Complex Load for Power Amplifier torture testing

[SIY]

It would still be interesting to see what you come up with if you have the time

Happy to. I built the load last night, will run some tests this weekend.

 

[DonH56]

Random comment whilst wondering why I continue to read this train wreck... Something to do while my test is running.

ESLs are driven through a transformer (with a few very rare exceptions) thus their input impedance looks inductive even though the panel itself is a big capacitor. I really doubt any of the other ESLs are that much different in impedance than ML.

Very low impedance at the top of the audio band is usually not a problem in the real world because the signal content is so low in amplitude. Didn't we just have another big debate about frequency content over 20 kHz? It is there, but at a low level. That, and the rule that allows just an average impedance to be specified, lets ESL (etc.) speaker manufacturers get away with specifying 8-ohm nominal loads when we know durn well the impedance is well below that at certain frequencies (or ranges of frequencies). Still it's nice to know the amp is stable driving a low (perhaps reactive) load at high frequency. Just low impedance can be an issue since feedback factor is dropping, BJT-type output devices start to look inductive (assuming the emitters are driving the loads), etc.

I've noticed a number of conventional speaker designs exhibit impedance magnitude of 4 ohms or below for much of the lower midrange to upper bass region, leading to thermal (and perhaps current-limiting) problems for some amplifiers (and probably many AVRs). So a large-scale signal at 100 Hz (sine wave) into 2-4 ohms is another test that may be worth performing.

I am unaware of anyone requesting an audio amplifier be driven by a test signal up to 10 MHz. Achieving ruler-flat in-band response (to 20 kHz) requires -3 dB in the 100's of kHz range so testing with a signal having that sort of content (not a full-scale sine wave) is reasonable.

I always used to slap the output of an amp on a spectrum analyzer to check for high-frequency oscillations and occasionally saw junk in the 0.1-1+ MHz range. I have not done that in years. Back when "ultra-wideband" was all the range plenty of amps had problems driving reactive loads. And a few DC-coupled amps would motorboat in response to very low-frequency signals so I would do a sweep down to 1 Hz or below.

As a designer I expect none of the current crop of amplifiers would exhibit severe problems like oscillations and imagine they'll all shut themselves down cleanly in response to a current overload or over temp condition. And as a designer I know it is all too easy for some little thing to slip through and lead to an unstable design under some unexpected real-world condition. I have designed more than one good oscillator that was meant to be a linear amplifier... And of course a few oscillators that refused to start and made great amplifiers.

 

[DonH56]

An SCR/Crowbar IIRC. Nasty.

Some old Quad amps did the same IIRC. Plenty of big PA power amps would take DC to ground in a DC event (to break the relay contact arc), figuring the amp is already toast, might as well finish it off.

Wish my old Phase (Blaze, Flame) Linear 700 did that... Instead, it applied the supply voltage across the speaker terminals when the outputs shorted, usually frying the speaker before the fuses blew. OTOH I had to repair an old guitar amp that shorted the outputs like that and, thanks to the wisdom and technical genius of the guitar player who replaced the fuses with bus wire to "crank it up", took out the main transformer. I had forgotten just how badly those things smelled...

 

[SIY]

I really doubt any of the other ESLs are that much different in impedance than ML.

But they are, nonetheless. See the curves shown earlier for Quads (the most popular ESLs of all time). Unfortunately, the impedance curves I ran on my old Acoustat 1+1s are on some archived hard disk in my garage, but I bet someone else has posted them. They do not much resemble the M-L impedance.

Your point about signal magnitude is exactly... on point.

 

[Rja4000]

Audiograph Powercube and it costs US $15,000 for a stereo unit

Wasn't that $55k?
$15k is way more reasonable.

 

[ajawamnet]

I've got plenty of these little b#stards. The glue binding fails and the pages go everywhere if you are not careful. And I need a magnifying glass these days to read them.

View attachment 42096

Back when I was doing it they were full size... I recall ORA and MCM selling various other equiv. parts too.

 

[ajawamnet]

Wish my old Phase (Blaze, Flame) Linear 700 did that... Instead, it applied the supply voltage across the speaker terminals when the outputs shorted, usually frying the speaker before the fuses blew. OTOH I had to repair an old guitar amp that shorted the outputs like that and, thanks to the wisdom and technical genius of the guitar player who replaced the fuses with bus wire to "crank it up", took out the main transformer. I had forgotten just how badly those things smelled...

Yea - I recall the old DC protection relays in things like Marantz receivers. Anyone here recall the old STK00xx monolithic amp modules?

Speaking of guitar amps - I've fixed plenty of the Line 6 things that are the size of a Marshall, but the amplifier is actually a TDA7293 tiny little thing that is rated at 100W but it better be 8 ohm speaker cabs. They state 4 ohm for the IC but man, I've seen what a 4x12 4 ohm cab does to that little PCB...

Unlike a Mesa or Marshall that you can almost weld with.

Alerts

 

[March Audio]

Random comment whilst wondering why I continue to read this train wreck... Something to do while my test is running.

ESLs are driven through a transformer (with a few very rare exceptions) thus their input impedance looks inductive even though the panel itself is a big capacitor. I really doubt any of the other ESLs are that much different in impedance than ML.

Very low impedance at the top of the audio band is usually not a problem in the real world because the signal content is so low in amplitude. Didn't we just have another big debate about frequency content over 20 kHz? It is there, but at a low level. That, and the rule that allows just an average impedance to be specified, lets ESL (etc.) speaker manufacturers get away with specifying 8-ohm nominal loads when we know durn well the impedance is well below that at certain frequencies (or ranges of frequencies). Still it's nice to know the amp is stable driving a low (perhaps reactive) load at high frequency. Just low impedance can be an issue since feedback factor is dropping, BJT-type output devices start to look inductive (assuming the emitters are driving the loads), etc.

I've noticed a number of conventional speaker designs exhibit impedance magnitude of 4 ohms or below for much of the lower midrange to upper bass region, leading to thermal (and perhaps current-limiting) problems for some amplifiers (and probably many AVRs). So a large-scale signal at 100 Hz (sine wave) into 2-4 ohms is another test that may be worth performing.

I am unaware of anyone requesting an audio amplifier be driven by a test signal up to 10 MHz. Achieving ruler-flat in-band response (to 20 kHz) requires -3 dB in the 100's of kHz range so testing with a signal having that sort of content (not a full-scale sine wave) is reasonable.

I always used to slap the output of an amp on a spectrum analyzer to check for high-frequency oscillations and occasionally saw junk in the 0.1-1+ MHz range. I have not done that in years. Back when "ultra-wideband" was all the range plenty of amps had problems driving reactive loads. And a few DC-coupled amps would motorboat in response to very low-frequency signals so I would do a sweep down to 1 Hz or below.

As a designer I expect none of the current crop of amplifiers would exhibit severe problems like oscillations and imagine they'll all shut themselves down cleanly in response to a current overload or over temp condition. And as a designer I know it is all too easy for some little thing to slip through and lead to an unstable design under some unexpected real-world condition. I have designed more than one good oscillator that was meant to be a linear amplifier... And of course a few oscillators that refused to start and made great amplifiers.

As SIY said, just look at the quad plots. They are only dropping to 3.8 ohms. Plenty of speakers do indeed go below 4 Ohms, but 0.5? Nah.

Stability is an issue that of course a designer is going to evaluate during design. I have done precisely this my buffer amp, headphone amp designs and power amps. You know you have screwed up when you see the power supply currents and feel the temp of the device. You know what it's doing before breaking out the scope or spec an.

However, as you allude to yourself, is it a significant problem people are experiencing in current crop of available amps? I think not. I think it's being over stated by the OP.

Serious question, what recent reports are there out on the net of hifi amps going unstable and melting themselves or speakers?

 

[anmpr1]

You got there before I did. I note a few other things: the capacitance in Pavel's load is much higher than even his pathological Martin Logan speaker, and significantly higher than Quads or Acoustats (which, although uncommon, were still made in higher quantities than the Martin Logan or Sound Lab). So yeah, if you're in the 0.01% who owns such things, the amp's tolerance of a capacitive load is of interest. It's just not for the other 99.99% of us who have better-engineered speakers.

I would not call either the Quad or Acoustat poorly engineered speakers. Perhaps they could have been made better, or different, in some respects. However the original Quad had/still has a most excellent sonic signature, holding its own, in its own way, with any speaker ever produced. Shortcomings existed (it couldn't play very loud and beamed), but was that due to 'poor engineering'? Hard to make a wide dispersion electrostat (Harold Beveridge's lens loaded panels overcame some of these problems--at a price--not only in dollars).

Acoustats were subjectively never as sonically pleasing (not bad, just had a kind of 'plasticky' sound), but could play very loud (for an electrostat), built quite sturdy, and priced relatively inexpensive, given their genre.

The Quad mated easily with the company's 303 and 405 amplifier, and at the upper end of tweak Julius Futterman (later, Harvey Rosenberg) built OTL amps specifically for them, with good results, we are told. Certainly Acoustats required a bit more beef. Jim Stickland's 'Trans Nova' 200 watt MOSFET amp worked well with the panels. Rosenberg claimed that his amps self destructed on Acoustats.

 

[SIY]

I would not call either the Quad or Acoustat poorly engineered speakers. Perhaps they could have been made better, or different, in some respects. However the original Quad had/still has a most excellent sonic signature, holding its own, in its own way, with any speaker ever produced. Shortcomings existed (it couldn't play very loud and beamed), but was that due to 'poor engineering'? Hard to make a wide dispersion electrostat (Harold Beveridge's lens loaded panels overcame some of these problems--at a price--not only in dollars).

Acoustats were subjectively never as sonically pleasing (not bad, just had a kind of 'plasticky' sound), but could play very loud (for an electrostat), built quite sturdy, and priced relatively inexpensive, given their genre.

The Quad mated easily with the company's 303 and 405 amplifier, and at the upper end of tweak Julius Futterman (later, Harvey Rosenberg) built OTL amps specifically for them, with good results, we are told. Certainly Acoustats required a bit more beef. Jim Stickland's 'Trans Nova' 200 watt MOSFET amp worked well with the panels. Rosenberg claimed that his amps self destructed on Acoustats.

If I gave the impression that Quad or Acoustat were poorly engineered, my apologies. I was trying to express the exact opposite- their engineering was excellent, far better than examples like those Martin Logans. I will confess to having a multi-decade lust for a pair of ESL-63s, and one day, perhaps...

I have both owned and built Futterman amps. Not impressed.

 

[anmpr1]

If I gave the impression that Quad or Acoustat were poorly engineered, my apologies. I was trying to express the exact opposite- their engineering was excellent, far better than examples like those Martin Logans. I will confess to having a multi-decade lust for a pair of ESL-63s, and one day, perhaps...

I have both owned and built Futterman amps. Not impressed.

Sorry for my misinterpretation.

The Futterman thing was strange. One man playing it very low key, devoting his life to one oddball circuit. Julius was, as far as I know and from everything I've read on him, more an obsessed hobbyist. He never charged much for his hand built gear. He wasn't a 'take the money and run' sort of opportunist. You could just never get one. Too long a waiting list. Self-proclaimed Dr. Gizmo (Rosenberg) was the PT Barnum of high-end. A carnival barker in his own way. But at least he was funny. There is always some redeeming value in humor. May both RIP.

 

[ajawamnet]

Sorry for my misinterpretation.

The Futterman thing was strange. One man playing it very low key, devoting his life to one oddball circuit. Julius was, as far as I know and from everything I've read on him, more an obsessed hobbyist. He never charged much for his hand built gear. He wasn't a 'take the money and run' sort of opportunist. You could just never get one. Too long a waiting list. Self-proclaimed Dr. Gizmo (Rosenberg) was the PT Barnum of high-end. A carnival barker in his own way. But at least he was funny. There is always some redeeming value in humor. May both RIP.

Futtermans... wow, that brings back memories... KLH 9's ... geez... no output transformer. What a beast to repair.

Kinda like the old broadcast finals...

... here's one - RCA BTF 10 - called the flame thrower

Final section:

Plate blocker:

Size of Tube (note this is not the tube used in the BTF-10 - shown for size only):

Before:

After:

This whole topic jogs all those old memories that I'd buried away... maybe for good reason.

But I'll bet that if you made an audio amp out of a 4CX5000 it'd sell... big glass dome back lit... wow.

I'm sure there'd be people that'd buy it

 

[anmpr1]

I never had any hands on experience with them. Julius argued that the output transformer was always the limiting sonic factor in amplifiers, and using that idea, he took it from there. I know at least one outfit attempted to commercialize his design (Harvard Electronics), and it was reviewed in, I think, both Stereo Review and High Fidelity. It was expensive, even back then.

But by that time solid state had pretty much taken over the consumer market. His was an application for an era past, and an solution to a problem few had. It was an electrical curiosity--the higher the speaker impedance, the better, so it didn't work well with a lot of more conventional loudspeakers. Anything much under 8 ohms and there was trouble. I remember that before Harvey Rosenberg took over, Julius was using television tubes. I don't know what Harvey used. I've got his book somewhere, and could look. It's probably on line.

Futterman's own amps looked like what you would expect--hand built in his apartment. NYAL OTLs were a different thing. Big, ugly, and industrial. Looked like something you'd see in an iron ore smelting operation control room, from Honeywell. Harvey expanded into the MOSFET-tube hybrid scene, but then went belly up. It's all just curious history, now.

 

[SIY]

@ajawamnnet The Futterman schematic doesn’t look right. If memory serves, in the Futterman circuit, the cathode resistor of the split load inverter was returned to the amp output rather than to ground.

The 4CX and 3CX transmitting tubes are not good for audio. Besides the high voltages, noise from forced air cooling, and grid current, they need a high impedance load, which makes output transformer design very challenging.

 

[DonH56]

Brings back memories from audio and ham days... I can count on one hand the number of Futterman's I saw and heard, and pretty much the same hand for the ones I had to repair. I think every one was just a little different.

I've said before I'm surprised those Eimac broadcast tubes have not made it into audio amps despite the obvious mismatch (on many levels).

 

[amirm]

Sorry, your assumption is wrong. Because resistive load 1,2,4 ohm or whatever has 0° phase angle of impedance and no imaginary component. You seem to share same misunderstanding of amplifiers stability with @amirm . The load in post #1 has capacitive impedance from 200Hz up to 20kHz and may result in amp instability, depending on amplifier loopgain amplitude and phase plots.

Your panache for insults is only exceeded by inability to read the answers give to you. Here is an answer I gave you already:

If you think people love to see such things, then setup the tests and run them. See how many people will loan you amplifiers to test for stability.

I clearly addressed the point regarding stability.

There are countless things that can be tested with amplifiers. I don't consider that my job. If you think it is so important, survey a dozen commercial amplifiers your way and let's see the instability problems you are so worried about.

 

[restorer-john]

I will confess to having a multi-decade lust for a pair of ESL-63s, and one day, perhaps...

My 85yo father still secretly wants a pair, as do I.

I'll even take a pair of ESL-57s if she comes with them...

 

[SIY]

I had 57s for several years. Wonderful and frustrating.

 

[Blumlein 88]

But they are, nonetheless. See the curves shown earlier for Quads (the most popular ESLs of all time). Unfortunately, the impedance curves I ran on my old Acoustat 1+1s are on some archived hard disk in my garage, but I bet someone else has posted them. They do not much resemble the M-L impedance.

Your point about signal magnitude is exactly... on point.

The old Acoustats in general are just like the Soundlab curves. They use the same basic method. A high and low ratio pair of transformers with a crossover network blending them somewhere between 700 and 1500 hz. I've measured both and also have Acoustat curves on an old computer somewhere. The published ones for the Soundlab are correct and very close to the larger Acoustats even in values obtained.

 

[Blumlein 88]

If I gave the impression that Quad or Acoustat were poorly engineered, my apologies. I was trying to express the exact opposite- their engineering was excellent, far better than examples like those Martin Logans. I will confess to having a multi-decade lust for a pair of ESL-63s, and one day, perhaps...

I have both owned and built Futterman amps. Not impressed.

Me neither. Well let me rephrase that, I'm impressed with how poorly Futterman and other OTL amps perform. They just can't handle much of any load at all.

I owned some ESL63s for a decade. You really should get some. I still think in some ways nothing is as good. I think most attention should be paid to the concentric quasi-point source design idea. I had some ESL-57s for awhile. They were good too though limited.