What makes a high end audio amp "better" then a low end unit?

[Alex Nikitin]

Regarding the output waveform from the LM3886 posted in your previous post: they look pretty bad, but we can't be sure genuine parts were used. The person conducting the test ordered some LM3886s off a random ebay seller, so could be fake. I'll reserve judgement until I see an example using genuine parts.

If you follow the link you'll find that these distortion were caused by an insufficient current drive on the MUTE pin. A nice picture to illustrate a concept of crossover distortion but it does not show real distortion of a properly operated LM3886.

Cheers

Alex

Ah that makes sense. Thanks for reading the whole thread.

P.S. - just FWIW, I do not consider "CD quality" (uncompressed 16bit 44.1kHz) as sufficient for Hi-Fi audio. As a fact I am rarely (maybe once a year) listen to  this kind of quality. My CD player is off almost all the time, I listen to vinyl and tape, occasionally to hi-res digital files (24bit 96kHz).

What's wrong with CD quality? It's much lower distortion and noise than vinyl and tape: hiss, hiss, flutter, flutter, wow, wow. If you like that sort of thing then good for you, but steer clear of any digital format, unless it's ripped from vinyl or tape.

Another interesting fact is, plenty of vinyl and tape recordings will have already been through a 16-bit ADC and DAC, before you get them. Many early digital recordings were actually initially sampled at just above CD quality, 16-bit 50kHz, then transferred to tape and vinyl for release. Tape was considered to be too noisy and distorted for the master/initial recording.
https://en.wikipedia.org/wiki/Hard_disk_recorder#History

24-bit 96kHz? That's just a waste of bandwidth! Good for recording and signal processing, but not worth it for playback.

Let's say my ears tell me otherwise. And yes, I am usually unhappy with vinyl or tape recorded from low resolution digital sources , so I'm trying to avoid these.

 Cheers

Alex

[Zero999]

The dynamic range on tape... depends on the tape, domestic tape is useless compared to high speed, wide, studio mastering tape for example.  The quality of vinyl is orders of magnitude lower than CD in particular due to noise and groove width.  To get the same range as CD you would have to make the groves very wide and well spaced and thus not get much on the record.  What vinyl users often refer to as "better sound quality and warmth" is actually just noise and imperfections not originally recorded on the medium, but added by the archaic technology.  Digital CDs are much cleaner, so if the recording is sterile and cold so will the play back be.  This means you hear it as the artist recorded it.

Vinyl is akin to TV pictures which are deliberately blurred and persisted compared to a reference monitor, the later makes normal TV and movies look cold, sterile and ugly.  Try watching a DVD on a good high end PC monitor and you'll see what I mean.  Then watch it on a good TV.  Reverse the experiment and view your PC on your TV with the same settings.  It will look blurred and softened etc.

Error:  Blurred, Warmth??? What are you talking about?  Vinyl has a 70KHz bandwidth.  Blurred? My 180 gram vinyl copy of Supertramp's Crime of the Century has cymbals strikes (I HD sampled and verified that signals past 50KHz exist) are so sharp and crystal clear compared to the cheap 22 Khz CD version, at full volume on my audio system, which is flat to 100KHz, the playback rips my ear drums to shreds with the most strikingly razor sound I've experienced except for 192KHz sampled audio...  Don't confuse a mastering choice when the vinyl version, or those writeups in audiophoolery magazines specifically cherry picking smoothly mastered/muffled junk as examples, with a properly full bandwidth pressed vinyl is created, the sound except for the needle traveling groove noise, is astounding for a physical medium.  That being said, 24/192k digital recordings free of all noise still cremates it.

I'm sure you're right about the bandwidth of vinyl, but it won't make any difference, when it exceeds the bandwidth of anyone's hearing, by more than an octave!

The distortion and noise of vinyl are much greater than CD, which is theoretically lower noise, than the microphones used to make the recording.

[coppice]

Error:  Blurred, Warmth??? What are you talking about?  Vinyl has a 70KHz bandwidth.

You only get that 70kHz bandwidth when using a Shibata or other complex stylus, originally developed to recover the sub-carrier signal on CD4 records. In the hey day of vinyl records only round and elliptical stylii were used in systems limited to stereo playback, so the bandwidth of replay systems was much lower.

[BrianHG]

I'm sure you're right about the bandwidth of vinyl, but it won't make any difference, when it exceeds the bandwidth of anyone's hearing, by more than an octave!

The distortion and noise of vinyl are much greater than CD, which is theoretically lower noise, than the microphones used to make the recording.

For hearing within first 20KHz, trying to reproduce the audio by sampling at 44.1k, it's the same effect of having a digital scope with a sampling rate of only 2x the max bandwidth.  With a repetitive signal, like most instruments, a quality up-sampling will do most of the time.  When it comes to metallic instruments which struck or hit,  the initial contact which might need to be a nasty squareish pulse only 1 to 4 sine cycles long will be messed up and it's location imaged by it's phase between the left and right channel might be missing the correct amplitude on one channel VS the other as you go beyond 15KHz.  This problem doesn't exist in analog formats or 96KHz recordings.  Once again, we are dealing with a complete audio train, (source quality, sampling quality, playback hardware, amp, speakers...) expected to perform a reproduction which sounds like being there in person.  Less than 0.1% of the music recorded today comes close to the attention and expertise and $$$ in mastering equipment required to create such mastered recordings.

In the past, because of the loss involved in all analog recordings, for a time, there was a real conscious effort in learning how to record to the best you can with whats available.  In the beginning of the digital age, this was still there when everything was new, a little hiss was still there and good practices were employed.  Now, the cheapest studio junk may be close to noise free & everyone records to maximum over-volume since this sound best on cheap consumer junk of 10-20 years ago which still had hiss on soft recordings when you had to turn the volume up, and in the car on FM radio where most of the music is first heard.  (Loudness War)  Today, even cheap consumer junk no longer has this hiss, but if new releases had proper volume recording levels, these songs would appear weak next to adjacent music.  So for now, we are stuck with maxed out clipped recordings though I hope this trend will begin to soften up a bit.  A lot of music I like has been forever trashed because every time the bass drum hits, or is a good vocalist stretches out their vocal chords, either the signal hits the flat line limits of +/-32767, or, a quick momentary volume lowering is heard as automatic sound compression kicks in.  And now, even though the recording equipment of the last 10 years could have easily handle the dynamic range easily without any hiss or need for compression tricks, we are left with a unnecessary permanent distorted loss of what the performance truly was.

[paulca]

Error:  Blurred, Warmth??? What are you talking about?  Vinyl has a 70KHz bandwidth.  Blurred? My 180 gram vinyl copy of Supertramp's Crime of the Century has cymbals strikes (I HD sampled and verified that signals past 50KHz exist) are so sharp and crystal clear compared to the cheap 22 Khz CD version, at full volume on my audio system, which is flat to 100KHz, the playback rips my ear drums to shreds with the most strikingly razor sound I've experienced except for 192KHz sampled audio...  Don't confuse a mastering choice when the vinyl version, or those writeups in audiophoolery magazines specifically cherry picking smoothly mastered/muffled junk as examples, with a properly full bandwidth pressed vinyl is created, the sound except for the needle traveling groove noise, is astounding for a physical medium.  That being said, 24/192k digital recordings free of all noise still cremates it.

I'm sorry but 70Khz-100KHz theoretical bandwidth is one thing, however playing that with a physical mechanical stylus and expecting true representation is another. Kinematics, momentum, resonance, secondary sympathetics, feedback and just a few of the issues with vinyl.  Claiming you can hear frequencies over 20Khz is another thing all together.  If you are anywhere older than 30 you are unlikely to hear anything even close to 20Khz.  Try downloading a test signal app and testing your ears, but, absolutely critically, have someone else push the button, including non-signal controls and record your responses.  You can look it up on charts that will guess you age from the results.  It will surprise you just how little you actually hear.  The test MUST be blind and controlled though to remove cognitive bias.

I don't doubt there are frequencies over 22Khz on your high rate sample of vinyl, what I would doubt is if those frequencies are actually coming from the record and if they were there because they were part of the recording.  Cymbals are basically pink noise, there would be very, very little way to determine if the frequencies you see are from the cymbal or are artifacts.  Even if you modeled the cymbal in the most complicated kinematics and material simulator on a super computer it would still probably be in-deterministic that those frequencies are from the cymbal and not secondaries or just foreign noise.

One test you could do is to record the same section from the vinyl multiple times and subtract the waveforms (or invert and add).  Do this with a CD and you might find the odd stray noise frequency, but do it with vinyl and you will get a lot of random non-sense in there.

[coppice]

If you are anywhere older than 30 you are unlikely to hear anything even close to 20Khz.  Try downloading a test signal app and testing your ears, but, absolutely critically, have someone else push the button, including non-signal controls and record your responses.  You can look it up on charts that will guess you age from the results.  It will surprise you just how little you actually hear.  The test MUST be blind and controlled though to remove cognitive bias.

There is a huge spread in the ages at which hearing bandwidth shrinks. I was still being driven crazy by a room full of EGA monitors whistling at 22kHz when I was over 40.

[paulca]

If you are anywhere older than 30 you are unlikely to hear anything even close to 20Khz.  Try downloading a test signal app and testing your ears, but, absolutely critically, have someone else push the button, including non-signal controls and record your responses.  You can look it up on charts that will guess you age from the results.  It will surprise you just how little you actually hear.  The test MUST be blind and controlled though to remove cognitive bias.

There is a huge spread in the ages at which hearing bandwidth shrinks. I was still being driven crazy by a room full of EGA monitors whistling at 22kHz when I was over 40.

Subjective and prone to bias.  Also possibly harmonics in lower frequencies.  However I do know what you mean.  I can remember way back in my 20s when I stopped being able to tell if the TV was on or not when I walked into a room, even if it was in hot standby with a black screen I could originally here it's whine...  then I couldn't anymore and never have again.

You can wind up parents using a dog whistle.  A lot of them are audible to young kids, but not you or their parents.  The kids will ask what the noise is when their parents think they are mad as they can't hear it.

[BrianHG]

Claiming you can hear frequencies over 20Khz is another thing all together.  If you are anywhere older than 30 you are unlikely to hear anything even close to 20Khz.  Try downloading a test signal app and testing your ears, but, absolutely critically, have someone else push the button, including non-signal controls and record your responses.

Just so we are clear, my fundamental hearing limit is currently at least 18Khz.  (Recently checked...)
As for the 70KHz claim of the vinyl.  Yes, I am wiling to do a double check and post the results here.  Note that I have already sampled the Supertramp album at 24/96 and will verify if I need to investigate further.  I know that as you go higher in frequency on the vinyl, the maximum signal level attainable does shrink.  70k was the highest I've seen on Audacity's spectrum analyzer.  (we are leaving the arena of what makes a good amp and entering what are the capabilities of vinyl...)

When I was young, I could not go into shopping malls because the ultrasonic motion detectors at 25KHz was deafening and even with my fingers stuck in my ears, it drove me mad.  My parents thought I was just loony and hated shopping.  It's a pain when no one else in the family has any clue about technology, sound, power, or anything more technical than accounting, law and some simple medical info.

(Thanks to this, I really beginning to think that truly, outside of medical hearing loss, there are those of us who just hear so much more than others and we are lacking the correct vocabulary to describe that difference since it's natural to each of us to hear what we always been hearing all of our life....  Sort of like a whole new set of colors which are invisible to 50% of everybody else, so, names were never given to them since they are meaningless.)

[coppice]

When I was young, I could not go into shopping malls because the ultrasonic motion detectors at 25KHz was deafening and even with my fingers stuck in my ears, it drove me mad.  My parents thought I was just loony and hated shopping.  It's a pain when no one else in the family has any clue about technology, sound, power, or anything more technical than accounting, law and some simple medical info.

(Thanks to this, I really beginning to think that truly, outside of medical hearing loss, there are those of us who just hear so much more than others and we are lacking the correct vocabulary to describe that difference since it's natural to each of us to hear what we always been hearing all of our life....  Sort of like a whole new set of colors which are invisible to 50% of everybody else, so, names were never given to them since they are meaningless.)

A few years ago I looked to see if I could find any studies into the variation of audio bandwidth across the population. I didn't find any. There are a number of studies into the rate at which that bandwidth shrinks with age, but they don't seem to document anything about the maxima they found, typically not mentioning what happens above 20kHz.

It puzzles most people in their early 20s when a 40 year old can tell them when the last 22kHz monitor in the room has finally been turned off.

[paulca]

clipped recording

Digital clipping (unfiltered).

[BrianHG]

To think I'm 50 and still can hear 18KHz.  Even recently been to a families house who still has a CRT TV and hearing that FN 15.7KHz wine.  Even in the early 90s, I got myself a super expensive 35inch multiscan Monitor with VGA in and made my own 31Khz scan de-interlacer just to be able to enjoy TV without that bloody sound, or interlace flicker.  I wonder if all the old extra crappy TV's destroyed anyone's hearing out there...

Thanks to mixing consoles with small CRT monitors, or, Atari ST computers for MIDI with their monitor involved in the sound studio's, and with coil based microphones, you will find that bloody 15.7KHz residue in many music recordings.  If it is not a .mp3, if you go through a number of uncompressed CD recordings, put them into Audition, set the spectrum analyzer for at least 4k point, or 16k point and look.  In pop music, at least 30% of released recordings will have weak solid tone at 15.768KHz.  The occasional rare release has it loud enough that I can heard it.

[Alex Nikitin]

To think I'm 50 and still can hear 18KHz.

I am 58 and can hear (checked recently using headphones) 16kHz OK and 17kHz if I push the level up. When I was 18 I could hear up to ~21kHz.

Cheers

Alex

[paulca]

I haven't tested myself in a while.  It really is hard to test yourself though.  You would be surprised how many things you hear that aren't there and I sure you "hear" things that you really didn't hear because you know they are there.

There are apps to self test that basically blind test you, asking you to hit a button when you hear a sound and they randomly play different frequencies with frequent random periods of silence.

[BrianHG]

I haven't tested myself in a while.  It really is hard to test yourself though.  You would be surprised how many things you hear that aren't there and I sure you "hear" things that you really didn't hear because you know they are there.

There are apps to self test that basically blind test you, asking you to hit a button when you hear a sound and they randomly play different frequencies with frequent random periods of silence.

Such apps are bs.  Just load Audition, or Audacity, make a new project at at 96k sample rate and generate a 12khz, then 14khz, then 15,16,17,18,19,20khz sine wave and listen.  Just be honest with yourself, if your head and ears are being driven into the ground, you can hear it.

Note that this type of test doesn't determine your sensitivity, just if you can hear that fundamental sine wave.  Any test software which tries to determine you sensitivity isn't worth much unless you have properly verified tuned headphones, proper amp and playback card.  Otherwise, you wont be certain of anything.  Like a youtube video I saw with a high frequency tone test.  I sampled the sound playing back through my PC and everything above 12KHz was missing, as seen in the comments that no one could hear anything above the 12k test though the presenter guaranteed that there was a sound playing there.

[BrianHG]

Enjoy my sound test example which has 2 seconds of 14khz,15khz,16khz,17khz,18khz,19khz,20khz.
Make sure your sound card is set to at least 96KHz sample rate, otherwise, down-sampling may filter out the higher frequencies.
(Yup, 18k is my upper limit, 17k is still easily heard.  This file doesn't sound on my laptop, but plays perfectly on my main PC with a ASUS motherboard 192khz sound...)

[paulca]

Just be honest with yourself, if your head and ears are being driven into the ground, you can hear it.

This is the problem.  This is why in actual hearing test assessments you are kept completely separated from the tester so you can't even see them pushing buttons.  The equipment they use costs 1000s as they can't even have a click when they initiate a sound.

If you know when the sound starts you might very well believe you can hear it when in fact you can't.

You must do it blind and you cannot blind test yourself.

[coppice]

Just be honest with yourself, if your head and ears are being driven into the ground, you can hear it.

This is the problem.  This is why in actual hearing test assessments you are kept completely separated from the tester so you can't even see them pushing buttons.  The equipment they use costs 1000s as they can't even have a click when they initiate a sound.

If you know when the sound starts you might very well believe you can hear it when in fact you can't.

You must do it blind and you cannot blind test yourself.

Its interesting how easily people are fooled into thinking they hear a tone, when all they heard is an initial wideband click. I think the professional hearing test systems actually soft start (gently ramp up) and soft stop the tones to completely avoid any broadband effects at the transitions.

[BrianHG]

Now you are talking about some sort of placebo effect, or, trying to imagine you hear something.  Like trying to guess the letters on a eye test chart 1 line below where you can actually read.

The flac files I made have a separator beep and click.  Trust me, starting at 14KHz, I ca clearly hear it and each frequency up, it gets fainter.  I hear nothing at all on the 19 and 20 Khz.  For 18Khz, I need to crank up the volume to the max on my headphones.  So, I can still detect 18k, but I'm sure 30 years ago, I could have done the same at 20k.  Over 40 years ago, even 25k, but, it is well know that quite a few children under the age of 10 can hear this high up as those bloody ultrasonic motion detector alarms used to drive me crazy at that age.

[paulca]

As coppice hinted to you are arguing against many decades of psychology.  Also you seem to actually believe our senses are empirical.  The truth is hugely different. 

You can study up on the senses and perception and find things that will deeply scare you as to just how inaccurate our perception is and just how much if it is made up.  Ball park figures are something like: of the 100% your physical sense devices take in 90% is filtered out.  Of the 100% of stimulus you perceive 80% is generated by your brain.  So your brain takes only 10% of the data from the world around then fills in 80% of what it sends to your conscious. 

As a random example.  Look at yourself in a mirror, now switch eyes, left to right, look at your left eye, look at your right eye.  Obviously your eyes move to do so.  Can you see your eye ball move?  No.  You see it in one position and then suddenly it's in it's new position.  The actual motion might take 30-40ms but you never see this motion because it's filtered out completely.  Even looking at the screen in front of you, switch from one corner to the other.  Obviously your eye needs to move through every point along the way but you won't see it.  It's filtered out.  I can't remember the name of this, but your eyes has two modes, something like point and track.  The above is an example of the "point" mechanism post-filtering.  The intermediate points are imperceptible to you.

Here is a question that might freak you out... where did the time go for the 30-40ms and why don't you perceive it passing?

[Buriedcode]

I knew this thread would get to the core of perception.  There's still plenty of electronics posts, which is something!

But I didn't think it would get to folks saying "well I can hear up to xxxkhz!" It seems people assume that human hearing cuts off at a certain frequency, it of course doesn't, its just attenuated, like any other filter. You can sense 20kHz if its loud enough regardless of age, you can even sense 40kHz, again if its loud enough, but as to whether one would consider that "hearing"? that is a different matter, after all we can sense 15Hz, but is it the ears sensing the stimulus?.   Also, being able to hear >20kHz doesn't seem to offer any advantage, and in fact must be quite irritating (like having perfect pitch).

Some people should also let go of the idea that their senses are consistent.  It doesn't matter how 'good' or clever, you think you are, or even being aware of ones own biases (which apparently makes us *less* accurate) you cannot escape this. As paulca alluded to, our minds do a hell of a lot of filtering to the raw signals our senses pick up, even to the point of pretty much creating 'experience' itself, using the real world stimulus as a guide to update its model.  This means there are several layers between what sound enters your ears, and what you 'hear', and its mostly dynamic. It's the reason behind optical and auditory illusions, it adjusts 'EQ' which is why we become accustomed to the sound of certain headphones, or sound systems (including the speakers and the comb filter effect of the room).

But referring to the topic..  I think we can all agree that the "sound" of an amplifier is just one of many many parameters that affect the sound we hear.  Even if we (try to) exclude cognitive bias, given the same speakers but a different room, will sound different.  Also.. it will sound different depending on what you have heard during the course of the day, loud bass-heavy noise will make you less sensitive to bass frequencies, likewise with higher frequencies - the mind naturally tries to flatten your "EQ" as constant stimulus isn't novel and so, filtered out.  The effect isn't particularly dramatic, but its definitely there, and not considered cognitive bias, perhaps its what"listening fatigue" is?

All this means doing A/B tests, even double blind controlled studies can only go so far.  It can highlight certain limits (perception of THD, max frequency, min frequency etc..) but these should not be considered overly accurate, so any conclusions drawn should be taken with a pinch of salt. Just because a study is 'double blind, placebo controlled' doesn't make it absolutely accurate, just that compared to other methods, it is the most accurate way to access things.

[onesixright]

What about all the plane crashes? Pilots that rather trust there senses and not the instruments 

Once you have been in true IFR conditions -btw simulator will do fine- you very, very quickly find out how not to trust your senses. Because whats up is actually down and vice versa.

Sorry, IMO the this topic kind of de-railed ;-) couldn't resist.

[Zero999]

I'm sure you're right about the bandwidth of vinyl, but it won't make any difference, when it exceeds the bandwidth of anyone's hearing, by more than an octave!

The distortion and noise of vinyl are much greater than CD, which is theoretically lower noise, than the microphones used to make the recording.

For hearing within first 20KHz, trying to reproduce the audio by sampling at 44.1k, it's the same effect of having a digital scope with a sampling rate of only 2x the max bandwidth.  With a repetitive signal, like most instruments, a quality up-sampling will do most of the time.  When it comes to metallic instruments which struck or hit,  the initial contact which might need to be a nasty squareish pulse only 1 to 4 sine cycles long will be messed up and it's location imaged by it's phase between the left and right channel might be missing the correct amplitude on one channel VS the other as you go beyond 15KHz.  This problem doesn't exist in analog formats or 96KHz recordings.  Once again, we are dealing with a complete audio train, (source quality, sampling quality, playback hardware, amp, speakers...) expected to perform a reproduction which sounds like being there in person.  Less than 0.1% of the music recorded today comes close to the attention and expertise and $$$ in mastering equipment required to create such mastered recordings.

It's questionable whether the difference in phase at higher frequencies is perceptible, but even if is, analogue recording equipment can mess it up completely, especially vinyl recordings which have to be passed through a filter first, to compensate for the dreadful non-linear frequency response. The phase of the lower frequency signals (much more important) can be altered, more than a CD quality recording.

In the past, because of the loss involved in all analog recordings, for a time, there was a real conscious effort in learning how to record to the best you can with whats available.  In the beginning of the digital age, this was still there when everything was new, a little hiss was still there and good practices were employed.  Now, the cheapest studio junk may be close to noise free & everyone records to maximum over-volume since this sound best on cheap consumer junk of 10-20 years ago which still had hiss on soft recordings when you had to turn the volume up, and in the car on FM radio where most of the music is first heard.  (Loudness War)  Today, even cheap consumer junk no longer has this hiss, but if new releases had proper volume recording levels, these songs would appear weak next to adjacent music.  So for now, we are stuck with maxed out clipped recordings though I hope this trend will begin to soften up a bit.  A lot of music I like has been forever trashed because every time the bass drum hits, or is a good vocalist stretches out their vocal chords, either the signal hits the flat line limits of +/-32767, or, a quick momentary volume lowering is heard as automatic sound compression kicks in.  And now, even though the recording equipment of the last 10 years could have easily handle the dynamic range easily without any hiss or need for compression tricks, we are left with a unnecessary permanent distorted loss of what the performance truly was.

I can't ever remember noticing any hiss on digital recordings, only old analogue tape and vinyl, which I won't touch. I agree with you about the poor mixing of some recordings and the loudness war ruining some music.

[David Hess]

A further problem is that the bias conditions change with temperature so after a period of high output power, the transistors are hotter and the bias voltage should track to maintain the same idle current.  Bias circuits are designed to track temperature but it is the temperature of the transistor dies which matters and there is a considerable time lag between them and the heat sink or even transistor cases.  Better designs mount the temperature sensor as close as possible to the transistor dies to minimize lag in the temperature measurement.  I usually epoxy the temperature sensing diode directly to the transistor's metal tabs.

This is one area where integrated devices like the LM3886 actually benefit, as they have direct temperature monitoring of the output device junctions. I remember Sanken (I think) producing audio output transistors with on-die monitoring too, unfortunately heavily faked.

An integrated design provides fast thermal feedback to control the output stage bias however this also creates thermal feedback between the output and input stages.  Thermal feedback is a if not the major factor limiting performance of precision integrated circuits.  In precision operational amplifiers, it limits open loop gain and for this reason, precision designs make an effort to keep output loading low.  I suspect a external error amplifier would improve the performance of the LM3886 but it is already quite good.

Thermal feedback issues affecting output stage bias can be avoided entirely.  Douglas Self wrote about various "non-switching" amplifiers (1) which used various automatic bias control schemes developed and produced during the 1970s and 1980 but they did not catch on.  The Linear Technology LT1166 - Power Output Stage Automatic Bias System is a modern and current example of this idea.

He also reported that the performance of these systems while good was not exceptional.  Automatic bias avoids the thermal issue but transconductance variation between the output devices still exists.  I suspect the benefit was understated because standard audio tests do not take into account this sort of distortion and nobody bothered to combine this with local feedback.

Bob Cordell published a design taking into account both of these aspects of thermal feedback but not local feedback as shown below.  Note that while the LT1166 was intended to be used with MOSFETs, it can also be used with bipolar transistors.

(1) "Non-switching" in this case means that neither output transistor is allowed to fall into cutoff.

[coppice]

It's questionable whether the difference in phase at higher frequencies is perceptible, but even if is, analogue recording equipment can mess it up completely, especially vinyl recordings which have to be passed through a filter first, to compensate for the dreadful non-linear frequency response. The phase of the lower frequency signals (much more important) can be altered, more than a CD quality recording.

Phase perception is not really open to question these days, as its an extensively researched topic. Its one of the areas where lossy music compression codecs achieve a lot of their bit rate gains. We are sensitive to phase at low frequencies, but above a couple of kilohertz nobody has demonstrated an ability to detect phase. However, there is another component to this. Some filters mess with the phase so much you can hear pre-echo effects. Its not the actual phase, but the serious temporal smearing which you can detect. This is largely a digital issue, as the kinds of filters which would cause that much temporal smearing aren't really used in analogue designs.

[Gyro]

This is one area where integrated devices like the LM3886 actually benefit, as they have direct temperature monitoring of the output device junctions. I remember Sanken (I think) producing audio output transistors with on-die monitoring too, unfortunately heavily faked.

An integrated design provides fast thermal feedback to control the output stage bias however this also creates thermal feedback between the output and input stages.  Thermal feedback is a if not the major factor limiting performance of precision integrated circuits.  In precision operational amplifiers, it limits open loop gain and for this reason, precision designs make an effort to keep output loading low.  I suspect a external error amplifier would improve the performance of the LM3886 but it is already quite good.

Thermal feedback issues affecting output stage bias can be avoided entirely.  Douglas Self wrote about various "non-switching" amplifiers (1) which used various automatic bias control schemes developed and produced during the 1970s and 1980 but they did not catch on.  The Linear Technology LT1166 - Power Output Stage Automatic Bias System is a modern and current example of this idea.

He also reported that the performance of these systems while good was not exceptional.  Automatic bias avoids the thermal issue but transconductance variation between the output devices still exists.  I suspect the benefit was understated because standard audio tests do not take into account this sort of distortion and nobody bothered to combine this with local feedback.

Bob Cordell published a design taking into account both of these aspects of thermal feedback but not local feedback as shown below.  Note that while the LT1166 was intended to be used with MOSFETs, it can also be used with bipolar transistors.

(1) "Non-switching" in this case means that neither output transistor is allowed to fall into cutoff.

Thanks David, I hadn't come across the LT1166.

Yes you're right of course, it does also make the input stage very vunerable to thermal feedback, especially on a die that can quick peak at 160'C. Hopefully they've made the geometry small and balanced and avoided serious gradients across it.

It would be interesting to see a die photo - there is a pdf available for the bare die version but unfortunately it only has a diagram showing bond-out pad coordinates.

Chris.