The ohms Trick

[Leporello]

I think I'll get as good a performance from the A31 dishing out 400W @4ohms

I think you're right.

 

[Leporello]

So, more about the ohms thing: regardless of what some people say about an ESL needing lots of current or, occasionally, voltage, most loudspeakers are designed so that a flat frequency response of voltage across the input terminals will result in the desired acoustical frequency response (whether that desired response is flat on-axis in an anechoic chamber is another long story). An amplifier rated at 200W continuous into 8 ohms will produce 40 volts RMS into that load. A typical solid state amp will be down a fraction of a dB from DC to daylight, so essentially puts out this 40 volts into 8 ohms across the entire audio spectrum. An ideal voltage source (a fictional construct) would put out the same 40 volts into 4 ohms, hence double the power, and again into 2 ohms. It would fuse a railroad spike across the terminals in order to maintain 40 volts. In the real world, something eventually has to give, whether it's the power supply or the inherent non-zero impedance of the amplifying devices.. That said, an amp that can produce quadruple the power into 2 ohms as it can into 8 is a good real world approximation of an ideal voltage source. The reason you want this is not to squeeze the maximum power out of your amp for power hungry speakers (though you probably want that too) but to maintain the designed frequency response. If your speakers have an impedance of 4 ohms at one frequency and 1 ohm at another, the response is going to dip down from what Martin Logan designed.

You may, in fact, like this softening of the response, in which case, go for it. Only you can decide. The Latino monoblocks, that I had a blast building, don't come close to ideal voltage sources. To my ears the effect is not so much a veil as a heavy curtain bundled around the speakers. Not all tube amps will have such a high effective output impedance, but probably all tube amps will compared to a competent solid state design. Thus, a vintage Haffler/Dynaco design is probably unsuitable for ML 'stats, unless you like a really rolled off sound.

Another factor is negative feedback, which many amps utilize in order to achieve such a low effective impedance as to approximate an ideal voltage source. The problem is under certain conditions, negative feedback can turn positive, resulting in oscillation--in the worst case, leading to self-destruction, but otherwise, producing "ringing" on transients and a corresponding harsh sound. So for 'stats you want an amp that's not only 2 ohm capable, but 2 ohm stable. Stability can be load dependent, and 'stats have the opposite voltage-current phase relationship at high frequencies from conventional speakers. This is not so much an issue with today's solid state designs, but if you're in the market for something vintage, it's something to check out.

 

[ejspain]

So, more about the ohms thing: regardless of what some people say about an ESL needing lots of current or, occasionally, voltage, most loudspeakers are designed so that a flat frequency response of voltage across the input terminals will result in the desired acoustical frequency response (whether that desired response is flat on-axis in an anechoic chamber is another long story). An amplifier rated at 200W continuous into 8 ohms will produce 40 volts RMS into that load. A typical solid state amp will be down a fraction of a dB from DC to daylight, so essentially puts out this 40 volts into 8 ohms across the entire audio spectrum. An ideal voltage source (a fictional construct) would put out the same 40 volts into 4 ohms, hence double the power, and again into 2 ohms. It would fuse a railroad spike across the terminals in order to maintain 40 volts. In the real world, something eventually has to give, whether it's the power supply or the inherent non-zero impedance of the amplifying devices.. That said, an amp that can produce quadruple the power into 2 ohms as it can into 8 is a good real world approximation of an ideal voltage source. The reason you want this is not to squeeze the maximum power out of your amp for power hungry speakers (though you probably want that too) but to maintain the designed frequency response. If your speakers have an impedance of 4 ohms at one frequency and 1 ohm at another, the response is going to dip down from what Martin Logan designed.

You may, in fact, like this softening of the response, in which case, go for it. Only you can decide. The Latino monoblocks, that I had a blast building, don't come close to ideal voltage sources. To my ears the effect is not so much a veil as a heavy curtain bundled around the speakers. Not all tube amps will have such a high effective output impedance, but probably all tube amps will compared to a competent solid state design. Thus, a vintage Haffler/Dynaco design is probably unsuitable for ML 'stats, unless you like a really rolled off sound.

Another factor is negative feedback, which many amps utilize in order to achieve such a low effective impedance as to approximate an ideal voltage source. The problem is under certain conditions, negative feedback can turn positive, resulting in oscillation--in the worst case, leading to self-destruction, but otherwise, producing "ringing" on transients and a corresponding harsh sound. So for 'stats you want an amp that's not only 2 ohm capable, but 2 ohm stable. Stability can be load dependent, and 'stats have the opposite voltage-current phase relationship at high frequencies from conventional speakers. This is not so much an issue with today's solid state designs, but if you're in the market for something vintage, it's something to check out.

Good read.

 

[Heyfey]

I understand the desire to have an amp that will double power output as resistance halves, but in the case of ML stats they are typically specified as reaching the 2-ohm range at/about 20kHz. For real-world music, how much program material is there around these (to me) ultrasonic frequencies? I know there are harmonics and overtones up there, but I can't help but wonder if this is a more theoretical issue than a real one for most of us. Am I off base?

 

[heuboda]

Some years ago I have measured the impedance of my old Summit X and the impedance was about 0,7 Ohms at 22,3kHz. I am afraid that the "new" speaker I got, the Renaissance, will even get lower than that. But I have a tube amp and it has only an 4 Ohm output and I know that it works well in this combination. I also tried MC 1,2kW non tube amps with 2 Ohms output and there was no difference in sound quality. At least not for my ears ...

 

[Nuri58]

And the wattage for high frequency is very low beside EL panels do not draw a lot of power. Again if you read ML's own statement there is no real issue for 8 ohm receivers. I had a Maranz AVR thousand of years ago, which in the beginning shot down frequently when playing too loud - got better over time. I used it for a couple of weeks while wating for my Arcam AVR 300 (only rated at 8 Ohm). I meanwhile had the AVR 600 (also only rated a 8 Ohm), then the AVR850 rated at both 4 and 8 Ohms (no switch to change the Ohm setting like most AVRs/power amplifiers). Now I moved the fronts ftom the AVR 850 to my p429 in bridge mode (8 Ohm /275W). Whilst always biamping I now run it biwired). Never had an issue, no matter how loud i play (have limited max output to 75% - anything louder is both too loud and starts sounding like too high a load for the speakers). Parties goes on for many hours at this volume, without the amplifier runs hotter than you can put yout hand on it without it's hotter than you can leave your hand on it.

 

[ejspain]

I understand the desire to have an amp that will double power output as resistance halves, but in the case of ML stats they are typically specified as reaching the 2-ohm range at/about 20kHz. For real-world music, how much program material is there around these (to me) ultrasonic frequencies? I know there are harmonics and overtones up there, but I can't help but wonder if this is a more theoretical issue than a real one for most of us. Am I off base?

Great question. For me, I'm not "chasing the ohms" persay...knowing my 11As have a 4ohm impedance, I just want to make sure I'm feeding them enough, efficient, lowest THD power (I can afford) so they're able to perform their best. I also know they cap out at 550W so I don't want to overdrive them by wasting (what I feel) money on an amp with too much power. I'm a retired Navy Chief so I have thick skin...tell me I'm crazy but I think an amp capable of 250W@8 / 400W@4 - all channels driven, should make my Impressions very happy.

 

[Robert D]

Great question. For me, I'm not "chasing the ohms" persay...knowing my 11As have a 4ohm impedance, I just want to make sure I'm feeding them enough, efficient, lowest THD power (I can afford) so they're able to perform their best. I also know they cap out at 550W so I don't want to overdrive them by wasting (what I feel) money on an amp with too much power. I'm a retired Navy Chief so I have thick skin...tell me I'm crazy but I think an amp capable of 250W@8 / 400W@4 - all channels driven, should make my Impressions very happy.

That should be fine. I'm using an Aragon that's 200w/400w on my Prodigy speakers. The Prodigy panels are larger and I also have to power the two woofers. Mine sounds great.

 

[Russr]

It's not really about "stated doubling of power"... i mean, really... anyone else remember the Sony "boomboxes" of the 80's rated at 1,300WattsPMPO and other such garbage?!
I think you could say it's a function of design philosophy, and how much power reserve there was (a function of power supply capacitors, etc.). It's not a function of "Well, it's labelled at 8 ohms, so that means it's good for 2x at 4ohms (or 2, or 0.2...). I've tried running ESLs on fairly decent receivers, and the summary is it doesn't work out too well.

Now here's a question for the boffin-types: In voice-coil speakers, a struggling (overdriven) amplifier will output signals approximating square waves which fries voice coils. What happens with ESLs in a similar situation? There are lots of stories here on the board about "painful" audio, etc., but I've never read about "fried panels" or the like... is that because we're all too civilized to mention anything like that?

Russ

 

[ejspain]

That should be fine. I'm using an Aragon that's 200w/400w on my Prodigy speakers. The Prodigy panels are larger and I also have to power the two woofers. Mine sounds great.

Exactly. Another great point...the Impressions are dawned with 275W amps to drive the woofers so we're just driving the panels.

 

[audioxcel]

Aren't we overlooking the fact that high frequencies require less power than low frequencies. So, maybe a dip to 2 ohms at say 20K Hz really won't require more power as say 200 Hz at 8 ohms?

 

[ejspain]

Idk but at the end of the day its all about what the panels want and what sounds good to your ears. For years I drove my Aeon i's (4ohms) with an Outlaw amp and they loved the power but I hated the hiss (not a quiet amp). It eventually died and I picked up a Marantz M7055 (150W over 5 chnls @8ohms)...pretty decent amp but doesn't "do-the-dirty" (4ohms)....actually wasn't bad until I brought in the Impressions. Now good and broken in and they're making it pretty obvious they atleast want you to match the impedance and give me some dang power.

 

[Robert D]

Idk but at the end of the day its all about what the panels want and what sounds good to your ears. For years I drove my Aeon i's (4ohms) with an Outlaw amp and they loved the power but I hated the hiss (not a quiet amp). It eventually died and I picked up a Marantz M7055 (150W over 5 chnls @8ohms)...pretty decent amp but doesn't "do-the-dirty" (4ohms)....actually wasn't bad until I brought in the Impressions. Now good and broken in and they're making it pretty obvious they atleast want you to match the impedance and give me some dang power.

Is this the one youre thinking about? This should do good. Thats a low price at only $2300, especially at todays wacked out prices.

https://www.anthemav.com/products-current/type=amplifier/model=mca-225-gen2/page=specs

 

[Robert D]

This Parasound would be good too.

https://parasound.com/a21+.php
Its a bit more than the Anthem, $3500. Still a low price these days for a good stereo amp.

https://www.overtureav.com/high-performance-audio/power-amplifiers/parasound/parasound-halo-a-21/
I think thats the price range youre looking for?

 

[ejspain]

Is this the one youre thinking about? This should do good. Thats a low price at only $2300, especially at todays wacked out prices.

https://www.anthemav.com/products-current/type=amplifier/model=mca-225-gen2/page=specs

Nope. About to pull the trigger on the Parasound A31

 

[Robert D]

Nope. About to pull the trigger on the Parasound A31

That's good. I thought you wanted a 2 channel amp. The A31 is about like the a21+ that I linked above, with an extra channel. The a21+ has slightly more power per channel, but you'd probably be hard pressed to hear it I bet.

 

[Leporello]

It's not really about "stated doubling of power"... i mean, really... anyone else remember the Sony "boomboxes" of the 80's rated at 1,300WattsPMPO and other such garbage?!
I think you could say it's a function of design philosophy, and how much power reserve there was (a function of power supply capacitors, etc.). It's not a function of "Well, it's labelled at 8 ohms, so that means it's good for 2x at 4ohms (or 2, or 0.2...). I've tried running ESLs on fairly decent receivers, and the summary is it doesn't work out too well.

Now here's a question for the boffin-types: In voice-coil speakers, a struggling (overdriven) amplifier will output signals approximating square waves which fries voice coils. What happens with ESLs in a similar situation? There are lots of stories here on the board about "painful" audio, etc., but I've never read about "fried panels" or the like... is that because we're all too civilized to mention anything like that?

Russ

Voice coils and 'stats just have different agents of destruction. An amp above clipping is putting out much more RMS voltage than its peak times 0.707 would suggest, resulting in more voice coil burning watts (amp specs are for undistorted power). The main destructive force for ESL's is peak over-volting, resulting in arcs. Once an arc starts, breaking down the dielectric of the air, it doesn't want to stop. But that's different from what fries voice coils. Clipping could actually save an ESL, if it occurs below the (transformer stepped up) breakdown voltage of the air between the electrodes. Note, though, that a high amplitude square wave could also burn out the step-up transformer primary.

ML's are relatively "fry proof". They have to be played abusively loud. Not so for earlier ESL's like original Quads and Infinity Servo-Statics. Those were forever burning out panels. I think Acoustat invented the first fry-proof ESL.

 

[Leporello]

And the wattage for high frequency is very low beside EL panels do not draw a lot of power. Again if you read ML's own statement there is no real issue for 8 ohm receivers. I had a Maranz AVR thousand of years ago, which in the beginning shot down frequently when playing too loud - got better over time. I used it for a couple of weeks while wating for my Arcam AVR 300 (only rated at 8 Ohm). I meanwhile had the AVR 600 (also only rated a 8 Ohm), then the AVR850 rated at both 4 and 8 Ohms (no switch to change the Ohm setting like most AVRs/power amplifiers). Now I moved the fronts ftom the AVR 850 to my p429 in bridge mode (8 Ohm /275W). Whilst always biamping I now run it biwired). Never had an issue, no matter how loud i play (have limited max output to 75% - anything louder is both too loud and starts sounding like too high a load for the speakers). Parties goes on for many hours at this volume, without the amplifier runs hotter than you can put yout hand on it without it's hotter than you can leave your hand on it.

I can't imagine what would make an amp shut down initially but get better over time. I've experienced one getting worse over time.

I initially drove my CLS II's with a Cambridge Azure 850A. The CLS II has an abusively low impedance not only at high frequencies but, it turns out, at DC, around 2 ohm. the Cambridge has a short circuit protection circuit that initially applies a test voltage to the speaker terminals and if it gets sucked down too low, goes into protection mode. The amp seemed happy enough driving the speakers once it started. I could unplug a speaker, turn it on, plug it back in, and I would be OK. At first this happened occasionally, then nearly every time I tried to turn on the amp with speakers attached. This got old. Thanks to a schematic I found online, I could see that the voltage comparator that detected the short circuit protection was set for about 2 ohms. I opened it up and tacked an additional resistor to the upper arm of the voltage divider. That worked great, for a while, then it reverted back to its old behavior. It could be I miscalculated what wattage resistor was needed for my mod. But installing it in the first place was a major PITA, requiring removal of both heat sinks. I decided it was time to just get another amp, first the fun Latino tube amp, then the Parasound. I am now done with looking for amps for the CLS, for the foreseeable future.

 

[Nuri58]

I can't imagine what would make an amp shut down initially but get better over time. I've experienced one getting worse over time.

Yes always puzzled my. Mind you this was at lower volumes not near the peak, but at least i had sound while waithig my new device.

 

[legarem]

I finally used an 'old' Classe DR15 which was recapped with my CLSII. These amps are indestructible and superbly well made. My Melos Single ended 75W mono blocks can drive the CLSII but treble is lacking. The mids are glorious. I will probably adjust the caps paralleled on the negative feedback resistor so I could perhaps get more air in the treble.
I know single ended tube amps are not the best to drive ESLs