Hi,
I'm not sure how long this problem has been there, but I guess about two or three weeks. My right DIY esl (a 68cm x 200cm full range panel) has much lower sensitivity than the left one: I measure 8.5 dB less at all frequencies. This is too big of a difference to be ignored. In the first months after completion there was not such a big difference. From what I remember, 1 dB difference.
I use second hand AudioStatic HV bias power supplies for the diaphragm. Not sure if this can be translated in a bias voltage by the schematic alone? Or is this also dependent on other things like diaphrahm coating? Because I would like the forumula posted by bolsert to calculate theoretical SPL to get an idea how efficient my esl is and how much possibly is leaking:
https://www.diyaudio.com/forums/pla...ns-low-sensitivity-range-esl.html#post3006565
Things I already tried
I connected my own bias power supply from a previous model as this one has a neon flash light in order to see flashing rate / leakage. I swapped everything: audio transformer, amplifier channel, HV bias power supply, segmentation resistor circuits, but the right esl stays much quieter. (the difference stays at 8.5 dB)
So yesterday I disassembled the right esl panel and cleaned the diaphragm on the outer edges with alcohol just to be sure it's not leaking here to the wooden frame. I checked the HV bias wire solder point where it contacts the copper charging ring because when putting one ear close to this point I heard a hissing noise. After coating that area the hissing disappeared and neon light flash rate decreased considerably. But as soon as I put both frames together (one frame has the copper charing ring and silk damping screen, the other has the mylar diaphragm) the neon light starts to increase flashing frequency.
I checked for dust, hairs etc. I found a little insect and removed it, but no difference.
I'm wondering what could be the cause - or what method would work to find the cause? Right now I used the neon flash light.
I'm not sure how long this problem has been there, but I guess about two or three weeks. My right DIY esl (a 68cm x 200cm full range panel) has much lower sensitivity than the left one: I measure 8.5 dB less at all frequencies. This is too big of a difference to be ignored. In the first months after completion there was not such a big difference. From what I remember, 1 dB difference.
I use second hand AudioStatic HV bias power supplies for the diaphragm. Not sure if this can be translated in a bias voltage by the schematic alone? Or is this also dependent on other things like diaphrahm coating? Because I would like the forumula posted by bolsert to calculate theoretical SPL to get an idea how efficient my esl is and how much possibly is leaking:
https://www.diyaudio.com/forums/pla...ns-low-sensitivity-range-esl.html#post3006565
Things I already tried
I connected my own bias power supply from a previous model as this one has a neon flash light in order to see flashing rate / leakage. I swapped everything: audio transformer, amplifier channel, HV bias power supply, segmentation resistor circuits, but the right esl stays much quieter. (the difference stays at 8.5 dB)
So yesterday I disassembled the right esl panel and cleaned the diaphragm on the outer edges with alcohol just to be sure it's not leaking here to the wooden frame. I checked the HV bias wire solder point where it contacts the copper charging ring because when putting one ear close to this point I heard a hissing noise. After coating that area the hissing disappeared and neon light flash rate decreased considerably. But as soon as I put both frames together (one frame has the copper charing ring and silk damping screen, the other has the mylar diaphragm) the neon light starts to increase flashing frequency.
I checked for dust, hairs etc. I found a little insect and removed it, but no difference.
I'm wondering what could be the cause - or what method would work to find the cause? Right now I used the neon flash light.
If I'm remembering correctly, isn't your spacing significantly smaller than Audiostatic's? Bias would normally be reduced proportionally based on spacing. If it has not been, you may be applying too much bias. One speaker could be stable with that voltage and the other not, due to minor differences in tension or spacing. This would typically show as a significant bow in the diaphragm when bias is applied, which can be seen looking at light reflecting off the diaphragm. Taken to extremes, the diaphragm can collapse into the stator, which causes significant reduction in output. If you have a variac around, you could gradually increase bias to see if you run into a point where output between left and right speakers diverge after a certain bias level.
Another thought is a point of high leakage due to stator insulation failure. This is often visible in a dark room as corona. It's not that common with wire stators, but not unheard of.
The area where the insect was may have some carbon tracking. It could be a point of high leakage for the bias. Or some other area may have the same problem.
High voltage probes are handy and don't cost that much anymore. They can help track down what's going on with bias where.
Another thought is a point of high leakage due to stator insulation failure. This is often visible in a dark room as corona. It's not that common with wire stators, but not unheard of.
The area where the insect was may have some carbon tracking. It could be a point of high leakage for the bias. Or some other area may have the same problem.
High voltage probes are handy and don't cost that much anymore. They can help track down what's going on with bias where.
Hi, thanks for your help and suggestions.
I swapped tranformers (I took a different pair) and discovered that the difference in SPL reduced to only 3 dB! So I grinded the thin copper wires of the first pair of transformers where I previously measured 8.5 dB SPL difference and I got a SPL difference of 3 dB again, and soundquality greatly improved!
So after all, it were corroded (secondary) transformer wire connections...
At least thanks for your help as it can still be of help in future.
I swapped tranformers (I took a different pair) and discovered that the difference in SPL reduced to only 3 dB! So I grinded the thin copper wires of the first pair of transformers where I previously measured 8.5 dB SPL difference and I got a SPL difference of 3 dB again, and soundquality greatly improved!
So after all, it were corroded (secondary) transformer wire connections...
At least thanks for your help as it can still be of help in future.
Audiostatic uses an 18 stage multiplier with 0.047uF caps to step up 220Vrms(50/60Hz) to roughly 5.5kV. However, because of the large number of stages and low frequency, the supply has an effective source resistance of about 220Meg. So, even small amounts of leakage will drop the voltage reaching the diaphragm substantially. As an example, trying to measure the HV wutg a 1000Meg HV probe would drop the level to 4.5kV. The full 5.5kV is a little higher than optimal for your 2mm gap size, but not unsafe. With the small amount of leakage indicated by flashing of the neon lamp, you are probably closer to 5kV.
When checking SPL with this formula, be aware that it is accurate only when driving the entire panel uniformly (ie not segmented)
One other method I have used to compare amount of leakage between panels, is to let them charge up and then while playing music disconnect the wire between HV supply and diaphragm. A good low leakage panel may take several minutes to lose 6dB in output while a leakier panel may fade away nearly completely by then.
That is good news! Funny, when I started reading your first post and saw the amount being close to 6dB difference, I was wondering if maybe one of the transformers was not functioning properly.
Last edited:
Again thanks for your reply - I will try measuring the -6 dB drop timeout when disconnecting HV supply - but a quick test suggests the panels have high leakage.
Yesterday the SPL difference was there again... I swapped transformers but the problem remained.
As I already put my ear close to the panels to detect any hissing sound from HV leakage I din't know what to do. My girlfriend suggested loosening the bolts that hold both stators together. I turned on the music on the problematic channel. Loosening the first bolts didn't make any audible difference - but when loosening 1 particular bolt SPL suddenly increased greatly!
First I checked if it was just that particular bolt which was the case. Then I removed the metallic bolt and pushed both stators at that point toghether by hand and the SPL dropped again. Then I noticed a hissing sound about 20 cm above the bolt hole.
So after all it seems to be an easy to localize HV leakage problem. I will keep you informed.
Yesterday the SPL difference was there again... I swapped transformers but the problem remained.
As I already put my ear close to the panels to detect any hissing sound from HV leakage I din't know what to do. My girlfriend suggested loosening the bolts that hold both stators together. I turned on the music on the problematic channel. Loosening the first bolts didn't make any audible difference - but when loosening 1 particular bolt SPL suddenly increased greatly!
First I checked if it was just that particular bolt which was the case. Then I removed the metallic bolt and pushed both stators at that point toghether by hand and the SPL dropped again. Then I noticed a hissing sound about 20 cm above the bolt hole.
So after all it seems to be an easy to localize HV leakage problem. I will keep you informed.
Hi again - I spent a lot of hours on the sensitivity issue now. I removed 1 cm outer edge of the mylar which is glued on a pvc isolation strip. (this trip is glued to the wooden statro frame on it's other side). Just to be sure bias voltage can't leak to the wooden frame at the outer edges. I cleaned sports with alchohol, removed some dust. I made a probe of a pvc strip and some copper tape to simulate a small part of the bias copper charging ring and connected it to the bias power supply and put it on different spots on the mylar to detect leakage sport without effect.
I made some test samples with coated mylar to check leagage. I discovered that the neon bulb flash rate is even higher when connecting earth from bias power suply to the wooden frame rather than to the wire stator.
I had no idea what could cause one panel to leak bias excessively after taking all my measurements. When swapping rear stators the problem is gone.
Yesterday I loosened the particular bolt again and sound pressure increased again. I heard a hissing noise and located it by putting one ear close to the stator. The I switched off light and this time (I tried visual inspection in the dark before), for the first time, I was able to see two very small blue 'dots' flickering. I marked the spoth with tape, but it seems to be at the pvc insulator strip edge on the inner side of the stator. So at the small side surface of the pvc strip, where the mylar is glued on to. My first, quick diagnosis is that maybe the cutted pvc strips have un-smooth side surfaces where bias is leaking to.
I will continue my research soon as even though it frustrated me a bit - at the same time I can learn something about bias leakage.
I made some test samples with coated mylar to check leagage. I discovered that the neon bulb flash rate is even higher when connecting earth from bias power suply to the wooden frame rather than to the wire stator.
I had no idea what could cause one panel to leak bias excessively after taking all my measurements. When swapping rear stators the problem is gone.
Yesterday I loosened the particular bolt again and sound pressure increased again. I heard a hissing noise and located it by putting one ear close to the stator. The I switched off light and this time (I tried visual inspection in the dark before), for the first time, I was able to see two very small blue 'dots' flickering. I marked the spoth with tape, but it seems to be at the pvc insulator strip edge on the inner side of the stator. So at the small side surface of the pvc strip, where the mylar is glued on to. My first, quick diagnosis is that maybe the cutted pvc strips have un-smooth side surfaces where bias is leaking to.
I will continue my research soon as even though it frustrated me a bit - at the same time I can learn something about bias leakage.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.