Thank you Dr. Mazzola for your revealing answers. I see devices, and their schematics as mere symbols which oversimplifies their reality; unlike your inherent abilities [and those of many others] to " be there and in touch" which enables(d) you to invent, develop and use semis in circuits, plus teach students all about them like you see them.
It may be plausible to suggest that Schade understood the value attendant to this "internal inverse voltage feedback"; now absent or negligible in his beam tetrode, which spurred him to enable it externally as in Fig. 33c of his paper. It may follow that he was simultaneously advertising his Beam Power Tetrode [as you suggested], and also expanding its value as a new Beam Power Triode [with emphasis on power absent in mere triodes] which truly satisfied his 5 requirements of the Ideal Power Tube.
Please include in part III like you already started doing in Part II, the marriage of the couple: the Sweet Spot and Schade Feedback. Is R100A a better or worse switching device than the same R100A utilizing external Schade Feedback? Is Schade Feedback even used or has value in switching circuits?
Best regards.
It may be plausible to suggest that Schade understood the value attendant to this "internal inverse voltage feedback"; now absent or negligible in his beam tetrode, which spurred him to enable it externally as in Fig. 33c of his paper. It may follow that he was simultaneously advertising his Beam Power Tetrode [as you suggested], and also expanding its value as a new Beam Power Triode [with emphasis on power absent in mere triodes] which truly satisfied his 5 requirements of the Ideal Power Tube.
Please include in part III like you already started doing in Part II, the marriage of the couple: the Sweet Spot and Schade Feedback. Is R100A a better or worse switching device than the same R100A utilizing external Schade Feedback? Is Schade Feedback even used or has value in switching circuits?
Best regards.
This is great. I have been working toward this moment for two years in crafting these three articles, namely the insight that you are showing to me in your posts.
Yes, it is my interesting opportunity to link internal device properties with the terminal characteristics that we know (or think we know) about our favorite devices and show how the circuit and the device influence each other to achieve linearity through negative feedback. One thing I might say differently: Schade feedback, if it is what is drawn precisely in Schade's article, is in my opinion not absent or negligible in the beam tetrode or pentode, or any other practical controlled electronic "valve" (i.e., transistor) that I am aware of. It's really about how well the designer uses it. "Ultra-linear" and "sweet spot biasing" are examples of that use.
In fact, at this point, I can't help but point out that anyone that read my Part I should have anticipated the coming discussion of this in Part III. In the concluding section of Part I "Searching for the linear transistor – Straightening the Curve" internal negative feedback is explicitly discussed and shown in the "Electrokinesis!" box to be observed in the transfer curve of the 2SC4004 BJT. To put it simply, it's internal Schade feedback sampled by the internal resistance of the collector drift region of the BJT. The lesson of Part II was that external resistance (i.e., the load) does the same thing. What's left? To wrap it all up in the general formalism of negative feedback. After we do that, we will be empowered to understand how to view device characteristics as properties influenced by how they were measured and to experimentally explore some features about our devices that perhaps we didn't know they have. It's gonna be fun.
Again, thanks for following along so well. I introduced negative feedback to my students in our intermediate electronics circuits course this morning. I can see it is going to take awhile for it to sink in with the students. It always does because this is a challenging subject even for engineers.
Since I got no response to post #419, let me suggest a hint to thuart the na sayers. You say you have so much gain, and you can afford to "throw some away" for use as distortion lowering global feedback? Or even local? What does the harmonic spectrum look like? What does your Gain Bandwidth product look like? What is the feedback at lower frequencies vs. higher frequencies? With and without feedback?
The excess gain can easily lower low order distortion at low frequencies but, you don't have the same gain (read feedback, you probably still have the same gain) at high frequency. So you have less feedback at high frequencies??? Causing a trend towards the same distortion as an amp without feedback at higher frequencies. So, what does the harmonic spectrum look like now? You've lowered the THD numbers but only in the region of the very tolerable consonant frequencies and not the intolerant dissonant region.
These philosophies have been studied and published. Where did I get this from? But, without going into psycho hearing etc. etc. The simple stages vs. the multi stages, produce a different harmonic spectrum. Our ears/brain are very acute. -90db of 7th harmonic might as well be 1% of 2nd or 3rd according to some. How much 7th harmonic does a descent SE stage produce? How much 7th harmonic does a 3 or 4 stage amp produce? And how much does that 7th harmonic get corrected by just applying feedback to a 3 or 4 stage amp...
I'm just babeling now... Well ZM might like it but...
The excess gain can easily lower low order distortion at low frequencies but, you don't have the same gain (read feedback, you probably still have the same gain) at high frequency. So you have less feedback at high frequencies??? Causing a trend towards the same distortion as an amp without feedback at higher frequencies. So, what does the harmonic spectrum look like now? You've lowered the THD numbers but only in the region of the very tolerable consonant frequencies and not the intolerant dissonant region.
These philosophies have been studied and published. Where did I get this from? But, without going into psycho hearing etc. etc. The simple stages vs. the multi stages, produce a different harmonic spectrum. Our ears/brain are very acute. -90db of 7th harmonic might as well be 1% of 2nd or 3rd according to some. How much 7th harmonic does a descent SE stage produce? How much 7th harmonic does a 3 or 4 stage amp produce? And how much does that 7th harmonic get corrected by just applying feedback to a 3 or 4 stage amp...
I'm just babeling now... Well ZM might like it but...
I did publish a piece which addresses some of this
2008 "Audio Distortion and Feedback"
FIRST WATT ARTICLES
I got some static on this because I used "coffee table" values to
get some pretty pictures, but I think I made my point, and its not
about feedback so much as cascaded complexity.
(I think my critics were just jealous because I know how to make Excel
do this and they don't )
2008 "Audio Distortion and Feedback"
FIRST WATT ARTICLES
I got some static on this because I used "coffee table" values to
get some pretty pictures, but I think I made my point, and its not
about feedback so much as cascaded complexity.
(I think my critics were just jealous because I know how to make Excel
do this and they don't
)
Thanks for chiming in Nelson. To be honest, I probably got much of this understanding from you! I just wanted to give Dr. Mike an angle to experiment and prove feedback is not as good as some would lead us to believe. I think you can see this phenomena in simulation. He has only up to now focused on a single stage and the gain device contributions though and that is an important starting place.
I did publish a piece which addresses some of this
2008 "Audio Distortion and Feedback"
FIRST WATT ARTICLES
I got some static on this because I used "coffee table" values to
get some pretty pictures, but I think I made my point, and its not
about feedback so much as cascaded complexity.
(I think my critics were just jealous because I know how to make Excel
do this and they don't
Thanks for this reminder. The students of Dr. Mazzola who are just learning about NFB, and audio DIYers have an opportunity to benefit from reading this easy-style technical article.
Since I got no response to post #419, let me suggest a hint to thuart the na sayers. You say you have so much gain, and you can afford to "throw some away" for use as distortion lowering global feedback? Or even local? What does the harmonic spectrum look like? What does your Gain Bandwidth product look like? What is the feedback at lower frequencies vs. higher frequencies? With and without feedback?
The excess gain can easily lower low order distortion at low frequencies but, you don't have the same gain (read feedback, you probably still have the same gain) at high frequency. So you have less feedback at high frequencies??? Causing a trend towards the same distortion as an amp without feedback at higher frequencies. So, what does the harmonic spectrum look like now? You've lowered the THD numbers but only in the region of the very tolerable consonant frequencies and not the intolerant dissonant region.
These philosophies have been studied and published. Where did I get this from? But, without going into psycho hearing etc. etc. The simple stages vs. the multi stages, produce a different harmonic spectrum. Our ears/brain are very acute. -90db of 7th harmonic might as well be 1% of 2nd or 3rd according to some. How much 7th harmonic does a descent SE stage produce? How much 7th harmonic does a 3 or 4 stage amp produce? And how much does that 7th harmonic get corrected by just applying feedback to a 3 or 4 stage amp...
I'm just babeling now... Well ZM might like it but...
This a really good point. When I look at the THD vs freq graphs of the very high feedback amps it looks like someone decided to do a treble boost on those 80's graphic equalizers. It doesn't seem to be the absolute values of the distortion (within reasonable limits of course) per-se as maybe how much it rises vs the mids. Maybe the low thd mids encourage you to crank things up only for the high order distortion to come back and annoy you. Or maybe it's something else who knows.
All said, they have a distinctive sound that appears quite 'neutral' initially but is not very enjoyable to listen to for too long.
Last edited:
It is not only high-feedback amps that do this.
Low feedback Mosfet designs have a rising distortion characteristic, even
when the front end has flat open loop bandwidth because the capacitance of
the output devices is non-linear. This would be distinguished from high
feedback designs which have limited bandwidth and less feedback is available
at higher frequencies.
It appears to make a difference, and some of Stereophile's high frequency
IM tests have shown lower than expected mid-band IM components for some
low feedback Mosfet designs.
I did publish a piece which addresses some of this
2008 "Audio Distortion and Feedback"
FIRST WATT ARTICLES
I got some static on this because I used "coffee table" values to
get some pretty pictures, but I think I made my point, and its not
about feedback so much as cascaded complexity.
No static from me. I read your excellent primers to understand what the issues are, thank you very much.
What you describe is undoubtedly happening in what has been engineered (at least typically) in the audio amp space. And I definitely understand that the issue is "cascaded complexity" and the transient intermodulation distortion therein. And FLG's observations in post 445 are more of that experience with what has been engineered and there's certainly no reason for someone like me to doubt the truth of the reported observations. Nor the basic truth about the difference in the performance of a negative-feedback system across the frequency range if the magnitude of the open-loop gain is significantly rolling off. All that is fully anticipated in the theory of negative-feedback systems. We teach it at the university level, for example, when discussing the non-ideal properties of real operational amplifiers.
My question is, are any of these limitations fundamental to the operation of an amplifier with "global" negative feedback? Or are they the symptoms of an amplifier which was engineered this way maybe through negligence, but more likely because of habit or of economics?
If I could wave a magic wand and make roll off of open-loop gain over the bandwidth of interest (audio) go away, and while were applying magic, correct perhaps the more problematic issue of delay, would all the outcomes of cascaded complexity be "fixed" by the magic of mixing the output of the cascade with a pure undistorted reference wave from your preamp at the input of the cascade? Certainly that was Harold Black's optimistic prediction in his patent specification.
At the risk of being incendiary, I think they would. Furthermore, I believe there are ways known to achieve the magic I speculated upon.
Ok, I've said enough. If I were talking to my wife like this, I would already have got the frying pan.
If I remember correctly isn't that pretty much what Bruno Putzeys at Hypex claims? At least in this article, and the F-word article: To make a good amp with feedback a large amount of feedback is used that must be global over the whole audio band and not fall off at high frequencies. With large I think I remember him discussing that he also found that if a moderate amount of feedback is used then higher order distortion rises, but if one uses a much more feedback the distortion falls again. I think he used ~ 70 dB in the NCores.
Thanks for the heads up on the Hypex article! My first comment is that Burno's article is a classic example of "sales collateral," a term I mentioned in an earlier post. No surprise there. But you remember correctly, as Bruno argues earnestly in his section 5.2 "A More Muscular Control Loop" the point of view I expressed in post 453. So earnestly that he may deserve the frying pan treatment more than me! This point of view is rooted in the deep study of the theory that has occurred over the past 150 years since the practical application of negative feedback theory became apparent to the entire physics and engineering community, beginning famously with one of the greatest of the scientific masters of the nineteenth century, James Clerk Maxwell.
Anyway, the article is written from the perspective of a specialist in applied control engineering who is well versed in the modern theory. The reference to "sliding mode control" is dead ringer for that. But what is missing from the article is a discussion of the role that "delay" plays in preventing the global control loop from being a superhero. The references to stability, cross-over frequency, oscillation criteria, etc. implies certain knowledge of the phase plot that always goes with the magnitude plot in a so-called Bode plot of the type shown in his article. No control loop engineer would ignore phase angle so I suppose it just wasn't "profitable" for him to discuss it in this bit of collateral. But, we need to discuss it if we are going to understand what negative feedback can and cannot do in a practical application.
there is a fine line between proper feedback and proper error-correction
considering my experience , lack of knowledge , exact amplification needs and age - even in case that I'm EE ........ I wouldn't waste my time on later one
Zen wisdoms! I love em. If I'm understanding you properly, I believe this translates to: "The goal is not zero THD, but to leave just enough of what you like."
well , sort of ...... who am I to dispute what J.Hiraga and Pa (and few more clever sports) are saying all these years
though - you know that , being one of them responsible for bottling 300B+OPT in SSouth dress ......... now , harder part - all you need (or want) is to explain to us , regular Greedy Boyz , how to achieve the same with 2N3055
though - you know that , being one of them responsible for bottling 300B+OPT in SSouth dress ......... now , harder part - all you need (or want) is to explain to us , regular Greedy Boyz , how to achieve the same with 2N3055
Attachments
Ah, yes. The PASS-SIT-1.
Responsible in a, well, distant sort of way. While cleaning out the bilge at SemiSouth, I stumbled on a small handful. We were thinking power switches, not linear gain devices, so at the time my helpers and I didn't know what I was holding and they were lost.
I'll see what I can do about your idea.
Responsible in a, well, distant sort of way. While cleaning out the bilge at SemiSouth, I stumbled on a small handful. We were thinking power switches, not linear gain devices, so at the time my helpers and I didn't know what I was holding and they were lost.
I'll see what I can do about your idea.
well , sort of ...... who am I to dispute what J.Hiraga and Pa (and few more clever sports) are saying all these years
though - you know that , being one of them responsible for bottling 300B+OPT in SSouth dress ......... now , harder part - all you need (or want) is to explain to us , regular Greedy Boyz , how to achieve the same with 2N3055
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.