Hello Bolserst,
One last question ( I know I am annoying)...If I use Metalitized 12 micron Aluminum Mylar (only because I have tons of it), and then spray Licron over the aluminized part (or Sodium Lauryl Sulfate hand soap, or turtle wax - or whatever coating de jour), won't the panel operate in Constant Charge Mode, and not Constant Voltage Mode??
One last question ( I know I am annoying)...If I use Metalitized 12 micron Aluminum Mylar (only because I have tons of it), and then spray Licron over the aluminized part (or Sodium Lauryl Sulfate hand soap, or turtle wax - or whatever coating de jour), won't the panel operate in Constant Charge Mode, and not Constant Voltage Mode??
no the metalized layer lets charge migrate over the entire diaphragm where a resistive coating acts to drastically slow the ability of the charge to move, the more resistive the better. forget about using metalized mylar that is not a good option unless you like the idea of blowing things up. Best regards Moray James.
No, as moray james already mentioned, it will operate in constant voltage mode.
Flow of electricity will always choose the path of least resistance. When you coat the aluminum part with Licron you are putting a high resistance(Licron) in parallel with a low resistance(aluminum). Essentially all of the charge will flow thru the aluminum rather than the Licron because it is "easier" for it do so.
You could also think about it in terms of the water analogy you brough up back in post #10.
With the HV supply modeled as a tank of water under pressure, the aluminum might be thought of as a 10" diameter pipe connecting the HV tank to the diaphragm. Large amounts of water(ie charge) easily flowing to or from the diaphragm, so NOT constant charge mode. Adding a 0.001" diameter tube (to represent the Licron) in parallel with it won't slow down the water transfer thru the 10" pipe one bit.
OK, but doesn't the metalitized aluminum (or copper charge ring on a conventional constant charge panel) offer the charge to the coating the same way?
I always hear having a better charge ring (wider copper foil ring, or make the ring go all the way around the panel, etc) to get that charge to the coating...tyu said it improved the spl quite a bit... so why not charge the coating from under the coating? It is still the charged coating between the mylar and stator...correct?
But I do agree, if coating not properly applied, it will have areas of low resistance.
Would having that high resistance coating reduce chance of flame up?
I always hear having a better charge ring (wider copper foil ring, or make the ring go all the way around the panel, etc) to get that charge to the coating...tyu said it improved the spl quite a bit... so why not charge the coating from under the coating? It is still the charged coating between the mylar and stator...correct?
But I do agree, if coating not properly applied, it will have areas of low resistance.
Would having that high resistance coating reduce chance of flame up?
No. The aluminum coating provides a low resistance path between the HV supply and every point on the diaphragm, or between any two different points on the diaphragm. The copper charge ring only provides a low resistance path to the edge of the diaphragm(which is clamped and can't move), not between points out on the diaphragm where motion is taking place.
This can be true, especially for sheet metal and foam tape ESLs that can develop leakage paths across the foam tape which is around the edges of the panel. Suppose you only put the copper foil up the right side of the panel and there is a leakage path on the left side foam tape spacer. The charge on the left side of the diaphragm would partially leak away since its supply comes across the high resistance coating from the copper strip on the right hand side which may be similar in magnitude to the resistance of the leakage path which is draining charge away. But, if you put a copper foil strip on the left side as well, the leakage path on that side has much less affect since there is now a low resistance path to the left side of the diaphragm.
Of course, the HV supply will also be supplying a steady current thru the leakage path which may drop the voltage level a bit, but the overall charge on the left side of the diaphragm will be much higher than when only a copper strip on the right side is used.
Applying a high resistance coating on top of the aluminum will not reduce the available current in the event of arcing because the aluminum still provides a low resistance path from the HV supply to every point on the diaphragm.
Great info on bias an charging the panels............thanks to all......
One thing i have just found after my 10th pr of ML... I have re-work the panels..........is the res. that are in the high V feed........4ea 15meg ohm 1/4 watt carbons........ that gives 60meg......
Can an well roll the top end way down on the panels.....just puting better res. in....can give better sound...........if it all you do well give better sound!
Higher SPL an better sound has been my quest.....but it took alot of work to find this one....it almost like the res. that ML uses are there to roll the topend of the panels down to match the bass driver...........but when i get 2-4DB more output of these panels....to my ear the bass driver match much better!.............
So get the right type of res. an get sweeter sound.... an make these great ML panel with these bass driver..... two ways speaker.... sound Much more like fullrang ESL.......
have fun ...this new year
One thing i have just found after my 10th pr of ML... I have re-work the panels..........is the res. that are in the high V feed........4ea 15meg ohm 1/4 watt carbons........ that gives 60meg......
Can an well roll the top end way down on the panels.....just puting better res. in....can give better sound...........if it all you do well give better sound!
Higher SPL an better sound has been my quest.....but it took alot of work to find this one....it almost like the res. that ML uses are there to roll the topend of the panels down to match the bass driver...........but when i get 2-4DB more output of these panels....to my ear the bass driver match much better!.............
So get the right type of res. an get sweeter sound.... an make these great ML panel with these bass driver..... two ways speaker.... sound Much more like fullrang ESL.......
have fun ...this new year
Hi Bolserst,
I'm having a little difficulty understanding the difference between a constant voltage-low resistance diaphragm vs a constant charge-high resistance diaphragm. Wish I could appreciate your wonderful color graphs but I'm afraid their meaning fly over my head higher than some of the jets you work on.
In the paper you sent me a while ago by Frank Verwaal, he used an example of an ESL trying to play a double bass and a flute at the same time(easier for me to relate). I think I understand(?) that in the constant voltage system as the bass plays, the diaphragm, especially at its center, extends closer to the stator attracting the bias charge to move more toward the center also. Does this mean the flute now does not have an equal charge to play as loud, or waver unnaturally in intensity because of the diaphragm charge favoring the lows? If this is the case, can the measurable IM distortion be heard, since even at low frequencies, the excursion is relatively small. Or am I not making any sense at all.
I'm assuming my stats are in the constant voltage category(200k-300k). Would it be beneficial for me to someday take the panels apart and wipe off some of the graphite until the diaphragms are in the several meg range to lessen the charge migration? Too many questions, as always.
Thanks
I'm having a little difficulty understanding the difference between a constant voltage-low resistance diaphragm vs a constant charge-high resistance diaphragm. Wish I could appreciate your wonderful color graphs but I'm afraid their meaning fly over my head higher than some of the jets you work on.
In the paper you sent me a while ago by Frank Verwaal, he used an example of an ESL trying to play a double bass and a flute at the same time(easier for me to relate). I think I understand(?) that in the constant voltage system as the bass plays, the diaphragm, especially at its center, extends closer to the stator attracting the bias charge to move more toward the center also. Does this mean the flute now does not have an equal charge to play as loud, or waver unnaturally in intensity because of the diaphragm charge favoring the lows? If this is the case, can the measurable IM distortion be heard, since even at low frequencies, the excursion is relatively small. Or am I not making any sense at all.
I'm assuming my stats are in the constant voltage category(200k-300k). Would it be beneficial for me to someday take the panels apart and wipe off some of the graphite until the diaphragms are in the several meg range to lessen the charge migration? Too many questions, as always.
Thanks
Hi,
basically the answer reduces to two easy formulas.
The first calculates the Capacitance of two parallel located electrodes (stator and Membrane).
C = epsilon x A/d, with epsilon =8.854xexp-12[F/m], A = diaphragm/stator-area in [m2] and d= distance stator --> membrane in [m].
Since epsilon an A are fixed parameters, the capacitance will vary by the distance d
The second formula also calculates the Capacitance, but from the amount of Charge and applied Voltage.
C = Q/V
This formula rearranged equals Q = CxV or V = Q/C
Both of the latter are constant for a constant-Q or a constant-V ESL.
If now the Capacitance changes due to a varying stator-membrane distance d, then either the voltage between the two electrodes changes (constant-Q) or the amount of charge changes (constant-V).
These formulas also reveal a seemingly lesser known advantage of constant-Q ESLs .... the reduced flashover probability.
When the disphragm approaches the stator, d becomes smaller, eventually even 0 when touching the stator.
Hence the Capacitance inceases.
With a constant-Q ESL then the Voltage term decreases also, becoming 0 when membrane and stator touch.
If it weren't for the polarizing voltage (not included in theses simple formulas) stator and membrane wouldn't theoretically need to be isolated at all.
The probability of a flashover reduces considerably and at the same prevents the high-ohmic membrane coating a spark from becoming 'hot'.
The constant-V ESL on the other hand shows a muuch higher probability to spark.
Not only remains the voltage differential between membrane and stator high, but also does the charge -which can move freely over the membrane area due to thar's low resistance- move towards the closest point between membrane and stator.
Due to the local increase in charge the force acting upon this membrane spot increases also, introducing a self-increasing non-linear component.
At the same the local increase in charge will generate a hot and visible flashover as a large discharge current can flow.
Aluminum sputtererd membranes can burn off in fullness and in a quite spectacular way.
jauu
Calvin
basically the answer reduces to two easy formulas.
The first calculates the Capacitance of two parallel located electrodes (stator and Membrane).
C = epsilon x A/d, with epsilon =8.854xexp-12[F/m], A = diaphragm/stator-area in [m2] and d= distance stator --> membrane in [m].
Since epsilon an A are fixed parameters, the capacitance will vary by the distance d
The second formula also calculates the Capacitance, but from the amount of Charge and applied Voltage.
C = Q/V
This formula rearranged equals Q = CxV or V = Q/C
Both of the latter are constant for a constant-Q or a constant-V ESL.
If now the Capacitance changes due to a varying stator-membrane distance d, then either the voltage between the two electrodes changes (constant-Q) or the amount of charge changes (constant-V).
These formulas also reveal a seemingly lesser known advantage of constant-Q ESLs .... the reduced flashover probability.
When the disphragm approaches the stator, d becomes smaller, eventually even 0 when touching the stator.
Hence the Capacitance inceases.
With a constant-Q ESL then the Voltage term decreases also, becoming 0 when membrane and stator touch.
If it weren't for the polarizing voltage (not included in theses simple formulas) stator and membrane wouldn't theoretically need to be isolated at all.
The probability of a flashover reduces considerably and at the same prevents the high-ohmic membrane coating a spark from becoming 'hot'.
The constant-V ESL on the other hand shows a muuch higher probability to spark.
Not only remains the voltage differential between membrane and stator high, but also does the charge -which can move freely over the membrane area due to thar's low resistance- move towards the closest point between membrane and stator.
Due to the local increase in charge the force acting upon this membrane spot increases also, introducing a self-increasing non-linear component.
At the same the local increase in charge will generate a hot and visible flashover as a large discharge current can flow.
Aluminum sputtererd membranes can burn off in fullness and in a quite spectacular way.
jauu
Calvin
Hi Calvin,
My grasp of simple formulas is on par with my ability to understand certain graphs....very limited. I'm basically an old garage tinkerer who appreciates good music and good reproduced music. Lots of trial and error. No engineering background. Thank you, though, for your thoughtful and detailed explanation.
I am beginning to understand the charge migration issues of a diaphragm with a low resistance coating. What happens when the double bass plays makes sense. But what's happening with the flute.....not sure if I correctly comprehend it. Basically I'm trying to visualize what is physically taking place between the diaphragm and the stator to cause the non linear distortion. Is it similar at all to doppler distortion?
I have made a few panels with aluminized mylar, and as you commented, they can put on quite a show. For some reason flying bugs were attracted to the music almost as much as a light bulb on a summer evening. I'd try to chase them down with a fly swatter before their electrocution, but wasn't always successful. You could count how many bugs had been fried by the number of holes in the mylar. Besides that, the aluminum coating over time would begin to have hairline fractures, eventually losing continuity and sound.
Thanks again for your help.
Bondsan
My grasp of simple formulas is on par with my ability to understand certain graphs....very limited. I'm basically an old garage tinkerer who appreciates good music and good reproduced music. Lots of trial and error. No engineering background. Thank you, though, for your thoughtful and detailed explanation.
I am beginning to understand the charge migration issues of a diaphragm with a low resistance coating. What happens when the double bass plays makes sense. But what's happening with the flute.....not sure if I correctly comprehend it. Basically I'm trying to visualize what is physically taking place between the diaphragm and the stator to cause the non linear distortion. Is it similar at all to doppler distortion?
I have made a few panels with aluminized mylar, and as you commented, they can put on quite a show. For some reason flying bugs were attracted to the music almost as much as a light bulb on a summer evening. I'd try to chase them down with a fly swatter before their electrocution, but wasn't always successful. You could count how many bugs had been fried by the number of holes in the mylar. Besides that, the aluminum coating over time would begin to have hairline fractures, eventually losing continuity and sound.
Thanks again for your help.
Bondsan
Hi Bondsan
Lets try with words.
Imagine that there is a voltage applied to the stators, positive voltage on one stator, negative on the other. This creates a uniform electric field between the two stators. If there is an electric charge between the stators, it experiences the same force, no matter its position between the stators.
If you put a charged membrane with a high resistivity and a large series resistance, then when the membrane moves, the charge stays the same and the forces on the membrane stay the same.
Now, if we use a highly conductive membrane, as the membrane moves close to either of the stators, the charge on the membrane changes, and therefore, the force on the membrane also changes.
In the ideal ESL, we want the forces to be directly proportional to the stator voltages - i.e., so that the electrostatic motor is linear.
If we use a conductive membrane, the motor is non-linear, causing distortion of many % at low frequencies where the membrane displacement is large. A few years ago I saw measurements on one ESL using such a membrane with distortion figures of 8%!
Another side effect of a conductive membrane is that the closer a conductive membrane gets to either of the stators, the greater the force of attraction. If it gets close enough, it will stick to the stator, and if large currents flow, you can blow quite large holes in it. If you use a high resistivity membrane and a decent series resistor connecting it to the HT supply, the charge movement is very limited and the damage to the membrane will be minor.
Hope that helps
R
Lets try with words.
Imagine that there is a voltage applied to the stators, positive voltage on one stator, negative on the other. This creates a uniform electric field between the two stators. If there is an electric charge between the stators, it experiences the same force, no matter its position between the stators.
If you put a charged membrane with a high resistivity and a large series resistance, then when the membrane moves, the charge stays the same and the forces on the membrane stay the same.
Now, if we use a highly conductive membrane, as the membrane moves close to either of the stators, the charge on the membrane changes, and therefore, the force on the membrane also changes.
In the ideal ESL, we want the forces to be directly proportional to the stator voltages - i.e., so that the electrostatic motor is linear.
If we use a conductive membrane, the motor is non-linear, causing distortion of many % at low frequencies where the membrane displacement is large. A few years ago I saw measurements on one ESL using such a membrane with distortion figures of 8%!
Another side effect of a conductive membrane is that the closer a conductive membrane gets to either of the stators, the greater the force of attraction. If it gets close enough, it will stick to the stator, and if large currents flow, you can blow quite large holes in it. If you use a high resistivity membrane and a decent series resistor connecting it to the HT supply, the charge movement is very limited and the damage to the membrane will be minor.
Hope that helps
R
Hi Golfnut,
A few words make a big picture. So in a constant voltage system(like my semi FR panels), as long as James Galway is playing a solo part of a flute concerto, the speakers will sound ok. However, once the rest of the orchestra, especially the double basses, join him, the diaphragm becomes non-linear and all the music, including Galway's flute, are reproduced with some distortion....Right? Hopefully not 8%!
This leaves me in a bit a of a quandary. Should I try to rub some of the graphite off with alcohol, which may be difficult to accomplish evenly, particularly around the membrane edges. Or tear the plastic frames apart(double sided tape holds them together) and rebuild them using a high resistance coating. We'll see.
Thanks for helping an advancing years dog learn some new tricks.
Bondsan
A few words make a big picture. So in a constant voltage system(like my semi FR panels), as long as James Galway is playing a solo part of a flute concerto, the speakers will sound ok. However, once the rest of the orchestra, especially the double basses, join him, the diaphragm becomes non-linear and all the music, including Galway's flute, are reproduced with some distortion....Right? Hopefully not 8%!
This leaves me in a bit a of a quandary. Should I try to rub some of the graphite off with alcohol, which may be difficult to accomplish evenly, particularly around the membrane edges. Or tear the plastic frames apart(double sided tape holds them together) and rebuild them using a high resistance coating. We'll see.
Thanks for helping an advancing years dog learn some new tricks.
Bondsan
The distortion is not related to Doppler distortion.
If you follow golfnut’s description that if you can keep the charge on the diaphragm constant the force on the diaphragm due to the stator voltages will have a linear relationship, I think you are almost there to getting a feel for what is physically taking place.
The piece missing is how/why does the charge on the diaphragm change if the coating is highly conductive.
1) The amount of charge that can be stored on the diaphragm is proportional to the bias supply voltage level, and the sum of the capacitance between the diaphragm and the two stators. If the diaphragm moves closer to one of the stators the capacitance increases, so it can store more charge and the force on the diaphragm increases as a result of this increased charge. (ie a nonlinear stator voltage to diaphragm force relationship which will affect all frequencies present during the portion of time the charge has increased)
2) The time it takes to move charge on the diaphragm is inversely proportional to any resistance in the circuit. Having a high resistance coating increases the time it takes for the charge to move on or off the diaphragm. If the time constant is long enough, there will not be time for the charge to change during the cyclic motion of the diaphragm closer and further from the stators. If the resistance is low and the time constant short, the charge on the diaphragm will have time to change in unison with the diaphragm motion. Obviously it is not an all or nothing situation; depending on the resistance there will be a frequency range above which operation is constant charge and below which operation transitions to constant voltage.
3) Note that the resistors you put in between your bias supply and the diaphragm are also in series with the charging circuit for the diaphragm. So, they will also increase the charging time constant and help the ESL operate in the constant charge mode.
Golfnut mentioned the issue of charge that is already on the diaphragm moving around. Obviously the resistor between your bias supply and diaphragm will not keep this from happening. But, it does keep the total charge on the diaphragm from changing which goes a long way toward keeping the distortion level in check. You can imagine that if the total charge on the diaphragm is constant and some of it moves toward the middle where the diaphragm is closer to the stator, it increases the force in the middle, but decreases the force in the areas from where the charge came from. This is a much better situation than if additional charge is pumped on and off the diaphragm from the bias supply.
BTW, this is the approach Acoustat took…fairly conductive diaphragm with high resistance in line with the bias supply to keep total charge on the diaphragm constant.
Last edited:
As Bolserst says, a resistor in series with the membrane will do most of the job - something greater than 10Mohm should do a nice job. A lot of HT supplies will already have an output impedance of this magnitude anyway.
The distortion due to change movement on and off the membrane is reduced as the resistance increases, so the larger the resistance the better. It really depends on how well made the panels are, if good insulating materials are used you may be able to use 100 Mohm,
and that would be heaps. I suggest modifying one panel at a time and checking it against the other using a mono (AM radio) rather than stereo source. If you make value too large, you'll notice a drop in the output, the mono image will move off center, due to poor resistance of insulating materials. If you have a microphone to check - even better.
When handling recently charged speakers, I suggest unplugging the speakers, wait a few minutes, discharge the panels/supply with a screwdriver (connect ground first), all the while keeping one hand in your pocket. ESL HT should not be fatal, but it will bite hard (several crocodiles on the Rutherford scale). Never touch them with audio connected - that is seriously dangerous.
R
The distortion due to change movement on and off the membrane is reduced as the resistance increases, so the larger the resistance the better. It really depends on how well made the panels are, if good insulating materials are used you may be able to use 100 Mohm,
and that would be heaps. I suggest modifying one panel at a time and checking it against the other using a mono (AM radio) rather than stereo source. If you make value too large, you'll notice a drop in the output, the mono image will move off center, due to poor resistance of insulating materials. If you have a microphone to check - even better.
When handling recently charged speakers, I suggest unplugging the speakers, wait a few minutes, discharge the panels/supply with a screwdriver (connect ground first), all the while keeping one hand in your pocket. ESL HT should not be fatal, but it will bite hard (several crocodiles on the Rutherford scale). Never touch them with audio connected - that is seriously dangerous.
R
Hi Bondsan
Lets try some maths
The distortion due to change movement is reduced in proportion to
1/4/pi/f/RC
where
f is the frequency
R is the resistance
C is capacitance
The capacitance of a panel is approximately 0.8pF x panel area (in cm squared) divided by the distance between the stators in mm.
Eg a panel 20cm x 40cm with 1.6 mm stator-membrane spacing (3.2mm between stators) = 200 pF
so with f = 100 Hz, R = 10M, C = 200 pF, the reduction in distortion is a factor of 4.
If your little panels are smaller than this, you might have to use a higher resistance to gain any benefit at all.
hope this helps
R
BTW If you think Marlborough Sav. Blanc is good, try any NZ wine from Central Otago, Pinot Noir a specialty, but all very good.
Lets try some maths
The distortion due to change movement is reduced in proportion to
1/4/pi/f/RC
where
f is the frequency
R is the resistance
C is capacitance
The capacitance of a panel is approximately 0.8pF x panel area (in cm squared) divided by the distance between the stators in mm.
Eg a panel 20cm x 40cm with 1.6 mm stator-membrane spacing (3.2mm between stators) = 200 pF
so with f = 100 Hz, R = 10M, C = 200 pF, the reduction in distortion is a factor of 4.
If your little panels are smaller than this, you might have to use a higher resistance to gain any benefit at all.
hope this helps
R
BTW If you think Marlborough Sav. Blanc is good, try any NZ wine from Central Otago, Pinot Noir a specialty, but all very good.
Yeah...they certainly do appear to be contradictory statements don’t they.
Let me try restating the two comments in reverse order to clarify.
2) For a given series resistance and ESL capacitance there is a finite time it takes to change the charge on the diaphragm. For higher frequencies, the stator voltages are changing too quickly for diaphragm charge to change. At lower frequencies, the stator voltages are changing more slowly and the diaphragm charge has time to change. The lower the membrane and series resistance the shorter the charging time constant and the higher the transition frequency between the diaphragm charge not having time to change and having time to change.
1) If stator voltages applied to your ESL have low frequency content that allows the diaphragm charge to change, nonlinearity results. If there is high frequency content present at the same time as the low frequency content, the high frequency content will also be affected in a nonlinear way by the diaphragm charge changing. If the same high frequency content was applied to the stators without the low frequency content present, the diaphragm charge would not change so no nonlinear effects.
"There are three kinds of people in the world....those that understand math and those that don't".
Hi Golfnut
My panels average around 3" x 14" in size. Converting to CM and dusting off my Arithmometer, brings them 85.34 square CM, resulting in a pF of 68.27(?). Not sure how accurate this is since the panels are sub divided into smaller cells too. Will probably order some 20 megs and give them a try.
Found a bottle of Otago Pinot Noir. Will give it a try with the appropriate dinner.
Hi Bolserst
By golly I think I actually understand what's taking place between the D/S! Your reverse explanation cleared up what did seem a contradiction and was most helpful. It would also seem as the volume is increased the greater the non linear distortion becomes.
This leads me to more diaphragm questions, but I will post those in another thread. Many thanks to you and Golfnut for your non engineer explanations.
Bondsan
Hi Golfnut
My panels average around 3" x 14" in size. Converting to CM and dusting off my Arithmometer, brings them 85.34 square CM, resulting in a pF of 68.27(?). Not sure how accurate this is since the panels are sub divided into smaller cells too. Will probably order some 20 megs and give them a try.
Found a bottle of Otago Pinot Noir. Will give it a try with the appropriate dinner.
Hi Bolserst
By golly I think I actually understand what's taking place between the D/S! Your reverse explanation cleared up what did seem a contradiction and was most helpful. It would also seem as the volume is increased the greater the non linear distortion becomes.
This leads me to more diaphragm questions, but I will post those in another thread. Many thanks to you and Golfnut for your non engineer explanations.
Bondsan