Negative feedback Amp=more faithful reproduction?

Negative feedback (NFB) vs zero negative feedback (ZNFB). There seems to be unsubstantiated contention that ZNFB amps sound more realistic. I know this is an age old debate favoring the ZNFB design, but I think most audiophiles have never bothered to look into this matter and believe the advertisements and proponents of zero NFB design. I have been in that camp until recently. My own experience and research into articles on this matter leads to me believe NFB is needed for faithful reproduction of music. I'm not saying NFB design is more "musical", which is a highly subjective term and usually means more euphonic or colored. I've posted a similar question awhile back, but I was hoping we can have a more evidence based discussion on this matter. Perhaps, we need clarification of descriptive terms we use to describe sound. My contention is, in general, NFB designs produces a more accurate or faithful reproduction of music than ZNFB designs. Here is a very good article on feedback and distortion:
Accuracy is, of course, a fiction, just the same as musicality is. That said, it's horses for courses. There is appropriate use of NF and there is inappropriate use. There are good sounding components that use NF and bad, and good sounding components that use no NF, and bad.

What you are saying is pretty much like, Ford and Chevy trucks use zinc plated bolts and Dodge trucks use cadmium plated bolts, so because the performance of Ford and Chevy trucks is better it must be the result of having used zinc plated bolts. Does that make any sense?
This has been discussed many times with same division of opinions. The accuracy position has no absolute definition, nor does musicality and natural camp proponents.In the end we all choose components based on how they sound (our ears). Those who prefer the presentation of NFB designs should obviously buy them just as others who prefer ZNFB components should buy those. It's subjective and there's no avoiding that point.Trying to determine which side is right is fruitless. Buy what sounds best to you, you'll be happier.What does it matter if someone has different listening preferences. I'm glad there're multiple topology alternatives avaliable. There is no objective standard that is going to solve any thing.
Just to go on a little further.....

There's no way to know what "faithful reproduction" is, it's just a catch phrase, but, let's for a moment, kid ourselves and pretend that there is. In that event, how could you, as a consumer, isolate the fact that the reason that one component is superior in "faithful reproduction" to another can be ascribed to that one difference in circuit design, the use of NF?
Technically faithful reproduction can be measured, no? You compared the input signal to the output signal.
Charles1dad and Viridian, did you read the entire article in the link? I think he raises valid points about nature of distortion and feedback in amplifiers. We can't all be complete subjectivists here. Afterall, it's mostly hard science and technology have produced these wonderful high end products we enjoy, with a little bit of snake oil sprinkled in of course :)
"You compared the input signal to the output signal." Just how simplistic is this? So, I'll make it easy, if the two signals are exactly the same but one has 3% second harmonic distortion, and the other has .8% of the, more jarring, 7th harmonic distortion, which is the more "faithful"? Please answer my question.

I did not read the article before, but just now read it, and I agree it is thoughtful and well reasoned, but how he values the subjective result of objective changes is very different than I. And that's fine, as I said before, there are good NF amps and bad. Good feedbackless amps and bad. Unlike the author, I have no fish to fry. And after over 40 years in the hobby, I now find that I am a complete subjectivist. But, as you said, "We can't all be complete subjectivists here." And that's fine with me as well. To each, her own.
Roger Modjeski has some interesting views on this topic and the application of negative feedback in designs. FWIW I have an amp he designed that uses negative feedback and another that doesn't. I enjoy them both quite a bit.
The conclusions drawn by the author regarding what quantities of distortion are audible and which are not, seem to be at odds with the work of Hiraga, who would also give more importance to the relative quantities of each order of distortion rather than to their absolute values.
Negative feedback became a 'catchphrase'.. so naturally when that got tired, "No negative feedback' became a catchphrase...
See where this is going?
It actually has littel to do with how a product sounds. It is a design bit, just like dozens of design bits which have NOT managed to become an advertising handle.
The pity is some folks glomm onto it as if it actually mattered as if it is really important
I personally do not give a rat's arse whether a design uses it or not.
The undisputed answer can be found here.
Learn it, love it, live it
So let's see if I have this right, umm.
Amp negative feedback good
Audiogon negative feedback bad.......umm.
Negative feedback reduces harmonic and intermodulation distortions, reduces output impedance and widens bandwidth. This can be easily measured. Negative feedback, especially a global one, can introduce Transient Intermodulation Distortion (discovered in 70's) that enhances odd order harmonics making sound bright and unpleasant. Loudness cues are in these harmonics and our ears are very sensitive to them. Transient Intermodulation can be completely avoided if bandwidth of the signal is limited at the input of the amplifier to value that amplifier had without feedback. For instance, if measured bandwidth is 30kHz without feedback then it has to be input limited to 30kHz after feedback is applied (that increased amps bandwidth). It is harder to design an amp that measures good with shallow feedback and often designer chooses to improve specifications at the cost of the sound. I would start width the best widest bandwidth most linear design and apply only enough feedback to reduce THD to 0.1%. After that I would temporarily disconnect feedback and measure bandwidth with very small signal. Finally I would limit bandwidth at the very input of the amp to this amount.
It is a tradeoff between THD, IMD and the bandwidth. If you want to lower THD, IMD by use of feedback then you need more gain and amp has lower bandwidth at the higher gain. Bandwidth will increase greatly with feedback but not limiting it at the input to original one will enable TIM.

Transient Intermodulation is practically an overshoot (in time domain) caused by the fact that amplifier, having limited bandwidth, introduces signal delay. This delayed signal is subtracted at the input reducing gain of the amplifier. For instance 40dB negative feedback means that amplifier has 100 times bigger gain without the feedback. Feedback lowers this gain to normal except for the moment of fast transition of the signal because feedback signal is subtracted a moment too late (being delayed) and amplifier has for this short moment 100x higher gain. It causes overshoot of transitions (enhanced odd harmonics) but doesn't show at sinewaves, normally used to test THD and IMD.
""You compared the input signal to the output signal." Just how simplistic is this? So, I'll make this easy..."

Oh no you didn't. You didn't just try to dis me? ;) I don't think you are being condescending, so I'll give you benefit of the doubt and assume you are asking a genuine question. If that is the case, I thank you for your response.

"...if the two signals are exactly the same but one has 3% second harmonic distortion, and the other has .8% of the, more jarring, 7th harmonic distortion, which is the more "faithful"? Please answer my question."

Sure I'll answer. The result is the result and can't be disputed assuming you measured it correctly. What you're talking about is the INTERPRETATION of the result, which can be disputed. I do not know how a 3% 2nd harmonic distortion sounds relative to a 0.8% 7th harmonic distortion. But if the 7th harmonic distortion is say 80 dB down, I doubt I could ever hear it. Now 3% 2nd harmonic distortion may be audible, so I would consider that less faithful reproduction than a "jarring" 7th harmonic distortion I can't hear. Your interpretation may be entirely different than mine.

I can see how some might consider my input/output comparison simplistic. So let me explain. Why is it that a square wave is one of the industry standard of measurement of how an amplifier performs? Music does not contain square waves as far as I know. And I doubt if anyone considers a square wave a close facsimile of a music signal, which what an amplifier is suppose to reproduce. Wouldn't a more appropriate test signal be that of an instrument at certain frequency (eg, 1 kHz piano, guitar, or sax tone). Or even an instance of music containing many instruments. Is this technically too difficult to perform? I am relatively new at looking into the technical side of amplifier circuitry and performance, so my questions may seem naive to those who have years of technical experience. But this is one of the reasons why I started this thread. I would appreciate input from experienced individuals who would take the effort to be informative. As in any controversial thread, there will be individuals who will have nothing better to add other than dumb sarcastic or condescending comments.
Hi Kijanki, thanks for your input. What do you think about the article I linked?
Hi Elizabeth, what is a current fad or "catchphrase" is not important to me, but the finding the underlying reason for "catchphrase" or current fad is. If I did't think NFB or ZNFB made a difference in sound, I wouldn't have made this inquiry.
Elizabeth, I have and owned amps that have adjustable NFB, and you can clearly hear a difference between different levels of feedback. IME, the higher feedback setting produced more realistic sound. So your contention feedback has little to do with how a product sounds is in direct contradiction to my experience. Do you have any amps with adjustable feedback? If you do, I would be surprised if you can't hear a difference.
IMHO, NFB and ZNFB are different tools (among many, many others) that an amplifier designer has at his/her disposal to get the best performance (or let's say, for our discussion, the best faithful reproduction) given the amp's design parameters.

Some amps are designed to work better with NFB while other amps are designed to work better with ZNFB. So I don't think anyone can categorically say one method is better than the other for faithful reproduction, whatever that means.

Having said that, I personally prefer amps with the simplest circuit designs because I subjectively believe (rightly or wrongly) that the less capacitors (not power supply caps!), resistors, boards, etc., in the signal path means there is potentially less chance that there is going to be any change to the signal from input and output at the amp. That, to me, means ZNFB. But that also means that the amp design has to be inherently linear so that it doesn't need NFB to begin with.

I tend to prefer Nelson Pass designed amps - from Thresholds, Pass Labs, and First Watt. Almost all are ZFB designs. However, even Nelson Pass uses NFB in some of his amp designs because it works for that particular amp.

So, if you like a particular amp that was designed with NFB or with ZNFB, and it sounds good to you, then listen and enjoy it.
Adjustable NFB is a great way to prove individual preference, not a technical absolute. You used the phrase 'more realistic sound'. To you. We're engaged (yet again) in a thread of trying to prove others to be less knowledgable, savvy and just downright wrong in their audio beliefs compared to the OP. Why? It's great you've seen the light. Is validation that important to you?

Love your post and reside on your side of the fence. Curious, Have you ever listened to SET amps (hopefully with the proper speaker)? Pass is my favorite brand of ss amps.

fwiw, I don't think there is a correct side of the fence. We have our bias and whatever design gets our attention wins at least for the time being (audiophiles have been known to change their mind).

One of my amps has a variable feedback dial that goes from 0-12 db. I hear the bass firm up when I dial it in but it loses something in the overall sound that I bought the amp in the first place for.

Dracule1, I glanced over article and it seems to cover pretty much everything. I will read it in detail.

Stereophile also conducted experiment with adjustable feedback amplifier and they liked sound at minimum feedback. It doesn't mean that feedback at minimum was zero. It is virtually impossible to make SS amp without any feedback since even emitter resistors are form of feedback. The issue is not to use too much feedback. It comes to good very linear design to start with. Testing with sine waves doesn't show real picture (presence of TIM). You can have two amplifiers that measure identical but one is sounding very harsh/bright. Often amplifier that measures better sounds the worse. In early 70's before TIM was discovered SS amplifier's had huge negative feedback and THD in order of 0.001% (absolutely no need for that). In cases like that TIM can get so bad that output transistors choke on big overshoots (charge trapped at the junction) creating small gaps in music. Our brain fills them up but it makes us tired. I agree with Elizabeth that ZNFB is a catch phrase since amplifier with some amount of feedback will most likely sound better. NFB is a valuable tool that should not be used to cover poor design, cheap components etc.
Dracule1 it was not my intent to condescend when calling the model that you used "simplistic", so please do not take it that way. We agree to disagree, and that's just fine. If you feel that this is not forwarding the conversation, I am happy to bow out.

When I posed the example, I meant I was making it easy because it would be unlikely to find such an obvious example in a real environment, distortion spectra being much more complex than what I posited. Not that it needed dubming down for any of the respondents.

Square waves are used because they contain multiple freqencies. Time and frequency response are related. You can even read the frequency response of the DUT by looking at the top of the square wave backwards.
"Curious, Have you ever listened to SET amps (hopefully with the proper speaker)?"

Onemug, I have listened to some SET amps and OTL amps, all operating in Class A (oh no, another amp design principle!). In my setup at the time, I didn't feel they had the overall extension at both extremes or bass slam I like. Perhaps I should have tried more expenses tube amps. =) But they were very good in the midrange and presence. I have since gone to a bi-amp system and eventually may use some SETs or OTLs amps for their benefits in the M/H frequencies (currently using Aleph 2s, which do have some SET characteristics)) and use SS amps for their bass slam in the L frequencies. It's all a balancing act.
I am by no means a technical expert and have enjoyed both NFB amps and ZNFB designs, but I did want to note the perhaps it is a good idea to bring the human brain into this conversation (I am a neuroscientist, so I can't resist). Namely, it is fine that we are able to measure amps on the bench, but we also need to acknowledge that our ultimate enjoyment of an amp-speaker combo is based on how our brains process the signals that reach our ears. Psychoacoustics plays a major role that can't be overlooked, such as determining what frequencies we can hear most acutely and which order of distortion is more noticeable.

Also lurking in each individual brain are personal preferences. Nelson Pass' First Watt line of amps is a great example of a wide variety of design choices that are intended to both give a good match with certain types of speakers and to provide different flavors to account for different listener preferences.
it's adjustable on my amp (Mesa Baron,) so it is easy to hear the differences. I have ears, so I do not really care about the technical rationale. Some music sounds better with a little NFB (primarily studio produced rock), but most music sounds better without any NFB, at least with my system. In general, the sound is a bit less "alive" with NFB, but if a record is made with lots of compression, that hardly matters, and other factors come into play.
I 'm glad to see most of the replies are based on common sense and actual listening experiences.Which ever amplifier type sounds best to you, that's the one to buy. No need to cite this paper or that article to strengthen your preferences. I don't know why people need validation for their choices. For me, I'll take a well designed ZNFB SET amplifier anytime.It matters none at all if someone prefers something completely different, so what.Make a choice and stop worrying if others disagree. Different ears (and brains)will choose accordingly, imagine that.
Many people like sound with some distortion that makes it more alive (distorted guitar vs clean guitar). My Benchmark DAC1 sounded too clean at first but it doesn't anymore (learned to listen). Many accused DAC1 of sounding sterile. When they compared different DACs in studio Benchmark was the closest to live sound but not favored by most of testers. One person expressed opinion that all instruments sound separately while he likes them more "together" (sound blob?). Reducing NFB increases distortion making sound more "alive". It also reduces bandwidth making bright systems sound more civilized. In addition it increases output impedance making impression of more bass on over-damped speakers with SS amp. All is a matter of taste. I prefer super clean highly resolving sound with very tight bass, while others might prefer some distortion or noise with rounder bass, calling sound more "organic".
Pass does indeed up simpler circuits. His XA.5 and X.5 series apparently have but 2 gain stages. So, in that case would feedback be 'global' or 'local'?
Also, Pass amps I see the specs for are about 26db gain...which is on the low side for people accustomed to 30db PLUS gain.

If I could afford it, I'd get Pass amps for my panels....I see a pair of XA30.5s at Reno which would be perfect for a biamp setup! (another can of worms)

Other posts address feedback without ever defining if they are speaking to Global or Local. It makes a big difference...
Kinjanki well said. That had been my experience with NFB as well.

I'm not looking for validation. I've been in this hobby for close to 30 years and I know what I like. I'm searching for technical explanation of why we hear what we hear.

I have listened to SET amps paired with high efficiency speakers. There's no denying the beautification of the midrange which can sound glorious and involving. But in the long run I know it's the way things sound in real life. And I do understand why some gravitate to this type of sound.
Meant to say its NOT the way things sound in real life.
There is no stereo system that sounds the way things sound in real life, tube or transitor, NFB or not, though I have not heard your hi-fi. No doubt, it is quite special.
You're having some difficulty 'getting it'. It's not a 'type of sound'. I have a ten watt push/pull tube amp feeding 95db fostex single driver speakers. I've owned many different combinations over my life (60 this year). It took me until about two years ago to realize how much more enjoyable (FOR ME) my 3500 LPs sounded. I have been listening for more than four decades to this music previously. I don't expect you to understand what I hear. Here's the part you fail to comprehend. Who cares? It's about personal enjoyment. Real's all perception...visual, auditory or tactile. Your search has no definitive end.
I’m pretty sure that I recall some speaker manufactures recommending ZNFB amplifiers for their speakers. I always took it for granted that there was a scientifically based rationale for this. Could these be instances where NFB vs. ZNFB isn’t a totally subjective decision, or are the speaker designers simply expressing their personal preference?
Mt, I'm sure your system sound awesome to you, but I've had my fill of single driver fostex speakers with low power tube amps which have significant drawbacks IMO. But you're getting off topic, and I do not see a reason to resort to personal attacks. Somehow you think you know more than I do and assume I am missing something important. Can you please stay on topic?
Phaelon, you may be thinking of Rethem speakers which have powered subwoofer drivers with a wide bander taking care of the rest of the freq. they are coming out with low power ZNFB amp for their speakers. I think with powered subwoofers the low freq limitation of most ZNFB amps is negated while preserving the upper range qualities some people seek.
I don't assume anything. You're the one arguing with everyone who states
personal preference is important to them. The topic is, and has been, is
that your viewpoint is right (in an absolute manner, backed by some written
babble which you ascribe to as being definitive). It's not a personal attack, I
don't know you, nor do I care to. This is just more of the same drivel which
shows up too often. Your 'having your fill of single driver fostex speakers' is
irrelevant. Again, who cares?
If you ge sucked into the music and it sounds bettef the longer you listen, you have a GOOD amp. Be happy. NF design is just one aspect. Alone, it means relatively little. No nfb designs in general might be better overall and there may be fewer nfb designs that sound "irritating", and ghere are certainly many mediocre products that use, nfb, but there are also many i have heard where fatigue whatever the cause is not an issue. No abolute good, bad, rights, wrongs here as is usually the case. If nf scares you do not buy products that use it. But if you have good ears you trust i would not let anybody scare yiu away from what sounds good based on limited theoretical arguments.
Mt, please start making sense. Start your own thread to vent whatever frustrations you have. You're just trolling here. You'll get no more response from me.
Mapman, I understand those who say just enjoy your gear, there is no best. I'm not looking for the best. And I am not looking for validation, nor am i unhappywith my gear. Some on this thread mistakenly make these assumptions about the intent of my thread. But that is not the point of this thread. I perceive NFB to be more realistic than ZNFB more often than not and I'm looking for a techical explanation. Perhaps I posted this in the wrong forum. Many have posted without reading the link which I think is pretty comprihensive on the subject. Tech Talk may be the more appropriate forum.
"NF design is just one aspect. Alone, it means relatively little."

And I would go so far as to say design elements other than NFB (or ZNFB) have a greater effect on how an amp sounds.

Take it all in perspective. Enjoy and happy listening!
Edwyun, I have an amp with adjustable feedback, so you can change feedback while keeping every other parameter constant. In this case, feedback can have a significant, if not drastic, effect on the sound. So stating feedback alone has little effect does not hold true in this case.
So you are looking for a technical explanation as to why you like and find more realistic your amp with NFB than with ZNFB?

I believe the technical explanation is in the article you provided a link to in your OP. Some may agree, others may disagree. I have read that article and others from the author. Some things I agree with and others I do not. He also states that there has been great disagreement on the issue. So why would you expect anything less?

If you agree with the article then fine. That's all there is to it.
Well then it appears that your particular amp needed NFB to begin with.
The degree of feedback needed is dependent on speaker that is driving. You make a valid point. The designer purposefully built feedback into the amp.

I don't necessarily understand the article fully as my technical knowledge is limited, but I can recognize a well thought out article when I see it. I wanted start a discussion on this subject, and see what others had to say. Like I said, this seems to be the wrong forum. Many here are subjectivists and do not care about objective analysis.
Mt10425 and others have given very thoughtful answers and points of view. Dracule,you say you are`nt seeking validation but you responses suggest otherwise.There is no technical explanation to support your position,it`s simply your own preference as to what sound you happen to like.There`s no objective determination to seek in subjective issues as this is.Regardless of what forum or subject heading you choose to post on, the results/replies won`t change.You prefer the sound of NFB amps for the reasons you cite, it`s nothing more or less than that.Being condescending(as though you`ve seen the light now and other have`nt) to those who prefer a sound different from what you consider to be right or more"accurate" is pointless.Mt10425 is`nt off topic at all.
That is a very informative article as are many of the replies, as usual.

Only thing of value I can think to add is yes, practically, it is how the whole system functions together that matters, not specific design paradigms applied to specific components. There are many ways to skin a cat. That phrase could never be more true than it is regarding the details of ways to deliver quality sound reproduction. Each will chose their preferred solution. VEndors will obviously also chose their approach and any vendor worth their salt will believe in their approach 110%. CEtain design approaches that are more unique or innovative, like no NF and one of my personal favorites, Walsh drivers, done well, help to make certain vendors stick out in the pack, which is important if you are in a business for the long haul.

When someone comes up with the undisputed mathematical equation for why I like what I like, please let me know! I am waiting patiently.....
So your amp sounds best to you when you dial in a little feedback big deal. It does not tell you that all amps should sound better that way. It is very hard to design an amp without feedback. My hutch tells me that if your amp were designed to sound its best without feedback and then give you the choice you would prefer no feedback. This goes back to the Manley amp that allowed you to listen to their mono blocks in single ended mode or push pull mode. I liked it better in push pull mode and then thought that single ended amps suck. Oh not true, you can't design a circuit to sound good doing both you need different transformers and parts. So you see my friend your conclusions on a simple rotary switch is not viable. You need to keep an open mind which at that time I did not and now do.
Mapman, the question here is if NFB makes reproduction MORE faithful. Some answered that there is no faithful reproduction, which doesn't answer this question, while others stated they prefer sound of ZNFB which doesn't answer this question either. The truth is, that there is no SS amp (and very few tube amps) with zero feedback. SS amp by nature is a voltage source, and as such needs some form of negative feedback, at least for the output stage. The question is how much and that might depend on the load, your preference etc. It is undisputed fact, that even small amount of NFB widens bandwidth, lowers output impedance and reduces THD/IMD distortions. Does it make reproduction more faithful on average? I think it might. What sound you like has nothing to do with original question.

I'm working on this equation, you asked for, but it can take some time.

You did read this in the article, right?

"Only by testing, working with the devices and taking careful measurements will you find out what really happens. Relying on maths formulae (regurgitated ad nauseam) or 'common wisdom' is not always the best way to get to the truth."

So what did the author use to test his theories?

"For the majority of the tests described, the effects were simulated rather than measured. There are some very good reasons for this, with the primary reason being that the simulator has access to 'ideal' amplifiers. These have infinite bandwidth, infinite input impedance, zero distortion and zero output impedance. Being perfect, they also contribute zero noise."

I know of no amplifier that has infinite bandwidth, infinite input impedance, zero distortion, zero output impedance, and contributes zero noise. If anyone has one, let me know, I would like to purchase it.

Also, you did read his qualifications, right?

"PROPERLY IMPLEMENTED, negative feedback will invariably reduce distortion to levels that are well below audibility. Not just harmonic distortion, but the much more intrusive intermodulation distortion. IF DONE INCORRECTLY the results can be awful."

I agree with this statement. But therein lies the rub. And you do realize that in the 70's NFB was used to get distortion below audibility. But most agree that, as implemented, it sounded pretty bad.
"What sound you like has nothing to do with original question. "

It does in that there is no concrete objective answer to the question possible.

It is a subjective question not an objective one to start with, so an objective/quantitative answer cannot be had.

Therefore one must practically rely on what they hear and prefer as data points towards obtaining a meaningful answer. All else is just a particular spin on the truth, and I do not mean that in a derogatory sense, only stating the facts.

Kijanki, I can think of nobody I would rather have solve that equation! Take your time! I can survive in the interim. :^)

I think you're talking about distinctions between local feedback and global feedback.