There are varying terminologies, that's just the one I latched on to.
Some just call it series or parallel feedback, but that doesn't tell the whole story because there are two ways to derive feedback and two ways to apply feedback. Series or parallel for both. The derivation tells you whether it is voltage or current feedback.
There are lots of terms for the same things.
Some just call it series or parallel feedback, but that doesn't tell the whole story because there are two ways to derive feedback and two ways to apply feedback. Series or parallel for both. The derivation tells you whether it is voltage or current feedback.
There are lots of terms for the same things.
For completeness, feedback can be derived by sensing voltage or current, can be applied in series or in parallel with input signal, and can be of polarity to make the feedback positive (increasing gain) or negative (decreasing gain).
This is true for any feedback. For example a cathode follower senses output voltage, is fed back in series with input signal, and has the polarity to reduce gain (negative feedback). An anode follower/ Stage Local/ modern-term-"Schade" senses output voltage, is fed back in parallel with input signal, and has the polarity for negative feedback. Feedback between the cathodes of two cascaded stages senses output current, is applied is series with input signal, and has the polarity for positive feedback.
All good fortune,
Chris
This is true for any feedback. For example a cathode follower senses output voltage, is fed back in series with input signal, and has the polarity to reduce gain (negative feedback). An anode follower/ Stage Local/ modern-term-"Schade" senses output voltage, is fed back in parallel with input signal, and has the polarity for negative feedback. Feedback between the cathodes of two cascaded stages senses output current, is applied is series with input signal, and has the polarity for positive feedback.
All good fortune,
Chris
Where I live its under a tree. Preferably in the Summer.
Google OH Schade, you will get many hits. 👍
Looking around the web, I don't think I've ever seen an electronics topic described in such a consistently poor way as feedback. Bits and pieces, those with some understanding discussing it among themselves (and even then, you can find EE's conferring with peers for assistance - without much in the way of good entry for those of us who get the basic idea of putting some output back into the input but don't follow the verbal descriptions without some sort of imagery to illustrate. (Lynn Olsen's easily followed images from his ETF presentation on signal current loops comes to mind)
Even go-to's like John Broskie's excellent (and very generous) endeavours don't really bring it all down to basics for those who don't have a technical background but would like to learn in one coherent package.
I am willing to wager that better than 50% of the all too familiar resistance to feedback of any sort expressed by diy dabbling audiophiles is due to the fact the way the subject is always presented is obscure and half-fassed. Go ahead, prove me wrong. Show me one page where all the generic forms of feedback are illustrated in a way that those without math skills and/or previously extant understanding will get.
What do you think of them amperes?
Even go-to's like John Broskie's excellent (and very generous) endeavours don't really bring it all down to basics for those who don't have a technical background but would like to learn in one coherent package.
I am willing to wager that better than 50% of the all too familiar resistance to feedback of any sort expressed by diy dabbling audiophiles is due to the fact the way the subject is always presented is obscure and half-fassed. Go ahead, prove me wrong. Show me one page where all the generic forms of feedback are illustrated in a way that those without math skills and/or previously extant understanding will get.
What do you think of them amperes?
I always recommend that folk learn op-amp theory and Bode diagrams. It's a simple, logical, general, and easily accessible model that applies to toob amps too. And Toto too.
All good fortune,
Chris
All good fortune,
Chris
Lynn Olson
The presentation is here : http://www.nutshellhifi.com/library/ETF.html
Part 2 : http://www.nutshellhifi.com/library/ETF2.html
One could argue the only real ultralinear (or UL for short) are amplifiers using Acrosound brand output transformers since they coined the term as their brand name for this. Anything else is just a screen tap transformer.
Is this a facial tissue shouldn't really be called kleenex unless it's Kleenex from Kimberly-Clark sort of problem or was there something about the Acrosound transformer's construction that was not included in so called UL transformers from other companies?
I've never heard an ultra-linear amplifier that I liked but being prompted by your post to go through the Hafler/Keroes article ( starting Page 15 here ) I have to ask if each UL winding ratio would have to be for a specific tube in order to net the characteristic benefits described in the artlcle, making it a size 14 boot for a size 10 foot sort of match if buying UL transformers off the shelf based solely for the desired plate load.
Also interesting to see in the article that the first example of UL transformer is not tapped but with a separate winding for the screen. None of the Acrosound drawings I've seen show that arrangement. Was such a transformer ever produced for 'everyday' use by anybody else?
I've never heard an ultra-linear amplifier that I liked but being prompted by your post to go through the Hafler/Keroes article ( starting Page 15 here ) I have to ask if each UL winding ratio would have to be for a specific tube in order to net the characteristic benefits described in the artlcle, making it a size 14 boot for a size 10 foot sort of match if buying UL transformers off the shelf based solely for the desired plate load.
Also interesting to see in the article that the first example of UL transformer is not tapped but with a separate winding for the screen. None of the Acrosound drawings I've seen show that arrangement. Was such a transformer ever produced for 'everyday' use by anybody else?
Partial trode / tapped OPT was invented (along with many, many other things, including stereo and the method of encoding it on records still in use) by Alan Blumlein before the War (which he didn't survive). After the War, Hafler and Keroes studied the interplay of varying tapping percentages on relative power outputs and distortion for the then-new "audio" output valves designed with bigger cathodes and stronger G2s, built for the Power marketplace arms race, and found a tapping percentage region with a good balance of results (into a resistive load only), chose a magic number and trademarked the number as "UL".
Later, after Acrosound, Hafler founded Dynaco, where various tapping percentages were used, optimized for their particular use. The model A-440 Dynaco OPT has a tertiary winding (center tapped, 43% of turns) but it's rare and expensive now. Not sure if there was ever a model A-441, meaning including a 4 Ohm secondary tap.
All good fortune,
Chris
Later, after Acrosound, Hafler founded Dynaco, where various tapping percentages were used, optimized for their particular use. The model A-440 Dynaco OPT has a tertiary winding (center tapped, 43% of turns) but it's rare and expensive now. Not sure if there was ever a model A-441, meaning including a 4 Ohm secondary tap.
All good fortune,
Chris
I like my screen tap transformers for PP KT150’s… I’m still not convinced triode operation is really better when listening.
Regards, Gerrit
Regards, Gerrit
My point was, language is based on words that convey ideas and meanings, and a word's meaning can change over time based on common usage.
I think the majority of people interpret the phrase "Schade feedback" as meaning resistive plate to plate local feedback and I also don't believe most people are referring to a paper written almost 100 years ago, to a circuit design no longer in common use. Like are people actually confused when they see that term used?
Just like when someone talks about a "UL" transformer, they aren't referring to a specific transformer brand that trademarked the phrase Ultra linear, nor are they talking the very specific circuit and % that would meet that specific design criteria, they are talking about an output transformer with a screen tap.
I see this discussion as arguing semantics and tilting against a common use definition of a term. That usage is already out of the gate, good luck reeling that back in.
I think the majority of people interpret the phrase "Schade feedback" as meaning resistive plate to plate local feedback and I also don't believe most people are referring to a paper written almost 100 years ago, to a circuit design no longer in common use. Like are people actually confused when they see that term used?
Just like when someone talks about a "UL" transformer, they aren't referring to a specific transformer brand that trademarked the phrase Ultra linear, nor are they talking the very specific circuit and % that would meet that specific design criteria, they are talking about an output transformer with a screen tap.
I see this discussion as arguing semantics and tilting against a common use definition of a term. That usage is already out of the gate, good luck reeling that back in.
People here on this forum certainly do use the term "Schade feedback" to mean parallel feedback around the output stage.
Issue #1 is that O. H. Schade used series feedback around the output stage in his paper, which he did not invent.
Issue #2 is that if any of us try to explain to anyone outside this forum that we are using "Schade feedback" we won't be understood. That's only a term understood here. If I were to try to explain the reasoning for the term to any of my Engineering colleagues, they would laugh because of Issue #1.
I'm a proponent of switching to terminology that is more universally understood and correct, but I feel like few here care about this.
I swing the other way on UL. Pretty much anyone who knows tube circuits will know what you are talking about if you say you designed a UL amplifier. Yes, the name is extremely misleading, but it is pretty universally understood.
Issue #1 is that O. H. Schade used series feedback around the output stage in his paper, which he did not invent.
Issue #2 is that if any of us try to explain to anyone outside this forum that we are using "Schade feedback" we won't be understood. That's only a term understood here. If I were to try to explain the reasoning for the term to any of my Engineering colleagues, they would laugh because of Issue #1.
I'm a proponent of switching to terminology that is more universally understood and correct, but I feel like few here care about this.
I swing the other way on UL. Pretty much anyone who knows tube circuits will know what you are talking about if you say you designed a UL amplifier. Yes, the name is extremely misleading, but it is pretty universally understood.
I will argue that I see this Schade feedback term used on forums and websites all over the internet, not just on this one forum. Here is an example used over a decade ago
https://rh-amps.blogspot.com/2013/07/criticism-and-superficiality.html
Like I said, this word use issue is way past trying to put back inside the gate. If you have engineering colleagues would would laugh over you using this term with them, then call it "parallel feedback around the output stage" when talking with them.
https://rh-amps.blogspot.com/2013/07/criticism-and-superficiality.html
Like I said, this word use issue is way past trying to put back inside the gate. If you have engineering colleagues would would laugh over you using this term with them, then call it "parallel feedback around the output stage" when talking with them.
Stephe,
I've been around all through that and watched it happen. I'm pretty sure Alex Kitic got that term here. I've debated his approach with him on this forum.
But you're probably right that others on the internet use the term now.
I will definitely refer to this feedback by its more widely-understood name with my colleagues and with all of the excellent people of this forum. 😉
I've been around all through that and watched it happen. I'm pretty sure Alex Kitic got that term here. I've debated his approach with him on this forum.
But you're probably right that others on the internet use the term now.
I will definitely refer to this feedback by its more widely-understood name with my colleagues and with all of the excellent people of this forum. 😉
A long while back I looked at UL N FDBK and concluded it was mostly a scheme to avoid a regulated screen supply. But seeing as it does lower output impedance and somewhat improves distortion versus loading, I suggested trying Active Ultra-Linear to see what screen V's really are needed to actually give high quality results. This was intended mainly as an experiment, but could actually be used for an Amplifier.
The idea uses a high gm LTP to drive the output tube screen grid voltages. Using scaled / attenuated voltages from the output tube plates, compared to the output tube drive signals. ( R divider networks to the LTP grids provide the comparison. ) If the output tubes are performing with linear gain matching the divider networks, then the output screen Vs are left as is. But if NOT, the output screen V's would get modified dynamically to maintain the constant gain. The experiment purpose was to see just what the output screen V's really needed for "Ultra" results, but could be used in an actual amplifier. ( normally, attenuated output plate N Fdbks to the driver stage cathodes would be used to the same effect. Two less tubes then. )
The idea uses a high gm LTP to drive the output tube screen grid voltages. Using scaled / attenuated voltages from the output tube plates, compared to the output tube drive signals. ( R divider networks to the LTP grids provide the comparison. ) If the output tubes are performing with linear gain matching the divider networks, then the output screen Vs are left as is. But if NOT, the output screen V's would get modified dynamically to maintain the constant gain. The experiment purpose was to see just what the output screen V's really needed for "Ultra" results, but could be used in an actual amplifier. ( normally, attenuated output plate N Fdbks to the driver stage cathodes would be used to the same effect. Two less tubes then. )
Last edited:
jan.didden,
I believe that Miller Effect Feedback is a term used in this forum.
It is plate to grid capacitance inside of one tube.
I believe that Schade Feedback is a term used in this forum.
It is Output plate, to Driver plate negative feedback. Usually there is a coupling capacitor from the Driver plate that connects to the Output tube grid.
I can Not be convinced to call Schade feedback by the name Miller Effect feedback.
Miller effect feedback path is short.
Schade feedback path is longer.
But this reminds me of the discussion of . . . is it a pentode?
Pentode Patents protected the use and concept of a Suppressor Grid.
Beam Power Pentodes, as some call them, do not have a suppressor grid. Instead they have beam formers.
It is more complex than that, because we also have Kinkless Tetrodes with a 5th element, the beam formers. A Tetrode with 5 elements???
What is in a name?
Correct original even patented meaning?
Or, sombody who arbitrarily wants to re-name an Alligator when he uses the name Crocodile for an Alligator . . .
Dead animal scientists are turning over in their grave.
I believe that Miller Effect Feedback is a term used in this forum.
It is plate to grid capacitance inside of one tube.
I believe that Schade Feedback is a term used in this forum.
It is Output plate, to Driver plate negative feedback. Usually there is a coupling capacitor from the Driver plate that connects to the Output tube grid.
I can Not be convinced to call Schade feedback by the name Miller Effect feedback.
Miller effect feedback path is short.
Schade feedback path is longer.
But this reminds me of the discussion of . . . is it a pentode?
Pentode Patents protected the use and concept of a Suppressor Grid.
Beam Power Pentodes, as some call them, do not have a suppressor grid. Instead they have beam formers.
It is more complex than that, because we also have Kinkless Tetrodes with a 5th element, the beam formers. A Tetrode with 5 elements???
What is in a name?
Correct original even patented meaning?
Or, sombody who arbitrarily wants to re-name an Alligator when he uses the name Crocodile for an Alligator . . .
Dead animal scientists are turning over in their grave.
If I understand your description, the only difference between what you call Schade and Miller is on which side of the coupling capacitor you connect the feedback: either on the driver plate, or on the output grid. There is only one cap between those.
So for AC there's no difference.
Jan
So for AC there's no difference.
Jan