There will always be always some small paths of leakage of the high voltage bias. Even the air gap will have some ionization occurring below hard break down. Other leakage paths can be surface dirt on spacers and dust bridging the gap. These factors vary a lot depending on humidity and age of the ESL. If there is enough leakage current to significantly reduce the bias voltage on the diaphragm due to voltage drop across the large series resistor, this might explain the improvement heard by some folks after reducing a very large series resistor to a smaller value. If an ESL used a 500M resistor and had a 5000V supply, all it would take is 2 micro-amps of leakage to drop the voltage down to 4000 volts. If this were the case, changing the 500M to a 20M might make a noticeable difference in sensitivity, which is sometimes heard as greater detail, etc. There is a real balancing act with this resistor between providing enough current to feed the leakage paths without dropping much voltage, and keeping a constant charge on the diaphragm. Fortunately we can make sure the resistivity of the diaphragm is high enough that we don’t need to rely on the series resistor for all the constant charge effect by itself.
We could envision a choke performing some of the “constant charge duty”, instead of a large resistor, while freely allowing DC current to flow. But the choke would have to have many thousands of henries of inductance. Since the leakage currents are minuscule, you would not need to worry about DCR, nor about air gaps. Then you could wind many thousands of turns of very fine wire on a high permeability core. You would have to take care that the shunt C was very tiny, which would prove very difficult, and you’d have to package the whole beast in an insulating enclosure since the fine wire won’t have enough voltage rating next to a grounded core. A series resistor of a few megs would still be helpful, if only for safety and to damp any resonances. It might be an idea worth pursuing. Any takers?
We could envision a choke performing some of the “constant charge duty”, instead of a large resistor, while freely allowing DC current to flow. But the choke would have to have many thousands of henries of inductance. Since the leakage currents are minuscule, you would not need to worry about DCR, nor about air gaps. Then you could wind many thousands of turns of very fine wire on a high permeability core. You would have to take care that the shunt C was very tiny, which would prove very difficult, and you’d have to package the whole beast in an insulating enclosure since the fine wire won’t have enough voltage rating next to a grounded core. A series resistor of a few megs would still be helpful, if only for safety and to damp any resonances. It might be an idea worth pursuing. Any takers?
tiny chokes
Brian: the one Henry chokes that I picked up surplus are one inch diametre by 3/8 inch thick outside dimensions. They appear to be wound with about 34 guage or smaller wire. This must be approaching what you have in mind no? They were only about $1.25 each. I do plan to series up a couple of these composite ten Henry chokes to compare to just one unit per supply. Let me know if this meets your criteria. Regards Moray James.
Brian: the one Henry chokes that I picked up surplus are one inch diametre by 3/8 inch thick outside dimensions. They appear to be wound with about 34 guage or smaller wire. This must be approaching what you have in mind no? They were only about $1.25 each. I do plan to series up a couple of these composite ten Henry chokes to compare to just one unit per supply. Let me know if this meets your criteria. Regards Moray James.
Moray,
No that’s not quite what I meant. I said I thought it might take THOUSANDS of Henries to ensure approximately constant charge, especially if you wanted to run the ESL down low. That’s why I struggle with the discussion about only 10H making a difference. Let’s say you wanted the choke to present 1Meg of impedance at 100 Hz (which is marginal at best), just for starters, to get a ball-park figure. Well, that’s 1600 Henries. Impossible? No, there are grid bias chokes in that range, for example. If you don’t have to pass any significant current, and DCR can be (and really SHOULD be) as high as possible, then it takes just lots and lots of turns of tiny gauge wire around a gapless high permeability core. There is no dissipated power, so the core would need to be only as big as necessary to fit so many turns around it. At 1 KHZ, such a choke would present 10MEG of Z - IF there were no shunt C, which there will be of course. This part is probably the hardest. I’m no choke winding expert, but I think that keeping the shunt C to few picofarads may be impossible without resorting to many separate series chokes, each of which must have hundreds or thousands of Henries. That might be crazy sounding, but perhaps worth trying. Got ten 500 H chokes on the shelf? (20 for a stereo pair) Not me. You could always add a large resistor in series with these chokes to gild the lily. Maybe 20M.
No that’s not quite what I meant. I said I thought it might take THOUSANDS of Henries to ensure approximately constant charge, especially if you wanted to run the ESL down low. That’s why I struggle with the discussion about only 10H making a difference. Let’s say you wanted the choke to present 1Meg of impedance at 100 Hz (which is marginal at best), just for starters, to get a ball-park figure. Well, that’s 1600 Henries. Impossible? No, there are grid bias chokes in that range, for example. If you don’t have to pass any significant current, and DCR can be (and really SHOULD be) as high as possible, then it takes just lots and lots of turns of tiny gauge wire around a gapless high permeability core. There is no dissipated power, so the core would need to be only as big as necessary to fit so many turns around it. At 1 KHZ, such a choke would present 10MEG of Z - IF there were no shunt C, which there will be of course. This part is probably the hardest. I’m no choke winding expert, but I think that keeping the shunt C to few picofarads may be impossible without resorting to many separate series chokes, each of which must have hundreds or thousands of Henries. That might be crazy sounding, but perhaps worth trying. Got ten 500 H chokes on the shelf? (20 for a stereo pair) Not me. You could always add a large resistor in series with these chokes to gild the lily. Maybe 20M.
keep the stock 500 meg ohm resistor..
you can look for a surplus one henry choke with a ceramic core and fine wire. The ones I use are about a cm thick and 2.5 cm in dia. the wire gage is quite fine. Don't worry about current capability of the choke as there is very little of that at the end of the supply and the big load resistor deals with what is left over. I use ten in series and I rotate the chokes 90 degees one to the next to keep the fields from coupling up. Let us know what you think about it when you have it all set up.
you can look for a surplus one henry choke with a ceramic core and fine wire. The ones I use are about a cm thick and 2.5 cm in dia. the wire gage is quite fine. Don't worry about current capability of the choke as there is very little of that at the end of the supply and the big load resistor deals with what is left over. I use ten in series and I rotate the chokes 90 degees one to the next to keep the fields from coupling up. Let us know what you think about it when you have it all set up.
having trouble on the web finding 1 H. inductors. Several chinese companies, but they sell only in huge lots. nothing at radio shack, and parts express only has speaker making inductors one 1H is about $20 (?) bucks.
Cant find anything about suppliers on this site either.
Where can I buy these bad boys?
Thanks.
Paul
Cant find anything about suppliers on this site either.
Where can I buy these bad boys?
Thanks.
Paul
My 2+2's have medallion and C modificatons done around 1990.
As long as I am pulling things out and soldering should I just replace all of the resistors, capacitors wiring diodes etc. or just the components I've seen recommended in these threads.
(input capacitors, change bias supply resistor from 200Mohm to 20 Mohm and add a .01 cap to bias supply as seen on the Izzy Wizzy (?) site.
While I am at it I was going to add the 10 H inductors after the resistor , and try separating the left and right halves of each speaker to each transformer.
Still having trouble finding 1H inductors tho/
Will start as soon as I can find suppliers.
As long as I am pulling things out and soldering should I just replace all of the resistors, capacitors wiring diodes etc. or just the components I've seen recommended in these threads.
(input capacitors, change bias supply resistor from 200Mohm to 20 Mohm and add a .01 cap to bias supply as seen on the Izzy Wizzy (?) site.
While I am at it I was going to add the 10 H inductors after the resistor , and try separating the left and right halves of each speaker to each transformer.
Still having trouble finding 1H inductors tho/
Will start as soon as I can find suppliers.
Does anyone have a recommendation for high voltage capacitors for the acoustat mod. I guess the Black Gates are no longer made. At the conventional speaker making sites there are no capacitors rated in the 5000 - 7000 V. range.
Also I was wondering about the Izzy wizzy mod, they recommended 5000 V rated capacitors for C4 & C5, yet these caps aren't in the power supply, but in the audio signal.
I was hoping to replace my capacitors with audio grade capacitors (Musicap, Auricap, etc. ) because these are known to have good sound, But I don't know about just finding a cap on the internet and using it because it's rated at 5000 V
Paul
Also I was wondering about the Izzy wizzy mod, they recommended 5000 V rated capacitors for C4 & C5, yet these caps aren't in the power supply, but in the audio signal.
I was hoping to replace my capacitors with audio grade capacitors (Musicap, Auricap, etc. ) because these are known to have good sound, But I don't know about just finding a cap on the internet and using it because it's rated at 5000 V
Paul
Hi,
You don´t need ´audio-grade´ caps (whatever is meant with that anyway) in the HV-supply. It´s not only wasted money but truly rubbish. There are often cheap HV-ceramic caps to find. Those are totally ok here. Film caps are already ´over the top´ but can be used too. But again..for the HV-supply You don´t need anything expensive and fancy.
jauu
Calvin
You don´t need ´audio-grade´ caps (whatever is meant with that anyway) in the HV-supply. It´s not only wasted money but truly rubbish. There are often cheap HV-ceramic caps to find. Those are totally ok here. Film caps are already ´over the top´ but can be used too. But again..for the HV-supply You don´t need anything expensive and fancy.
jauu
Calvin
Hi,
the only reason for a good cap to be changed could be that it has developed defects because of wrong design or faulty working conditions. Film caps e.g. that suffered from internal flashovers (overvoltage). If that is not the case, than there is no reason for change.
I´m not familiar with C4C5Itsywitsyteenybeeny... but i assume a schematic would help
. If C4 and C5 are on the primary side (amplifier side) of the audio tranny than there is no need to use 5kV-rated types. 100V types would be ok. If they however are situated on the secondary side (Stator side, HV-signal side) their voltage rating must be high very high (I´d opt for at least 2x p-p signal voltage).
jauu
Calvin
the only reason for a good cap to be changed could be that it has developed defects because of wrong design or faulty working conditions. Film caps e.g. that suffered from internal flashovers (overvoltage). If that is not the case, than there is no reason for change.
I´m not familiar with C4C5Itsywitsyteenybeeny... but i assume a schematic would help
. If C4 and C5 are on the primary side (amplifier side) of the audio tranny than there is no need to use 5kV-rated types. 100V types would be ok. If they however are situated on the secondary side (Stator side, HV-signal side) their voltage rating must be high very high (I´d opt for at least 2x p-p signal voltage).jauu
Calvin
If I wanted to change the signal path capacitors in my 2 +2's with audiophile grade caps, the highest working voltage I've seen are some Hovland's rated at 1200 V. For the .01 uF caps, should I just parallel 5 or six to get to the 5000 V range .
Same with the resistors. I would like to replace them with the best I can get, but don't see the audiophile types for sale in these High voltage ranges.
Paul
Same with the resistors. I would like to replace them with the best I can get, but don't see the audiophile types for sale in these High voltage ranges.
Paul
Hi,
...to get to the 5000V range......you will have to connect them in series! Paralleling doesn´t increase voltage rating but the capacitance value. Series connecting increases voltage range but decreases the capacitance value. You have to keep in mind that the voltage range may be a specified as DC-value! The AC-value is typically just 1/3 of the DC-value. And if you look at the datasheets You´ll learn that the AC-value is specified for very low frequencies 50Hz/400HZ/1kHz because these caps were often intended for usage in power-management systems. With higher frequencies the voltage rating drops considerably. The WIMA MKP10, 0.01µF is rated at 2500Vdc and 900Vrms (AC). This later calculates to ~2.500Vpp. The value decreases with a constant rate above 2kHz and reaches 90Vrms (250Vpp) at 20kHz. Sounds slightly different, ey?
Luckily the high frequency content of music decreases too, but if you happen to test with higher power and a white noise signal -which You find on basically each and every test CD- you very probably blow the caps! With regard to this I wouldn´t trust the info on the izzywizzy site that just one cap with a 5000V rating (AC? DC?) will work reliably (I´ve got the impression that they don´t know exactly what they are dealing with anyway, since their parts choice looks rather like a try-and-error game than solid engineering)
You will rather need a string of series connected caps or a specialized HighVoltage cap with ~20kV rating. Series connected WIMA MKP10 or FKP1 will be very ok, as well as competitor´s caps in a comparable build and with comparable specs.
This only relates to C4 and C5 which must have high voltage ratings!
If the maximum signal voltage is far below 100V -as it is the fact for all the other caps- it is utterly nonsense and a lot of wasted money to use vastly voltage overrated caps!!
High voltage rating with capacitors just makes them big and expensive, not better. But in the end that´s what the HighEnder wants.....impressive, vastly oversized and preferably terribly costly looking stuff! Should it come out that its crappy stuff, don´t worry. Just tell everybody that it´s sounding better than You´ve ever heard before.
"Audio Grade" is only a marketing term but implies a superior quality of such a part over ´standard´parts. It also seems to imply that this part is superior in any case, regardless of its position within a circuit. This simply does not hold true! Since the demands of Audio are fairly low compared to other technical areas the so called ´Audio Grade´parts don´t need to be high quality in a technical sense. Just think of carbon resistors, oil-filled caps, electrolytics and especially some cables which are grossly inferior to standard parts.
The claimed superiority only relates to unmeasurable parameters like sonic behaviour. And that is a subjective and very debatable parameter.
So, apart from C4 and C5 all other caps in the circuit may happily be specified far below the kV-range.
For R2 and R3 (HV-Side) noninductive wound wire resistances are ok
Similar to the caps it may be necessary or even needed to series connect a couple of resistors. I wouldn´t use thickfilm resistors here and the probabely best Rs -made from metal foil- will be very expensive
On the lowvoltage side R1 and R4 can be of Metal foil build. The probabely best resistors here are the Isabellenhuette PBH and PBV and PSB. http://www.isabellenhuette.de
The HV-resistor R5 can be made from a chain of metal film resistors (0.5-0.6W, 1%-5%, size 0207). They are spec´d around 250V-300V, but in practise use them not above ~150V. For 5kV bias you need app 30pcs of 1MegOhm.
The original value of 500Megs is much higher than I´d use if a high ohmic coating is used on the diaphragm. 50Megs at maximum. You should be careful not to touch any HV-point, since the smaller resistance means higher currents and therefore more pain when getting shocked.
If someone reports sonic differences because of using a different resistor for R5 of same value, I´d say this would rather be a hint that the circuit itself is of minor design quality or psychology is playing games on him.
I´d rather make the HV stiffer by using larger Cap-values for C6 to C10 (>10nF) and use a negative polarity output instead of the positive. And adding a blinker cicuit (see Quad) will isolate the HV from the membrane.
The 10nF cap directly at the output of the cascade does no good. If you feel the need -a need that is actually nonexistent in praxis- that you have to smooth out the output than add a series resistor of 10M between output und 10nF-cap.
I don´t believe that a inductor could do anything positive here. The series resistance of R5 added to the high resistance value of the membrane coating works as ripple and noise filter already.
Besides the restrictions Mr. Beck mentioned (though I don´t see ´noise signals´ of sufficiently high frequency anyway) we are talking here about microscopically small currents. Which leads to the Q: If there were any ripple and noise above 100/120Hz.....would it supply for enough current in the inductor to make any difference to a resistor of same DC-ohmic-value? Or in other words:" Does the inductor work as inductor, or merely like a low-ohmic wire wound resistor?" Replacing R5 alltogether by a inductor shouldn´t be done because of safety reasons (current limiter).
jauu
Calvin
...to get to the 5000V range......you will have to connect them in series! Paralleling doesn´t increase voltage rating but the capacitance value. Series connecting increases voltage range but decreases the capacitance value. You have to keep in mind that the voltage range may be a specified as DC-value! The AC-value is typically just 1/3 of the DC-value. And if you look at the datasheets You´ll learn that the AC-value is specified for very low frequencies 50Hz/400HZ/1kHz because these caps were often intended for usage in power-management systems. With higher frequencies the voltage rating drops considerably. The WIMA MKP10, 0.01µF is rated at 2500Vdc and 900Vrms (AC). This later calculates to ~2.500Vpp. The value decreases with a constant rate above 2kHz and reaches 90Vrms (250Vpp) at 20kHz. Sounds slightly different, ey?
Luckily the high frequency content of music decreases too, but if you happen to test with higher power and a white noise signal -which You find on basically each and every test CD- you very probably blow the caps! With regard to this I wouldn´t trust the info on the izzywizzy site that just one cap with a 5000V rating (AC? DC?) will work reliably (I´ve got the impression that they don´t know exactly what they are dealing with anyway, since their parts choice looks rather like a try-and-error game than solid engineering)
You will rather need a string of series connected caps or a specialized HighVoltage cap with ~20kV rating. Series connected WIMA MKP10 or FKP1 will be very ok, as well as competitor´s caps in a comparable build and with comparable specs.
This only relates to C4 and C5 which must have high voltage ratings!
If the maximum signal voltage is far below 100V -as it is the fact for all the other caps- it is utterly nonsense and a lot of wasted money to use vastly voltage overrated caps!!
High voltage rating with capacitors just makes them big and expensive, not better. But in the end that´s what the HighEnder wants.....impressive, vastly oversized and preferably terribly costly looking stuff! Should it come out that its crappy stuff, don´t worry. Just tell everybody that it´s sounding better than You´ve ever heard before.
"Audio Grade" is only a marketing term but implies a superior quality of such a part over ´standard´parts. It also seems to imply that this part is superior in any case, regardless of its position within a circuit. This simply does not hold true! Since the demands of Audio are fairly low compared to other technical areas the so called ´Audio Grade´parts don´t need to be high quality in a technical sense. Just think of carbon resistors, oil-filled caps, electrolytics and especially some cables which are grossly inferior to standard parts.
The claimed superiority only relates to unmeasurable parameters like sonic behaviour. And that is a subjective and very debatable parameter.
So, apart from C4 and C5 all other caps in the circuit may happily be specified far below the kV-range.
For R2 and R3 (HV-Side) noninductive wound wire resistances are ok
Similar to the caps it may be necessary or even needed to series connect a couple of resistors. I wouldn´t use thickfilm resistors here and the probabely best Rs -made from metal foil- will be very expensive
On the lowvoltage side R1 and R4 can be of Metal foil build. The probabely best resistors here are the Isabellenhuette PBH and PBV and PSB. http://www.isabellenhuette.de
The HV-resistor R5 can be made from a chain of metal film resistors (0.5-0.6W, 1%-5%, size 0207). They are spec´d around 250V-300V, but in practise use them not above ~150V. For 5kV bias you need app 30pcs of 1MegOhm.
The original value of 500Megs is much higher than I´d use if a high ohmic coating is used on the diaphragm. 50Megs at maximum. You should be careful not to touch any HV-point, since the smaller resistance means higher currents and therefore more pain when getting shocked.
If someone reports sonic differences because of using a different resistor for R5 of same value, I´d say this would rather be a hint that the circuit itself is of minor design quality or psychology is playing games on him.
I´d rather make the HV stiffer by using larger Cap-values for C6 to C10 (>10nF) and use a negative polarity output instead of the positive. And adding a blinker cicuit (see Quad) will isolate the HV from the membrane.
The 10nF cap directly at the output of the cascade does no good. If you feel the need -a need that is actually nonexistent in praxis- that you have to smooth out the output than add a series resistor of 10M between output und 10nF-cap.
I don´t believe that a inductor could do anything positive here. The series resistance of R5 added to the high resistance value of the membrane coating works as ripple and noise filter already.
Besides the restrictions Mr. Beck mentioned (though I don´t see ´noise signals´ of sufficiently high frequency anyway) we are talking here about microscopically small currents. Which leads to the Q: If there were any ripple and noise above 100/120Hz.....would it supply for enough current in the inductor to make any difference to a resistor of same DC-ohmic-value? Or in other words:" Does the inductor work as inductor, or merely like a low-ohmic wire wound resistor?" Replacing R5 alltogether by a inductor shouldn´t be done because of safety reasons (current limiter).
jauu
Calvin
substitution of cap values 2+2 mod
Hi thanks again for your help.
I think I've found the parts I want to get.
The closest I've found for the 5000V ..01 uF is a 6000V (went over instead of under Voltage) .0047 uF. Polypropylene isnt available in this high of a voltage, but polyester is. Are either or both of these changes ok?
Also, the high frequency caps in the manual are 220uF,10uF, and .01uF.
I've seen in this forum to replace them with 50uF Blackgate, which I understand are discontinued. I have medallian and C updates. What values are people recommending.
Thanks again,
Paul
Hi thanks again for your help.
I think I've found the parts I want to get.
The closest I've found for the 5000V ..01 uF is a 6000V (went over instead of under Voltage) .0047 uF. Polypropylene isnt available in this high of a voltage, but polyester is. Are either or both of these changes ok?
Also, the high frequency caps in the manual are 220uF,10uF, and .01uF.
I've seen in this forum to replace them with 50uF Blackgate, which I understand are discontinued. I have medallian and C updates. What values are people recommending.
Thanks again,
Paul
Where are the measurements folks?!
I_Forgot did what folks should do, which is to use circuit theory to help guide the design process. The next step is to make the change guided by the math and then measure the change. Lets for a moment believe that the the choke and the cap before it made a difference, then that difference should be measurable. If you don't have the tools to measure it, then how to you know that you actually improved it? Ears are hardly reliable. And how do you optimize the change? Is 7H the right number? How about 5 or 10?
So until there are actual measurements, my advise is to stick with the stock acustat circuit.
Also 405man is right, the neon bulb circuit isn't my design, it's been around for decades.
Sheldon
I_Forgot did what folks should do, which is to use circuit theory to help guide the design process. The next step is to make the change guided by the math and then measure the change. Lets for a moment believe that the the choke and the cap before it made a difference, then that difference should be measurable. If you don't have the tools to measure it, then how to you know that you actually improved it? Ears are hardly reliable. And how do you optimize the change? Is 7H the right number? How about 5 or 10?
So until there are actual measurements, my advise is to stick with the stock acustat circuit.
Also 405man is right, the neon bulb circuit isn't my design, it's been around for decades.
Sheldon
Measurements... right!
You can make a string of HV high resistance resistors (say, 10 50-meg-ohm units from your local surplus place) and maybe a 10 meg resistor on the bottom. Then you can read proportional AC and DC voltages or 'scope the signal off the bottom resistor (the one with one end attached to ground, that is).
Glad to see that some sound back-of-the-envelope electrical judgment made the silly choke idea go poof.
The izzy wizzy HV cap, on the other hand, raises the bias supply. Which makes ESLs louder and hence sound better (until you put the SPL back for a fair A-B comparison and then they might sound the same).
Don't forget the old rule for working around HV: always keep one hand in your pocket.
You can make a string of HV high resistance resistors (say, 10 50-meg-ohm units from your local surplus place) and maybe a 10 meg resistor on the bottom. Then you can read proportional AC and DC voltages or 'scope the signal off the bottom resistor (the one with one end attached to ground, that is).
Glad to see that some sound back-of-the-envelope electrical judgment made the silly choke idea go poof.
The izzy wizzy HV cap, on the other hand, raises the bias supply. Which makes ESLs louder and hence sound better (until you put the SPL back for a fair A-B comparison and then they might sound the same).
Don't forget the old rule for working around HV: always keep one hand in your pocket.
some find it unbelievable....
but it can be done, has been done and is done. Kind of like paint by numbers without the numbers. Difficult yes that goes without saying but it is done. The best of all worlds is (I think) when both measurements and listening are used together as long as the ears have veto power. All the great masters built their instruments with their ears and their hands. That tradition holds to todays best as well.
Lets face it you can find good examples of both methods and this is not about right or wrong. Either method can suggest experimentation be done as long as the ears rule the day I dont think you can go wrong. You know what happens when the "bean counters" start to tell the engineers how to design a bridge.
but it can be done, has been done and is done. Kind of like paint by numbers without the numbers. Difficult yes that goes without saying but it is done. The best of all worlds is (I think) when both measurements and listening are used together as long as the ears have veto power. All the great masters built their instruments with their ears and their hands. That tradition holds to todays best as well.
Lets face it you can find good examples of both methods and this is not about right or wrong. Either method can suggest experimentation be done as long as the ears rule the day I dont think you can go wrong. You know what happens when the "bean counters" start to tell the engineers how to design a bridge.
stokessd said:
Sheldon, I agree in terms of designing anything that measurements have to be used to get in the ballpark so to speak. However, In this case I agree that I cannot think of a mechanism explaining what I hear. I had it demonstrated to me (blind) in a friends system and it is very obvious.
That said, it is easy to decide whether different is better if you live with the change for a while, and IMO making a good design better requires the use of the ears.
Rob.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.