Obviously speakers are not perfect resistors. This, however, presents a bit of a challenge when testing amplifier stability, as running full power into a loudspeaker is hardly realistic- nobody wants 110dB sound pressure levels in their shop, especially when that 110dB is a 5kHz sine wave.
What's the recommended test procedure? My initial idea was to rip the crossover out of an old PA speaker and put dummy load resistors in place of the drivers, but that doesn't account for the inductance and resonant frequency of the actual drivers (or speaker cabinet). Beyond that, not all speakers have the same crossovers.
I've seen some recommendations to put a .22uF cap across the output and feed in a small signal, but this seems unrealistic as well.
As I start working on my own designs more and more, I'd really like a little input on the best way to make sure it's stable into speakers with oddball crossover networks.
What's the recommended test procedure? My initial idea was to rip the crossover out of an old PA speaker and put dummy load resistors in place of the drivers, but that doesn't account for the inductance and resonant frequency of the actual drivers (or speaker cabinet). Beyond that, not all speakers have the same crossovers.
I've seen some recommendations to put a .22uF cap across the output and feed in a small signal, but this seems unrealistic as well.
As I start working on my own designs more and more, I'd really like a little input on the best way to make sure it's stable into speakers with oddball crossover networks.
There is some stuff about this on Patrick Turners site, but it's down this morning, showing a 502 "bad gateway" When I've had stability issue's it's shown up at low volume, so no full power testing was needed.
Andy.
Andy.
I'm sorry to see a lack of response to your post because it's an important question. There isn't a general answer because the variety of loudspeaker loads is so great. Some idiotic speaker cables have even been known to cause amplifier burst oscillations.
I can't recommend Bob Cordell's book (Designing Audio Power Amplifiers, 2011) too highly. Although written about modern amplifiers, it's tremendously useful to anyone who wants to learn, and is very accessable to amateurs, like us. This topic is discussed, along with recommendations. Also a parasitic oscillation sniffer.
Since some oscillations can be fatal to the output transformer windings' insulation, I can't really recommend that you try to see how big a capacitor you can hang off the output terminals. Arf!
But, the variety of various loudspeakers' crossover networks and usually ultimately inductive tweeters doesn't really seem to matter that much in a reasonably competant amplifier. The old rules of thumb about stopping resistors, layout, power supply bypassing, fear of loop feedback, output Zobel networks, etc. mostly work in lieu of adequate modeling.
All good fortune,
Chris
I can't recommend Bob Cordell's book (Designing Audio Power Amplifiers, 2011) too highly. Although written about modern amplifiers, it's tremendously useful to anyone who wants to learn, and is very accessable to amateurs, like us. This topic is discussed, along with recommendations. Also a parasitic oscillation sniffer.
Since some oscillations can be fatal to the output transformer windings' insulation, I can't really recommend that you try to see how big a capacitor you can hang off the output terminals. Arf!
But, the variety of various loudspeakers' crossover networks and usually ultimately inductive tweeters doesn't really seem to matter that much in a reasonably competant amplifier. The old rules of thumb about stopping resistors, layout, power supply bypassing, fear of loop feedback, output Zobel networks, etc. mostly work in lieu of adequate modeling.
All good fortune,
Chris
Obviously the techniques you mentioned are good for dealing with oscillation issues, but what doesn't get talked about as much as it probably should is a proper testing procedure for "worst realistic case". I can throw oscillation-inhibiting parts into a circuit until I'm blue in the face but at the end of the day, I want to actually test into a realistic load to be sure.
One idea I have thought about is to take a big PA speaker (one that's rated at like 200+ Watts) and cut the cone off the woofer and maybe do something to dampen the tweeter (block of wood around it, maybe?) and use that as a test so I can drive to higher powers without making so much noise.
My hope wasn't necessarily a theory based discussion thread even, but a place where the "old timers" so to speak could chime in on how *they* test for instability.
I'm trying to remember what amplifier it was that specified a maximum and minimum speaker cable length... I believe it was a solid state design. I thought that sounded awfully suspicious.
One idea I have thought about is to take a big PA speaker (one that's rated at like 200+ Watts) and cut the cone off the woofer and maybe do something to dampen the tweeter (block of wood around it, maybe?) and use that as a test so I can drive to higher powers without making so much noise.
My hope wasn't necessarily a theory based discussion thread even, but a place where the "old timers" so to speak could chime in on how *they* test for instability.
I'm trying to remember what amplifier it was that specified a maximum and minimum speaker cable length... I believe it was a solid state design. I thought that sounded awfully suspicious.
I know my answer wasn't to your actual question, but let me pose you one back: What reliability could you put in any answer I, or anyone, would give? You could choose the Stereophile recommended LCR load or 2uF or a particular actual loudspeaker, but what confidence can you have in its generality?
And really more important: what is the amplifier's margin of stability into any chosen load? The threshold between not oscillating and actually oscillating is narrow and we'd like to be a comfortable distance away. How do we know how far away we are? Knowing only that something doesn't oscillate in a given circumstance tells us nothing about its margin of stability, even in that particular circumstance.
Certainly not trying to be negative or obtuse - your goal is worthy and admirable - but some simple questions don't have simple answers, and this is one of 'em.
So, assuming that all precautions have been taken for VHF parasitics, and we want to study stability, the old school way is to make very wideband (1 Hz to 500K Hz or so) measurements of output magnitude and phase, with a load of our choice. Old school, slow going by hand, but contains useful information ("actionable intelligence").
All good fortune,
Chris
And really more important: what is the amplifier's margin of stability into any chosen load? The threshold between not oscillating and actually oscillating is narrow and we'd like to be a comfortable distance away. How do we know how far away we are? Knowing only that something doesn't oscillate in a given circumstance tells us nothing about its margin of stability, even in that particular circumstance.
Certainly not trying to be negative or obtuse - your goal is worthy and admirable - but some simple questions don't have simple answers, and this is one of 'em.
So, assuming that all precautions have been taken for VHF parasitics, and we want to study stability, the old school way is to make very wideband (1 Hz to 500K Hz or so) measurements of output magnitude and phase, with a load of our choice. Old school, slow going by hand, but contains useful information ("actionable intelligence").
All good fortune,
Chris
Last edited:
H713, maybe some more insight in to what testing capabilities you have or intend to have, and what amplifiers you make or intend to make (power, frequency span, feedback scheme, quality of layout and parts, construction restrictions).
MJ Magazine from Japan has articles on amplifiers. Schematics, frequency response graphs, distortion graphs, etc.
But MJ also has scope photos of the amplifier square wave tests, many of the amps use negative feedback. 10kHz square, with resistive loads, and then with various value capacitors across the resistive load. It often is done at low power, but you get a pretty good idea of the amplifier stability (or instability).
But MJ also has scope photos of the amplifier square wave tests, many of the amps use negative feedback. 10kHz square, with resistive loads, and then with various value capacitors across the resistive load. It often is done at low power, but you get a pretty good idea of the amplifier stability (or instability).
Last edited:
As to what loudspeaker simulator to use, that is a tough one. Whatever you do, someone will come up with a worse situation.
I once had a nominal 8 Ohm loudspeaker that low impedance above 5 kHz, and was still decreasing as the frequency went higher. I do not know what was in the crossover. But it certainly was not due to the tweeter inductance.
I tested that loudspeaker with the same $55,000 10Hz to 4 GHz Vector Network Analyzer that I also used to test amplifiers phase and frequency response into resistive loads, and Bifilar wound vacuum tube interstage transformers (No I did not test them beyond 1MHz).
I also tested the phase and frequency response of amplifiers, but this time into loudspeaker loads. Of course the testing was done at low power. But one test was to compare the resulting curves of the loudspeaker load, versus the resistive load. Very interesting.
I no longer have access to the vector network analyzer. I was tempted to purchase a completely different German vector network analyzer that 'only cost' $5000, but that is too expensive for me.
I once had a nominal 8 Ohm loudspeaker that low impedance above 5 kHz, and was still decreasing as the frequency went higher. I do not know what was in the crossover. But it certainly was not due to the tweeter inductance.
I tested that loudspeaker with the same $55,000 10Hz to 4 GHz Vector Network Analyzer that I also used to test amplifiers phase and frequency response into resistive loads, and Bifilar wound vacuum tube interstage transformers (No I did not test them beyond 1MHz).
I also tested the phase and frequency response of amplifiers, but this time into loudspeaker loads. Of course the testing was done at low power. But one test was to compare the resulting curves of the loudspeaker load, versus the resistive load. Very interesting.
I no longer have access to the vector network analyzer. I was tempted to purchase a completely different German vector network analyzer that 'only cost' $5000, but that is too expensive for me.
Last edited:
I'll talk about stability. Physical stability ! Make sure the amp you have designed can be standing on a workbench partially
disassembled when measuring during repair. Have the chassis designed that the amp could be standing upside down ( or whatever is needed to access internal
points of interest in a live amp). An amp that is unstable and might drop down and destroy tubes in the fall is not what you want on your workbench.
Radford is a positive example where built-in handles may be used to support an upside down amp that is live.
disassembled when measuring during repair. Have the chassis designed that the amp could be standing upside down ( or whatever is needed to access internal
points of interest in a live amp). An amp that is unstable and might drop down and destroy tubes in the fall is not what you want on your workbench.
Radford is a positive example where built-in handles may be used to support an upside down amp that is live.
petertub,
You are so right! You gave good "feedback"
I have found myself mounting old amplifiers upside down on books, to support the choke, power transformer, and output transformer. Not very stable. If the amplifier is live, and it tumbles, you have to remember not to chase after it with your hands . . . that could be shocking! Tubes are cheaper than a coffin.
That is similar to the rule that if a hot soldering iron falls, do not chase after it with your hands.
I put a single handle on old amplifiers, but not two, so that will not work. I will have to remember when I do a new amplifier to use two handles.
You are so right! You gave good "feedback"
I have found myself mounting old amplifiers upside down on books, to support the choke, power transformer, and output transformer. Not very stable. If the amplifier is live, and it tumbles, you have to remember not to chase after it with your hands . . . that could be shocking! Tubes are cheaper than a coffin.
That is similar to the rule that if a hot soldering iron falls, do not chase after it with your hands.
I put a single handle on old amplifiers, but not two, so that will not work. I will have to remember when I do a new amplifier to use two handles.
I always make it a rule to have at least 3 points of contact with an amplifier on the bench. Generally the transformers are one of them, and then something in the front (block of wood works great).
Some amps are a real PITA and have tubes that are taller than the transformers, and I've been known to make a custom "cradle" to hold amps like the AA-121, which is kind of awkward to balance.
Some amps are a real PITA and have tubes that are taller than the transformers, and I've been known to make a custom "cradle" to hold amps like the AA-121, which is kind of awkward to balance.
When I'm working on an amp upside down (or building it for that matter) I just use one of these covers as a holder. Hammond Mfg. - Perforated Chassis Covers (1451 Series)
Patrick Turners site is back up, he has stuff to say about stablity testing in this section, further into the article - amplifier-frequency-tests
A.
A.
The power response test is performed by measuring maximum power and THD at various points across the audio band -- including the 20 Hz and 20 kHz measurement points -- when properly loaded. It is typically done at the onset of clipping, or if poorly defined, at the point where deviation from a sine waveform begins.
HF stability tests are performed using a 1 watt 10 kHz square wave (as measured when normally loaded) with scope hung on the output, and then the output shorted, opened, or operated in a capacitance only load ranging from .005 uF to 1.0 uF. At some value, the cap only load will certainly produce a maximized ring on the tops and bottoms of the square waveform, but should never cause it to break into full blown oscillation.
LF stability tests are most easily performed by again hanging a scope on the output of the amp, and lightly pulsing the input of the amplifier when operating first into a normal resistive load, and then into no load. In both cases, the baseline of the scope should settle rapidly, with little after bounce. A baseline that is slow to settle or actually is stimulated into full blown oscillation obviously indicates trouble. You can pulse the amplifier with any convenient DC source, including the use of a simple battery. Just be sure to use a pulse voltage appropriate for the sensitivity of the amplifier, and an output level that would otherwise produce an output of about 1 watt.
HF stability tests are performed using a 1 watt 10 kHz square wave (as measured when normally loaded) with scope hung on the output, and then the output shorted, opened, or operated in a capacitance only load ranging from .005 uF to 1.0 uF. At some value, the cap only load will certainly produce a maximized ring on the tops and bottoms of the square waveform, but should never cause it to break into full blown oscillation.
LF stability tests are most easily performed by again hanging a scope on the output of the amp, and lightly pulsing the input of the amplifier when operating first into a normal resistive load, and then into no load. In both cases, the baseline of the scope should settle rapidly, with little after bounce. A baseline that is slow to settle or actually is stimulated into full blown oscillation obviously indicates trouble. You can pulse the amplifier with any convenient DC source, including the use of a simple battery. Just be sure to use a pulse voltage appropriate for the sensitivity of the amplifier, and an output level that would otherwise produce an output of about 1 watt.
I would only caution ...
NO-LOAD testing with pulsing works fine for SS amps... Tube amps with OT's may experience inductive over-shoot and cause arcing in the output stage...
NO-LOAD testing with pulsing works fine for SS amps... Tube amps with OT's may experience inductive over-shoot and cause arcing in the output stage...
Agreed in usual condition, but with only 1 watt, the risk is low.
I never had it, but I guess it depends of the amplifier under test.
I never had it, but I guess it depends of the amplifier under test.
If the amp is triggered into oscillation the transformer might be in danger.
Never use no load, or be prepared for disaster.
- Status
- Not open for further replies.