Douglas Self on Negative feedback and distortion


I've been reading Douglas Self book on amplifier design and something he said that really makes me think twice.

As you have seen most amplifier makers claim that their amps either does not use global NFB at all or very little of it to improve dynamic (or transient response).

According to Self, the only parameter that matters is distortion and nothing else. I supposed he measures the extra harmonics that the amp produces given a sinusoidal input. In other words, distortion is measured in the frequency domain.

If I remember correctly in my Control Theory course way back in my college days, the frequency domain reponse cannot tell how the amp will response for a given step input. And the STEP RESPONSE is what can tell a lot about the behavior of an amp dynamic and transient response.

In his book, he is very adamant about his position that the only thing that matters is the amp frequency response.

I don't thing frequency response contains information about how any amp would respond to a step input but I could be wrong. Frequency response is only a steady state behavior of the amp. It cannot tell how much the amp would over-shoot, under-shoot, tendency to ringing, and so and so, given a step response. I don't think you can look at the frequency response and make any conclusion about the amp tendency to overshoot, undershoot, ringing and so on...

What do you think?

By the way, I think his book is excellent read into the theory an amplifier design if you can ignore some of his more dogmatic position.
andy2
Newbee: It was the original SS model i.e. the non-balanced version. Given that i'm using my Sunfire's in my HT system, and the Marsh doesn't have a processor loop, it would have been a pain to try and make it work there. I tried the Marsh with other amp / speaker combo's and i didn't really like it. In that specific system though, it sounded very, very good. That's why i had originally called up Marsh and asked about the availability of their Pre-Pro. Since this was probably about two years ago, they were advertising such a unit, but it wasn't available for a LONG period of time after that. As such, i purchased another Pre-Pro but can't find any reliable sources pertaining to the sonics / performance of the Marsh unit.

Tedbu: That's a good observation. Glad you shared it with us. If the feedback circuit itself isn't fast enough or cover a wide enough bandwidth, there's no doubt that it will hinder performance. While this should be a matter of common sense design, i bet that a lot of engineers / designers overlook this factor. Sean
>
In some of the am[plifiers I have designed and built on my own I have founf that the consistency of feedback with frequency makes the biggest difference.I fin d that if a negative feedback loop is usede,if a squarewave is applied to the input a sqarewave should be observes at the feedback terminal of tghe fronr end differential amplifier.Anotherwards,the amou8nt of feedback should be constant with ffrequency and roll off at the same rate as the open loop frequency response rolls off.Using this parameter,open loop freq.resonse and closed loop are equal.
El: Thanks for summing that one up in fewer words than i would have : )

Marakanetz: I was referring to "older" Bob Carver designs i.e. Phase Linear, Carver Corporation, etc..., not his current Sunfire offerings. Having said that, even some of the spec's on the Sunfire aren't all that fabulous and it can be heard in the sonics too. That is, the lack of bandwidth results in a lack of high frequency air, clarity and detail as compared to the finest of amps. While the Sunfire's are no slouches, i don't consider them to be "show-stoppers" either. They do offer very solid bang for the buck though, especially where GOBS of power are required and thermal losses are a concern. Sean
>

PS... A nice sounding and "somewhat reasonably priced" combo that i stumbled into by accident was a Marsh preamp running into a Sunfire amp. It was one of those combo's where each piece on it's own is a very reasonable performer, yet something "magical" happens when you put them together. Can't remember what i had for interconnects and power cords though, so i won't be much help in trying to "re-create the magic" if someone else were to try a similar set-up.
Andy2...If the specs that describe the good amp were properly chosen, the laws of physics would DICTATE that a circuit having the same specs would perform in the same way. The electrons would not know whose circuit they were traveling through.
Sean: Let's do a thought experiment. Suppose there is an amp that everyone would agree it has nice and tight bass. A set of measurements are then made on this amp and these measurements are to be taken as benchmark specs for tight and tuneful bass. So by definition, any amp that has these exact specs must then have tight and tuneful bass.

1.Then comes along an amp designer who sets out to disprove those specs by designing a amp with those exact specs but has subjectively very lousy bass. According to the law of physics, I don't see anything that can prevent this from happening.

Of course, this is highly theoretical but I guess you can see my points.

Intuitively, I understand what you say but as someone on this thread pointed out that every set of specs has a loop hole.
Well, I never tried it, as I am a former musician and liberal artsy/ fartsy guy. My ears have never failed me- but they have cost me a bundle. Probably the most spot- on post in this whole affair was:
Hammy & Marakanetz: I'm somewhere between the two of you. That is, i believe that you CAN hear specs, if the tests are performed in the proper manner and ALL the spec's are taken into consideration and properly interpreted.

Yes, all things in moderation....something less and less common in politics. I suppose if you can interpret sound from specs. they can be a handy aid.
Hammy: Bob Carver's older design's wouldn't pass the spec test and that's why they don't pass the sonic test either. If you think that they did / would pass the spec test, then you're not interpreting the spec's right.

Having said that, you can believe / interpret things the way that you want, but i've already proven to more than a few here that my interpretation of sonics based on measured performance is pretty close. After all, some of these folks have emailed me with nothing but spec's, yet i was able to describe to them what they were hearing, why and how to at least partially correct it. I've also made similar statements publicly about a couple of different amps, their sonic traits and their lack of ability to integrate into various systems. Others that own / have owned those amps supported those comments. The comments that i made were based on interpretation of technical data derived from test results. I'm not saying that everyone can do this, but obviously, some folks can do it better than others. Sean
>
Sean, we could get you a list of highend amps, give you the specs- I am sure you couldn't answer andy's question. Especially more subjective things like listener fatigue. Look, Eldartford's answer makes perfect sense. Of course Andy and I don't ignore specs. I need to know the impedance of a cartridge or speaker, whether the designer uses negative feedback to cover a crap design (Gives great spec though), what the sensitivity of a speaker is ...

A lot of what you say has merit. It's not as black or white as I have portrayed it... I just know from experience that you hear music, not paper. It may be the bees knees on paper, and sound like crap. A lot of bob carver's products were like that. The specs from the lab were impressive, but in the real world the amps needed more current than a twenty amp residential circuit could provide. Oops! Worked great in the industrial lab though- loved those specs. My previous example of ultrasonic harmonic distortion is a good one as well- you can't test for something if you don't understand the science yet. Clearly hi-fi is not perfect. Perhaps twenty years from now digital amps will reproduce the waveforms perfectly, and my faith in specs will be renewed. In the meantime, as science has proved ultrasonic H.D. has audible effects due to interaction with audible freqs., T.H.D. probably should be renamed total harmonic distortion up to twenty thousand cycles, or T.H.D.U.S.S. (THD using sixties science). How many other sacred cows are flawed? Why is analog experiencing a renaisance? Why do tube amps sound so good? (yes, aside from bass). Clearly Solid State specs better . I maintain the human brain is a thousand fold more complex at processing sound as a PC. The human ear is much more precise than the lab analyzer- partially because it works exactly like the analyzer I use to process my music- my ear!

In conclusion, use specs. to help compare products for final evaluation....but don't try to tell us you can judge two products exclusively on paper.
Andy: How's this for a list? Slew Rate, Rise & Fall Time, Rated Power Bandwidth, Rated Frequency response, Current Capacity ( steady state and peak ), Output Impedance, Square Wave Tilt @ 20 Hz, Power Output @ clipping into various impedances, THD, IMD, S/N Ratio, Quantity of Global Feedback used, Phase Response, Time Delay, etc...

Like i've said before, if one has access to a wide variety of spec's and those spec's were derived using proper testing procedures, one can have a good idea of what to expect from an amp and / or how consistent it will be with different loads. Some amps are sonic chameleons because they don't achieve a high level of electronic stability and this will be reflected in the various spec's / test results. If one familiarizes themselves with the proper interpretation of these spec's, they can get by with just a portion of the above information and make a relatively well educated "guesstimate" in terms of sonic performance. Sean
>

PS... Since some spec's are influenced by others, it's possible that a manufacturer could leave out specific spec's that would explain why others would look so good on paper, yet sound so bad in real life. That's why more info is always a good thing as it gives you more of an explanation as to what is going on in the grander scheme of things.
Sean: what kind of specs that can tell you how good tight and tuneful the bass of an amp?
Hammy: That's why i said ALL the spec's have to be kept in perspective and properly obtained, not just "some of the spec's have to be good" or "some of the spec's were obtained in a manner that isn't representative of real world conditions", etc...

El: Nice explanation.

Andy2: Some of the things that you say aren't available from spec's are "kinda sorta" interpretable via spec's. I agree about soundstage width and depth though. Sean
>
Just want to add my two cents:

So far, as far as I can tell, there are no specs that could tell me how well an amp will perform in the following categories:

1. Soudstage width
2. Soundstage depth
3. Instrument separation
4. Treble sweetness
5. Tight and tuneful bass
6. Non-fatige
6. Well, I can go on and on.

My conclusion is: specs are pretty much useless. For example, every amp that I know of have a spec of frequency response from 20hz - 20khz. Now, that tells a lot.

That was my original point that distortion is not the only parameter to judge an amp. There are so many variables that it is impossible to come up a set of parameters that can tell you everything about an amp.

That's why those guys are making big bucks. I mean there are more secrets in amp design than those surrounding the whole Irag WMD things.
As a test design engineer (now happily retired)I have a different understanding of the purpose of specs.

Performance of the equipment, we all agree, is determined by its design. A prototype is built and its performance is good. Now we want to put it into production. How can we be sure that each unit that comes off the line is as good as the prototype? We certainly cannot do, on each production unit, the exhaustive performance evaluation that was done on the prototype.

The designer identifies the parameters that he believes are critical to his design, and limits (maximum values) are defined for these parameters. Now, when production units are to be evaluated it is straightforward to measure these parameters. It can even be automated.

So, specs are important...they assure that the unit you buy performs like the one that the designer evaluated (and liked well enough to put into production).
Welllll...I'll agree to disagree, because I know that "The one that truely specs out great MUST sound great" is not true. It only means the unit performs well on that test. Music is much more complex than audio industry specs are good for. Furthermore, many companies don't test specs uniformally. Some test 20-20. Some with a true load. Some some with a test resistor. Specs show CDs sound better than LPs. They don't. They are more convenient. Researchers recenly found 2nd and third order harmonics above the range of human hearing affect audible tones and thus are perceived by listeners. Specs don't test for that. How many other unresearched areas of acoustic and psychacoustic interactions are not addressed by "slew rate" , "damping factor" and "Total Harmonic Distortion". Plus, you DID miss my point that in designing to ACE spec A, a designer might screw up spec B, C, and D.
Hammy & Marakanetz: I'm somewhere between the two of you. That is, i believe that you CAN hear specs, if the tests are performed in the proper manner and ALL the spec's are taken into consideration and properly interpreted. Having said that, Pro gear typically won't give you the spec's that you need, so they too are about as useless as the lack of spec's that most "hi-end" manufacturers offer. Sean
>
Hammy,
and you're missing mine:
The one that truely specs out great MUST sound great.
If the one that specs out great and doesn't sound great it means the specs are NOT true.
Nowdays specs mean completely NOTHING unless you're purchasing a professional equipment.
The marketing assumption is that consumer is dumb, therefore needs nothing to know about specs.
Marakanetz- you are missing my point. The Yamaha doesn't sound as good at 20 watts. Or 5 watts. Or 1 watt. I'm trying to point out you can't hear specs, but IF YOU ENGINEER A PRODUCT TO SPEC OUT WELL (because that is what the buying public is looking for) YOU MAY SACRIFICE MUSICALITY.

For example, there may be trade-offs in ANY design. If there are 100 things that make up a musical amplifier, a good designer, like a good physician, first "tries to do no harm." When you sacrifice 40 elements of good design to get low THD (as you said, on a resistor at one freq.) and a low pricepoint, you are going to wind up with poor equipment. Even at the same price, the product that is engineered to sell well will not sound as good as the one that is engineered to sound great. One engineer spend his money on bells, whistles remote gadgets- the other spends it on good power transformers, mil. spec. transistors/ resistors/ capacitors/...point to point wiring....The first guy covers up his deficiences with lots of zero negative feedback and buys good articles in Stereo Review. Marketing over substance. Specmania is a good way to fall into that trap.
I think that what Audioengr was getting at is that there are loop-holes to every generalization. That is, something can deliver flat response over a wide bandwidth at a given amplitude on a steady-state basis, but that doesn't mean that it will respond correctly to changes in amplitude in a linear manner. While i agree with that, it all boils down to speed and impedances. If you've got the speed, the circuitry can respond as fast or slow as needed. If the impedances are right, power transfer is maximized and dynamic variations in amplitude don't present a problem. It really is simple if you think about it and that's why i can't understand why most of the products on the market are the way that they are i.e. under-designed and over-priced.

All of this stuff was talked about and dealt with 30 years ago by a designer that never got the credit he deserved. That person's name is David Spiegel. Too bad he only ever marketed one product as i'm sure that he would be an "audio legend" had he kept at it. His ideas were light-years ahead of most, yet he was humble enough to admit that his ideas weren't original at all when i spoke to him about them. Sean
>
Audioengr...You may be clever enough to design "an amp that has bandwidth to 1 MHz, but has lousy step reponse (overdamped). " But why would you, or anyone else, do that. In practice there is correllation.
It's been a while since college. But I'll put in my 2 cents here.

For a close-loop amplifier design, one is mostly concerned with GBW (gain-bandwith product), which is a constant for an amp. You wouldn't run an amp in open-loop where the gain is max and bandwidth is min (half-power bandwidth). You extend the bandwidth by reducing the gain (dB) with measured amount of feedback (i.e. close-loop). Of course, bandwidth enhancement may not be an issue for audio frquency range.

Remember that there is always some parasitic cap in solid state devices that causes Miller-Feedback which is unavoidable. As for local negetive feedback in a multi-stage amp, it is usually employed for, amongst other reasons, attaining certin amount of gain per stage as well as slew-rate control. How the step response will look like can be predicted by the slew rate of the amp, or by how fast the amp is.

Global feedback, which I guess is taken from the very last stage, can be seen as error-correction from a Control System point of view. It is also a way of reducing the effective output impedence of the amp.
Eldartford wrote:
"Frequency response is not the same thing as step function response, but it is correlated. And a lot easier to measure. If an amp is 0.5 dB down at 100 Kc it probably has good step response."

You cannot make this inference. Step response and impulse response are dependent on short-term transient power delivery. I can design you an amp that has bandwidth to 1 MHz, but has lousy step reponse (overdamped). To achieve critical damping, the decoupling must be done right (topology), and using components with optimum characteristics for that. This is what is wrong with 99% of consumer gear. The designers understand circuit design, but not transient power delivery.
I heard a prototype amp based on Slone's OPTI-MOS amplifier circuitry in my system. It was class AB with some level of NFB. It was the most transparent amp with superb Bass control and decent treble extension that I have heard in my system. It did not add anything and did not take anything away from the sound. I'm waiting for the production version to emerge. Watch www.zusaudio.com for details - they are just starting up.
Hammy,
If you carefully read specs of cheap 100W Yamaha you'll realize that distortion level is measured at 1kHz @1W of output power. Whatever reaches 100W you may only assume what kind of junk you're realy getting. Normally this cheap Yamaha will be in OK shape upto 20W.
Self actually says in his book that one should use as little of nfb as possible. He advices that each stage (which in itself has its own natural feedback) should be designed to give as small a distortion as possible. Then after each gain stage has been designed, as a final step, a nfb loop should be added to suit one's needs.

His argument is that global feedback is unfairly criticized by amplifier makers since each gain stage already has its own natural feedback built in. For example, an output stage complementary pair of PNP/NPN transister operating in class A/B has an inherent feedback with its emiter output resister ... So since you cannot avoid feedback in the first place, then why it is so bad?
My takes on this is that global feedback does have by far a larger impact on the behavior of the entire amp as compared to each gain stage feedback.

My one problem (among others) main problem is with his position on distortion measurement.
Step response can tell you everything you need to know about an amplifier, but criteria are not well defined. From what points do you measure rise time? 10 percent to 90 percent of max amplitude? How many overshoots? How big does the overshoot have to be to be counted? Etc. Etc. You can look at a scope trace and say "that's good" or "thats bad", but that is not very definitive.

Frequency response is not the same thing as step function response, but it is correlated. And a lot easier to measure. If an amp is 0.5 dB down at 100 Kc it probably has good step response.

Negative feedback is most often thought of as a way to reduce distortion, and it does do that. However, it also extends bandwidth, most notably in the case of pentode tube amp output transformers, where the feedback is provided by dedicated windings of the transformer that are connected to auxillary grids of the output tubes.
Re: Randy Sloans book (High Power Amplifier Design) I have read it a few times but have not seen the Self book (I have read a few of his articles). Sloans book is a good introduction to most all issues in classic power amplifier design but is reported to cover the same ground as Self. It appears to be more organized than it is and I am continually digging out my tattered copy of Horowitz and Hill to remind myself how many of the basic circuit elements Sloane throws about really work. His explanations of dominant pole capicators and amplifier stability are rather confusing to say the least. I hope Self does a better job.

BTW, Sloan is also dogmatic in his own way claiming that only Class B amps make any sense. Ironically he does include a small section on Class A biasing schemes and includes a supposed 'high-performance' Class A design. He is also a strong proponent of the wide-band frequency response/low distortion is everything school. I personally agree with these guys in that it is much much harder (nay impossible) to design a high performance amplifier without using feedback. In no other field of electronics would this even be discussed seriously.
I don't trust ANYONE who is dogmatic about specs. We're not trying to reproduce sine waves at 1000Khz. It's ridiculous- all you have to do is compare a cheap yamaha with 100 watts and ultra low distortion to a 25 watt audiophile amp. How many compromises did yamaha engineer in to get those specs right? A little Naim Nait sounds better and plays louder- but the yamaha is better on paper.
You're right,
The control theory states that negative feedback MAINLY exists to keep amplifier stable.
An amplifier at the same time can have multiple stages of amplifications which we also can consider as an amplifiers.
A global negative feedback i.e. over the whole amplifier not only should exist to decrease distortions over the larger freequency bandwidth but to prevent an amp from self-oscillations.
You're asking the right questions. I haven't read his books but i've seen enough quotes from them and read enough on his website to know that he and i don't see eye to eye on too many subjects.

I will say that to achieve ultra-wide bandwidth with excellent linearity typically entials fast response out of the circuitry. The faster the circuitry, the less likely you are to suffer from ringing and / or under / overshoot. Combining this with the proper impedances lends stability to the circuit. When you've got speed and stability, you've got good performance that is consistent without the need for a lot of error correction. The more error correction that you have has to do with the lack of speed or consistency in processing the signal. Rather than make the circuit slower and more complex, why not just design the circuit right in the first place? Wouldn't that be better than trying to band-aid it after the fact???

As a side note, i've never read but have also heard good things about Randy Sloan ( or Sloane's ) books. Anyone familiar with those and if so, how do they compare to Self's books? Sean
>