|
|
Post by donh50 on Aug 31, 2022 14:35:28 GMT -5
Was it the 400 or 700 that had .......problems?......I thought it was the 400, and I know I've seen 'kits' of repair goodies...new boards and such of a newer design t 'fix' them. Heard it first many years go in a Pacific Stereo driving BOSE speakers.......The 901 from an early series.....Hung from the ceiling, of all things. I agree it SHOULD be in an amp manufacturers best interest to test into a real speaker. And I've seen many proposed 'dummy loads' online, but NO standards of which I'm aware. I suspect further that such a test would be for 'internal use only'......since nobody else publishes. I would suggest thee Audiograph System......Please see link... www.audiograph.se/wp-content/uploads/2018/10/PowerCube_12p_brochure_complete.pdfAs for 1 ohm Apogee? That would be the Scintilla. I think later they came out with one of a more reasonable impedance. The only amp capable of the original was from Lincoln Electric or Miller. I'm well aware of how awful many 'Stats can be. I'm not sure the very low power demand is not a red herring. After all, it is basically using your amp to charge a capacitor and by the time you do the power factor calculation, your amp is really not happy or in some cases.....capable...... The 700 had lots of problems, not fully resolved by the "A" version though I heard the "B" version was a little better. The main failure mechanism shorted the power rails to the speaker outputs so all too often the speakers went with the amp. The 400 I am less sure about, did not see as many of those, and in my limited experience were more reliable. That was early 1980's and I graduated into a different career after that, losing touch by 1990 or so with the start of our family (little time for audio). Phase Linear and Bose was a very popular combo for a while; I heard several PL700/Bose 901 systems back then. The 901's were pretty insensitive and the 700 was the biggest amp around for many folk (Ampzilla was another choice, though I saw far fewer, and it was more a "cult" thing despite its good performance -- and it had some issues as well...) The Power Cube is fairly well-known and AP offers it as an option, but it is very expensive. Amir (at ASR) looked into getting one but it was ~$30k IIRC. The FTC amplifier rules are under review for updating (finally!) but last I heard were still going to be somewhat watered down and no specific load included (a couple were proposed, including the one Stereophile uses). I put in my comments, as did many others, but I think the comment period is closed. The original Apogee was something like 0.1 ohms direct drive with an option for 1 ohms. Scintilla came after, and included the coupling transformer (no direct option) so was 1 ohms (with a 4-ohm option, I think). I'm a little fuzzy after that. Mark Levinson's little 25 W class A beasts would drive them, and I saw a Krell or two driving them. Been a long time. I have not seen amps self-destruct driving ESLs in many years, though some have overheated, I think due more to their low sensitivity than the wicked-low HF load. I never thought of calculating the power factor, an interesting idea, but again IIRC the phase angle is positive (inductive) not capacitive (negative) at the highest frequencies due to the transformer used for DC bias and to couple the signal to the panels. That should somewhat compensate the panel capacitance (I've no idea the actual PF). I used to throw a large cap across a 4-ohm resistor to check for stability before saying it would work on ESLs, but I was younger and didn't think of the transformer coupling. Still, provided some interesting displays from some amps... Sometimes people would throw an extra small (low-value) series resistor in the line as a band-aide but that also hosed the frequency response, the panels would tend toward "boomy", and panel modes were a bigger problem. That said I have not had ESLs in my system in years, and never actually owned a pair (though lived with many for a while either in my system or friend's). I just figured if they were blowing amps right and left there would be a bigger stink about them. FWIWFM - Don |
|
KeithL
Administrator 
Posts: 10,256
|
Post by KeithL on Aug 31, 2022 16:41:47 GMT -5
The bottom line is that you are correct... and amplifier manufacturers always test their products with a variety of real world speakers. One of the main reasons why nobody has ever been able to agree upon a specific "standard speaker load" is that there is so much variation in actual real-world speakers. It's also worth mentioning that, not only do various speakers sometimes have unusual impedance characteristics, but speaker drivers also generate back-EMF which an amplifier must handle appropriately as well. The result is a load that is far more complex than any simple combination of passive impedance characteristics. (Even the power cube, which addresses the range of possible impedance variations well, fails to address this aspect of the situation at all.) The upshot of all this is that, over the years, amplifier designers have learned to produce designs that are "stable into all relatively normal speaker loads". And, in turn, speaker manufacturers have learned to avoid producing designs that cannot be driven satisfactorily by most well designed amplifiers. And, to respond to your final point, there are very few modern solid state amplifiers that have trouble with "excessively low power demands" or "excessively high load impedances". (If anything I might suspect that with some early Class-D designs.) It's more likely to involve a combination of an unusually high or low load impedance that is also highly reactive as well. (Many tube amps used to have problems if run with insufficient loading - but this is rarely an issue with solid state amps.) To be fair here... If you're going to sell a speaker that will only run well with one or two specific amplifiers... Then you should sell them as a set... Or, at the very least, publish a list of the amplifiers that are known to work well with your speakers. I'm all in favor of electrostatic speakers that include an amplifier specifically optimized to work with them. (There are all sorts of potential benefits to doing it that way.) Was it the 400 or 700 that had .......problems?......I thought it was the 400, and I know I've seen 'kits' of repair goodies...new boards and such of a newer design t 'fix' them. Heard it first many years go in a Pacific Stereo driving BOSE speakers.......The 901 from an early series.....Hung from the ceiling, of all things. I agree it SHOULD be in an amp manufacturers best interest to test into a real speaker. And I've seen many proposed 'dummy loads' online, but NO standards of which I'm aware. I suspect further that such a test would be for 'internal use only'......since nobody else publishes. I would suggest thee Audiograph System......Please see link... www.audiograph.se/wp-content/uploads/2018/10/PowerCube_12p_brochure_complete.pdfAs for 1 ohm Apogee? That would be the Scintilla. I think later they came out with one of a more reasonable impedance. The only amp capable of the original was from Lincoln Electric or Miller. I'm well aware of how awful many 'Stats can be. I'm not sure the very low power demand is not a red herring. After all, it is basically using your amp to charge a capacitor and by the time you do the power factor calculation, your amp is really not happy or in some cases.....capable...... |
|
KeithL
Administrator
Posts: 10,256
|
Post by KeithL on Aug 31, 2022 16:55:29 GMT -5
The main problem with the Phase Linear 700 was heat sinks... not nearly enough of them... They were apparently able to deliver plenty of power when playing music... But were unable to survive the preconditioning period required by the FTC test spec current at that time without overheating... Was it the 400 or 700 that had .......problems?......I thought it was the 400, and I know I've seen 'kits' of repair goodies...new boards and such of a newer design t 'fix' them. Heard it first many years go in a Pacific Stereo driving BOSE speakers.......The 901 from an early series.....Hung from the ceiling, of all things. I agree it SHOULD be in an amp manufacturers best interest to test into a real speaker. And I've seen many proposed 'dummy loads' online, but NO standards of which I'm aware. I suspect further that such a test would be for 'internal use only'......since nobody else publishes. I would suggest thee Audiograph System......Please see link... www.audiograph.se/wp-content/uploads/2018/10/PowerCube_12p_brochure_complete.pdfAs for 1 ohm Apogee? That would be the Scintilla. I think later they came out with one of a more reasonable impedance. The only amp capable of the original was from Lincoln Electric or Miller. I'm well aware of how awful many 'Stats can be. I'm not sure the very low power demand is not a red herring. After all, it is basically using your amp to charge a capacitor and by the time you do the power factor calculation, your amp is really not happy or in some cases.....capable...... The 700 had lots of problems, not fully resolved by the "A" version though I heard the "B" version was a little better. The main failure mechanism shorted the power rails to the speaker outputs so all too often the speakers went with the amp. The 400 I am less sure about, did not see as many of those, and in my limited experience were more reliable. That was early 1980's and I graduated into a different career after that, losing touch by 1990 or so with the start of our family (little time for audio). Phase Linear and Bose was a very popular combo for a while; I heard several PL700/Bose 901 systems back then. The 901's were pretty insensitive and the 700 was the biggest amp around for many folk (Ampzilla was another choice, though I saw far fewer, and it was more a "cult" thing despite its good performance -- and it had some issues as well...) The Power Cube is fairly well-known and AP offers it as an option, but it is very expensive. Amir (at ASR) looked into getting one but it was ~$30k IIRC. The FTC amplifier rules are under review for updating (finally!) but last I heard were still going to be somewhat watered down and no specific load included (a couple were proposed, including the one Stereophile uses). I put in my comments, as did many others, but I think the comment period is closed. The original Apogee was something like 0.1 ohms direct drive with an option for 1 ohms. Scintilla came after, and included the coupling transformer (no direct option) so was 1 ohms (with a 4-ohm option, I think). I'm a little fuzzy after that. Mark Levinson's little 25 W class A beasts would drive them, and I saw a Krell or two driving them. Been a long time. I have not seen amps self-destruct driving ESLs in many years, though some have overheated, I think due more to their low sensitivity than the wicked-low HF load. I never thought of calculating the power factor, an interesting idea, but again IIRC the phase angle is positive (inductive) not capacitive (negative) at the highest frequencies due to the transformer used for DC bias and to couple the signal to the panels. That should somewhat compensate the panel capacitance (I've no idea the actual PF). I used to throw a large cap across a 4-ohm resistor to check for stability before saying it would work on ESLs, but I was younger and didn't think of the transformer coupling. Still, provided some interesting displays from some amps... Sometimes people would throw an extra small (low-value) series resistor in the line as a band-aide but that also hosed the frequency response, the panels would tend toward "boomy", and panel modes were a bigger problem. That said I have not had ESLs in my system in years, and never actually owned a pair (though lived with many for a while either in my system or friend's). I just figured if they were blowing amps right and left there would be a bigger stink about them. FWIWFM - Don |
|
|
|
Post by leonski on Aug 31, 2022 18:41:23 GMT -5
Was it the 400 or 700 that had .......problems?......I thought it was the 400, and I know I've seen 'kits' of repair goodies...new boards and such of a newer design t 'fix' them. Heard it first many years go in a Pacific Stereo driving BOSE speakers.......The 901 from an early series.....Hung from the ceiling, of all things. I agree it SHOULD be in an amp manufacturers best interest to test into a real speaker. And I've seen many proposed 'dummy loads' online, but NO standards of which I'm aware. I suspect further that such a test would be for 'internal use only'......since nobody else publishes. I would suggest thee Audiograph System......Please see link... www.audiograph.se/wp-content/uploads/2018/10/PowerCube_12p_brochure_complete.pdfAs for 1 ohm Apogee? That would be the Scintilla. I think later they came out with one of a more reasonable impedance. The only amp capable of the original was from Lincoln Electric or Miller. I'm well aware of how awful many 'Stats can be. I'm not sure the very low power demand is not a red herring. After all, it is basically using your amp to charge a capacitor and by the time you do the power factor calculation, your amp is really not happy or in some cases.....capable...... The 700 had lots of problems, not fully resolved by the "A" version though I heard the "B" version was a little better. The main failure mechanism shorted the power rails to the speaker outputs so all too often the speakers went with the amp. The 400 I am less sure about, did not see as many of those, and in my limited experience were more reliable. That was early 1980's and I graduated into a different career after that, losing touch by 1990 or so with the start of our family (little time for audio). Phase Linear and Bose was a very popular combo for a while; I heard several PL700/Bose 901 systems back then. The 901's were pretty insensitive and the 700 was the biggest amp around for many folk (Ampzilla was another choice, though I saw far fewer, and it was more a "cult" thing despite its good performance -- and it had some issues as well...) The Power Cube is fairly well-known and AP offers it as an option, but it is very expensive. Amir (at ASR) looked into getting one but it was ~$30k IIRC. The FTC amplifier rules are under review for updating (finally!) but last I heard were still going to be somewhat watered down and no specific load included (a couple were proposed, including the one Stereophile uses). I put in my comments, as did many others, but I think the comment period is closed. The original Apogee was something like 0.1 ohms direct drive with an option for 1 ohms. Scintilla came after, and included the coupling transformer (no direct option) so was 1 ohms (with a 4-ohm option, I think). I'm a little fuzzy after that. Mark Levinson's little 25 W class A beasts would drive them, and I saw a Krell or two driving them. Been a long time. I have not seen amps self-destruct driving ESLs in many years, though some have overheated, I think due more to their low sensitivity than the wicked-low HF load. I never thought of calculating the power factor, an interesting idea, but again IIRC the phase angle is positive (inductive) not capacitive (negative) at the highest frequencies due to the transformer used for DC bias and to couple the signal to the panels. That should somewhat compensate the panel capacitance (I've no idea the actual PF). I used to throw a large cap across a 4-ohm resistor to check for stability before saying it would work on ESLs, but I was younger and didn't think of the transformer coupling. Still, provided some interesting displays from some amps... Sometimes people would throw an extra small (low-value) series resistor in the line as a band-aide but that also hosed the frequency response, the panels would tend toward "boomy", and panel modes were a bigger problem. That said I have not had ESLs in my system in years, and never actually owned a pair (though lived with many for a while either in my system or friend's). I just figured if they were blowing amps right and left there would be a bigger stink about them. FWIWFM - Don When I bought my MG1 panels, the owners were testing a homebuilt amp with some stats. It emitted an ultraonic chirp and that was IT. Power Factor is a simple calculation. Know the phase angle? I think at that point a hand calcuator with Cosine helps........It's been years for me..... Other than that? For a power amp, maybe a Kill-A-Watt meter which is still inexpenve and reads Watts and KvA. ......and a direct calculated PF as a result.... I'd start with a 1khz test tone......and go on from there. For an amp / speaker, the power cube is good. Can it be LEASED? You don't need to buy it and when I was in-fab, much of our equipment was leased..... The ONLY people concerned with ESL speakers having problems with amps are the few it happens to. If I was going for Pot-Limit? I'd consider the Sanders Mag Tech amp. Bulletproof.....And $$$ |
|
|
|
Post by donh50 on Sept 4, 2022 9:15:59 GMT -5
When I bought my MG1 panels, the owners were testing a homebuilt amp with some stats. It emitted an ultraonic chirp and that was IT. Power Factor is a simple calculation. Know the phase angle? I think at that point a hand calcuator with Cosine helps........It's been years for me..... Other than that? For a power amp, maybe a Kill-A-Watt meter which is still inexpenve and reads Watts and KvA. ......and a direct calculated PF as a result.... I'd start with a 1khz test tone......and go on from there. For an amp / speaker, the power cube is good. Can it be LEASED? You don't need to buy it and when I was in-fab, much of our equipment was leased..... The ONLY people concerned with ESL speakers having problems with amps are the few it happens to. If I was going for Pot-Limit? I'd consider the Sanders Mag Tech amp. Bulletproof.....And $$$ Guess I should have said I have not seen "commercial or competent DIY amps" in many years that had problems with ESLs. I understand PF just never thought to apply it to speakers. The Kill-A-Watt is for wall power; I am not sure its bandwidth but suspect it is not very high, probably not enough for audio use. Also not sure how sensitive it is so if it would handle the low voltages most speakers receive at nominal loudness. No idea if the Power Cube can be leased. Not something I would do (not my day job, not doing audio measurements these days). And @keithl makes a good point that back-EMF is not well-modeled by most speaker loads. I would expect low enough amplifier output impedance would handle back EMF just fine, though amplifier impedance rises with frequency as feedback falls off. I do not know the magnitude of the effect vis-a-vie ESLs, e.g. how much the large panel and transformer coupling contribute. Not something I have researched in many years (my day job is GHz stuff, not audio). I remember trying to measure it decades ago but do not remember the result. And far too lazy to dig up my grad acoustics book that had a chapter on ESLs and the physics behind them (though I think it did discuss back-EMF). Of course, the people concerned about any problem are the ones to whom it happens.  But the people with problems also tend to post about it; it is the ones with no problems you rarely hear from. I think these days if a large number of folk were blowing up amplifiers driving ESLs we'd hear about it. Obviously your opinion is different, no worries. I first met Roger Sanders when he was in GA; he actually lives just an hour or two away on CO now, but I have not visited though have exchanged a few emails. His ESLs (and amps) were on my short list but I went a different direction, away from panels, the last time I bought speakers. |
|
|
|
Post by leonski on Sept 7, 2022 21:57:02 GMT -5
Ill-a-watt for wall use only.
It is setup for 60hz only and as such doesn't really need 'bandwidth'.....
PF IS of importance to speakers. 60 degrees is 1/2 the power generated is NOT actually delivered to the load.
In commercial power distribution, a factory which has a bad power factor.....say below .9 or so? may be charged a surcharge on their bill.....
EU wall wart I think are mandated to be 'power line friendly' so the load stays in sync.
Yes, if Rdon of an amp (resistance of the device when ON) was ZERO than damping would be very high indeed.... As a test? Bridge your speaker outputs with a paper clip.
Flick the woofer with your finger. Listen an observe the dampin. Remove the clip and flick again. Much different sound caused by LACK of damping.....except th internal damping
generated by the speaker.....
|
|
KeithL
Administrator
Posts: 10,256
|
Post by KeithL on Sept 8, 2022 9:33:38 GMT -5
I'm going to jump in here... just to clarify a distinction. When you're discussing AC line power - Power Factor (PF) is indeed a big deal. There are several reasons for this - all more or less related to "efficiently getting the power to the load". And, yes, virtually all big industrial machinery is either designed to have a good power factor, or has built in "power factor correction". And, at the level of a national power grid, all of those little wall warts add up to a major part of the overall load. (The better the PF, the more efficiently power is delivered to the load, so it can be both used and billed for, and losses in the transmission lines are reduced... ) (Generators also have issues driving loads with unusual or varying PF.) HOWEVER, this is NOT why damping factor matters with loudspeakers. (To put it bluntly, the most efficient moving coil loudspeaker is less than 5% efficient, so "electrical efficiency" is a non-issue with amplifiers and speakers.) There is really one major reason why Damping Factor matters with loudspeakers. And it can indeed be demonstrated by shorting your speaker terminals with a paper clip. A moving coil loudspeaker cone acts as "both a motor and a generator". As you drive the speaker, current in the voice coil makes a magnetic field, which pushes against the fixed magnet, and moves the cone (the "motor"). However, when you stop driving the speaker, and the signal voltage drops to zero, the cone has a tendency to keep moving due to stored energy (momentum). And, at that point, the cone is pushing the voice coil through the magnetic field created by the fixed magnet, and you have "a generator". And, as you may or may not know, when you short circuit the output of a generator, you oppose its movement, causing it to slow down. The energy of the momentum is converted into electrical energy and eventually heat - in a process sometimes referred to as "dynamic braking". When you short the speaker terminals with that paper clip you are "enabling dynamic braking for the cone of the loudspeaker" by "shorting the output of the generator". This effect applies mostly to the woofer - although it can apply to a lesser degree to other drivers. (The woofer is the heaviest driver, so has the most stored energy, and it is also connected most directly to the speaker terminals.) THIS is the effect that causes some speakers to deliver "clean tight bass" with an amplifier with a high DF and "sloppy bass" when connected to one with a low DF. Damping Factor is defined as the ratio between the output impedance of the amplifier and the impedance of the speaker. So, for example, if you have an 8 Ohm speaker, and the output impedance of the amplifier is 0.08 Ohms, then the DF = 100 ( 8 / 0.08 ). However, as usual, things get a bit more complicated. First, for the purposes of this calculation, both the resistance of your speaker cables, and the wiring inside the speaker itself, are ADDED to the output impedance of the amp. And, second, the amplifier itself is an ACTIVE DEVICE, so you cannot "just measure the resistance across the output terminals", but must instead perform a dynamic measurement under power. (You are not simply "shorting the terminals"; instead you are "actively driving them to zero volts".) Without going into detail - when you have an amplifier with feedback the effective output impedance of the amplifier is lowered by a ratio equivalent to the amount of feedback applied. So, if your output devices have a measured Rdon of 0.1 Ohms, and you have 100x feedback, the effective output impedance of your amp becomes 0.001 Ohms. This is why modern solid state amps, which tend to use significant amounts of feedback, can so easily deliver really excellent DF's, and so control loudspeakers so well. Note that the amount of effect this all has on the sound of the loudspeaker will depend on several factors... Including the mass of the cone, the efficiency of "the motor mechanism", how much mechanical damping it also has, and the crossover circuitry, to name just a few. That other reason why load impedance matters with amplifiers is both simpler in concept and more complex in practice. It is simply that some amplifiers have issues when driving loads with unusual power factors. These issues range from slew rate current limiting, to basic stability issues, and can affect the amplifier in various ways. But they all work out to "distortion, ringing, or other unusual behavior, when presented with excessively reactive loads". (This all comes under the broad heading of "stable when operating into reactive loads".) Ill-a-watt for wall use only. It is setup for 60hz only and as such doesn't really need 'bandwidth'..... PF IS of importance to speakers. 60 degrees is 1/2 the power generated is NOT actually delivered to the load. In commercial power distribution, a factory which has a bad power factor.....say below .9 or so? may be charged a surcharge on their bill..... EU wall wart I think are mandated to be 'power line friendly' so the load stays in sync. Yes, if Rdon of an amp (resistance of the device when ON) was ZERO than damping would be very high indeed.... As a test? Bridge your speaker outputs with a paper clip. Flick the woofer with your finger. Listen an observe the dampin. Remove the clip and flick again. Much different sound caused by LACK of damping.....except th internal damping generated by the speaker..... |
|
KeithL
Administrator
Posts: 10,256
|
Post by KeithL on Sept 8, 2022 10:00:32 GMT -5
Something like a Kill-A-Watt meter is designed specifically to work with 50/60 Hz power. It's also probably designed and calibrated to work with voltages in the normal range of line voltages. It would be a VERY bad idea to connect one to the output of an amplifier. You could actually damage something - and would almost certainly get inaccurate results. You can make that sort of measurements using a proper shunt resistor and a properly rated DVM - if you know what you're doing. Note that, unless your meter specifies a rated frequency range, it is NOT rated for 20 - 20 kHz. MOST DVMs are only accurate over a relatively narrow range around 50/60 Hz. Meters that are accurate over the full range of audio frequencies will say so in the specs... If you really want to measure speaker impedances your best bet is to buy one of those "little black box speaker tester gadgets". (The ones that connect to your computer... and usually include other useful stuff as well.) Electrostatic speakers tend to be a particularly nasty load. The panels themselves are pure capacitors (which is not a happy load for most amplifiers). But, since the panels require several thousand volts of audio drive, they are almost always used with a step-up transformer. (Electrostatic panels require both a high-voltage DC bias voltage and a high-voltage audio drive signal.) The transformer on its own is inductive, but it reacts with the panel, with the result being a much more complex load. (Often purely capacitive at some frequencies and purely inductive at others; and the reactance of the panel is multiplied by the turns ratio of the transformer.) In addition to that, if you accidentally cause the amp to clip, the transformer may produce dangerously high voltage transients. (The sort of transients that can destroy the amplifier and even the transformer itself.) Most commercial electrostatic loudspeakers are designed to provide at least a more or less sane load for an amplifier. (But you DO NOT want to muck around inside them, or modify them, or build your own, unless you REALLY know what you're doing.) The 700 had lots of problems, not fully resolved by the "A" version though I heard the "B" version was a little better. The main failure mechanism shorted the power rails to the speaker outputs so all too often the speakers went with the amp. The 400 I am less sure about, did not see as many of those, and in my limited experience were more reliable. That was early 1980's and I graduated into a different career after that, losing touch by 1990 or so with the start of our family (little time for audio). Phase Linear and Bose was a very popular combo for a while; I heard several PL700/Bose 901 systems back then. The 901's were pretty insensitive and the 700 was the biggest amp around for many folk (Ampzilla was another choice, though I saw far fewer, and it was more a "cult" thing despite its good performance -- and it had some issues as well...) The Power Cube is fairly well-known and AP offers it as an option, but it is very expensive. Amir (at ASR) looked into getting one but it was ~$30k IIRC. The FTC amplifier rules are under review for updating (finally!) but last I heard were still going to be somewhat watered down and no specific load included (a couple were proposed, including the one Stereophile uses). I put in my comments, as did many others, but I think the comment period is closed. The original Apogee was something like 0.1 ohms direct drive with an option for 1 ohms. Scintilla came after, and included the coupling transformer (no direct option) so was 1 ohms (with a 4-ohm option, I think). I'm a little fuzzy after that. Mark Levinson's little 25 W class A beasts would drive them, and I saw a Krell or two driving them. Been a long time. I have not seen amps self-destruct driving ESLs in many years, though some have overheated, I think due more to their low sensitivity than the wicked-low HF load. I never thought of calculating the power factor, an interesting idea, but again IIRC the phase angle is positive (inductive) not capacitive (negative) at the highest frequencies due to the transformer used for DC bias and to couple the signal to the panels. That should somewhat compensate the panel capacitance (I've no idea the actual PF). I used to throw a large cap across a 4-ohm resistor to check for stability before saying it would work on ESLs, but I was younger and didn't think of the transformer coupling. Still, provided some interesting displays from some amps... Sometimes people would throw an extra small (low-value) series resistor in the line as a band-aide but that also hosed the frequency response, the panels would tend toward "boomy", and panel modes were a bigger problem. That said I have not had ESLs in my system in years, and never actually owned a pair (though lived with many for a while either in my system or friend's). I just figured if they were blowing amps right and left there would be a bigger stink about them. FWIWFM - Don When I bought my MG1 panels, the owners were testing a homebuilt amp with some stats. It emitted an ultraonic chirp and that was IT. Power Factor is a simple calculation. Know the phase angle? I think at that point a hand calcuator with Cosine helps........It's been years for me..... Other than that? For a power amp, maybe a Kill-A-Watt meter which is still inexpenve and reads Watts and KvA. ......and a direct calculated PF as a result.... I'd start with a 1khz test tone......and go on from there. For an amp / speaker, the power cube is good. Can it be LEASED? You don't need to buy it and when I was in-fab, much of our equipment was leased..... The ONLY people concerned with ESL speakers having problems with amps are the few it happens to. If I was going for Pot-Limit? I'd consider the Sanders Mag Tech amp. Bulletproof.....And $$$ |
|
|
|
Post by leonski on Sept 8, 2022 19:23:03 GMT -5
First? Agreed. A Kill-A-Watt meter as a test on your speakers is a bad idea. Electrically and of questionable value as to just WHAT you are reading.
But it SHOULD work for your amp, especially if you play a single frequency test tone at a time.
I once found some 'stats at a Swap Meet. They were in Rough Shape and one of the power supply boxes had been opened and 'fiddled with'.
The owner wanted an astronomical 400$ for what, IMO were a pair of speakers good for parts. Maybe. No Sale.
If memory serves, they were Acoustat.
The reason PF for a speaker is of note is that it just is not 'IT'. The amplifier must be considered. And any amp not capable into the combination of phase and impedance
should be discarded.
That's where the Power Cube system comes in. It stressed an amp at (example) 3 impedances into a RESISTOR. this is a standard test. But than? The test is repeated
at +-30 degrees and +-45 degrees AND at all tested impedences. This represents Capacitiva and inductive loads. Once you know what the amplifier is capable of, you
can compare to the measured data for the speaker. It's tough to come up with 'one number' for a speaker, but in general? Look for worst case. Or a trend, like 2 or 3 octaves
below 3 ohms. Or in the case of what Stereophile looks for? A Combination of low impedance and high phase angle.
The one-number idea is important. Not really meaningful in this case. You kind of need to know highs and lows.
Once you know what you are looking at, the graph procduced in the power cube system makes perfect sense.
The worst amps are HT amps, in general. Sure, power increases as impedance drops. But some of 'em include protection featuers like an impedance switch or a simply incapable
power supply. My Magnepan speakers are low sensitivity, but not really a bad load. Yet I don't know that any HT receiver would make friends with them. Last time I visited a store
selling HT receivers, I approached the display and could FEEL the heat radiating. A few of 'em were really too hot to touch on top. I'd give 'em a week with a 5.1 panel setup.
|
|
KeithL
Administrator
Posts: 10,256
|
Post by KeithL on Sept 9, 2022 10:23:28 GMT -5
Magneplanars are really a very friendly load - and almost purely resistive... They just like lot of power. And you're quite right... most AVRs don't like relatively low impedance loads. The reason is quite simple... and it has to do with efficiency at relatively low power levels. In general music is quite dynamic. As a quick approximation we say that, with typical music content, the average signal level will be between 1/10 and 1/20 of the level of the highest peak. This means that, if you have an amplifier that clips at 200 watts, and you play it as loudly as you can without any clipping, on average you are using between 10 and 20 watts. But here's the catch... A typical Class A/B amplifier is about 70% efficient AT FULL POWER. However, as you lower the output power, the efficiency of a Class A/B amplifier is also reduced. NOTE that this reduction in efficiency at reduced power levels is NOT some sort of design problem or deficiency that certain amplifiers suffer from. It is INHERENT and UNAVOIDABLE in the way linear output stages work (which includes Class A/B and Class B designs). It can be made even worse by poor design - but it cannot be improved or eliminated. (And, in case you were wondering, Class A amplifiers are MUCH less efficient over their entire operating range.) So, at lower power outputs, the amplifier is delivering less power, but may well be "burning" more power, and the difference ends up as heat. As it turns out, with a typical Class A/B amplifier, running into an 8 Ohm load, the AMPLIFIER will generate the most HEAT when delivering about 1/3 of its rated maximum power. And this situation is even more extreme if that same amplifier is connected to a 4 Ohm load. (You can basically optimize the efficiency of a design for either 4 Ohms or 8 Ohms... but not for both.) And THIS is why an amplifier is going to run so hot when driving a low impedance speaker that also happens to need a lot of power. And it's why AVRs, which tend to NOT have an awful lot of heat sinks, tend to run so hot and have other problems in situations like that. (The only "solution" is to provide enough heat sinks, or air flow using fans or other methods, to effectively dissipate the heat that is generated.) Also note that our Class H output topology does in fact reduce this effect to some degree - although not entirely. It's also worth mentioning that this has nothing whatsoever to do with the efficiency of the power supply. It is strictly related to how the amplifier circuitry itself does what it does. (And Class-D amplifiers operate in an entirely different way - which is not subject to this relationship between power and efficiency.) First? Agreed. A Kill-A-Watt meter as a test on your speakers is a bad idea. Electrically and of questionable value as to just WHAT you are reading. But it SHOULD work for your amp, especially if you play a single frequency test tone at a time. I once found some 'stats at a Swap Meet. They were in Rough Shape and one of the power supply boxes had been opened and 'fiddled with'. The owner wanted an astronomical 400$ for what, IMO were a pair of speakers good for parts. Maybe. No Sale. If memory serves, they were Acoustat. The reason PF for a speaker is of note is that it just is not 'IT'. The amplifier must be considered. And any amp not capable into the combination of phase and impedance should be discarded. That's where the Power Cube system comes in. It stressed an amp at (example) 3 impedances into a RESISTOR. this is a standard test. But than? The test is repeated at +-30 degrees and +-45 degrees AND at all tested impedences. This represents Capacitiva and inductive loads. Once you know what the amplifier is capable of, you can compare to the measured data for the speaker. It's tough to come up with 'one number' for a speaker, but in general? Look for worst case. Or a trend, like 2 or 3 octaves below 3 ohms. Or in the case of what Stereophile looks for? A Combination of low impedance and high phase angle. The one-number idea is important. Not really meaningful in this case. You kind of need to know highs and lows. Once you know what you are looking at, the graph procduced in the power cube system makes perfect sense. The worst amps are HT amps, in general. Sure, power increases as impedance drops. But some of 'em include protection featuers like an impedance switch or a simply incapable power supply. My Magnepan speakers are low sensitivity, but not really a bad load. Yet I don't know that any HT receiver would make friends with them. Last time I visited a store selling HT receivers, I approached the display and could FEEL the heat radiating. A few of 'em were really too hot to touch on top. I'd give 'em a week with a 5.1 panel setup. |
|
KeithL
Administrator
Posts: 10,256
|
Post by KeithL on Sept 9, 2022 10:39:50 GMT -5
When it comes to electrostatic speakers the DIY situation is quite interesting. You have three main parts: - the panels themselves - the audio signal portion (usually special signal transformers) - the bias circuitry The bias circuitry is relatively trivial. It generates a very high DC voltage, but at very low current, and can be replaced or repaired pretty easily (and cheaply). The signal transformers are extremely specialized parts that tend to be quite expensive and may be difficult or impossible to repair or replace. but (#1) - some DIY folks would want to use a custom-built amplifier that directly drives the panels at high voltage without transformers. but (#2) - this means that, if they are intact, the transformers may be "the one part that really matters". The panels are a more interesting subject. They are both easily damaged and subject to failure or deterioration over time. (They're also rather delicate, and quite large, which combination makes them difficult to ship.) There's a good chance that it may be impossible to purchase replacements and you may or may not be able to pay someone to repair them. (Replacement panels for some brands and models are readily available... as are "rebuild services" for some... but both tend to be expensive.) Many of the panels CAN be repaired or "rebuilt" by someone with the knowledge to do so. (Rebuilding panels requires a lot of labor but the materials themselves are readily available and quite cheap.) Many DIY folks WANT to build their own panels (and so would prefer to purchase "a set of working or easily repairable electronics" to go with homebuilt panels). (....) I once found some 'stats at a Swap Meet. They were in Rough Shape and one of the power supply boxes had been opened and 'fiddled with'. The owner wanted an astronomical 400$ for what, IMO were a pair of speakers good for parts. Maybe. No Sale. If memory serves, they were Acoustat. (...) |
|
|
|
Post by leonski on Sept 9, 2022 15:14:50 GMT -5
I've looked IN DETAIL at the specs for 'D' amps. Yes....they are rated to 90% + efficiency but At Full Power.
Efficiency goes down with power consumed. So even though the housekeeping power input may be 2 or 3 watts at Zero output,
that also means Zero efficiency.
I'd guess that at 1 or 2 watts, the efficiency is limited by the power supply, not the output or its class or operation...
I had no way on the spot to evaluate the 'Stats I saw. But they were in rough shape. If they guy had said 100$, I might have
taken a chance. But the 'ask' was ridiculous.
My experinnce with really HV power supplies makes me leery of any DIY approach. the Glassman gear we had in our implanters once
failed. A LOT of space inside to prevent arcing, Very special and HIGH voltage parts / capacitors. We once blew UP a Beckman with the
HV probe. We had it sitting inside the implanter and had defeated the safeties. We stationed a couple guys at a distance FROM the implanter
and would let nobody NEAR. But when we pressed 'go' it ARCED and blew the meter to charcoal. Quite a noise, too.
I don't even blink at Vacuum Tube amps with a trivial 500 to 600 volts maximum......
In an implanter? Extraction was fixed at 20kv, IIRc whle the acceleration voltage was variable up to nearly 200kv. There is circuitry to measure
Beam Current but its been a while so I don't remember much. Certainly in the MA region.
|
|
But the people with problems also tend to post about it; it is the ones with no problems you rarely hear from. I think these days if a large number of folk were blowing up amplifiers driving ESLs we'd hear about it. Obviously your opinion is different, no worries.