place and the sensitivity is increased to its initial level. Same as before, however, the sensitivity falls off after a few minutes. Plugging the power supply back in restores everything to its initial state and the cycle continues.
I realize this isn't the preferred way to construct an ESL, but I would nonetheless like to understand why the static charge cannot be indefinitely maintained between the diaphragm and the stator as long as the bias supply is not disconnected. I would be very grateful for any explanations or suggestions.
Sincerely,
Nick
I have constructed the simplest type of ESL consisting of an aluminized mylar / perforated steel condenser sandwich in series with an audio transformer secondary and a 300v dc bias supply (voltage doubled AC with two .1 uf caps in series). A small all-in-one desk stereo readily drives the primary of the transformer.
When the power supply is initially plugged-in, the mylar adheres very tightly to the perforated plate - so much so that the indentations from the stator perforations are clearly visible in the mylar surface. Sensitivity is very high and the music plays cleanly.
After a few minutes (with or without music signal), the mylar starts to lose its attraction and eventually starts to peel away from the stator. If music is playing during this time, the volume can be heard to decrease significantly as the sensitivity is reduced.
Unplugging the power supply from the wall does not do anything, but placing a resistor across the two series .1 uf capacitors of the bias supply causes the mylar to snap back into place and the sensitivity is increased to its initial level. Same as before, however, the sensitivity falls off after a few minutes. Plugging the power supply back in restores everything to its initial state and the cycle continues.
I realize this isn't the preferred way to construct an ESL, but I would nonetheless like to understand why the static charge cannot be indefinitely maintained between the diaphragm and the stator as long as the bias supply is not disconnected. I would be very grateful for any explanations or suggestions.
Sincerely,
Nick
This is very interesting I have similar strange effects with my little panels.
From the sound panning left to right across 3 panels as I vari the Bias voltage to the sound appearing to be coming form the back of the panel as the bias voltage is raised until it is fully charged.
To stop the panning I had to feed each Diagphrams bias thru its own individual high value resistor.
I the second scenario once the panel was evenly charged (typically at the highest voltage level) the sensitivity level would change as I varied the bias voltage as it should without it sounding like the sound was coming from the back of the panel.
I have only messed with a single ended configuration for testing of the diagphram coating and never tried running a signal in to them.
I was quite amazed how the diagphram sucked completely into the stator at such a low voltage.
I am guessing these are the effects of charge migration.
Is your stator coated with anything?
Also have you tried a Diagphram with a high resistance coating on it as well?
What is your Diagphram to Spacer (D/S) width?
I have to types of coated stators,
One that is just paint,
And the other one is more of a high performance panel using Powder Coating and sealed with clear acrylic enamel and it is this stator the produces the weird effects.
This panel withstands the highest bias voltage my supply can produce at 13.8Kv without arcing even with a .070" D/S.
The plain painted stator just arcs when it gets above 3.5kv to 4.5kv as it uses a white Titanium Dioxide pigmented spray can enamel paint.
I don't recall any of these strange effects with just the painted stators, But then I again I never got them much above 3Kv or so either.
I didn't use them much because of this as getting the most efficiency was the name of the game and using a very high bias voltage is the key.
Especially for a small panel.
From this experience I have learned to not use any Titanium Dioxide pigmented paints.
However one was used in the three panel setup with two Powder Coated panels when the panning effect had occurred.
I have not tried my new method of stator coating as of yet but will be doing so very soon and can be found here,
http://www.diyaudio.com/forums/plan...tric-coatings-fact-fiction-2.html#post2893839
and the next post,
http://www.diyaudio.com/forums/plan...tric-coatings-fact-fiction-2.html#post2894427
Cheers !!
jer
From the sound panning left to right across 3 panels as I vari the Bias voltage to the sound appearing to be coming form the back of the panel as the bias voltage is raised until it is fully charged.
To stop the panning I had to feed each Diagphrams bias thru its own individual high value resistor.
I the second scenario once the panel was evenly charged (typically at the highest voltage level) the sensitivity level would change as I varied the bias voltage as it should without it sounding like the sound was coming from the back of the panel.
I have only messed with a single ended configuration for testing of the diagphram coating and never tried running a signal in to them.
I was quite amazed how the diagphram sucked completely into the stator at such a low voltage.
I am guessing these are the effects of charge migration.
Is your stator coated with anything?
Also have you tried a Diagphram with a high resistance coating on it as well?
What is your Diagphram to Spacer (D/S) width?
I have to types of coated stators,
One that is just paint,
And the other one is more of a high performance panel using Powder Coating and sealed with clear acrylic enamel and it is this stator the produces the weird effects.
This panel withstands the highest bias voltage my supply can produce at 13.8Kv without arcing even with a .070" D/S.
The plain painted stator just arcs when it gets above 3.5kv to 4.5kv as it uses a white Titanium Dioxide pigmented spray can enamel paint.
I don't recall any of these strange effects with just the painted stators, But then I again I never got them much above 3Kv or so either.
I didn't use them much because of this as getting the most efficiency was the name of the game and using a very high bias voltage is the key.
Especially for a small panel.
From this experience I have learned to not use any Titanium Dioxide pigmented paints.
However one was used in the three panel setup with two Powder Coated panels when the panning effect had occurred.
I have not tried my new method of stator coating as of yet but will be doing so very soon and can be found here,
http://www.diyaudio.com/forums/plan...tric-coatings-fact-fiction-2.html#post2893839
and the next post,
http://www.diyaudio.com/forums/plan...tric-coatings-fact-fiction-2.html#post2894427
Cheers !!
jer
Last edited:
Thanks for the thoughtful reply. It sounds like we may be chasing the same ghost.
In my test setup, an oversized square of aluminized mylar is simply draped over the uninsulated, perforated stator. The only "spacer" is the thickness of the mylar.
Even if you take the audio modulation out of the equation, I can't figure out why a static power supply can't maintain the attraction between the stator and diaphragm. I'm sure the power supply still measures the same regardless of what the diaphragm is doing. It's just losing its static attraction as a function of time.
I'm not sure I entirely understand the concept of "charge migration." The common workaround seems to be to use a high resistance in series with the diaphragm. This would make this otherwise elegant situation unworkable as I would need much higher voltages to achieve the same result. I wonder if a "charge ring" would help things.
For what it's worth, the panel is not being overdriven in this scenario. I have had this happen previously and it does similarly reduce the sensitivity of the panel and can sometimes result in arcing, but my current problem is the loss of sensitivity when the panel is being operated within the range of its capabilities.
-Nick
In my test setup, an oversized square of aluminized mylar is simply draped over the uninsulated, perforated stator. The only "spacer" is the thickness of the mylar.
Even if you take the audio modulation out of the equation, I can't figure out why a static power supply can't maintain the attraction between the stator and diaphragm. I'm sure the power supply still measures the same regardless of what the diaphragm is doing. It's just losing its static attraction as a function of time.
I'm not sure I entirely understand the concept of "charge migration." The common workaround seems to be to use a high resistance in series with the diaphragm. This would make this otherwise elegant situation unworkable as I would need much higher voltages to achieve the same result. I wonder if a "charge ring" would help things.
For what it's worth, the panel is not being overdriven in this scenario. I have had this happen previously and it does similarly reduce the sensitivity of the panel and can sometimes result in arcing, but my current problem is the loss of sensitivity when the panel is being operated within the range of its capabilities.
-Nick
I have only found one scenario where a charge ring is of any use and that is where for any reason that the conductive coating might become an open connection.
I have had this happen usually around the perimeter of the diagphram mounting frame.
When I first started using Licron it was a bit brittle and it would crack right at the edge where the diagphram is allowed to start flexing, Rendering the diagphram not being able to charge.
I solved this by using a thicker coat along these edges and never had a problem with them after that,even after 9 years.
I when I rebuilt the panels I added a charge ring, But do to a design flaw of not having one in the first place there was not ample space for it to be there.
It was my original cause of arc over to a weak spot on the edge of the stator and arced through to one of the mount screws that is used to sandwich the stators together.
I then did another complete refurbishing of the panel and this proved very difficult to repair.
It lasted for a while until I really pushed the voltages.
I would have been better off with a new ground up build for the time I spent on it and is were I am at now.
I have not witnessed any reduction of sensitivity by using any resistors to feed the bias to the diagphram unless there is a lot of leakage somewhere, Theoreticaly there should be none.
Any oil from your hands and fingerprints can cause this!
I wash everything down with denatured alcohol when I am handling the components before assembly.
All of my panels will play for quite a while after the bias voltage has been reduced to zero.
Infact most of the time I can't get all of the charge to dissipate by manually discharging by shorting them out.
Only time or taking them apart will get the job done.
I have used double sided transparent tape on my micro esl and it worked great.
But again I took every measure to assure that there was no path for leakages.
I ran that one on a 500v Bias and work perfectly with no issues from the time a assembled it to when I tore it apart for some pictures.
I still have it and it can be re-assembled at anytime.
I just need to make a second one so that I will have a pair to make some headphones out of them.
I have some aluminized mylar that I was going to try, but, studies have shown that there is less THD in the lower frequency's by using a high resistance coating.
So I never used it but I still have it as well.
I may get it out try it just for kicks.
My plan was to use it for a direct drive type of transducer as only a single amplifier would be required to drive the panel.
But aluminised mylar doesn't handle current to good at much higher voltages.
On a single ended device I can see that it is possible that maybe on some peaks this might cancel out the bias voltage but since you are using such a low resistance coating the charge time is very fast and should not be an issue.
Check the connection to your diagpram as if there has been any arcing it could have vaporized the aluminium around the connection to the diagphram causing it to become open.
I did experiment with this and is one of the reasons I never used the material as well.
This could be simply checked with an ohm meter from the feed wire to another point on the coated side of the mylar.
You can also try coating the Stator with some clear acrylic enamel and see if that solves your leakage problem.
I don't remember off hand what the dielectric strength of mylar is, But look for any pinholes or spots were the aluminium coating is gone.
As this will be evidence of any voltage breakdown of the mylar from the uncoated side.
A charge ring may help in this situation as the sharp current spikes will be distributed over a larger area connection to the diagphram.
Like the function of a large buss bar would do in a power distribution point.
jer
I have had this happen usually around the perimeter of the diagphram mounting frame.
When I first started using Licron it was a bit brittle and it would crack right at the edge where the diagphram is allowed to start flexing, Rendering the diagphram not being able to charge.
I solved this by using a thicker coat along these edges and never had a problem with them after that,even after 9 years.
I when I rebuilt the panels I added a charge ring, But do to a design flaw of not having one in the first place there was not ample space for it to be there.
It was my original cause of arc over to a weak spot on the edge of the stator and arced through to one of the mount screws that is used to sandwich the stators together.
I then did another complete refurbishing of the panel and this proved very difficult to repair.
It lasted for a while until I really pushed the voltages.
I would have been better off with a new ground up build for the time I spent on it and is were I am at now.
I have not witnessed any reduction of sensitivity by using any resistors to feed the bias to the diagphram unless there is a lot of leakage somewhere, Theoreticaly there should be none.
Any oil from your hands and fingerprints can cause this!
I wash everything down with denatured alcohol when I am handling the components before assembly.
All of my panels will play for quite a while after the bias voltage has been reduced to zero.
Infact most of the time I can't get all of the charge to dissipate by manually discharging by shorting them out.
Only time or taking them apart will get the job done.
I have used double sided transparent tape on my micro esl and it worked great.
But again I took every measure to assure that there was no path for leakages.
I ran that one on a 500v Bias and work perfectly with no issues from the time a assembled it to when I tore it apart for some pictures.
I still have it and it can be re-assembled at anytime.
I just need to make a second one so that I will have a pair to make some headphones out of them.
I have some aluminized mylar that I was going to try, but, studies have shown that there is less THD in the lower frequency's by using a high resistance coating.
So I never used it but I still have it as well.
I may get it out try it just for kicks.
My plan was to use it for a direct drive type of transducer as only a single amplifier would be required to drive the panel.
But aluminised mylar doesn't handle current to good at much higher voltages.
On a single ended device I can see that it is possible that maybe on some peaks this might cancel out the bias voltage but since you are using such a low resistance coating the charge time is very fast and should not be an issue.
Check the connection to your diagpram as if there has been any arcing it could have vaporized the aluminium around the connection to the diagphram causing it to become open.
I did experiment with this and is one of the reasons I never used the material as well.
This could be simply checked with an ohm meter from the feed wire to another point on the coated side of the mylar.
You can also try coating the Stator with some clear acrylic enamel and see if that solves your leakage problem.
I don't remember off hand what the dielectric strength of mylar is, But look for any pinholes or spots were the aluminium coating is gone.
As this will be evidence of any voltage breakdown of the mylar from the uncoated side.
A charge ring may help in this situation as the sharp current spikes will be distributed over a larger area connection to the diagphram.
Like the function of a large buss bar would do in a power distribution point.
jer
Last edited:
Again, thanks for the detailed response. There is a lot to chew on here.
While it is altogether possible that there are tiny pinholes in the mylar that would represent a leakage path between the aluminum deposition layer and the stator, I still don't understand why the power supply does not instantaneously compensate for any leaked current. It is as though the mylar is exhibiting a memory effect and only charging to a certain level.
For what it's worth, freshly cut diaphragms seem to maintain their attraction longer than well-worn ones, but they eventually lose sensitivity and peel away from the stator, as well. Also, I am currently using a thin strip of copper folded over a corner of the diaphragm to keep the alligator clip lead from puncturing the aluminum layer on the mylar.
Ignoring the audio aspect of it completely, shouldn't it be possible to maintain the static attraction between the stator and diaphragm as long as the power supply isn't compromised? What is the mechanism behind the diaphragm becoming re-energized when the power supply is shorted? I am so pleased with the sound when everything is working, but the loss of sensitivity makes it totally unusable.
-Nick
While it is altogether possible that there are tiny pinholes in the mylar that would represent a leakage path between the aluminum deposition layer and the stator, I still don't understand why the power supply does not instantaneously compensate for any leaked current. It is as though the mylar is exhibiting a memory effect and only charging to a certain level.
For what it's worth, freshly cut diaphragms seem to maintain their attraction longer than well-worn ones, but they eventually lose sensitivity and peel away from the stator, as well. Also, I am currently using a thin strip of copper folded over a corner of the diaphragm to keep the alligator clip lead from puncturing the aluminum layer on the mylar.
Ignoring the audio aspect of it completely, shouldn't it be possible to maintain the static attraction between the stator and diaphragm as long as the power supply isn't compromised? What is the mechanism behind the diaphragm becoming re-energized when the power supply is shorted? I am so pleased with the sound when everything is working, but the loss of sensitivity makes it totally unusable.
-Nick
Thinking back to 2003 when I built my very first panel I remember I to had experienced such actions as no sound.
Some of this I found to be a crack in the Diagphram coating.
But even on my final panel of today I notice a similar action as I begin to charge the panel, it starts to get louder and then it drops sharply at one point (gets quieter) then as a keep raising the voltage starts to get loud again.
To the point that there was very little sound and sometimes absolutely no sound at the same input voltages that it should be producing well over 100db!
It is when the drop occurs is when the sound appears to be coming from behind the panel(very strange phase cancellations from a single panel).
Once the panel gets fully charged then varying the bias voltage only effects its loudness after that, unless the diagphram losses its charge and then we start from the beginning again.
Back then I was using 900v bias as that was all that I had until I built a better supply.
I was also experimenting with whether or not one should coat both sides of the diagphram.
And was when I had discovered some unusual actions,But they were intermittent and I could not always get them to repeat.
One of them was loss of sound and and diminishing of the sound after time and after taking the panel apart, finding nothing wrong, putting it back together again, it would work fine and every few times the same thing would happen again, as you have described.
Looking back on Volta's pith ball experiment sheds a little light on some things.
But here are some more things to think about.
Remember that we are dealing with a capacitor, not only one, but two of them!
First as you already know we have a capacitor consisted of the stator and the diagphram as this is common knowledge.
The second one is within the diagphram itself as it is an insulator with a conductive surface on one side.
And just because there is (lets say) positive charge on the conductive side doesn't necessarily mean that there will be the same polarity of charge on the opposite side.
In fact there very well could be an equal negative charge on the opposite side because it is a capacitor therefore canceling out the effect of the bias voltage.
I had discovered this once when I had both sides of the diagphram coated but they were two separate circuits.
The idea was "Will I have more efficiency having both side coated as opposed to one?".
It turned out yes it was but the difference was so small that it hardly made a difference at all, Like maybe .5db to 1db (very nil).
I had no measuring equipment at the time but I could hear the difference as I connected and disconnected one side of the diagphram while it was playing.
Then all of a sudden it stopped playing all together and I was like WTF just happened,but when I reconnected the other side of the diagphram the started playing again.
So after that I stuck to just coating one side and never really had a problem with it since (until I built my highly efficeint panels then I started to see some strange things as I had already explained).
I had forgotten all about this until now, So thank you for reminding me!!!!
This explains the strange actions when I was testing my diagphrams in a single ended configuration as to why they acted as if the were not working but I could feel the static charge as I handled them.
Although I did have this in the back of my mind and wrote it off and ignored it.
Here is what I think is happening,
The reason as to why they kept playing after I had shorted out the conductive side is,
The plain side still had a charge on it and as I reapplied the voltage they would fade out until I got past the equal opposite voltage on the other side of the diagphram and then they would get loud again and act normal.
Once they were fully charged ( above about 4Kv or so)then both sides would have the same polarity of charge and the sound level would just change in volume as the bias was varied through its complete range of 0V to 10Kv.
Although my supply can go to 13.8Kv leakages in the panel prevent it from getting that high.
Very interesting indeed,I may have to have another look at coating both side of the diagphram again!!
jer
Some of this I found to be a crack in the Diagphram coating.
But even on my final panel of today I notice a similar action as I begin to charge the panel, it starts to get louder and then it drops sharply at one point (gets quieter) then as a keep raising the voltage starts to get loud again.
To the point that there was very little sound and sometimes absolutely no sound at the same input voltages that it should be producing well over 100db!
It is when the drop occurs is when the sound appears to be coming from behind the panel(very strange phase cancellations from a single panel).
Once the panel gets fully charged then varying the bias voltage only effects its loudness after that, unless the diagphram losses its charge and then we start from the beginning again.
Back then I was using 900v bias as that was all that I had until I built a better supply.
I was also experimenting with whether or not one should coat both sides of the diagphram.
And was when I had discovered some unusual actions,But they were intermittent and I could not always get them to repeat.
One of them was loss of sound and and diminishing of the sound after time and after taking the panel apart, finding nothing wrong, putting it back together again, it would work fine and every few times the same thing would happen again, as you have described.
Looking back on Volta's pith ball experiment sheds a little light on some things.
But here are some more things to think about.
Remember that we are dealing with a capacitor, not only one, but two of them!
First as you already know we have a capacitor consisted of the stator and the diagphram as this is common knowledge.
The second one is within the diagphram itself as it is an insulator with a conductive surface on one side.
And just because there is (lets say) positive charge on the conductive side doesn't necessarily mean that there will be the same polarity of charge on the opposite side.
In fact there very well could be an equal negative charge on the opposite side because it is a capacitor therefore canceling out the effect of the bias voltage.
I had discovered this once when I had both sides of the diagphram coated but they were two separate circuits.
The idea was "Will I have more efficiency having both side coated as opposed to one?".
It turned out yes it was but the difference was so small that it hardly made a difference at all, Like maybe .5db to 1db (very nil).
I had no measuring equipment at the time but I could hear the difference as I connected and disconnected one side of the diagphram while it was playing.
Then all of a sudden it stopped playing all together and I was like WTF just happened,but when I reconnected the other side of the diagphram the started playing again.
So after that I stuck to just coating one side and never really had a problem with it since (until I built my highly efficeint panels then I started to see some strange things as I had already explained).
I had forgotten all about this until now, So thank you for reminding me!!!!
This explains the strange actions when I was testing my diagphrams in a single ended configuration as to why they acted as if the were not working but I could feel the static charge as I handled them.
Although I did have this in the back of my mind and wrote it off and ignored it.
Here is what I think is happening,
The reason as to why they kept playing after I had shorted out the conductive side is,
The plain side still had a charge on it and as I reapplied the voltage they would fade out until I got past the equal opposite voltage on the other side of the diagphram and then they would get loud again and act normal.
Once they were fully charged ( above about 4Kv or so)then both sides would have the same polarity of charge and the sound level would just change in volume as the bias was varied through its complete range of 0V to 10Kv.
Although my supply can go to 13.8Kv leakages in the panel prevent it from getting that high.
Very interesting indeed,I may have to have another look at coating both side of the diagphram again!!
jer
Last edited:
I'm still trying to wrap my head around your capacitor-within-a-capacitor explanation, but I'm glad my question jogged your memory.
I've been out of town for several days, but I had a chance to run some trials last night. Worth noting is the fact that replacing the aluminized mylar in my test setup with thin aluminum foil and a paper insulator completely resolved the intermittent charge issue. The sensitivity wasn't nearly as good as the mylar due to the extra weight and gap thickness, but it held a uniform charge for as long as it was plugged in. Also worth noting is that it discharged itself within seconds of unplugging the power supply, whereas the mylar would hold a charge for a much longer time once the bias was removed. Upon removing the mylar sheet completely from the test circuit, it retained a static charge and happily adhered itself to the metal backsplash of my kitchen range for the better part of an hour before releasing itself. While its sensitivity is superior when it is working, it is very frustrating that it won't hold a charge predictably.
-Nick
I've been out of town for several days, but I had a chance to run some trials last night. Worth noting is the fact that replacing the aluminized mylar in my test setup with thin aluminum foil and a paper insulator completely resolved the intermittent charge issue. The sensitivity wasn't nearly as good as the mylar due to the extra weight and gap thickness, but it held a uniform charge for as long as it was plugged in. Also worth noting is that it discharged itself within seconds of unplugging the power supply, whereas the mylar would hold a charge for a much longer time once the bias was removed. Upon removing the mylar sheet completely from the test circuit, it retained a static charge and happily adhered itself to the metal backsplash of my kitchen range for the better part of an hour before releasing itself. While its sensitivity is superior when it is working, it is very frustrating that it won't hold a charge predictably.
-Nick
A capacitor needs two conducting surfaces separated by a dielectric which can be air, some material, or a combination of both. So, I don't think a capacitor within a capacitor is really relevent to your situation since you don't have conducting surfaces on both sides of the mylar the way geraldfryjr had in his test panel.
See the attached figures for a simplified explanation of what is going on.
(A) When you first turn on the HV supply, the HV capacitors charge and their voltage levels are transferred to the ESL stator and diaphragm. These voltages levels placed in close proximity to each other result in attaction between the metalized layer of the diaphragm and the stator. The figure shows the diaphragm a distance away from the stator(for ease of drawing), rather than stuck to it which is what happens.
(B) Over time, the mylar starts to pick up some charge from the stator it is in contact with. Once it builds up to a level similar to that of the charge on the metalized side of the diaphragm, the attraction between the metalized layer and the stator are offset by the repulsion of the static charge collected on the mylar. Turning off the HV supply will make little difference in the situation until the supply capacitors are discharged.
(C) Discharging the supply capacitors with a resistor removes the charges on the stator and metalized side of the mylar as well. All that remains is the static charge on the mylar. This static charge will drain off over time depending on the humidity level. As you found out, the retained static charge will cause it to be attracted to pretty much any object.
If you had the required test equipment, you could also see that the phase of the sound output is reversed in (C) relative to (A).
So, why did the paper insulator fix the problem?
Because it is not as good an insulator as mylar and charge was bled off about as quickly as it built up.
Similar charge retention problems will be encounted with more traditional push-pull ESLs with insulated stators if the insulation material is too good of an insulator. Strickland experienced this when he was experimenting with different types of insulation for the Acoustat stator wires like PE/Teflon. The fix was to switch back to the Janszen patent recommendation of PVC.
“…Strickland told me once that he'd seen lab speakers that actually played louder after power down of the bias supply than before (for a while anyway), and sometimes he observed audio phase reversals at that time too. This was all with PE before he switched to nice leaky PVC. Weird science…”
http://www.diyaudio.com/forums/planars-exotics/58407-insulated-vs-bare-metal-stators.html#post658044
Note that even PVC wire insulation retains a small amount of charge that will "power" the ESL for a while at much reduced sensitivity if the HV supply is removed or disconnected after having been left on for 24 hours. As with your single ended experiment, the phase of acoustic output is reversed. If you have a variable supply, it is interesting to turn it back on and slowly increase the voltage. Instead of SPL increasing, it first decreases, reaches a null point, and then increases asthe HV is brought up to normal operating level.
See the attached figures for a simplified explanation of what is going on.
(A) When you first turn on the HV supply, the HV capacitors charge and their voltage levels are transferred to the ESL stator and diaphragm. These voltages levels placed in close proximity to each other result in attaction between the metalized layer of the diaphragm and the stator. The figure shows the diaphragm a distance away from the stator(for ease of drawing), rather than stuck to it which is what happens.
(B) Over time, the mylar starts to pick up some charge from the stator it is in contact with. Once it builds up to a level similar to that of the charge on the metalized side of the diaphragm, the attraction between the metalized layer and the stator are offset by the repulsion of the static charge collected on the mylar. Turning off the HV supply will make little difference in the situation until the supply capacitors are discharged.
(C) Discharging the supply capacitors with a resistor removes the charges on the stator and metalized side of the mylar as well. All that remains is the static charge on the mylar. This static charge will drain off over time depending on the humidity level. As you found out, the retained static charge will cause it to be attracted to pretty much any object.
If you had the required test equipment, you could also see that the phase of the sound output is reversed in (C) relative to (A).
So, why did the paper insulator fix the problem?
Because it is not as good an insulator as mylar and charge was bled off about as quickly as it built up.
Similar charge retention problems will be encounted with more traditional push-pull ESLs with insulated stators if the insulation material is too good of an insulator. Strickland experienced this when he was experimenting with different types of insulation for the Acoustat stator wires like PE/Teflon. The fix was to switch back to the Janszen patent recommendation of PVC.
“…Strickland told me once that he'd seen lab speakers that actually played louder after power down of the bias supply than before (for a while anyway), and sometimes he observed audio phase reversals at that time too. This was all with PE before he switched to nice leaky PVC. Weird science…”
http://www.diyaudio.com/forums/planars-exotics/58407-insulated-vs-bare-metal-stators.html#post658044
Note that even PVC wire insulation retains a small amount of charge that will "power" the ESL for a while at much reduced sensitivity if the HV supply is removed or disconnected after having been left on for 24 hours. As with your single ended experiment, the phase of acoustic output is reversed. If you have a variable supply, it is interesting to turn it back on and slowly increase the voltage. Instead of SPL increasing, it first decreases, reaches a null point, and then increases asthe HV is brought up to normal operating level.
Attachments
Yes, Thank you for verifying this for me Bolserst!!
It was the only logical thing that I could figure.
I have experienced all of the things that you and Strickland had described.
I now have to determine if running a 9kv Bias as oppossed to running a 4.5kv Bias is worth doing as it does make a difference in sensitivity.
I know that this has been discussed before, But I guess I am going to have to setup a test on this some how at a later time.
Now that I have figured out a stator coating that can handle such voltages I have to compare the sensitivity to one that can't, under the same driving voltage conditions.
jer
It was the only logical thing that I could figure.
I have experienced all of the things that you and Strickland had described.
I now have to determine if running a 9kv Bias as oppossed to running a 4.5kv Bias is worth doing as it does make a difference in sensitivity.
I know that this has been discussed before, But I guess I am going to have to setup a test on this some how at a later time.
Now that I have figured out a stator coating that can handle such voltages I have to compare the sensitivity to one that can't, under the same driving voltage conditions.
jer
I hate to belabor the point, since I'm well aware that the general consensus is that I should just throw another stator into the mix and use a high resistance diaphragm, but there is a great simplicity to the single stator approach that is attractive to me and I would very much like to resolve the stability issues as the audio performance is quite satisfactory for my purposes.
The diaphragm action as described by bolserst makes complete sense to me, but I have a couple of additional observations that might shed some light on my continuing problems and may be helpful in some way to others experiencing issues with their more conventional designs:
1) I have been using the term "mylar" as a generic. I am actually using plastic foil of indeterminate origin sold in hobby stores for gift wrapping. It is remarkably thin and light with a mirrored appearance that is slightly transparent when held up to a bright light. My stator is a small piece of very smooth perforated metal (fairly magnetic, presumably steel). There are no rough spots or jagged perforations, which I realize can be an issue with higher-voltage designs. As far as I can tell, this setup is very similar to the early ESL tweeters that used to come with console radios.
2) I noticed that my bias supply (voltage doubled wall current coming off of a 1:1 isolation transformer) causes a slight hum and some static when using either the plastic diaphragm or the foil/paper combination, but it is much more pronounced with the foil/paper diaphragm. This surprises me in either case since I wouldn't expect any appreciable amount of current to be drawn by the bias supply.
3) As predicted by bolserst, I can "reset" the plastic diaphragm to a neutral charge by exposing it to a humid environment (steam from the sink).
4) After a few days of handling the plastic diaphragm, reversing polarity, discharging the supply, and generally abusing it, it seems to become inert, hardly responding to bias voltage of any polarity and suffers dramatically reduced sensitivity to the audio signal. Holding the diaphragm to the light sometimes reveals "ghosting" of the stator holes, presumably where the aluminum deposition has thinned for some reason.
5) I experimented with replacing with the DC bias supply with a self-biasing arrangement consisting of a second step-up transformer connected to secondary of the audio transformer that is subsequently voltage-doubled. This actually worked quite well for louder passages, but turning down the volume quickly caused the bias to droop (the slackening of the diaphragm was clearly visible). I would have expected the bias to stay at the level set by the loudest passage, at least for a while. If my plastic diaphragm is such a good insulator that it is becoming polarized by simply being in contact with the charged stator, as in bolserst's explanation, I would have expected it to happily hold the momentary bias charge for a considerable amount of time.
None of these observations are really in any logical order, they are basically just my lab notes. At the end of the day, my goal is to stabilize the system so that every time I plug it in, the plastic diaphragm sucks itself strongly into the stator and stays there until the bias voltage has been removed. If I really am barking up the wrong tree and my materials simply aren't up to the task, I'd be willing to accept that, but I feel like I am very close to figuring this out.
Thanks in advance for any insight people can share. This community really is an incredible resource.
-Nick
The diaphragm action as described by bolserst makes complete sense to me, but I have a couple of additional observations that might shed some light on my continuing problems and may be helpful in some way to others experiencing issues with their more conventional designs:
1) I have been using the term "mylar" as a generic. I am actually using plastic foil of indeterminate origin sold in hobby stores for gift wrapping. It is remarkably thin and light with a mirrored appearance that is slightly transparent when held up to a bright light. My stator is a small piece of very smooth perforated metal (fairly magnetic, presumably steel). There are no rough spots or jagged perforations, which I realize can be an issue with higher-voltage designs. As far as I can tell, this setup is very similar to the early ESL tweeters that used to come with console radios.
2) I noticed that my bias supply (voltage doubled wall current coming off of a 1:1 isolation transformer) causes a slight hum and some static when using either the plastic diaphragm or the foil/paper combination, but it is much more pronounced with the foil/paper diaphragm. This surprises me in either case since I wouldn't expect any appreciable amount of current to be drawn by the bias supply.
3) As predicted by bolserst, I can "reset" the plastic diaphragm to a neutral charge by exposing it to a humid environment (steam from the sink).
4) After a few days of handling the plastic diaphragm, reversing polarity, discharging the supply, and generally abusing it, it seems to become inert, hardly responding to bias voltage of any polarity and suffers dramatically reduced sensitivity to the audio signal. Holding the diaphragm to the light sometimes reveals "ghosting" of the stator holes, presumably where the aluminum deposition has thinned for some reason.
5) I experimented with replacing with the DC bias supply with a self-biasing arrangement consisting of a second step-up transformer connected to secondary of the audio transformer that is subsequently voltage-doubled. This actually worked quite well for louder passages, but turning down the volume quickly caused the bias to droop (the slackening of the diaphragm was clearly visible). I would have expected the bias to stay at the level set by the loudest passage, at least for a while. If my plastic diaphragm is such a good insulator that it is becoming polarized by simply being in contact with the charged stator, as in bolserst's explanation, I would have expected it to happily hold the momentary bias charge for a considerable amount of time.
None of these observations are really in any logical order, they are basically just my lab notes. At the end of the day, my goal is to stabilize the system so that every time I plug it in, the plastic diaphragm sucks itself strongly into the stator and stays there until the bias voltage has been removed. If I really am barking up the wrong tree and my materials simply aren't up to the task, I'd be willing to accept that, but I feel like I am very close to figuring this out.
Thanks in advance for any insight people can share. This community really is an incredible resource.
-Nick
Very Good, Thanks for your report !!!
1.
Mylar is what is called a PET material and there are many variations as "Mylar" is basically a brand name.
There are also Tensilized (pre stretched) and regular Untensilized that resembles to seem a little stretchy and soft feeling like model airplane coverings such as Monokote and such.
They are all PET films, But there are several variations of them.
The tensilised versions are what is desirable to use as it will not stretch any more on its own and/or lose its set tension except with extreme tempreture variations athough it will return to its original set tension when the temp is restored to were it was set at.
Polyethylene terephthalate - Wikipedia, the free encyclopedia
Stuff like Kapton and Kynar are Polymides and are much more tempurature resistent and can handle a very high heat without shirnking.
Kapton is some amazing stuff, I have a few samples of it.
Kapton - Wikipedia, the free encyclopedia
Check the Dupont website for a wealth of info on this.
2.
Your observations seem to show that there is some amount of current flowing.
You should try putting an ammeter in series with your power supply to find out exactly how much current is flowing.
Or if you have a scope you can see how much ripple there is.
Being that it is a Single Ended setup the hum caused by the ripple will be more pronounced because you don't have the second stator counter acting the force of the ripple voltage on the opposite side of the diagphram.
3 and 4.
This has basically already been discussed.
However, I should note that in a study that Beveredge had written, the corona does tend to eat away the aluminium coating on the surface of the film.
Another example that I have found was on TAC (The Audio Circuit) back in 2000 to 2003 by Willy Lefebvre ( I think was his name).
He used two Diagphrams one on each side of a single central stator kind of like an Isoberic setup if you will, This worked well.
http://www.audiocircuit.com/DIY/Electrostatic-Speakers/Project:Central-stator-ESL-by-Willy-Lefebvre
Later on I had found his websight and his update was that part of the aluminium coating had been eaten away after a while and so he had choose not to rebuild them using this method construction and choose to follow traditional methods after that.
It took a lot of digging to find that info quite a while ago and recent searches for another follow up has not been fruitful, as many of those older sites and links are now defunct.
Also recheck your connection points as I had mentioned earlier for the above reasons.
A visual inspection is not enough I am talking about continuity here.
Also any kind of bodily fluid by handling the panel is sufficeint to cause a path for the charge to leak through.
Especially along the edge sides and believe it or not this is where most all of the failures of even the commercially built panels occur, due to leakage paths along the edge.
5.
The self biasing techniques have been used in the past and been tried by many, It simply does not work well.
IMHO, In order to make it even remotely work well you would need to take the signal and clip it to no more than a .1v p-p signal and multiply that into your required bias voltage.
Although, Even a .1v p-p input signal can be heard very well.
Not to mention the clipping action will probably be heard too as it is can get imposed on the original passing signal.
FWIW
jer
1.
Mylar is what is called a PET material and there are many variations as "Mylar" is basically a brand name.
There are also Tensilized (pre stretched) and regular Untensilized that resembles to seem a little stretchy and soft feeling like model airplane coverings such as Monokote and such.
They are all PET films, But there are several variations of them.
The tensilised versions are what is desirable to use as it will not stretch any more on its own and/or lose its set tension except with extreme tempreture variations athough it will return to its original set tension when the temp is restored to were it was set at.
Polyethylene terephthalate - Wikipedia, the free encyclopedia
Stuff like Kapton and Kynar are Polymides and are much more tempurature resistent and can handle a very high heat without shirnking.
Kapton is some amazing stuff, I have a few samples of it.
Kapton - Wikipedia, the free encyclopedia
Check the Dupont website for a wealth of info on this.
2.
Your observations seem to show that there is some amount of current flowing.
You should try putting an ammeter in series with your power supply to find out exactly how much current is flowing.
Or if you have a scope you can see how much ripple there is.
Being that it is a Single Ended setup the hum caused by the ripple will be more pronounced because you don't have the second stator counter acting the force of the ripple voltage on the opposite side of the diagphram.
3 and 4.
This has basically already been discussed.
However, I should note that in a study that Beveredge had written, the corona does tend to eat away the aluminium coating on the surface of the film.
Another example that I have found was on TAC (The Audio Circuit) back in 2000 to 2003 by Willy Lefebvre ( I think was his name).
He used two Diagphrams one on each side of a single central stator kind of like an Isoberic setup if you will, This worked well.
http://www.audiocircuit.com/DIY/Electrostatic-Speakers/Project:Central-stator-ESL-by-Willy-Lefebvre
Later on I had found his websight and his update was that part of the aluminium coating had been eaten away after a while and so he had choose not to rebuild them using this method construction and choose to follow traditional methods after that.
It took a lot of digging to find that info quite a while ago and recent searches for another follow up has not been fruitful, as many of those older sites and links are now defunct.
Also recheck your connection points as I had mentioned earlier for the above reasons.
A visual inspection is not enough I am talking about continuity here.
Also any kind of bodily fluid by handling the panel is sufficeint to cause a path for the charge to leak through.
Especially along the edge sides and believe it or not this is where most all of the failures of even the commercially built panels occur, due to leakage paths along the edge.
5.
The self biasing techniques have been used in the past and been tried by many, It simply does not work well.
IMHO, In order to make it even remotely work well you would need to take the signal and clip it to no more than a .1v p-p signal and multiply that into your required bias voltage.
Although, Even a .1v p-p input signal can be heard very well.
Not to mention the clipping action will probably be heard too as it is can get imposed on the original passing signal.
FWIW
jer
Last edited:
Good find! Thanks for sharing this. It is similar to the various Colin Kyle patent filings from around that time and pretty accurately describes my setup. I have had some luck of late experimenting with different diaphragm materials - not all metalized plastics are created equal, it would seem. I'm just glad that I'm not stuck with varnished silk and gold leaf! 8]
While I was searching for something on another subject I ran across this document that may be of interest as well,
http://recordist.com/ampex/aes-prep...9_single_ended_electrostatic_loudspeakers.pdf
Enjoy !
jer
http://recordist.com/ampex/aes-prep...9_single_ended_electrostatic_loudspeakers.pdf
Enjoy !
jer
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.