|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
84.73.113.255
In Reply to: Re: On a different note: Response to Morricab on Amplifiers posted by theaudiohobby on February 10, 2007 at 01:10:37:
"It is best you read the whole text, cos he is referring to the concept of benign distortions."And you missed the point of me putting the Keith Howard paper into the mix. I, like him, believe that there is no such thing as "benign distortions", IF they are audible then they lead to degradation of the signal.
"Why is the a problem, when the the high feedback amplifier starts out which much lower, more than 100 times lower, THD in the first place"
If you look carefully you will see that this is not necessarily true. Besides it is YOUR conception that 100 times lower is a significant difference because to you 100 is a big number. If you can't hear 2% 2nd harmonic but can hear 0.01% 11th harmonic then who cares about factor of 100? This is where you fail to grasp the audibility issue.
"Do have data to back this up, cos the archives at the usual suspects make nonsense of this claim?"
That the amps measure this way is easy to see. The explanation is adapted from what Crowhurst found out about what feedback does to distortion and distortion harmonic products.
"do not have access to Norman Crowhurst article or book on the subject, that being said, the level of distortion is still much lower than an equivalent zero feedback amplifier, whose distortion worsens dramatically, and it ain't just lower order, when an actual musical signal. If the amplifier is of the SET variety, then the non-linearities of the transformer also come into play."
You are like a broken record. The levels are lower, the levels are lower, the levels are lower!! SO WHAT! What about audibility and offensiveness of those distortions! It has been clear for at least 50 years that THD is NOT a good correlation with subjective sonic quality. So now THAT being said, why do you continue to believe that the level of the distortion is not as important as the type?
Tell me, if you can, what are those distortions from the transformer? How would they manifest themselves in the distortion spectra? I don't assume you know so prove it to me you have some idea you know even a bit what you are blathering on about.
"Moreover, a zero feedback amplifier (especially one of the tube) variety by definition has a much higher output impedance, "
Depends on the speaker impedance and how well damped that speaker is. Besides, Audioholics showed that if the damping factor is over 5 the effect is minimal (less than the speaker itself). If the speaker has a high or constant impedance (like an Apogee, which has a low but constant impedance) then damping factor has a much lower relevance. For a typical box speaker with a high Q bass alignment (ie. bass reflex) there could be issues with control because the speaker itself has none.
Follow Ups:
"And you missed the point of me putting the Keith Howard paper into the mix. I, like him, believe that there is no such thing as "benign distortions", IF they are audible then they lead to degradation of the signal. "No, I did not miss the point, as far as this discussion is concerned, it is simply hairsplitting, a side point as far as the original post is concerned! You do not believe in benign distortions, but in your own words, I "fail to grasp the audibility issue.", in other words, you certainly believe that some distortion is less harmful than others.i.e. benign as in less harmful...
"If you look carefully you will see that this is not necessarily true. Besides it is YOUR conception that 100 times lower is a significant difference because to you 100 is a big number. If you can't hear 2% 2nd harmonic but can hear 0.01% 11th harmonic then who cares about factor of 100? This is where you fail to grasp the audibility issue."
Nope, go back my post and bwaslo post, the point being made is lost on you, these amplifiers do not produce only low order harmonics, just that the low order harmonics are much higher in level than the higher order harmonics under certain circumstances, the further point being made by Jim Lesurf is that the distortion being made by these amplifiers becomes progressively worse or more complex, as should be expected as there is no feedback to maintain linearity, when signals more complex than simple sine waves are applied to the input.
"That the amps measure this way is easy to see. The explanation is adapted from what Crowhurst found out about what feedback does to distortion and distortion harmonic products."
Sigh...since it is trivial to present masurements showing a number of zero feedback SETs whose noise floor is worse than many high-feedback amplifiers, where does that leave your analogy?
"You are like a broken record. The levels are lower, the levels are lower, the levels are lower!! SO WHAT! What about audibility and offensiveness of those distortions! It has been clear for at least 50 years that THD is NOT a good correlation with subjective sonic quality. So now THAT being said, why do you continue to believe that the level of the distortion is not as important as the type?"Really, re-read the thread from the first post, it said level AND anhormonics components i.e. it always referred to level and ORDER. You are contending a point never made.
"Tell me, if you can, what are those distortions from the transformer? How would they manifest themselves in the distortion spectra? I don't assume you know so prove it to me you have some idea you know even a bit what you are blathering on about."
Sigh....transformers are totally linear within the passband of a SET amplifier, what a laugh! I suggest you look at the 50Hz/100Hz distortion measurements and compare them to the 1KHz measure to get the real picture, why are good transformers important?
"Depends on the speaker impedance and how well damped that speaker is. Besides, Audioholics showed that if the damping factor is over 5 the effect is minimal (less than the speaker itself). If the speaker has a high or constant impedance (like an Apogee, which has a low but constant impedance) then damping factor has a much lower relevance. For a typical box speaker with a high Q bass alignment (ie. bass reflex) there could be issues with control because the speaker itself has none."
Most (if not all) SET amplifiers with zero feedback have output impedance less than 5. Secondly, even the Apogees did not have constant impedance, the variance from the nominal was smaller than the typical bass reflex speaker but it was not constant, at any rate, it small variance is partially negative by it's very low impedance.
The broken record here is you afterall I predict your respond in original thread. As far as I can see you did take time to understand the issues highlighted in the original post before responding. And for most part you have either contended points I did not make or attempted to confusicate certain points that were stated clearly from the very first post. I suggest you get hold of a copy of the article, read it then start a new thread for us to discuss it, Why? Because after three posts, you are yet to address the main issues stated in the first post in this thread.
Music making the painting, recording it the photograph
Well since I can't get a copy of the article for Hifi News (it isn't even available yet on their website), I went in search of the man Jim Lesurf. I found a surprising number of intersting articles on his websites.One that shows that music has lots of assymmetrical waveforms:
http://www.st-and.demon.co.uk/AudioMisc/asymmetry/asym.htmlNow he also has an article on Asymmetry in amplifiers on his site (I assume a sort of precursor to what he published in hifi news):
http://www.st-and.demon.co.uk/AudioMisc/distortion/page1.htmlWhat I find interesting here is that this discussion of Asymmetry centers around push pull or complementary amplifiers and not single ended amplfiers.
Here are some interesting comments from this article:
"The purpose of what follows is to outline some of the amplifier effects which might need to be probed to identify non-linearities which affect musical reproduction. Then go on to consider an alternative method that might provide measured results more relevant to musical performance."
"In general, musical waveforms often have the following medium-term and short-term properties:
The frequencies involved tend to have a relatively simple harmonic (or near-harmonic) relationship.
The amplitude tends to vary, with brief transient events appearing as well as a steadier level.
The waveforms often show distinct asymmetries
The form factor (peak to mean level ratio) tends to be somewhat higher than for a sinusoid, even during the body of a sustained note.
"This is his argument why using sine waves or even square waves to measure distortion behavior in amplifiers is ultimately inappropriate.
"It is a common observation in the audio world that “Class A” amplifiers often “sound better”. Similar claims are often made for valve (tube) amplifiers. There may be various reasons for these beliefs. However it also seems to be the case that some amplifiers that are neither Class A nor employ valves are also sometimes judged to “sound good’’."
"The diagram above shows a pair of transistors used as an output stage which I’ll assume is “Class AB” with negligable quiescent current...This means that if these rail fluctuations couple or influence the amplifier output they may well introduce distortion. Ideally, no such unwanted couplings will occur. However in some cases, when the signal waveforms are symmetric, the coupling contributions from the positive and negative rails might tend to cancel each other...The consequence may be that – when testing with simple symmetric waveforms like a single sinewave – these various distortion contributions largely cancel, hiding the effects from detection. The consequence may be that there is a risk when developing or testing using simple symmetric waveforms that we get a low apparent non-linearity value, but when asymmetric waveforms are present, higher levels of overall distorion may arise.""The key point to draw from the above examples is that Class AB designs can have various distortion mechanisms whose contributions may depend upon the details of the signal. In particular, upon its symmetry. Tests using entirely symmetric waveforms may not expose these problems in a specific design. Class A systems may be able to avoid some significant signal distortion mechanisms, but this may not be evident when comparing various amplifier topologies simply using sinewave tests. Hence employing symmetric waveform tests alone may not help us assess the relative merits of the different amplifier arrangements."
This is about your typical Class AB feedback amplifier. As you can see the normal hifi amp also will produce "anharmonic" distortions with real music. Likewise, it will produce its own special kind of distortion called Crossover distortion. This is MOSTLY eliminated with Class A amps but not completely because matching of transistors is never perfect. According to Lesurf it is not fixable with negative feedback.
Here is a more detailed analysis of one possible cause for this asymmetry:http://www.st-and.demon.co.uk/AudioMisc/distortion/page3.html
His comments on Negative feedback:
"That the main limitations are of two kinds
The risk of oscillations or instability
Class B type distortions may not be reduced
We can use feedback with care to improve the performance of an amplifier, but it should be clear that although some feedback helps it is no substitute for designing or using an amplifier whose open loop behaviour is as good as possible. ""In practice the situation where this arises is when we have used a pure class ‘B’ amplifier which will exhibit cross-over distortion. In these cases the feedback will reduce the distortion level for reasonably powerful signals, but small signals will experience high levels of gain error, and hence high distortion even when feedback is used. For this reason class ‘B’ should be avoided – especially when we are dealing with small signals – if we wish to avoid distortion."
"so if we allow the system to be ‘near’ oscillation we may find that it does oscillate under some circumstances when certain loads are attached...An amplifier system which has a flat, unpeaked, response, with no risk of unwanted oscillations irrespective of the output load is said to be Unconditionally Stable. For obvious reasons this is highly desirable if it can be achieved...To avoid these problems and ensure stability we have to adopt two approaches. One is to build an amplifier that works as well as possible without feedback, and then only apply a moderate amount. (Note that the terms “well as possible” and “moderate” here are matters of personal judgement not strict engineering.)"
If the amp is stable enough and distortion low enough then none is likely the right amount.
I have seen several measurements in Stereophile where amps exhibit crossover distortion and according to Lesurf this is not fixed with negative feedback.It appears that PP amps have problems with asymmetric behavior as well, particularly if they are not running true Class A, which is nearly all of them.
A few points for you to clarify, first"I have seen several measurements in Stereophile where amps exhibit crossover distortion and according to Lesurf this is not fixed with negative feedback."
which ones? Secondly crossover distortion is totally different mechanism, it is never addressed with feedback. Since you have raised the issue of crossover distortion, I will refer you back to what Earl Geddes said to you on the subject a while back.
"But I do agree that some forms of crossover distortion caould be quite audible even on "poor" colored loudspeakers and yield a fatigueing situation. But I would describe this amp as "broken" in that case. In any decent amp this would more than likely not occur. "
An amplifier with audible crossover distortion is broken, broken in terms of design.
"If the amp is stable enough and distortion low enough then none is likely the right amount. "
What is not in dispute is that the open loop linearity of any amplifier is no where near its closed loop linearity, so to a large extent you comments are academic, if objective accuracy is a goal for the amplifer designer, as the article you quote states, that are guidelines to using feedback and as long those guidelines are observed, the beneficial effects of implementing feedback in the amplifier design overshadows the disadvantages.
Well for starters crossover distortion is an Anharmonic type of distortion that is mostly non-existent in Class A and totally non-existent in single ended amplifiers.However; the main point of the post was that Class AB amps are asymmetric
http://www.st-and.demon.co.uk/AudioMisc/distortion/page2.html
http://www.st-and.demon.co.uk/AudioMisc/distortion/page3.htmlwhich results in anharmonic distortions. If SET and no feedback amps have anharmonic distortions then they are not alone.
"What is not in dispute is that the open loop linearity of any amplifier is no where near its closed loop linearity, so to a large extent you comments are academic, if objective accuracy is a goal for the amplifer designer, as the article you quote states, that are guidelines to using feedback and as long those guidelines are observed, the beneficial effects of implementing feedback in the amplifier design overshadows the disadvantages."
Again, to whom? The oscilloscope or the listener. Linearity for the sake of linearity is NOT the goal. Audible distortion should be minimized not ALL distortion, the elimination of which is impossible with inherently non-linear amplification devices.
You seem to have your wires crossed with the basics and such makes me wonder why hold such a bigoted view of amplification."Well for starters crossover distortion is an Anharmonic type of distortion that is mostly non-existent in Class A and totally non-existent in single ended amplifiers."
This statements is wrong way round, a Class A circuit by definition cannot have crossover distortion and since a single-ended amplifier is necessarily Class A, the phrase "mostly non-existent" is effectively meaningless.
Moving on, a Class AB circuits are symmetric not asymmetric, the article you cited simply notes that asymmetric signals are capable of exposing more distortion than a symmetric signal. Crossover distortion is easily suppressed by correctly biasing the circuit, with good circuit design, crossover distortion can be suppressed to vanishingly low levels, i.e. levels at which it is barely measureable.
"Again, to whom? The oscilloscope or the listener. Linearity for the sake of linearity is NOT the goal. Audible distortion should be minimized not ALL distortion, the elimination of which is impossible with inherently non-linear amplification devices. "
What do you expect me to say to this comment? At any rate, it is an irrelevant here and a non-sequiter in the context of this thread. I started out by saying that zero feedback single-ended amplifiers are inherently less linear as they suffer from more distortion products and at higher level than P-P circuits with feedback, you raised the issue of crossover distortion in Class AB, however that is easily addressed with good design. So as it stands, I do not see that you have added anything to the discussion so far that undermines my original comments.
Music making the painting, recording it the photograph
"Class A circuit by definition cannot have crossover distortion "Wrong again. Sure a single ended Class A amp cannot have crossover distortion...its single ended but a complementary design can still have some distortion at zero crossing if the transistors are not perfectly matched (and they are never perfectly matched). If they were perfect complements of each other and running true Class A then there would be no crossover distortion but we are talking about real amplifiers, right?
"Moving on, a Class AB circuits are symmetric not asymmetric, the article you cited simply notes that asymmetric signals are capable of exposing more distortion than a symmetric signal. Crossover distortion is easily suppressed by correctly biasing the circuit, with good circuit design, crossover distortion can be suppressed to vanishingly low levels, i.e. levels at which it is barely measureable."Wrong again. Please look at page 3 of the article.
http://www.st-and.demon.co.uk/AudioMisc/distortion/page3.html
He states that asymmetric waveforms can heat transistors differently for each half of the wave cycle leading to dynamic differences in their bias. This leads to asymmetry in the output devices thus crossover distortion. I don't know what effects lead to asymmetry in SET amps or amps without feedback but it seems that music itself can lead to asymmetry in complementary Class AB amplifiers and that asymmetry is distortion. Bipolar transistors are especially temperature sensitive.
So the point is simple here, even with a PERFECTLY matched class AB design that measures Perfect with sine waves it will likely have distortions with real music due to asymmetry of the bias of the devices.
"Wrong again. Sure a single ended Class A amp cannot have crossover distortion...its single ended but a complementary design can still have some distortion at zero crossing if the transistors are not perfectly matched (and they are never perfectly matched). If they were perfect complements of each other and running true Class A then there would be no crossover distortion but we are talking about real amplifiers, right?"Okay.
"Wrong again. Please look at page 3 of the article.
http://www.st-and.demon.co.uk/AudioMisc/distortion/page3.html
He states that asymmetric waveforms can heat transistors differently for each half of the wave cycle leading to dynamic differences in their bias. This leads to asymmetry in the output devices thus crossover distortion. I don't know what effects lead to asymmetry in SET amps or amps without feedback but it seems that music itself can lead to asymmetry in complementary Class AB amplifiers and that asymmetry is distortion. Bipolar transistors are especially temperature sensitive.
So the point is simple here, even with a PERFECTLY matched class AB design that measures Perfect with sine waves it will likely have distortions with real music due to asymmetry of the bias of the devices."
You have your knickers in a twist, Certain non-linearities, such as what you have noted cause a departure from symmetry (see the previous page of the link you posted", however the The transfer curve of a complementary P-P circuit is symmetric. On the other hand, the transfer curve of a SE Class A is asymmetric, non-linearity simply results in yet more asymmetric distortion, its asymmetric transfer curve means, that is it unable to cancel out any distortion components, harmonic or anharmonics, in a zero feedback design, the non-linearity is even worse, as the open loop linearity is simply nowhere near what is achievable with a closed loop design. This where LeSurf original article comes in, a single-ended design is inherently less linear than a P-P design if both both are run in Class A, the SE with zero feedback even less linear than than an equivalent than equivalent P-P design with negative feedback, musically complex signal introduce new anharmonics components that an asymmetric design cannot cancel out. As you note, a Class AB circuit introduces the possibility of a new distortion products, however good design can reduce crossover distortion to barely traceable levels. Also a good design will ensure that a design the active devices are run at optimal temperature. At any rate, I saw a funny link for you to peruse, which mentions the some of the distortion products that may contribute to the 'unique tone' of zero feedback SE amplifiers. Be that as it may, the original still, a zero feedback SE amplifier is definitely less linear than a complementary Class A P-P circuit with negative feedback.
"So the point is simple here, even with a PERFECTLY matched class AB design that measures Perfect with sine waves it will likely have distortions with real music due to asymmetry of the bias of the devices. "
Strictly speaking, this is a non-sequiter issue and the statement is not clear at all, at least since the document you quote a very important assumption that is "quiescent bias current level of the AB output is small enough to be ignored" in all scenarios, firstly because Class AB circuit with negligible current defeats the whole point of having a Class AB circuit in the first place, in a decent Class AB, these distortions mechanisms (which includes crossover distortions) will be vastly reduced, because of the optimal biasing, for one bias current is never negligible.
Music making the painting, recording it the photograph
"The transfer curve of a complementary P-P circuit is symmetric"Perhaps only with a symmetric waveform as demonstrated by Lesurf. With real music there is a very real chance of the output stage of a Class AB amp to become asymmetric and produce crossover distortion due to bias drift. Who cares about "with a perfect signal"? We are talking about real world performance.
Look, I haven't read the article and I will assume, from reading Lesurf's other articles, that the man knows what he is talking about to the point that the asymmetry exists.
My point of looking at the other articles is simple, other amp topologies also can suffer from asymmetry, especially the most common amp of all the Class AB complementary transistor amplifier. The fact that it is signal correlated is worrisome. I think I have sucessfully made this point using Lesurf's article as a point of reference (same as you are using the hifi news article by the same author).
However, the question now becomes: which asymmetry is more obnoxious? Everyone knows that crossover distortion, even in very small amounts, is particularly bad sounding and engineers like Bob Cordell and John Curl both consider it to be the most important to eliminate because it can't be really treated with feedback (I read this from both of them posting in DIYaudio forum). Crossover distortion consists of lots of high order components and often gives the sound an edge or sterile sound. It can also make the highs sound grainy and low level resolution "veiled". What does the asymmetry of the SET produce? Does Lesurf say?
If you look at the the measurements of decent measuring SETs, like the one I showed further up in the thread, you don't see any anharmonic components in the 1Khz FFT spectrum...at least above the noise floor. You see neat multiples of the fundamental and only low order (at 10 watts there are some higher order as well but very low in level). So if there are anharmonic components, they are far lower in level than the harmonic components, which are not at such a harmful level to begin with.
Now you also see neat multiples in Class AB amps but of course the test is with a symmetric waveform. Put an asymmetric waveforms through the amp and maybe the tidy distorton spectrum changes as crossover distortion becomes a factor. Put the same asymmetric waveform through a no feedback SET and maybe it behaves the same.
There are other forms of asymmetry related to amps with a high negative feedback. One relates to what Otala described where back EMF from the speaker can go through the feedback loop and get reamplified. This will be highly asymmetric as the original signal gets heavily distorted by the speaker before going back into the amp and then being reamplified. Not only that it heats up the output stage transistors asymmetrically, especially with highly reactive speakers.
I know of one tube amp manufacturer who does Class A triode differential amplfiers. It is basically like push/pull but he has a special circuit that keeps dynamic track of the bias of the two tubes to keep them VERY close together. He told me that one of the reasons why push/pull tube amps don't sound as good as SETs at low levels was because of this bias mismatch. To this day his is the only PP like amp I have heard that has the low level resolution to go toe to toe with the best SET amps. Oh he uses no feedback, btw.
This post is made possible by the generous support of people like you and our sponsors: