| Home |
| Search |
| Today's Posts |
|
#1
Posted to rec.audio.tubes
|
|||
|
|||
Electrostatic speaker ( ESL ) sensitivity.
Nowhere can I find an answer to my simple question,
" Why does the ESL speaker sensitivity rise when the EHT membrane charge in increased ?" While making tests of an ERAudio kit I am building, I have found that the sound level increase with increased EHT voltage applied. If I have a small ac signal at the stators, say +/- 500 peak volts, and EHT = -1,000V, then I increase the EHT to -2,000V, but leave the same ac stator signal, why does the SPL rise? With EHT = -1,000V, the max peak voltage differences on each side of the membrane are 500V and 1,500V, so the net force effect is the same as 1,000V applied by one stator. If the EHT is -2,000V, the two voltages between membrane and stator become 1,500V and 2,500V and there is the same net voltage difference of 1,000V, so you'd think the same force is applied to the membrane regardless of the EHT, as long as it exceeds the peak voltage changes in the ac signal applied. I have Ronald Wagner's book on DIYer ESL construction and Ron even has a graph showing a linear increase in SPL with EHT voltage, so that if I could use -5kV instead of only -2.5kV as I am so cautiously using now, I'd get a very welcome rise of +6dB sensitivity for the ERA speakers, something which would place them closer to Quad ESL57 etc. But in the book Ronald does not say why the rise in sensitivity occurs, and I must always ask why, as the answer may lead so some other question, and then another, and another, and everyone around me goes quite mad while I get much wiser. And at present I am also wondering what effect doubling the membrane tension would have, apart from raising the bass resonant frequencies. If Quad could use the same material with 3 times the ERA tension, I should be able to go double without trouble. When someone my age hears a comment, "twice would be nice", one has to put on the hat of youth, and hope that everything works, rather than assume it will like it did at 30. Stretching membranes is nervous nelly territory; one prick and its ****ed! :-)) ( Someone my age rarely will ever hear "twice would be nice". I don't have to try avoid Temptation; she will avoid me.) Patrick Turner. |
|
#2
Posted to rec.audio.tubes
|
|||
|
|||
|
Electrostatic speaker ( ESL ) sensitivity.
On Fri, 20 Apr 2007 06:24:23 +0000, Patrick Turner wrote:
Nowhere can I find an answer to my simple question, " Why does the ESL speaker sensitivity rise when the EHT membrane charge in increased ?" [...] Because an ESL is basically a huge capacitor, and in a capacitor Q=CV so its charge Q increases proportionaly to the tension V. The force F between the plates of the capacitor = kq1q2/r² with k a constant, r the distance between the plates, q1 and q2 the respective charges of each plate. When you increase the tension, you increase the charge and thus the force and thus the sensitivity. http://en.wikipedia.org/wiki/Capacitor http://en.wikipedia.org/wiki/Coulomb%27s_law |
|
#3
Posted to rec.audio.tubes
|
|||
|
|||
|
Electrostatic speaker ( ESL ) sensitivity.
didier gaumet wrote: On Fri, 20 Apr 2007 06:24:23 +0000, Patrick Turner wrote: Nowhere can I find an answer to my simple question, " Why does the ESL speaker sensitivity rise when the EHT membrane charge in increased ?" [...] Because an ESL is basically a huge capacitor, and in a capacitor Q=CV so its charge Q increases proportionaly to the tension V. The force F between the plates of the capacitor = kq1q2/r² with k a constant, r the distance between the plates, q1 and q2 the respective charges of each plate. When you increase the tension, you increase the charge and thus the force and thus the sensitivity. I am sill in doubt. Ppl have quoted the Coulomb Law at me but the only non linear element is the bottom line quantity of distance squared. If the distance does not substantially change between two charged elements, or plates of a capacitor, then the d squared does not change, and does not explain the doubling of sensitivity for a doubling of applied EHT. The electrostatic force equation is kc x /q1/ /q2/ F = --------------- d x d F is the electrostatic force, q1 is the charge of one plate, q2 the charge of the other, kc is a constant, d is the distance between the plates. kc is a constant for all equations to work. x means 'multiplied by' What does the encloser of the q1 and q2 mean with a / line drawn before and after? Are the charges in volts multiplied by each other? If so where you have -1,000 x -2000, the product is 2,000,000. If there was -2,000 x -3,000, same voltage difference, but higher voltages, the product is 6,000,000. So the doubling of EHT results in 3 times the force. This does not seem correct, so can anyone explain exactly what the equation means and how to interpret it and please could you be so kind to give a worked example to proove you are correct. Patrick Turner. http://en.wikipedia.org/wiki/Capacitor http://en.wikipedia.org/wiki/Coulomb%27s_law |
|
#4
Posted to rec.audio.tubes
|
|||
|
|||
|
Electrostatic speaker ( ESL ) sensitivity.
didier gaumet wrote: On Fri, 20 Apr 2007 06:24:23 +0000, Patrick Turner wrote: Nowhere can I find an answer to my simple question, " Why does the ESL speaker sensitivity rise when the EHT membrane charge in increased ?" [...] Because an ESL is basically a huge capacitor, and in a capacitor Q=CV so its charge Q increases proportionaly to the tension V. The force F between the plates of the capacitor = kq1q2/r² with k a constant, r the distance between the plates, q1 and q2 the respective charges of each plate. When you increase the tension, you increase the charge and thus the force and thus the sensitivity. http://en.wikipedia.org/wiki/Capacitor http://en.wikipedia.org/wiki/Coulomb%27s_law Carying on from my previous answer, and after a search I found http://teacher.pas.rochester.edu/PHY...lombs_Law.html A lab experiment was to be done to ........ """"To verify the proportionality of Coulomb's Law that the electric force between two point charges is directly proportional to the product of the charges and is inversely proportional to the square of the distance between them.""""" No example of a simple capacitor was given. If the charge is the result of a constant difference between two charges, or two applied voltages, then would not the q1 x q2 product be constant?? Patrick Turner. |
|
#5
Posted to rec.audio.tubes
|
|||
|
|||
|
Electrostatic speaker ( ESL ) sensitivity.
On Fri, 20 Apr 2007 23:50:33 GMT, Patrick Turner
wrote: I am still in doubt. Ppl have quoted the Coulomb Law at me but the only non linear element is the bottom line quantity of distance squared. And this description doesn't even really apply to ESL's, which are (in modern, post-Walker, times) symmetrically (differentially) driven with a constant charge diaphragm. I've got a paper by Baxandall around here somewhere that describes it in the depth that I can't. When I find it I'll post back, hopefully with an accessable reference. Otherwise I'll mail ya a Xerox'd copy. Much thanks, as always, Chris Hornbeck "For one swallow does not make a summer, nor does one day; And so too one day, or a short time, Or a great deed, does not make a man Blessed or happy." - Aristotle, _Nicomachaen Ethics_ |
|
#6
Posted to rec.audio.tubes
|
|||
|
|||
|
Electrostatic speaker ( ESL ) sensitivity.
I think you are confusing charge (Q in Coulombs) and voltage (V in Volts):
in a capacitor Q=CV so Q1=CxV1 and Q2=CxV2 and F = kc x C²x V1 x V2 / d² so the force F increases when: - d decreases or - V1 or V2 or both V1 and V2 increase. or both And a push-pull ESL is basically two huge capacitors in serie, that's why its capacity is half the one of a simple ESL. Hope this helps :-) |
|
#7
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coating andtension, etc.
didier gaumet wrote: I think you are confusing charge (Q in Coulombs) and voltage (V in Volts): in a capacitor Q=CV so Q1=CxV1 and Q2=CxV2 and F = kc x C²x V1 x V2 / d² so the force F increases when: - d decreases or - V1 or V2 or both V1 and V2 increase. or both And a push-pull ESL is basically two huge capacitors in serie, that's why its capacity is half the one of a simple ESL. Hope this helps :-) I don't follow your math above, but let me say I think I have it all sussed. The force between one stator and the membrane = k x q1 x q2 / d x d. If q1 is the field due to xvolts in stator and membrane, and the q2 is between membrane and stator, the two are multiplied in the figure for total force due to the two field effects. Take the ERA ESL, membrane is at -2.5kV, idle voltage at each stator = 0V, so the q1 = q2 = 2.5kV so the force is prorotional to 2.5 x 2.5 = 6.25 and as these forces act each side equally in opposite directions with zero signal, the membrane remains central. Consider the distance changes to be so small they are negligible, so d squared does not have to be included in the equation, certainly not at 1 kHz. Suppose we apply +1,000V peak to one stator, and -1,000V to the other. On one side beatween membrane and stator we have q1 x q2 product = 1.5 x 1.5 = 2.25, so F = 2.25 units and between the other stator and membrane we have 3.5 x 3.5 = 12.25 units. The difference in the forces actiong in opposite directions is the resultant on the membrane = 12.25 - 2.25 = 10. When the EHT is increased to 3.5kV, the forces remain equal but are larger on each side of the membrane, so it remains central. But with -3.5kV, apply the same +/-1,000V peak ac signal to each stator. We get F on one side = 2.5 x 2.5 = 6.25u, and on the other 4.5 x 4.5 = 20.25u. Resultant = 20.25 - 6.25 = 14.0u. So the force acting on the membrane increases from 10u to 14u with the increase of EHT from 2.5kV to 3.5kV. This amount of EHT voltage increase applied to the membrane raises the force level by +3dB and audio power power is doubled, and I measured exactly 3 dB rise in the recovered microphone signal level after raising EHT. So the more EHT one can apply, the higher SPL level one can get for a given input ac signal. Or the higher the EHT, the lower the ac signal needs to be for the same SPL. Ie, the sensitivity rises with higher EHT. The peak ac signal must not exceed the EHT voltage lest there be disotrtions. I think this is how it works; the q1 and q2 charges are a product, not a sum. ****Have I got it all wrong???**** In ERA bass panels, the bass membrane to stator distance = about 2.4mm. In QuadESL57, its 4mm. For similar electro static forces to exist in the two speakers, if ther Quad panels have EHT = 6,000V, then the ERA should have about 3.6kV. The ESL57 have a 1:290 step up transformer so 28Vrms input becomes 8,120Vrms max at the secondary, or +/-5,724Vpeak at each end of the sec, which is just less than the EHT voltage. In ERA, the tranny ratio is on;y 1:90, so for 28Vrms, we get 2,520Vrms across the sec, so +/- 1,776 peak V, so EHT needs to be at least 2,000V, but to get the same ESL57 membrane forces from such a small voltage, the EHT needs to be higher, and and it helps to have the stators closer to the membrane, so F is larger. However, as I have so painfully pointed out to all and to ERA if they are reading, using closer spacing and lower EHT and signal voltages than ESL57 mean that sensitivity is chucked out the window, and likehood of discharge from membranes to stators is far worse than in ESL57. All I am trying to do is make the ERA panels reliable, and acceptable, and perhaps they will give some long service without me having to constantly fix them. I will also try the higher tension of 2kG/60mm along each edge on the spring balance supplied by ERA instead of their recomendation of only 0.8kG. The '57 bass membrane travel maximum = 2mm, because the stator is 2mm thick, and plastic, with its coating on the outside 4mm away from the membrane, so the membrane can only travel 2mm before being arrested by the plastic. In the ERA, nothing prevents the membrane travelling the 2.4mm beteen stator and membrane except the paint on the metal stator, and when the membrane DOES TRAVEL the large distance, the force of attraction begins to rapidly increae because F becomes significantly affected by the large distance shortening. As d reduces, F increases at a rate = 1 / d squared, and if the force is larger than the tension in the membrane, wham, the membrane becomes stuck to the stator. If the EHT is regulated, and the coating resistance in only hundreds of megohms instead of thousands of megohms, any drop in EHT voltage due to discharge from membrane to stator is replenished, and the this helps stick the membrane to a stator. I suspect many owners of ERA ESL speakers would have this problem and not even be aware of it!!! The tension in the membrane should theoretically be able to pull the membrane away from the stator, but it sure cannot do this at EHT = -3.5kV. One would suspect that once stuck to one stator, that when an ac voltage reduces the electrostatic force during a signal wave, and with the pull of the tension, the membrane would simply spring free again, but no, we are not that damn lucky. Once stuck, the membrane want to stay stuck and the panel PS must be turned off and the membrane allowed to slowly pull free, or be prodded free. Quad ESL used a liquid nylon solution applied each side to a membrane, and it takes a long time for the charge to flow into the membrane from PS, and out when turned off, and resistance of the coating is thousands of megohms. So if the '57 membrane travels a long distance and an arc is formed, the current will be tiny due to spark discharges. But nevertheless arcs can become larger with time, and burn large holes in the speakers. My Quad repairer mentor in Mebourne suggests that the problems I am having with ERA panels is because the membrane tension specified by ERA is way too low, and the high resistance membrane coating does not have a high enough resistance. There is no way the ERA panels can be easily made with conductive coating to each side of the membrane. The stack of frames and stator parts do not allow the double sided approach, and to use the Quad system the panels would need to be of different physical construction. ERA say there are benefits in having metal stators, and no barrier to prevent membrane movement such as the 2mm plastic material used in ESL57. I have to disagree with ERA. To prevent the membrane travelling too far, and in fact not more than the 2mm of travel that occurs in the ESL57, some polyester or nylon cord or string about 0.5mm thick be threaded horizontally around the slots in the stators to form ridges at 20mm centres vertically. The membrane would then hit the string barriers, and then have to bend very sharply to fully suck against the stator like shrink wrap around a cabbage. I hope the higher membrane tension will overcome the electrostatic forces when it tries to stick to a stator. The fact that stiction occurs so easily at low EHT and signal voltages means there is a serious limitation to the speaker's sensitivity. As I see it, I am compelled to build what ends up being a hobbled speaker, unable to get beyond a canter, and never gallop. I was at my customers houe this afternoon to listen to a few recordings from his rather fabulous turn table on which he has fitted a new tone arm. I doubt we ever used more than 3 peak volts from the amps into their 5 ohms, and I had to say perhaps the ERA ESL might perform as well, but the amp will have to provide a very very much higher voltage output. The ESL load isn't all that difficult to drive. But estimating just what the average Z is is a bit of a mytsery. I plotted the ERA ESL-IIIB Z with its input filters and required crossover resistors and found it was :- 10hz, 15.4 ohms, 50hz, 13.3, 100Hz, 15.8, 240Hz, 32.0, 500Hz, 18.0, 1kHz, 13.2, 3kHz, 10.0, 10kHz, 5.0, 20kHz, 2.5, 50kHz, 1.6. I have yet to draw up an electronic copy of the exact schematic for ppl to try for what i think is the optimum result. If we divide up the audio band from 20ha to 20kHz into 6 bands, and eyeball the max/min Z for each band, we get 6 resulting average = 14, 24, 22, 13, 9, 5. Adding them up and dividing by 6 gives 14.5 ohms. Its not quite this simple, as music doesn't have high voltages applied to above 1 khz compared to below 1 khz, so in fact the total Z coulbe be above 14.5 ohms, but this figure will do for very rough calculations of amplifier requirement. I don't think more than 28Vrms could ever be safely applied to EAR speakers, so that means 54 watts. A 100 watt amp rated for 100W into 8 ohms would perform OK, because most well built ones will handle the lower signal voltages needed for the lower Z at HF. So if the 100W amp can make 10 watts into 3 ohms at F above 3kHz, it will be able to drive the speakers if they are actually able to take 28Vrms maximums without stiction or arcing or displaying any other problems. A 50 watt amp would be fine if it is rated to work into 15 ohms. Most amps are not specified to be able to do so, and have say 100W rating for 8 ohms, and say 180 watts into 4 ohms. Usually the same amp can make 56 watts into 15 ohms, and can be seen as 56 watt amps if that's the load we are using. I wish to use about 27 watts max into the 14.5 ohm load from an SEUL amp with 13E1 tube. This will give about 19Vrms amx, and SPL would be 3 dB lower that what the "100 watt" 56 watter amp could provide. So the procedures to try to overcome difficulties are :- Remove membrane no4 and place in the bin, and sand off the spacers clean for a no5 membrane. Search for a suitable string, perhaps builder's line, and tightly thread through the stators to above description and dope lightly with polyurethane varnish to stop any movement. stretch out another membrane but with 2kG used instaed of 0.8kG. Glue down the stator to the membrane on the table with polyurethane glue, lay a damp towel over the whole panel, and weigh down with 20 solid house bricks of 4kG each. wait a day for the glue to cure, then carefully cut the membrane around the panel to remove it from the bench, and observe carefully for any defects. Do it all again if there are defects. If not, apply a very thinly spread out conductive coating to barely wet the surface, and drag the sponge off alnog the length. Next day, do the same, but lay off across the panel, use the coating material as specified and diluted with distilled water and applied very thinly. When applied by a damp rather than soaking wet sponge, the coating liquid will not tend to dribble over the edges of panel memrane. Don't let the conductive coating spread over outside edges of the panel. Wait another day, and make sure your cat does not decide to take a nap on top of your work! The panel will be delicate, and prone to being torn if touched with anything sharp. So keep all things sharp well away in case you drop something on the panel. Clean out the bolt holes for plastic bolts; they will have glue in them. I carefully drill them out, rather than use a hot soldering iron. Be careful, because one prick in a panel and its ****ed. Then place the 12 bolts to hold panel halves together. Tighten all mildly, when all are in the panels. I already have all the wires connected to stators and EHT supply left fitted while re-doing a panel. Place the PVC channels and spacers around the four sides tightly, and you have the panel ready for testing. I test mine sitting on a chair and temporarilly taped to the chair back rest. Don't let the panel fall over. Connect wiring to the crossover board I made which ERA does not supply. Connect PSU and transformer temporarily sitting on the chair, and doble check all connections. Adjust EHT regulator if there is one to the lowest setting. Set up all response testing gear, amplifier, preamp, pink nise source volt meters, and reference speakers. turn on the PS and adjust EHT to 2,500V. Check that no discharge noises occur. Have a cup of tea, or wait 1/2 an hour for the EHT charge to soak in. Spend a few hours plotting at least 6 response measurements with signals not more than 80 dB SPL, and as the SPL meter indicates, at various distances in the room. the mic should be between 3M to 4M away and on axis and 0.9M to 1.2M off the floor, to simulate being seated in a listening chair. In my case I have a 50 watt amp capable of 21Vrms maximum output voltage, and when music signals reach this level, I will want the speakers to handle it OK without stiction or any obvious over load problems. I try to use some music I am familiar with to have my reference speakers going at a normal listening level, and then take an average SPL reading, this will be at about 1/4 of a watt. Then pink noise is used to give the same meter reading for SPL as with music. The reference speaker voltage is recorded, and power calculated. Then the ESL is set up with pink noise, again after checking for stiction by close inspection, and the pink noise run up generate the same SPL as the reference speakers. The ESL speaker voltage is measured, and power calculated. If the ESL power is say 4 times more than the reference speakers which are known to make 88dB/W/M then the ESL are 6dB less efficient, or will give 82dB/W/M Try the ESL with music at normal music levels and check the speaker signal voltage and have CRO connected to ensure the level is well clear of clipping. If the civilised music levels can be produced in ESL at marginally higher power levels than need for 88dB speakers and a 30 watt amp, and without audible amp distortions, all is well. I always operate in doubt, until I proove to myself all is really OK if I can make it right. These ESL speakers may perform better with some acoustically absorbant material place behind them I have some 50mm thick heavy polyester wool bats which should do the trick. Patrick Turner. |
|
#8
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingand tension, etc.
Patrick Turner wrote:
My Quad repairer mentor in Mebourne suggests that the problems I am having with ERA panels is because the membrane tension specified by ERA is way too low, and the high resistance membrane coating does not have a high enough resistance. Keep at it Patrick, and please dont be put off by the doof doof disco speaker crowd. Once you have a good condition, properly set up pair of electrostatics they will give you pure joy. I agree with your mentor this is the area you should be focusing on, something is definitely not right if you have to employ thick blobs or string to prevent stiction. Have you tried the hairdryer heatshrinking method? (details below) I know it sounds a bit hit and miss but IMO the mylar isn't that expensive that an experiment or two is going to break the bank. But how tight is too tight? I haven't read anything stating too tight is not good...but it seems plenty of problems occur with flapping membranes. I guess the optimum would be to get as much movement as possible in the diaphragm without stiction occuring, and if movement is restricted would it be possible to up the power to compensate? I would also look at very closely at your coating properties and measurements, have you tried Calaton CB or Elvermide? There's also a DIY nylon coating here made from chopped up fishing line.: http://www.quadesl.org/Hard_Core/Pan...alcoating.html Once again all this applies to ESL57 but the basic principles are worth following Diaphragm Replacement [Top] OK, you've got this far, and you're feeling pretty pleased with yourself. Let's see what we can do about that, then!! By this stage you should have two clean, repaired stators, with no tape breaks, corrosion, or conductive layer discontinuities. 1. Cut a piece of 12 mm Mylar which is about 10 cm larger than the bass panel stator in each dimension. 2. Clean a smooth laminex or glass topped table with ordinary household detergent. Dry well, and then clean with either isopropyl alcohol or acetone. N.B. Both of these materials are flammable. Acetone in particular!! Do not allow anyone to smoke or bring a naked flame into the same HOUSE with acetone. Both materials are relatively innocuous, otherwise. Your objective here is to remove anything of a lumpy, or gritty nature from the work area. 3. Lay out the diaphragm film smoothly on the table top, and fix the corners with masking tape, stretching the film slightly as you do this. Then attach masking tape to the centres of each side, stretching the Mylar ® again, in each direction. Continue taping each side by "splitting the difference" between previous tape points, until you have the Mylar® reasonably taught. Don't try to get it very tight at this point, but do tape it firmly all around. 4. Using the inner stator dimensions as a guide, mark, or mask off, a rectangle to be coated. 5. If using graphite (not recommended), rub the powdered graphite into the Mylar® as hard as you can. Try to really grind it in, and get a uniform coating. Test the surface for uniform conductivity with a surface resistivity meter or DVM that can read to at least 100 MW . This will produce a diaphragm which is far too low in resistance to be really useful in a Quad Electrostatic. Don't panic!! N.B. Soluble NYLON is the best coating to use, and it just wipes on in alcohol solution. 6. Try to rub off all the graphite, using a paper towel soaked in isopropyl alcohol. A graphite glaze will be produced, and the colour of the diaphragm will appear a very light gray. Do try to rub off all the graphite. IF you ground it into the film well enough, it won't come off completely. 7. Check the resistance with your meter. If it is not at least 100 MW then try rubbing the graphite off a little more. You also need to avoid making large discontinuities in the membrane coating. This is why this method is such a pain, even if the process is, practically speaking, very simple. When the membrane measures as you would like it, clean the general surroundings, but leave the Mylar® taped to the table top. 8. If you want to avoid this rubbing and grunting and carrying on, just wipe on some DIY soluble nylon, or CALATON CB, or ELVAMIDE any of which duplicate the original Quad diaphragm coating, and wait for it to dry off. Diaphragm coated - no effort. 9. Mix up some two part epoxy resin glue - the PLAIN kind - no metal fillers!! 10. Choose a stator to glue the membrane to. 11. Apply a thin (1/8th inch) bead around the perimeter where the old brown glue was. Do not put epoxy on the metal tape. The bead should be run between the tape and the outer edge of the stator. 12. Place the stator, glue side down(!) over the Mylar® film being very, very careful to align the inner rectangle of the stator with the coated area. 13. Press down with both thumbs all around the perimeter of the stator where the bead of glue was run. The idea is to squeeze the glue into a very thin film. N.B. Placing weights on the stator at this time, alone, will not do a good enough job. 14. Place a layer of books, a sheet of steel, or something solid, over the stator to spread the pressure, and weight the whole thing with bricks. Wait until the glue is well cured. I leave the stator overnight, usually, no matter which epoxy I use. 15. When the glue is well cured, remove all the weights and other paraphernalia, and lift up the stator with diaphragm attached. Trim the edges very carefully with a very sharp knife, (e.g. Stanley Knife). 16. Heat Shrink the diaphragm with a heat gun set at about 400 Watts, working about 20 to 30 cm from the diaphragm. There are so many variables involved in this, that it is impossible to give precise instructions. If you have not done this before, then you must practice with a spare piece of Mylar®. Tape a 20cm x 20cm piece of Mylar® film to a table top and practice heat shrinking until you can shrink the film tightly (no little creases) without melting a hole in the film! If you melt a hole in the diaphragm, then you can start again. 17. Coat the exposed side of the diaphragm with DIY soluble Nylon if you are using the original coating. This is what Quad did both sides at the factory, and what you should do if you want the speaker restored to its original condition. See the FAQs for an explanation of why this is so. 18. Melt holes in the diaphragm for each stator hole (60 holes) with a fine-pointed tip on a small wattage soldering iron. 19. Double check that the heat shrink job is OK. If not, carefully heat shrink again. Leave overnight and repeat the heat shrinking if not sure. The diaphragm should be perfectly smooth, although this is not as critical as it is for treble panels. 20. Using 12mm x 3mm (M3) bolts, hex nuts, and two washers (one each side), bolt the stators together. N.B. Do not forget to run the front stator high voltage wire back through the holes in each stator and re-solder it to the tab on the front stator. Tighten each nut and bolt firmly, but not too tight. 21. Run a line of PVC tape around the perimeter of the panel, covering the outermost bolts. This is added insulation to prevent leakage to the frame. Everything conducts (it seems some times) at 6 000 Volts. 22. Take the refurbished Dust Covers and identify the rear cover. This will need holes drilled in it to accommodate the hex nuts on the rear stator. Otherwise the panel, overall, will be thicker than the original and will not fit into the frame without other unpleasant maneuvering. 23. Re-fit the small phenolic connection board to the bottom of the rear dust cover - it bolts to the bottom of the frame with two small (12mm x 2mm) bolts. You may have to re-drill one mounting hole to do this if the new framing arrangements overlap the original holes. 24. Melt holes in the dust cover film to match the holes in the small board, and run the HT and signal wires through the appropriate holes. LEFT - REAR Stator (white). MIDDLE - EHT Connection (red). RIGHT - FRONT Stator (black). N.B. The older speakers have all white wires, so watch what you're doing! 25. Tape the dust covers in a “sandwich” with the rebuilt panel between them using 2" PVC tape. Make sure there are no gaps, or there will be panel leakage somewhere. 26. The panel is now ready to be refitted to the frame and the appropriate solder connections to the EHT and Audio transformer can be made now or later. regards Sean |
|
#9
Posted to rec.audio.tubes
|
|||
|
|||
|
Electrostatic speaker ( ESL ) sensitivity.
On Sat, 21 Apr 2007 03:12:14 GMT, Chris Hornbeck
wrote: I've got a paper by Baxandall around here somewhere that describes it in the depth that I can't. When I find it I'll post back, hopefully with an accessable reference. Otherwise I'll mail ya a Xerox'd copy. Ah, Ok. It's from an older edition of John Borwick's compendium _Loudspeaker Handbook_ : http://www.amazon.com/Loudspeaker-He...7210119&sr=1-7 If your library can't get a copy by inter-library loan, let me know and I'll Xerox and snail-mail ya one. It's great stuff and surprising. Much thanks, as always, Chris Hornbeck "Second star to the right, Then straight on 'til morning." |
|
#10
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingandtension, etc.
Sean wrote: Patrick Turner wrote: My Quad repairer mentor in Mebourne suggests that the problems I am having with ERA panels is because the membrane tension specified by ERA is way too low, and the high resistance membrane coating does not have a high enough resistance. Keep at it Patrick, and please dont be put off by the doof doof disco speaker crowd. Once you have a good condition, properly set up pair of electrostatics they will give you pure joy. I agree with your mentor this is the area you should be focusing on, something is definitely not right if you have to employ thick blobs or string to prevent stiction. Have you tried the hairdryer heatshrinking method? (details below) I know it sounds a bit hit and miss but IMO the mylar isn't that expensive that an experiment or two is going to break the bank. But how tight is too tight? I haven't read anything stating too tight is not good...but it seems plenty of problems occur with flapping membranes. I guess the optimum would be to get as much movement as possible in the diaphragm without stiction occuring, and if movement is restricted would it be possible to up the power to compensate? I would also look at very closely at your coating properties and measurements, have you tried Calaton CB or Elvermide? There's also a DIY nylon coating here made from chopped up fishing line.: http://www.quadesl.org/Hard_Core/Pan...alcoating.html The coating resistance seems very high, and beyond my measurement gear. I can place two coins on the surface only 1mm apart, and still get an OL reading in the meter which can measure about up to 50M. The physical arrangement of the plastic frames of the kits and spacers prevent coating both sides of the membrane, and I doubt its necessary. Once again all this applies to ESL57 but the basic principles are worth following Indeed. I am studying calculations for tension and membrane forces from the applied EHT and signals to see what limiting effects ocur at LF where "flapping" occurs, or rather oscillations or excess membrane movements about between 20hz and 70 Hz. I won't be using heat to tension the panel. Its too hit and miss. I've read all what you have below, but its dificult to implement exactly in the same way with the ERA panels I am building. Regards, Patrick Turner. Diaphragm Replacement [Top] OK, you've got this far, and you're feeling pretty pleased with yourself. Let's see what we can do about that, then!! By this stage you should have two clean, repaired stators, with no tape breaks, corrosion, or conductive layer discontinuities. 1. Cut a piece of 12 mm Mylar which is about 10 cm larger than the bass panel stator in each dimension. 2. Clean a smooth laminex or glass topped table with ordinary household detergent. Dry well, and then clean with either isopropyl alcohol or acetone. N.B. Both of these materials are flammable. Acetone in particular!! Do not allow anyone to smoke or bring a naked flame into the same HOUSE with acetone. Both materials are relatively innocuous, otherwise. Your objective here is to remove anything of a lumpy, or gritty nature from the work area. 3. Lay out the diaphragm film smoothly on the table top, and fix the corners with masking tape, stretching the film slightly as you do this. Then attach masking tape to the centres of each side, stretching the Mylar ® again, in each direction. Continue taping each side by "splitting the difference" between previous tape points, until you have the Mylar® reasonably taught. Don't try to get it very tight at this point, but do tape it firmly all around. 4. Using the inner stator dimensions as a guide, mark, or mask off, a rectangle to be coated. 5. If using graphite (not recommended), rub the powdered graphite into the Mylar® as hard as you can. Try to really grind it in, and get a uniform coating. Test the surface for uniform conductivity with a surface resistivity meter or DVM that can read to at least 100 MW . This will produce a diaphragm which is far too low in resistance to be really useful in a Quad Electrostatic. Don't panic!! N.B. Soluble NYLON is the best coating to use, and it just wipes on in alcohol solution. 6. Try to rub off all the graphite, using a paper towel soaked in isopropyl alcohol. A graphite glaze will be produced, and the colour of the diaphragm will appear a very light gray. Do try to rub off all the graphite. IF you ground it into the film well enough, it won't come off completely. 7. Check the resistance with your meter. If it is not at least 100 MW then try rubbing the graphite off a little more. You also need to avoid making large discontinuities in the membrane coating. This is why this method is such a pain, even if the process is, practically speaking, very simple. When the membrane measures as you would like it, clean the general surroundings, but leave the Mylar® taped to the table top. 8. If you want to avoid this rubbing and grunting and carrying on, just wipe on some DIY soluble nylon, or CALATON CB, or ELVAMIDE any of which duplicate the original Quad diaphragm coating, and wait for it to dry off. Diaphragm coated - no effort. 9. Mix up some two part epoxy resin glue - the PLAIN kind - no metal fillers!! 10. Choose a stator to glue the membrane to. 11. Apply a thin (1/8th inch) bead around the perimeter where the old brown glue was. Do not put epoxy on the metal tape. The bead should be run between the tape and the outer edge of the stator. 12. Place the stator, glue side down(!) over the Mylar® film being very, very careful to align the inner rectangle of the stator with the coated area. 13. Press down with both thumbs all around the perimeter of the stator where the bead of glue was run. The idea is to squeeze the glue into a very thin film. N.B. Placing weights on the stator at this time, alone, will not do a good enough job. 14. Place a layer of books, a sheet of steel, or something solid, over the stator to spread the pressure, and weight the whole thing with bricks. Wait until the glue is well cured. I leave the stator overnight, usually, no matter which epoxy I use. 15. When the glue is well cured, remove all the weights and other paraphernalia, and lift up the stator with diaphragm attached. Trim the edges very carefully with a very sharp knife, (e.g. Stanley Knife). 16. Heat Shrink the diaphragm with a heat gun set at about 400 Watts, working about 20 to 30 cm from the diaphragm. There are so many variables involved in this, that it is impossible to give precise instructions. If you have not done this before, then you must practice with a spare piece of Mylar®. Tape a 20cm x 20cm piece of Mylar® film to a table top and practice heat shrinking until you can shrink the film tightly (no little creases) without melting a hole in the film! If you melt a hole in the diaphragm, then you can start again. 17. Coat the exposed side of the diaphragm with DIY soluble Nylon if you are using the original coating. This is what Quad did both sides at the factory, and what you should do if you want the speaker restored to its original condition. See the FAQs for an explanation of why this is so. 18. Melt holes in the diaphragm for each stator hole (60 holes) with a fine-pointed tip on a small wattage soldering iron. 19. Double check that the heat shrink job is OK. If not, carefully heat shrink again. Leave overnight and repeat the heat shrinking if not sure. The diaphragm should be perfectly smooth, although this is not as critical as it is for treble panels. 20. Using 12mm x 3mm (M3) bolts, hex nuts, and two washers (one each side), bolt the stators together. N.B. Do not forget to run the front stator high voltage wire back through the holes in each stator and re-solder it to the tab on the front stator. Tighten each nut and bolt firmly, but not too tight. 21. Run a line of PVC tape around the perimeter of the panel, covering the outermost bolts. This is added insulation to prevent leakage to the frame. Everything conducts (it seems some times) at 6 000 Volts. 22. Take the refurbished Dust Covers and identify the rear cover. This will need holes drilled in it to accommodate the hex nuts on the rear stator. Otherwise the panel, overall, will be thicker than the original and will not fit into the frame without other unpleasant maneuvering. 23. Re-fit the small phenolic connection board to the bottom of the rear dust cover - it bolts to the bottom of the frame with two small (12mm x 2mm) bolts. You may have to re-drill one mounting hole to do this if the new framing arrangements overlap the original holes. 24. Melt holes in the dust cover film to match the holes in the small board, and run the HT and signal wires through the appropriate holes. LEFT - REAR Stator (white). MIDDLE - EHT Connection (red). RIGHT - FRONT Stator (black). N.B. The older speakers have all white wires, so watch what you're doing! 25. Tape the dust covers in a “sandwich” with the rebuilt panel between them using 2" PVC tape. Make sure there are no gaps, or there will be panel leakage somewhere. 26. The panel is now ready to be refitted to the frame and the appropriate solder connections to the EHT and Audio transformer can be made now or later. regards Sean |
|
#11
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingand tension, etc.
On Sun, 22 Apr 2007 03:20:10 GMT, Patrick Turner
wrote: The coating resistance seems very high, and beyond my measurement gear. I can place two coins on the surface only 1mm apart, and still get an OL reading in the meter which can measure about up to 50M. As I've posted before, this is *very* different than any current practice. have you read the various _Speaker Builder_ articles, especially by the Danish builders? They might be useful. I won't be using heat to tension the panel. Its too hit and miss. Again, it's a common (not to say standard) practice. At this long distance it seems like you're trying to both build a flawed kit and engineer a (technically complex) loudspeaker. These two goals are fundamentally in conflict. My very strong recommendation is to FIRST read Baxandall's discussion of the Walker ESL's, then to set fire to the kit. But that's just me. All the best fortune, as always, Chris Hornbeck "Second star to the right, Then straight on 'til morning." |
|
#12
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingandtension, etc.
Patrick Turner wrote:
The physical arrangement of the plastic frames of the kits and spacers prevent coating both sides of the membrane, and I doubt its necessary. I wouldn't entirely dismiss this Patrick, even though as you say it might be impossible with the (flawed?) ER kit spacings....Quad did it for very good reason. Its all about "Nett Charge separation" and something called Tribo-electric theory it could well be the cause of all your problems. quoting from the website: http://www.quadesl.org/Hard_Core/Pan...#IMPORTANTNEWS **** "If a Quad membrane was coated with Nylon on one side only, then one side of the diaphragm would tend to be in slight negative charge deficit with respect to the other side."**** The Theory Firstly, unlearn, as I had to, the idea that static electricity means "stationary electricity". Think in terms of "nett charge separation", since this is actually the phenomena and also what we require. It is also nice if the charge does not move about on the membrane creating spurious capacitance variation and other bad happenings. F.V. Hunt's original mathematics shows that if capacitance variation is kept to 8% then distortions of all types are 0.5% at audio frequencies. Impressive, in a speaker. The next basic lesson to learn (or re-learn) is about the Triboelectric Series, (see below). + NYLON ---- Acquires More Positive Charge - MYLAR (P.E.T. in treble panels) -- Acquires More Negative Charge - SARAN (P.V.C. in bass panels) -- Acquires More Negative Charge We can see that Nylon is near the top of the series, and both Mylar and Saran are much further down. In effect, this means that electrons tend to leave the Nylon and the Mylar tends to accept them. This produces a definite negativity in the Nylon. A close contact between Nylon and Mylar, as in a coating, guarantees that the electrons will stay very, very close to the Mylar "side" of the pair. Quad coated the diaphragm on BOTH sides. Perhaps now, we can see why. The Nylon surface tends to be positive and the Mylar tends to be negative. There is NO bulk current flow to speak of because of the generally insulative nature of each substance. It is a "nett separation of charge ". The migration of charge through the bulk of the membrane is slow by any normal electrical standard. If a Quad membrane was coated with Nylon on one side only, then one side of the diaphragm would tend to be in slight negative charge deficit with respect to the other side. Quad then provided a small trickle charging source at 1500 and 6000 Volts for the respective diaphragms to provide the "make up" of charge carriers in the system. I am sure that this is more obvious to us now, in an age of semi-conductors. In 1954-55, it was very clever indeed. Regards Sean |
|
#13
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingandtension, etc.
Chris Hornbeck wrote: On Sun, 22 Apr 2007 03:20:10 GMT, Patrick Turner wrote: The coating resistance seems very high, and beyond my measurement gear. I can place two coins on the surface only 1mm apart, and still get an OL reading in the meter which can measure about up to 50M. As I've posted before, this is *very* different than any current practice. have you read the various _Speaker Builder_ articles, especially by the Danish builders? They might be useful. I won't be using heat to tension the panel. Its too hit and miss. Again, it's a common (not to say standard) practice. At this long distance it seems like you're trying to both build a flawed kit and engineer a (technically complex) loudspeaker. These two goals are fundamentally in conflict. My very strong recommendation is to FIRST read Baxandall's discussion of the Walker ESL's, then to set fire to the kit. But that's just me. All the best fortune, as always, Well If I don't get a best fortune, a fraction of some lesser fortune may be managed. I have not come to any final solution or opinion, and I think ERA deserves a fair go. As everyone reading here should always realize, opinions here at r.a.t. or plans of action mentioned are valid as thought allows now and are subject to improvement via experience and gained wisdom. ERA would be the winner if they are reading because they have the option to improve their product. And I would win if I get the kit speakers I have to just work just a little bit better than they do. I have to study ahead about the amount of tension I should have in relation to the EHT and panel size, so I have to learn about the calculations involved. Its an electro mechanical problem and I have only been trained at the School Of Life. Let us not be too hard on anyone and think positive!!!! Chris Hornbeck "Second star to the right, Then straight on 'til morning." |
|
#14
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membranecoatingandtension, etc.
Sean wrote: Patrick Turner wrote: The physical arrangement of the plastic frames of the kits and spacers prevent coating both sides of the membrane, and I doubt its necessary. I wouldn't entirely dismiss this Patrick, even though as you say it might be impossible with the (flawed?) ER kit spacings....Quad did it for very good reason. Its all about "Nett Charge separation" and something called Tribo-electric theory it could well be the cause of all your problems. quoting from the website: http://www.quadesl.org/Hard_Core/Pan...#IMPORTANTNEWS I have read all of that and to change the way to make the ERA speakers with both sides of the membrane coated means there must be access to both sides during the build up process, so the sizes of frames and spacers have to all change. **** "If a Quad membrane was coated with Nylon on one side only, then one side of the diaphragm would tend to be in slight negative charge deficit with respect to the other side."**** Apparently the difference in charge of each side of the membrame coated on one side only is negligible, and the membrane says central in the presence of EHT up to about -2kV. More than this much EHT, the membrane tends to become stuck to stators. And the side that tends to stick to the stator is the uncoated side, which one might think was more positive than the -2kV side. The stiction occurs when once the membrane travels 2.4mm, the electrostatic force becomes huge do to the distance reducing. F is proportional to 1 / d squared. Unless the membrane tension creates enough force to overcome the electrostatic force, the membrane just sticks. Then when EHT is turned off, adhesive effects about which I know nothing occur, because the membrane does not want to spring free from the stator when no EHT is present, which suggests a charge is retained in the interface between membrane and stator. Its the uncoated side of the membrane which sticks to a stator. so perhaps when EHT drops at turn off, there is a charge in the mylar, but not in the coating. Its a weird effect. The Theory Firstly, unlearn, as I had to, the idea that static electricity means "stationary electricity". Think in terms of "nett charge separation", since this is actually the phenomena and also what we require. It is also nice if the charge does not move about on the membrane creating spurious capacitance variation and other bad happenings. F.V. Hunt's original mathematics shows that if capacitance variation is kept to 8% then distortions of all types are 0.5% at audio frequencies. Impressive, in a speaker. The next basic lesson to learn (or re-learn) is about the Triboelectric Series, (see below). + NYLON ---- Acquires More Positive Charge - MYLAR (P.E.T. in treble panels) -- Acquires More Negative Charge - SARAN (P.V.C. in bass panels) -- Acquires More Negative Charge We can see that Nylon is near the top of the series, and both Mylar and Saran are much further down. In effect, this means that electrons tend to leave the Nylon and the Mylar tends to accept them. This produces a definite negativity in the Nylon. A close contact between Nylon and Mylar, as in a coating, guarantees that the electrons will stay very, very close to the Mylar "side" of the pair. Quad coated the diaphragm on BOTH sides. Perhaps now, we can see why. The Nylon surface tends to be positive and the Mylar tends to be negative. There is NO bulk current flow to speak of because of the generally insulative nature of each substance. It is a "nett separation of charge ". The migration of charge through the bulk of the membrane is slow by any normal electrical standard. If a Quad membrane was coated with Nylon on one side only, then one side of the diaphragm would tend to be in slight negative charge deficit with respect to the other side. Quad then provided a small trickle charging source at 1500 and 6000 Volts for the respective diaphragms to provide the "make up" of charge carriers in the system. I am sure that this is more obvious to us now, in an age of semi-conductors. In 1954-55, it was very clever indeed. I have been sent all this to read before. Nothing explains the phenomena I am witnessing. I can only increase tension and place some physical motion restriction barier, ie, string "bars" 0.6mm in dia approx in rows across the bass panels at 20mm centres to restrict bass panel excursion. I should be about to have a 50 watt amp run up into clipping with music and the 30 peak volts generated sould not cause stiction, and not even if it is a 50Hz square wave. There is no evidence that ERA actually have carried out the kind of very reasonable tests I am attempting. So I have no clue to the potential of these kits, but I remain positive. To modify the existing ERA kits to incorporate the Quad ESL57 techniques would mean a complete redesign, and maybe 2 months farnarcling around, and nobody will pay me. I can't afford to do all the other folks R&D. Some, yes, but not all. If I did, I'd be making TURNER ESL. And I have months of other contracts to fulfil, all being delayed by the struggle with an ERA kit. I get this to work without stiction, or I give up, and I don't quit easily. Patrick Turner. Regards Sean |
|
#15
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membranecoatingandtension, etc.
Sean wrote: Patrick Turner wrote: The physical arrangement of the plastic frames of the kits and spacers prevent coating both sides of the membrane, and I doubt its necessary. I wouldn't entirely dismiss this Patrick, even though as you say it might be impossible with the (flawed?) ER kit spacings....Quad did it for very good reason. Its all about "Nett Charge separation" and something called Tribo-electric theory it could well be the cause of all your problems. It possibly could be, but a practical solution is the only one I am interested in. Stuff about membrane tension is formulated at http://www.quadesl.org/Hard_Core/ESL...esltheory.html However, nothing at this address talks me through the everyday means of calculating a typical panel tension as kG / cm of length along a panel side which would be required to overcome the electrostatic force present when the membrane is pulled over almost against the stator. The website display is only a very vague and poor attempt to teach anyone anything except that the creator of the website can create a website, and one that does not have a list of steps to take to get from no panel to a panel with correct tension. In contrast, I explain about 65 definite steps to make an OPT on my website, and I leave nothing out, and perhaps that's why my ISP says I had 118,000 hits in February alone. ( I replied to my ISP, Melbourne IT, saying they must be kidding me; how come not a single hit resulted in a sale? ) But the quadesl site write one confusing page and expect us to be wowed. Ronald Wagner's book is equally uninformative, telling only a fraction of the story of how to and why. The way I see it is that the membrane is loosely coupled to air. neverthe less air mass each side of a stretched membrane form a resonant structure, as is every drum, and when struck in the centre, the drum skin vibrates up and down at its fundmental F until it settles. If you have a conductive coating to the skin, and suitable means of applying varying voltages above and below the skin, you can make the skin of the drum easily vibrate at its resonant F. In the same way, a tuning fork can be used as a device in an oscillator. It is at resonance that the stretched membrane likes to swing back and forth with ease, and hence if this resonance is not suppressed, the motion can be very easily be made to make a membrane go all the way towards its nearby source of exciting energy, in this case the ESL stator plates. But once the membrane comes close to stator, the electrostatic force suddenly becomes very strong as the distance reduces to the thickness of a few coats of paint. Wham, the membrane just sticks within the time of a half cycle of frequency of resonance. So we must never let this huge electrostatic force ever develop under any conditions. Quad achieved this in the 57 with 2mm thick plastic stators, and minimum membrane to stator distance was not less than the maximum distance the membrane could possibly move. The membrane could pull tight against the inside of the stator, but was still well away from the low resistance conductive stator surface. The only way I see to keep the membrane from pulling tight to the stator is to use electrically inert string. The slots in the perforated steel stators are 20mm long, so one or two strings per slot can easily be placed. Let me say I make membrane tension = 2,000 grams per 60mm of panel side length, and since the panel width is 180mm, half that is 90mm from edge to centreline along the length. If you have a 1.8mm displacement of membrane, the force required to pull this = ( 1.8mm / 90mm ) x 2 x 2,000gms per 60mm. = 80 grams per 60mm of panel length. Once the membrane gets to the strings 20mm apart, each of which would be 0.6mm dia, then to get another 0.3mm would need a force = ( 0.3 / 10mm ) x 2 x 2,000gms per 60mm = 120 grams per 60mm of length. At 0.6mm travel the force would be 240 grams/60mm. At this point the membrane is against the stator paint, and I am not sure if the electrostatic force would be greater than 240grams / 60mm . Trouble is that the membrane could just touch the stator then "wrap" down onto the stator and stay stuck even if there is considerable force trying to pull it off, because of the strenth of electrostatic force. Then I may still have the weird adhesive effect which lasts even when the EHT is removed. I have found that when EHT is present, and the membrane is stuck, the signal voltages are not enough to overcome the electrostatic forces. If you have EHT = -2.5kV, and amplifier signal is at an absolute maximum 30peak volts, then +/- 1,350 peak volts is applied to the stators, and the mimimum stator to membrane voltage is still 2.5kV - 1.35kV = 1.15kV, and probably enough to maintain the high force between the membrane and stator once the distance has become less than 0.08mm, which is the present stator paint thickness. So to be on the safe side, I may need string at 10 mm centres, thus about doubling the forces needed to draw the membrane close to the stator. The amount of closure of holes in the stators with string will be about 7%, and since the stators could have 40% less open factor for clear rendition of bass frequencies, the string won't affect the sound if held tight with a layer of varnish. I have no idea if all this might work, but rather than spend 5 days labouring at the physics and maths containing calculus and reading all the suggestions in this thread that have not yielded any result from an experienced mind, it'd be much easier for me to spend one day simply removing the existing membrane, and wind in lengths of string, and to replace the membrane yet again, tensioned up to close to Quad63 values and hope for the best. I should then be able to say why the idea doesn't work and guage what I am really up against, and theorizing and calculations won't tell me that. I first learnt to wind OPTs without all the formulas for response and got fairly good results and winding geometry and testing told me what did and didn't work, and what gave rise to serious resonance problems. During the process I spent many night till dawn doing things repeatedly as i read more and found formula that actually worked when applied. Unless you try to reach the moon, you'll never know what a moon shot involves. I may contact my chemist who recommended the Isonel 642, The coating I have applied because of all the arcing/stiction problems I got even with EHT = -2.2kV, and ask him what would stop the mylar sticking to his product once an initial stiction pressure has been applied. And I'd want to know why the stiction can remain after EHT is removed. Maybe something additional needs to be painted onto the stators. Maybe silcone. Ever noticed how sticky tape just won't stick to anything coated in the slightest amount of silicone? I can make silcone paint; just thin down normal roof and gutter silicone with white spirit and mix well until fluid. I have used this solution to dope woven fabric used for foam replacement around dynamic speakers. It lasts, it is cheap, and is non toxic. No doubt the Quad processes are better than the ERA methods. To implement them as Quad did in a kit would make the kit much harder to build, and ERA would have a lot more troubled kit builders IMHO. Patrick Turner. quoting from the website: http://www.quadesl.org/Hard_Core/Pan...#IMPORTANTNEWS **** "If a Quad membrane was coated with Nylon on one side only, then one side of the diaphragm would tend to be in slight negative charge deficit with respect to the other side."**** The Theory Firstly, unlearn, as I had to, the idea that static electricity means "stationary electricity". Think in terms of "nett charge separation", since this is actually the phenomena and also what we require. It is also nice if the charge does not move about on the membrane creating spurious capacitance variation and other bad happenings. F.V. Hunt's original mathematics shows that if capacitance variation is kept to 8% then distortions of all types are 0.5% at audio frequencies. Impressive, in a speaker. The next basic lesson to learn (or re-learn) is about the Triboelectric Series, (see below). + NYLON ---- Acquires More Positive Charge - MYLAR (P.E.T. in treble panels) -- Acquires More Negative Charge - SARAN (P.V.C. in bass panels) -- Acquires More Negative Charge We can see that Nylon is near the top of the series, and both Mylar and Saran are much further down. In effect, this means that electrons tend to leave the Nylon and the Mylar tends to accept them. This produces a definite negativity in the Nylon. A close contact between Nylon and Mylar, as in a coating, guarantees that the electrons will stay very, very close to the Mylar "side" of the pair. Quad coated the diaphragm on BOTH sides. Perhaps now, we can see why. The Nylon surface tends to be positive and the Mylar tends to be negative. There is NO bulk current flow to speak of because of the generally insulative nature of each substance. It is a "nett separation of charge ". The migration of charge through the bulk of the membrane is slow by any normal electrical standard. If a Quad membrane was coated with Nylon on one side only, then one side of the diaphragm would tend to be in slight negative charge deficit with respect to the other side. Quad then provided a small trickle charging source at 1500 and 6000 Volts for the respective diaphragms to provide the "make up" of charge carriers in the system. I am sure that this is more obvious to us now, in an age of semi-conductors. In 1954-55, it was very clever indeed. Regards Sean |
|
#16
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingandtension,etc.
Patrick Turner wrote:
Stuff about membrane tension is formulated at http://www.quadesl.org/Hard_Core/ESL...esltheory.html However, nothing at this address talks me through the everyday means of calculating a typical panel tension as kG / cm of length along a panel side which would be required to overcome the electrostatic force present when the membrane is pulled over almost against the stator. The website display is only a very vague and poor attempt to teach anyone anything except that the creator of the website can create a website, and one that does not have a list of steps to take to get from no panel to a panel with correct tension. You are right, there's not a lot of information on perfect tensioning procedures...but maybe thats because its something thats not super critical? However, there is plenty of information extolling the virtues on coating both sides of the mylar diaphragms. Everything I've read on the subject points to coating both sides of the mylar, this results in high resistance on both sides while maintaining a constant charge....ie "incapable of changing" Only having one side coated seems a bit hit and miss, resulting in more push than pull and possibly fluctuating charge. A few snips and required reading from http://www.quadesl.org/Hard_Core/Pan...lcoatings.html "Provided the charge on the diaphragm is substantially incapable of changing during the period of one half cycle of the lowest frequency being handled, a simpler method of connection may be adopted, and the polarizing supply may be returned to either end of the transformer secondary instead of the centre tap." The reason for rendering the charge on the diaphragm "incapable of changing" is taken from Hunt's paper and book in which he shows that constant charge operation reduces harmonic distortion (in a nut shell). "The electrical resistance of the diaphragm is determined by three considerations. In the first place the resistance must be high enough to maintain substantially constant any charge carried by any elemental area of either side of the diaphragm when the diaphragm is vibrating at an audio frequency." So, here we have a hint that the diaphragm was coated on each side, and that treating the resistivity as a "lumped" quantity , and inserting series resistance in the EHT circuit is not on, as far as the designers were concerned. This attention to detail is probably one reason why modern ESLs that use low resistance diaphragms and high value series (charging) resistors never seem to have the mid-range purity of the Quad Electrostatic Loudspeaker. Moving on in the next paragraph of the patent mentioned above: "This reduces non-linearity due to the quadratic nature [F.V. Hunt] of the forces which are involved if the diaphragm potential is fixed. Due to the fact that in any practical loudspeaker the diaphragm does not operate as a rigid piston because of the necessity for supporting members and because acoustic impedances are not always uniformly distributed over its area, the same degree of improvement is not obtainable by the insertion of resistance external to a conducting diaphragm." So, here is a fairly damning indictment of the use of external (to the diaphragm) charging resistors to achieve a high resistance path and a low distortion sound. A high resistance membrane ensures good efficiency and a quiet background also. The second reason for the resistive properties of the diaphragm to be as they a "The resistance must also be high enough to prevent injurious sparking in normal operation. So that if a spark should tend to occur between a small area of the and one of the plates, the discharge current flowing in the areas of the diaphragm adjacent to the discharging area causes the potential of the discharging area to be reduced below the value required to maintain the discharge. Only a very small charge is dissipated before the voltage becomes thus reduced and the spark is accordingly rendered harmless." If the ER kit says only coat one side with nylon then they must be employing series resistance in the EHT???, according to the above site this is not a good thing. I wish you luck Patrick, this is all very interesting to me as I have a pair of ESL57 that undoubtedly will one day need refurbishing. They are still in excellent nick but like anything approaching 50 things start sagging and performance eventually drops : ) Luckily for me there's plenty of helpful information out there, unfortunately the same can't be said for your ER kit. Imo these kit sellers should have VERY detailed information available on their website, some kind of user forum and good support available. $2k is not chicken feed. regards Sean |
|
#17
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membranecoatingandtension,etc.
Sean wrote: Patrick Turner wrote: Stuff about membrane tension is formulated at http://www.quadesl.org/Hard_Core/ESL...esltheory.html However, nothing at this address talks me through the everyday means of calculating a typical panel tension as kG / cm of length along a panel side which would be required to overcome the electrostatic force present when the membrane is pulled over almost against the stator. The website display is only a very vague and poor attempt to teach anyone anything except that the creator of the website can create a website, and one that does not have a list of steps to take to get from no panel to a panel with correct tension. You are right, there's not a lot of information on perfect tensioning procedures...but maybe thats because its something thats not super critical? You can see that the person who wrote the website hasn't ever built an ESL. He has never done more than sit in a chair and type. I think the tension IS CRITICAL for optimum perfornance The ERA panels I have use 800 grams per 60mm of panel side length. The resonant F of the bass panel is around 32Hz. Quad63 use the same thickness of membrane material and use 2,500 grams/60mm of side length. bass resonance is higher. From what i see, ESL panels don't send such a strong signal back to the amplifier as a feedback signal. They do of course act like gigantic capacitor microphones, and could indeed be used as mics. So an error or distortion signal at the ESL is fed to the step up tranny sec via the series resistances,and transformed down to appear at the primary as low voltage high current and this then appear at the amplifier output and is also fed back into the amp for reverse amplification. Just how well this actually works is anyone's guess; my view is that at 50Hz, the ESL has such a high Z that distortion currents are very low at the speaker, so when transformed to the primary, they are miniscule, and the NFB is very low, so there is little error correction. So the bass response is peaked and only passive techniques can prevent its worst sonic effects, boomy bass coloured by the frequency of resonance. so even if music with no content at Fo, the panel may still be excited and wobble away at Fo whether we like it or not. Resonances that are sharp are usually not at a frequency of music, so will sound discordant. Tension affects the Fo, and the higher the tension is to prevent bass movements, the higher the Fo. However, there is plenty of information extolling the virtues on coating both sides of the mylar diaphragms. Er no. The guy who wrote all that hasn't built a panel, and hasn't focused on curing bugs nor made any real world measurements of real world speakers and with various distances used, various EHT levels, tensions and stator techniques. He can conclude what he likes, and all of what he says could be right, but I can't change my canoe while I am so far out to sea. If I was up the creek without a paddle, I could stop by the river bank and fashion a new one from a branch, or build a new canoe from bark off a better tree, but being where I am out at sea, I don't have the options of changing much. I have not concluded if in fact my canoe is one constructed from woven barbed wire. I'm basically limited to tying string around the ER Audio panels and increasing tension to make them work at all. Its like plugging a leak in the canoe with chewing gum. If then I get a canoe which is otherwise easy to paddle along, I'll be happy. Everything I've read on the subject points to coating both sides of the mylar, this results in high resistance on both sides while maintaining a constant charge....ie "incapable of changing" Only having one side coated seems a bit hit and miss, resulting in more push than pull and possibly fluctuating charge. The charge on ONE side of a very thin membrane does not result in a substantial difference in the two equal forces of tension either side of the membrane towards each stator. In the ERA panels I see no evidence that the membrane is substantially off centre with no signal due to unequal electrostatic forces which are the product of the charge in coulombs and the field strengths involved. basically, you have perhaps a kilogram of force from each stator spread over the membrane area acting towards each stator. By altering the voltage on the stators, the two forces are alterd in magnitude, with increasing attractive force on one side, and reducing force of attraction on the other, so the membrane moves as a result. Any 2H distortion in this action is mainly cancelled. A few snips and required reading from http://www.quadesl.org/Hard_Core/Pan...lcoatings.html "Provided the charge on the diaphragm is substantially incapable of changing during the period of one half cycle of the lowest frequency being handled, a simpler method of connection may be adopted, and the polarizing supply may be returned to either end of the transformer secondary instead of the centre tap." The reason for rendering the charge on the diaphragm "incapable of changing" is taken from Hunt's paper and book in which he shows that constant charge operation reduces harmonic distortion (in a nut shell). "The electrical resistance of the diaphragm is determined by three considerations. In the first place the resistance must be high enough to maintain substantially constant any charge carried by any elemental area of either side of the diaphragm when the diaphragm is vibrating at an audio frequency." So, here we have a hint that the diaphragm was coated on each side, and that treating the resistivity as a "lumped" quantity , and inserting series resistance in the EHT circuit is not on, as far as the designers were concerned. This attention to detail is probably one reason why modern ESLs that use low resistance diaphragms and high value series (charging) resistors never seem to have the mid-range purity of the Quad Electrostatic Loudspeaker. I see not the slightest link between "midrange purity", whatever that means, to coating resistance. By "midrange purity" do you mean low THD/IMD? For the panel to work well, we don't want the charge to be easily able to flow in or out of the membrane. Once we have drained out squillions of electrons from one or both sides of the membrane to make it positive, when a signal change occurs, we want the charge level to stay constant. How darn constant? Constant enough so charge won't change even at 20Hz. From what i see, ER panels don't suffer dynamic charge change to any great extent, and if the membranes did change their charge dynamically say +/- 100V at say 50Hz when there was a lot of bass content, then indeed there would be serious IMD because sensitivity varies with EHT, and if this varies dynamically, you'd get IMD. Hence unlike ERA, I am running larger than recommemded series R from TX to the bass panels to get a flattter response level, and using C&R filter drive to the centre treble strip to keep bass F out, and thus avoid bass being able to modulated treble amplitudes. Quad do this as well in ESL57.... I have not measured how important this is though. Moving on in the next paragraph of the patent mentioned above: "This reduces non-linearity due to the quadratic nature [F.V. Hunt] of the forces which are involved if the diaphragm potential is fixed. What the **** does this mean?????? Pure undecypherable boffin babble!! Due to the fact that in any practical loudspeaker the diaphragm does not operate as a rigid piston because of the necessity for supporting members and because acoustic impedances are not always uniformly distributed over its area, the same degree of improvement is not obtainable by the insertion of resistance external to a conducting diaphragm." Now he talks about dynamic drivers, then lurches sideays to ESL. Incomprehensible to someone like me who needs all statements to be fully explained, and be logical, and tell the whole well defined story. So, here is a fairly damning indictment of the use of external (to the diaphragm) charging resistors to achieve a high resistance path and a low distortion sound. A high resistance membrane ensures good efficiency and a quiet background also. No reasons are given for this claim. Could be right, could be BS. The second reason for the resistive properties of the diaphragm to be as they a "The resistance must also be high enough to prevent injurious sparking in normal operation. So that if a spark should tend to occur between a small area of the and one of the plates, the discharge current flowing in the areas of the diaphragm adjacent to the discharging area causes the potential of the discharging area to be reduced below the value required to maintain the discharge. Only a very small charge is dissipated before the voltage becomes thus reduced and the spark is accordingly rendered harmless." When I first encountered the appalling amount of arcing in ERA panels when I turned up the EHT above 2,500V, and which then got worse when the membrane became stuck to a stator, there were blue sparks all over the place. I turned everything off, lest i burn a hole in a membrane. When I stripped the arcing panel down, and put on a second pair of glasses, I could see little black beads of **** where the panel had arced repeatedly. These were the results of fried paint or coating or mylar, not sure, but had the arcing been left to occur, the membrane surely would have become punctured, and fuct. Now I have painted the stator insides with two coats of Isonel 642, rated to withstand 2,700V across 1 thou" or 0.025mm, when the stiction occurred again there was much less arcing. If I can keep the membrane away from the stator and not able to stick against a stator, arcing will NEVER be a problem. If the ER kit says only coat one side with nylon then they must be employing series resistance in the EHT???, according to the above site this is not a good thing. I am not so sure. It appears the ERA EHT supply has a diode plus capacitor "ladder" to produce a very high EHT, able to be adjusted easily by varying a low voltage supply ahead of the EHT supply. The highest EHT I could measure was -5.4kV with a special HV probe whose input resistance is above the range my Fluke DVM will measure. So the probe isn't causing near total EHT reduction whe I measure. There is a series R at the EHT supply output using what looks like a 33 megohm resistance, and again it reads OL with the Fluke. There is also a space and tracks for a neon bulb across the resistor, but the neon isn't fitted. After charge up, current flow in the 33M would drop, and the neaon wouldn't light up. But this leads me to think the membrane coating is very high resistance. Naturally enough, we are not told the ERA trade secret of what is in the coating, but it appears it is an acrylic solution and fairly stable; I tried removing it with metho and water to no avail once it had dried and after a week of curing. I couldn't measure the resistance between two coins only 1mm apart when resting on the membrane. I wish you luck Patrick, this is all very interesting to me as I have a pair of ESL57 that undoubtedly will one day need refurbishing. They will indeed need work inevitably. That they are in one piece 40 years after manufacture is a miracle. They are still in excellent nick but like anything approaching 50 things start sagging and performance eventually drops : ) Luckily for me there's plenty of helpful information out there, unfortunately the same can't be said for your ER kit. I am rapidly making up for the deficit of public knowledge about ER Audio kits. I know the ER Audio salesman in the UK, Colin Topps is listening, because he emailed me privately when i first posted on this subject. He would not listen to what I was saying, didn't understand me, underestimated my analytical abilities, and wanted to somehow talk away my problems with the bull**** of denial and sales talk, which did not succeed. He's never had any problems he says. Well, he sells the bloody things, so of course there are no problems, and then when I said privately that sorting things out with him or ERA in Perth may not be possible based on what he said, he told me to sod off quite rudely, and to never ever to email him again. I'd never employ someone like Colin. You just never should tell anyone to sod off. I don't expect any apologies. Colin became quite frustrated about not being able to provide a solution to a problem he maintained didn't exist, and I'm the minority, so he thinks he can afford to tell me to sod off. No doubt Rob Macinlay knows all about whats happening here; just one man out of many who have attempted to build the panels have found great difficulties in getting them to work as well as website claims suggest. But nobody from EAR or representing ERA in the UK has the courage to come from their hideout to join in the discussions here about the panels they produce. In contrast, Peter Walker debated his products publically and benefitted from the experience with smiles all the way to the bank. Plenty of ppl thought the 405 amp idea of 'current dumping' was plain wrong. The reliablity of the 405 and its build quality and its sales suggested the critics were wrong. Imo these kit sellers should have VERY detailed information available on their website, some kind of user forum and good support available. $2k is not chicken feed. I agree, but ppl wanna make a dollar, and you can't make a dollar if you spend time making sure all the info is there for everyone and you spend time with ppl who have problems. Its takes time, and time is money. I bet that if my customer who paid for the speakers wanted a refund due to his dissatisfaction, he'd not get one. So there is plenty of reason why I have to try hard to make these speakers work. I don't like to see a customer waste his money, nor myself suffer financial loss because I attempted to build what has even been claimed by some to be "better than Quad ESL" and failed in the attempt because of stiction and sensitivity problems. If the folks at ERA are not impressed positively by what I am saying, they should have no trouble defending their product in public, and submitting it to rigourous testing and analysis by an independant technician and professional qualified person. Hardly anyone exists who would do the work of evaluation properly, and without being paid a professional fee. Hire of an anechoic chamber isn't cheap either. Just because ppl say they are professional, and have always an ethical honest transparent approach does not men they have it, unless they proove and demonstrate it. Here in the discussion groups anyone making a claim can be deemed to be wrong, a liar perhaps, until evidence suggests otherwise. Its guilty until innocence is proven. We don't jail or execute when we find so many false claims made each day on the Net, we just invite discussions of the claims, and leave everyone to judge for themselves. This way our discussions act as fertilizer in which to grow better ideas. I look forward to seeing a post here from ERA. Patrick Turner. regards Sean |
|
#18
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membranecoatingandtension,etc.
In addition to my last post,
I share how I measure the EHT. The HV probe I have is specially able to measure up to 40kV, and has an internal resistance which is very high, and has a 1,000:1 voltage reduction ratio. The ERA EHT supply has a series R of what seems to be 33 megohms and beyond my Fluke digital multimeter to read which can measure only to about 30 megohms . Without EHT connected to the panel, I measured -3.67kV at the top of the diode + capacitance ladder. At the output of the 33Meg the voltage measured -3.5kV, so there is only a 0.17kV sag due to the loading effect of the probe, so indeed the probe internal resistance must be very high, and I calculate about 600 megohms, allowing for some sag in the supply before the 33M. The 1,000:1 reduction ratio means its output resistance is 600k, so the high input resistance of the Fluke should not affect what is being measured. The probe I am using is listed forsale at http://www.wescomponents.com.au See item 'high voltage probe' code no HV40 Cost aud $149.00 Patrick Turner. |
|
#19
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingandtension,etc.
Patrick Turner wrote:
Sean wrote: Patrick Turner wrote: Stuff about membrane tension is formulated at http://www.quadesl.org/Hard_Core/ESL...esltheory.html However, nothing at this address talks me through the everyday means of calculating a typical panel tension as kG / cm of length along a panel side which would be required to overcome the electrostatic force present when the membrane is pulled over almost against the stator. The website display is only a very vague and poor attempt to teach anyone anything except that the creator of the website can create a website, and one that does not have a list of steps to take to get from no panel to a panel with correct tension. You are right, there's not a lot of information on perfect tensioning procedures...but maybe thats because its something thats not super critical? You can see that the person who wrote the website hasn't ever built an ESL. He has never done more than sit in a chair and type. Good grief, I dont know the guy from a bar of soap but he's devoted a lot of time and many webpages to help and guide others (just like your website) and you dish that out? With that sort of attitude I'm not surprised people aren't more forthcoming helping you solve your problem. Good luck to you. ps there's more here on imbalanced diaphragms, it might be valuable reading for some. http://www.quadesl.org.uk/ - Do I coat one or both sides of the diaphragm ? Sean |
|
#20
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingandtension,etc.
Sean wrote:
Patrick Turner wrote: You can see that the person who wrote the website hasn't ever built an ESL. He has never done more than sit in a chair and type. Sean wrote: Good grief, I dont know the guy from a bar of soap but he's devoted a lot of time and many webpages to help and guide others (just like your website) and you dish that out? I couldn't agree more Sean - a case of pot,kettle??. Turner hasn't built one either - he is 'attempting' to build one from a kitset which is a whole lot different to building one from scratch. Gary Jacobsen has done rebuilds on Quad ESL panels - which involves essentially the same procedures as the kitset job that Turner is messing about with, wouldn't you say?. |
|
#21
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membranecoatingandtension,etc.
Sean wrote: Patrick Turner wrote: Sean wrote: Patrick Turner wrote: Stuff about membrane tension is formulated at http://www.quadesl.org/Hard_Core/ESL...esltheory.html However, nothing at this address talks me through the everyday means of calculating a typical panel tension as kG / cm of length along a panel side which would be required to overcome the electrostatic force present when the membrane is pulled over almost against the stator. The website display is only a very vague and poor attempt to teach anyone anything except that the creator of the website can create a website, and one that does not have a list of steps to take to get from no panel to a panel with correct tension. You are right, there's not a lot of information on perfect tensioning procedures...but maybe thats because its something thats not super critical? You can see that the person who wrote the website hasn't ever built an ESL. He has never done more than sit in a chair and type. Good grief, I dont know the guy from a bar of soap but he's devoted a lot of time and many webpages to help and guide others (just like your website) and you dish that out? Good grief indeed. I'll say whatever I think is fair here and where I think is appropriate. IMHO there are extremely few websites anywhere that have any real depth about them, and none are places where someone can pick up all the theory plus its applications plus worked out samples of designs. Ppl build stuff, then crow all about it like a rooster who just made with a chook. Unlike the average punter, I like to see MORE everywhere I go. With that sort of attitude I'm not surprised people aren't more forthcoming helping you solve your problem. Armchair experts are just that, they like to regurgitate the scientific jargon about ESL which originated from far more prolific, intelligent, resourceful and practical minds who lurked here on Earth 50 years ago, such as Peter Walker, D.T.N. Williamson, and Peter Baxandal. If somebody is going to tell us something about ESL, could they be so kind as to give us fully worked examples as they go? In my OPT design pages I revolve the explanations and theory and formula all around worked examples. I spent 6 mths unpaid last year to establish my 18MB website. And even after all that people want more, they still have questions. I have always tried to explain things when they emailed me weekly. From what i see in websites on ESL around the Web, there are extremely few ppl who have ever bothered to sit down and analyse the and explain and reproduce the accurate equivalent LCR circuit for the Quad ESL57, as explained by Peter Baxandal. But once you see Peter's circuit, and if you have learnt basic LCR theory, then the myths about ESL electronics dissappear, along with a lot of BS notions. On today's websites there are no detailed explanations with calculations conveyed for the ordinary person to understand with regard to membrane tension, membrane to stator distance, EHT ideal settings, and certainly ZERO explanations of why I am getting stiction problems. There is not a huge number of people attracted to r.a.t, which has, like so many public access forums, the reputation of a sewer where rats and cockroaches gather to throw turds at each other. But when one does a Google search on ERA kit speakers, our discussions will come up, and anyone reading our discussions will think twice about an ERA kit, and not be so disturbed when they realise at least one observant man had a few bothers which took time to fix. So I can understand why the the sales people and CEOs of companies making boutique audio gear stop short of appearing here, they'd get too dirty. But Peter Walker wasn't a shy little geranium now was he! He'd have the answers, and he'd done the homework, and he had the courage, and I expect ppl making speakers which they claim to equal Walker's wonders to be the same. I hoped to perhaps share experiences with others who have had either success or failures, and certainly real friendship is where problems can be shared, but we never take discssions so personally that we couldn't share a beer. But it appears that about nobody who is now building ESL panels and is aware of r.a.t wants to talk about the issues. Like ribbon or horn speakers, ESL have their rightful place in the audio world to offer sonic joys when properly set up. Good luck to you. ps there's more here on imbalanced diaphragms, it might be valuable reading for some. http://www.quadesl.org.uk/ - Do I coat one or both sides of the diaphragm ? I don't have enough experience to advise you. I am struggling with my first ESL panel building experience. If I get the ERA panels working with complete freedom from stiction, and as reliably as i think is necessary, maybe you may have more to think about. I have scraped off the glue and last membrane from the panel i am working on, and am ready for threading string and silicone doping tommorrow to hopefully stop the stiction bothers. This will take a day or two to cure. Maybe thursday I stretch another membrane and glue it to the prepared frame with stator complete with motion limiting string barriers, and silicone doping. Doing this part of the process will take many hours, but if I proove the technique works, and the EHT can higher at up to say -4,000V to give more sensitivity without stiction bothers then I have effectively R&D'd a better panel than ERA has so far invented, IMHO. The silicone tabs which I did use as stipulated by the ERA manual for node damping points to suppress bass resonances were found to have well bonded to the membrane, and very well bonded to the stators which have two coats of Isonel 642 on them. So the silicone used as a paint will work fine and never peel off, and dissallow the sticky bond between Isonel 642 and mylar that occurs when stiction has been sustained for some time, and it won't release after the power is turned off. I have been using silicone for 25 years, and its a wonderful bit of stuff, and when undoing or altering roof and gutter work I did 25 years ago not the slightest degradation seems to have occured; the metals will rust away before the silicone degrades. Patrick Turner. Sean |
|
#22
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membranecoatingandtension,etc.
Mike Coatham wrote: Sean wrote: Patrick Turner wrote: You can see that the person who wrote the website hasn't ever built an ESL. He has never done more than sit in a chair and type. Sean wrote: Good grief, I dont know the guy from a bar of soap but he's devoted a lot of time and many webpages to help and guide others (just like your website) and you dish that out? I couldn't agree more Sean - a case of pot,kettle??. Turner hasn't built one either - he is 'attempting' to build one from a kitset which is a whole lot different to building one from scratch. Gary Jacobsen has done rebuilds on Quad ESL panels - which involves essentially the same procedures as the kitset job that Turner is messing about with, wouldn't you say?. Don't be silly, rebuilding Quad ESL57 is not the same as my building my kit although some procedures are similar. The fact is that quite a number of people know about ESL57 which have been around for 50 years, and I have no arguments with them, but where anyone has presented their thoughts to the Net they have not provided a one stop shop for all the theory and practices generic to ESL enabling anyone to build their own from scratch. My experience with the ERA kit has so far been dissapointing. I never intended to become an ESL expert, and am not an expert yet, not by any means, but with dissapointment comes the urge to try harder while learning more. I am utterly non expert about repairs to Quad ESL57. I am however, full of questions, and some remain quite unanswered, due to the lack of comprehensible advice from others on the Net. If I were to try to build an ESL from scratch, I'd probably save myself a lotta bother and just copy well known Quad techniques which everyone has advised is superior to anything in a kit. Its also a lot more work. I have applied a similar approach to amp building but implemented the 60 years year old ideas somewhat better, and unlimited by bean counters. I don't care that I am unpopular, I'd rather ask 100 questions, and be asked to leave the party, than get drunk and stupid. Patrick Turner. |
|
#23
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membrane coatingandtension,etc.
Patrick Turner wrote:
Sean wrote: http://www.quadesl.org.uk/ - Do I coat one or both sides of the diaphragm ? I don't have enough experience to advise you. You didnt even read it...right??? Pay particular attention to the spacing of the quad treble panels....look familiar? "I now believe that there is an extra reason which I had overlooked. As soon as you apply the coating to one side only you immediately have an imbalance. You should after all do the same to one side of the diaphragm as the other to maintain balance." "The question is why was the diaphragm hitting the back stator hard ? I now realise that this was being caused by the coating only being applied to one side." "Because the charge is an electric field any conductive material in close proximity with it will have a voltage induced onto it. Inside the treble panel the stators are only 0.75mm away from the diaphragm. With a conductive coating on one side only it meant that the front stator had a voltage induced on it and because these voltages were the same charge they repelled each other. The effect of this is to push the diaphragm hard against the back stator. ****The simple cure for this is to coat both sides of the diaphragm thus restoring balance to the panel**** "The only question left is why do so many people only coat one side ? In the case of other ESL designs I suspect the reason is that the diaphragm is a sufficient distance from the stators as to make it impossible to induce a voltage on them. Don't forget that other designs including the ESL 63 use stator spacings of at least 6.5mm whereas the Quad treble panel has a stator gap of only 1.5mm. This means the stator to diaphragm distance is only 0.75mm so its not surprising that a voltage is induced. As to how other ESL 57 repairers manage to only coat one side and make them work is a mystery. Best to ask them." I am struggling with my first ESL panel building experience. Yes, and it seems you are hell bent on fixing them with string and plumbers silicone. I'm sure the proud new owner will be tickled pink. Sean |
|
#24
Posted to rec.audio.tubes
|
|||
|
|||
|
ER Audio ESL sensitivity, stiction problems, membranecoatingandtension,etc.
Sean wrote: Patrick Turner wrote: Sean wrote: http://www.quadesl.org.uk/ - Do I coat one or both sides of the diaphragm ? I don't have enough experience to advise you. I could repeat exactly what I said above, after reading Andrew's dissertation about repairing the Quad ESL57 panel where he says the following... His commonsense approach is a welcome refreshment of words.... """""In my early days of electrostatic speaker repair I coated one side of the diaphragm only. Many people will tell you that you need only coat one side. There are many internet sites which say this. Also there are books available which describe how to make your own ESL speaker and these also tell you to coat one side only. I will try to explain why many people suggest coating one side only and why I don't think this works in the ESL 57 treble panel. What has to be understood is that the music signal which is being applied to the stators is attracting/repelling the charge on the diaphragm not the diaphragm itself. It is only because the charge is held captive on the surface of the diaphragm that means the diaphragm will vibrate. Consider a diaphragm with a coating on one side. So, you have a charge on the surface of a diaphragm which is being driven by a musical signal. This charge is an electric field. As such it is similar to a magnetic field. If you put a piece of paper in front of a magnet it won't stop it working. This is the case with the diaphragm. The stator on the coated side of the diaphragm can directly see the coating and hence the charge. But between the rear stator and the charge is the diaphragm itself but this doesn't matter because the diaphragm is invisible as far as the electric field is concerned. This is fine as far as it goes and this is why most people only coat one side but with the Quad treble panel there is another consideration. Anyone who has taken a bass or treble panel apart will notice that the Quad diaphragms are coated on both sides. I tried to come up with an answer as to why this was so. I could only guess that this was because the original Quad coating was applied so carelessly that it needed to be coated both sides to ensure the whole surface of the diaphragm was driven. I now believe that there is an extra reason which I had overlooked. As soon as you apply the coating to one side only you immediately have an imbalance. You should after all do the same to one side of the diaphragm as the other to maintain balance. I have learnt the following after repairing many diaphragms so this is all practical experience. When I repair an arced treble panel I clean the burned area and re-insulate it with corona dope. This insulates the exposed area but the dope is such that it never quite hardens. It always remains soft to the touch and slightly sticky. I would then put the new diaphragm in the panel and try it out. It was not unusual to notice that when the high voltage supply was removed from the diaphragm you would get a nasty crackling noise. I always put this down to some minor leakage causing the diaphragm to crackle but I now realise that this crackling was the diaphragm pulling away from the area which has been repaired with corona dope. This meant that the diaphragm was hitting the back stator and sticking there until the high voltage supply was removed at which point the diaphragm would gradually become unstuck from the corona dope and thus cause the crackling. The question is why was the diaphragm hitting the back stator hard ? I now realise that this was being caused by the coating only being applied to one side. Because the charge is an electric field any conductive material in close proximity with it will have a voltage induced onto it. Inside the treble panel the stators are only 0.75mm away from the diaphragm. With a conductive coating on one side only it meant that the front stator had a voltage induced on it and because these voltages were the same charge they repelled each other. The effect of this is to push the diaphragm hard against the back stator. The simple cure for this is to coat both sides of the diaphragm thus restoring balance to the panel. The only question left is why do so many people only coat one side ? In the case of other ESL designs I suspect the reason is that the diaphragm is a sufficient distance from the stators as to make it impossible to induce a voltage on them. Don't forget that other designs including the ESL 63 use stator spacings of at least 6.5mm whereas the Quad treble panel has a stator gap of only 1.5mm. This means the stator to diaphragm distance is only 0.75mm so its not surprising that a voltage is induced. As to how other ESL 57 repairers manage to only coat one side and make them work is a mystery. Best to ask them."""""" You didnt even read it...right??? Notice how he has not come to a conclusion why other repairers get ESL to work OK with a membrane coating on one side only of the membrane. He talks about the exact same stiction problem I have found in the ERA speakers, and that is occurs where he has used anti corona paint just like I have to stop arcing. He doesn't say what exact anti corona paint he used but Isonel 642 I have used indeed not a hard varnish and it does remain very slightly sticky. I though without the anti corona paint that I was getting stiction at least while the panel was turned on and EHT was present. This gave rise to lots of arcs, and the anti corona paint seemed like the best way to stop it. The stiction prolem arrose again after applying the A.C paint, but without much arcing, but then the same slow release problem occurred as he has described. So at least one guy out of 3 billion men has all my problems, and you found him. But I am not sure I understand his reasoning.... """"With a conductive coating on one side only it meant that the front stator had a voltage induced on it and because these voltages were the same charge they repelled each other. The effect of this is to push the diaphragm hard against the back stator. The simple cure for this is to coat both sides of the diaphragm thus restoring balance to the panel.""""" He is talking about the mid-treble panel in a Quad ESL57. The treble stators are held at the same 0V dc potential because they are biased to 0V by either being connected directly to the half way taps on each side of the 1:290 step up tranny, or else biased to 0V with resistors and driven wirh capacitors fromthe transformer. Thus the mid/treble panels have RC crossovers to prevent the high bass voltage appearing at the mid-treble stators and arcing. The way Quad have done the crossovers doesn't look right to me because the the RC used to stators each side of the membrane are asymetrical. Under severe overload, pssibly a rectification effect occurs, causing temporary dc bias imbalance of the stator with series resistance, so the membrane gets pulled more to one side than the other. Nothing is "induced". Induction is a term for magnetic applications; currents can for example ne induced in wiring running past a nearby transformer, but never does this happen in electrostatic things like ESL. He observes what he thinks causes a problem, but it could be dc bias change due to rectification. I can't be sure though myself, because I have not measured the dc voltage across the resistors in series with the driven stators. In my case, I should measure the dc voltage across the driving resistors to my bass panels which are suffering the stiction problem To stop the stiction remaining even when the panels are turned off, I am trying strings tied at 10mm across the bass panels and doped with silicone paint, which is thinned down clear caulking silcone from the hardware store using white spirits. But why one would get rectifying effects is beyond me. I think an over driven membrane can begin touching both stators if driven hard, and because one side isn't coated its less sticky in contact with the AC paint, so it does not stick, and the side that sticks slightly more due to slight tackiness is the uncoated side, and that is the side which is chosen for the stiction when the attaction overcomes the membrane tension and it then flattens against the stator. It only takes a minute for the glue effect to cause more stiction than that caused only by the EHT. In a test here my panels remained stuck over night, and had to be prised off the stator with a bent bit of wire. I would say the Quad method of assymetrical crossovers is BS. OK, I know Walker had more brains than I have, but I like symmetry to reign over assymetry, and I have tried 0.0033uF and 50k resistors to feed each treble stator in my ERA panels. I don't seem to have a problem with stiction in the treble at all. Distance between stators and membrane is about 2.2mm. ERA said the perforated steel used for treble is thicker, but I couldn't see much difference. The thicker steel would give the closer distance to the membrane. Pay particular attention to the spacing of the quad treble panels....look familiar? Its not like the ERA which have much bigger spacing. "I now believe that there is an extra reason which I had overlooked. As soon as you apply the coating to one side only you immediately have an imbalance. You should after all do the same to one side of the diaphragm as the other to maintain balance." I think he believes something without reason. He spent time explaining before that that the charge on the coating applied to one side only is interacting equally with each stator; the thickness of mylar is completely electrostatically transparent. "The question is why was the diaphragm hitting the back stator hard ? I now realise that this was being caused by the coating only being applied to one side." Nah, I was hitting BOTH sides, tackiness between the uncoated side and the painted stator initiated stiction. "Because the charge is an electric field any conductive material in close proximity with it will have a voltage induced onto it. Inside the treble panel the stators are only 0.75mm away from the diaphragm. With a conductive coating on one side only it meant that the front stator had a voltage induced on it and because these voltages were the same charge they repelled each other. The effect of this is to push the diaphragm hard against the back stator. ****The simple cure for this is to coat both sides of the diaphragm thus restoring balance to the panel**** It just does not ring true to me. I don't yet see enough reasons to coat both sides One may have to, but I cannot see reasons why that are elecronically or correct in line with observed/described phenomena. I'll be able to say more after I put in place the means to stop the sticion in the ERA pass panels. One guy has coated his perforated steel stators by dipping them in molten polythene to give an insulation layer of 1mm on the surface of the steel. So the membrane can't get closer than a mm to the steel, and polythene isn't sticky, and at 1mm thick its not a bad anti corona treatment. I doubt it could be applied in ESL57 speakers because its too thick a coating. http://www.ele.tut.fi/~artoko/audio/speakers/esl.html. "The only question left is why do so many people only coat one side ? In the case of other ESL designs I suspect the reason is that the diaphragm is a sufficient distance from the stators as to make it impossible to induce a voltage on them. Don't forget that other designs including the ESL 63 use stator spacings of at least 6.5mm whereas the Quad treble panel has a stator gap of only 1.5mm. This means the stator to diaphragm distance is only 0.75mm so its not surprising that a voltage is induced. As to how other ESL 57 repairers manage to only coat one side and make them work is a mystery. Best to ask them." I am struggling with my first ESL panel building experience. Yes, and it seems you are hell bent on fixing them with string and plumbers silicone. I'm sure the proud new owner will be tickled pink. He sure will if I can produce higher SPL, and with less input voltage, and without stiction. And the improvement should be something ERA should addopt for their future efforts. Due to the way the ERA speakers are constructed, it is impossible to easily make their panels with conductive coating applied to both sides of the membrane. Say I could stretch out a membrane on a frame larger than the panel size of 1,200mm x 600mm, like a window frame. Then I have access to both sides which could be coated before gluing on the support frames to each side which would have to be 2mm thick plastic, and have copper tracking within to make contact with the coating of both sides. This would make it 4 times more fiddly-diddly and difficult to get right. At present with ERA ESL-IIIB, which is a combined bass-mid-treble unit, you stretch the membrane on a flat suface at least 900mm x 1500mm; an old door isn't a bad bench, and then you glue a prepared support frame with a stator attatched onto the stretched out membrane. Next day the excess membrane is cut around the outside of the frame, weights removed and you have a membrane nicely stuck to a frame with a stator one side. This open side is then coated. The other stators are in an identical second frame, and have a copper foil track which glues to the frame, leaving the bare metal side to contact the membrane coating of the first half frame with membrane on it. Nylon bolts hold the two frame haves together in the panel centre area, and PVC channels snap fit around the external perimeter. I cannot fault the way ERA have presented their physical construction of their speakers. The CEO at ERA is a little shy, and doesn't talk much about what must have been his failures when he R&D'd the first ESL he tried to make. I see room for improvement to sensitivity and reliability. I can be a little caustic about it all, but its all for a good cause, and in favour of the people affected, ERA themselves, my customer, and myself, and the world at large. I am glad you posted what you have. Patrick Turner. Sean |
| Reply |
| Thread Tools | |
| Display Modes | |
|
|
Similar Threads
|
||||
| Thread | Forum | |||
| FA tube electrostatic speaker amp like pignose or ampex | Marketplace | |||
| Sensitivity Numbers for a Box? | Car Audio | |||
| I need help repairing an electrostatic speaker... | Pro Audio | |||
| Electrostatic Speaker Crossover | Tech | |||
| Speaker sensitivity and fs in multiples. | Tech | |||
Linear Mode
