So current feedback amplifier is marketing poof .... ?
No, it is the term invented for certain topology that people often confuse with type of feedback. The same word, different meanings.
Here is the thread about confusing definitions: http://www.diyaudio.com/forums/loun...errors-technical-documents-5.html#post3080408
Last edited:
Ideal BJT!
(Waiting for theoreticians to argue that BJT is controlled by voltage)
Douglas Self I think did . I will find the page number when alcohol is not my master . It was about his so called VAS transistor . He said it is not exactly trans-conductance and should be called a trans-impedance stage . As usual I am sure to be wrong . At least I am brave when I am .
I must say we all seem to be a bit challenged here . That is good as we all might find some new ideas . This feedback stuff is not easy . Linearity is mostly about feedback and how fast it can be done . The echo effect producing higher harmonics although disputed seems to me to be unavoidable .
Last edited:
I knew what he was saying from the start. One word to describe it (v-vs-i feedback, nested vs. not nested) is orthogonal. You can have one with or without the other.Finally an answer .....
Would you like fries with your feedback?
Nigel said a very clever thing on another forum, which I obviously liked very much and I quote it here.
Somebody there asked what was the key difference between Nige and me. He said I love simple complexity and he loved complex simplicity.
To apply that to feedback, it seems that each and every one of us has his choice and blend of local versus global, of nested or not nested, of current or voltage, etc. One man loves espresso, another Nescafe/Maxwell House, the third a traditional brew, etc.
There is no absolute answer to all these questions, it's every man for himself.
Somebody there asked what was the key difference between Nige and me. He said I love simple complexity and he loved complex simplicity.
To apply that to feedback, it seems that each and every one of us has his choice and blend of local versus global, of nested or not nested, of current or voltage, etc. One man loves espresso, another Nescafe/Maxwell House, the third a traditional brew, etc.
There is no absolute answer to all these questions, it's every man for himself.
DVV I will answer that in coffee terms . My friends always say would you like a cup of coffee . I always say sorry to be rude , only if it is real coffee . The next answer is . Oh that is too difficult . I then say I will prove I can make coffee as fast my way as yours and as cheaply . I always drink it black and tea white . I was told by the French only traditional Nescafe is good instant coffee . It must be premixed with a little cold water .
I think a feedback discussion is very useful . I have noticed that the supposed advantages of feeding the inverting input of an amp never seem to have materialized for me . I always try it as I feel it must be right that it is better . I will one day when I make my mind up use it for an MC cartridge . Jan Alertz ( spelling ) suggests 845 R as optimum load . At least with that one the starting point is defined .
The Maxwell House factory was in Banbury close by . It might still be there . The smell is like it tastes . I was told dry cleaning fluid was used to extract the last residuals . I doubt it is true .
Bootstrapping an input appeals to me . Also bootstrapping valve anodes . Don't see that very often . Then we have feed forward to consider .
I noticed one thing about taste . No one likes bad rhythm . I suppose because it underpins speaking ?
I think a feedback discussion is very useful . I have noticed that the supposed advantages of feeding the inverting input of an amp never seem to have materialized for me . I always try it as I feel it must be right that it is better . I will one day when I make my mind up use it for an MC cartridge . Jan Alertz ( spelling ) suggests 845 R as optimum load . At least with that one the starting point is defined .
The Maxwell House factory was in Banbury close by . It might still be there . The smell is like it tastes . I was told dry cleaning fluid was used to extract the last residuals . I doubt it is true .
Bootstrapping an input appeals to me . Also bootstrapping valve anodes . Don't see that very often . Then we have feed forward to consider .
I noticed one thing about taste . No one likes bad rhythm . I suppose because it underpins speaking ?
The phrasing is, let's be gentle here, vague.
It's like saying it uses "conditional electricity conducting devices", meaning transistors.
As for "unique", be extra careful with that. In its day, Norwegian Tandberg had a "unique" patented recording system, called "Actilinear", which turned out to be a copy of Studer/reVox' system of many years before that. A battle ensued via the magazines, ended up in court, and Tandberg lost, only to be bankrupted a few months after that.
The initial fighting broke out on the pages of HFN&RR, much to their credit.
I phoned Ross Walker of Quad one day about a possible patent infringement if I produced something . He was so pleased that someone was honest about it that he offered help me . He said as long as I didn't call it by the same words he used he wouldn't mind . I asked how come ? he said " there are only three ways to hook up a transistor and one of them is wrong " . He said went on to say if it works usually it has been tried and published .
In slight contradiction to Ross , plate to plate feedback as it is perhaps wrongly called ( note I say perhaps ) with valves has some interesting consequences . It almost on first look seems to the wrong one being used ( anode ) . Never tried it with transistors . I imagine it would have some merit . Zealots of plate to plate seem to get a bit carried away . Things they say don't work in fact work well .
In slight contradiction to Ross , plate to plate feedback as it is perhaps wrongly called ( note I say perhaps ) with valves has some interesting consequences . It almost on first look seems to the wrong one being used ( anode ) . Never tried it with transistors . I imagine it would have some merit . Zealots of plate to plate seem to get a bit carried away . Things they say don't work in fact work well .
I would hope that our armed forces are taught the truth about electronics, but military and science don't always mix too well.nigel pearson said:
You may suspect it. That doesn't make it true. The 82 has a very different grid from the 81, because the 82 was derived from a power triode 6C4/EC90. I suspect that the 82 grid suffers greatly from 'island effect', which is much less of an issue for the 81.
If you are talking about re-entrant feedback distortion producing higher order terms then this has nothing whatsoever to do with 'echo' or signal delays. It is a simple consequence of algebra and occurs even if things happen instantaneously (as they effectively do in most audio amps anyway).
The root issue can be simply summed up (sorry!) as
1/(1+x) = 1 -x +x^2 -x^3 +x^4 . . . (off to infinity)
e.g. the inverse of a polynomial is usually an infinite series.
If it means anything, it probably means that the driver BJT has its base current fed from a high impedance (e.g. signal-controlled CCS). It could be as simple as the collector of the previous stage.Charles Darwin said:
It means you swap exponential law non-linearity from voltage feed for beta droop non-linearity etc. for current feed. "Unique" means we want you to believe that nobody else has ever thought of feeding a BJT base directly from a BJT collector. "Floating" means high impedance, and may imply good PSRR at that point in the circuit. "Current-driven" I have explained above.
If you mean so-called Schade connection this is merely a way of doing an anode follower. The signal feedback actually goes to the grid after the first anode; it merely shares the coupling cap from the previous stage and uses the previous stage output impedance as the effective input impedance for the valve analogue of the opamp inverting configuration. Nothing special.nigel pearson said:
If you mean so-called Schade connection this is merely a way of doing an anode follower. The signal feedback actually goes to the grid after the first anode; it merely shares the coupling cap from the previous stage and uses the previous stage output impedance as the effective input impedance for the valve analogue of the opamp inverting configuration. Nothing special.[/QUOTE] DF96
I think that is right . I spent a month testing it to be sure . However by a gnats whisker I think it is valid . The cathode of the driver shows unexpected complexities when analyzed ( too long even for me ) . I am not convinced valve people should use it . It seems against what they believe in ? Me, I am agnostic as it works and that's all there is to say . It certainly isn't as dependent on valve type as some say . That is not to say unimportant . Someone said it offers a cheap lunch but not a free one . That's about right .
I think that is right . I spent a month testing it to be sure . However by a gnats whisker I think it is valid . The cathode of the driver shows unexpected complexities when analyzed ( too long even for me ) . I am not convinced valve people should use it . It seems against what they believe in ? Me, I am agnostic as it works and that's all there is to say . It certainly isn't as dependent on valve type as some say . That is not to say unimportant . Someone said it offers a cheap lunch but not a free one . That's about right .
Hi Nigel;
I use such feedbacks in my amps. Actually, as DF96 said, it is feedback from anode to the first grid of the tube. It is feedback by voltage, applied in parallel with input signal. In output stage it goes directly from anode, instead of from secondary of the output transformers, so does not involve frequency-dependent phase and resonant properties of the transformer from primary to secondary windings. It is like to use a tube with lower Rp and lower distortions.
Feedback ratio is defined by resistances of feedback resistor and anode load of the driving stage. It decreases effective load resistance of that driver, so non-linearity of it's output resistance causes increased distortions. That's why I use either pentodes, or triodes with local feedback in cathode, as drivers, so variations of their output resistance have less impact on resulting distortions. Global feedback then decreases distortions of the driver as well.
I use such feedbacks in my amps. Actually, as DF96 said, it is feedback from anode to the first grid of the tube. It is feedback by voltage, applied in parallel with input signal. In output stage it goes directly from anode, instead of from secondary of the output transformers, so does not involve frequency-dependent phase and resonant properties of the transformer from primary to secondary windings. It is like to use a tube with lower Rp and lower distortions.
Feedback ratio is defined by resistances of feedback resistor and anode load of the driving stage. It decreases effective load resistance of that driver, so non-linearity of it's output resistance causes increased distortions. That's why I use either pentodes, or triodes with local feedback in cathode, as drivers, so variations of their output resistance have less impact on resulting distortions. Global feedback then decreases distortions of the driver as well.
Hi Anatoliy .
I until recently I treated tube amps like steam locomotives . Great but of a bygone age . My brother left me some unfinished ones . My father did teach me about them when I was 4 ( 1960 ) . College also , that was a long time ago .
I have just ordered new transformers so as to know where I am going as adding global feedback was disappointing . I have confidence it will work so I will not give up . The measurements are good so it must be just details to get right . Like repairing motorcycles I have to learn it for myself . I like not knowing as it can be exciting to find something working differently to expectations .
I have enjoyed doing this and now feel qualified to speak when friends describe their efforts to make tube amps . Not an expert , however I feel I have my driving license now .
I was promised some Gu 50 tubes . That's the last bit to try . I think from what you say I am on the right track as all of it is familiar .
I until recently I treated tube amps like steam locomotives . Great but of a bygone age . My brother left me some unfinished ones . My father did teach me about them when I was 4 ( 1960 ) . College also , that was a long time ago .
I have just ordered new transformers so as to know where I am going as adding global feedback was disappointing . I have confidence it will work so I will not give up . The measurements are good so it must be just details to get right . Like repairing motorcycles I have to learn it for myself . I like not knowing as it can be exciting to find something working differently to expectations .
I have enjoyed doing this and now feel qualified to speak when friends describe their efforts to make tube amps . Not an expert , however I feel I have my driving license now .
I was promised some Gu 50 tubes . That's the last bit to try . I think from what you say I am on the right track as all of it is familiar .
In 1960 I liked look and smell of a tube radio.
Looking inside I imagined that it is a big plant, and I was imagining walking inside, between tall tubes.
Seriously, Gu-50 are dangerous tubes, they require at least 600V on anode in pentode mode to be usable. I use them with 800V on anodes. However, in triode mode you can go with 400V, but anyway it may kill.
Looking inside I imagined that it is a big plant, and I was imagining walking inside, between tall tubes.
Seriously, Gu-50 are dangerous tubes, they require at least 600V on anode in pentode mode to be usable. I use them with 800V on anodes. However, in triode mode you can go with 400V, but anyway it may kill.
Even worse, i can smell it to this day.
(though snorting valve fume in 1960 is a wee early for me)
Sounds right up my street . I have a 440V transformer that will start me off .
When I was 4 my dad caught me with the back off of the radio ( Sobel with 10 x 6 Goodmans speaker 28 Gns or £29.40 ) . Instead of telling me off he explained how it worked . He said that is the part you must be careful of pointing to the transformer . When he was at work I would used his hi fi . I even made needles for 78's on a little machine he had . He caught me and was furious . My mother defended me and said I had been doing it for ages . That night he gave me his hi fi .
One day dad was overlooking a valve amp my brother made and asked about how it worked . We explained and he said where did you boys learn this ? From you dad ! He was a radar instructor .
When I was 4 my dad caught me with the back off of the radio ( Sobel with 10 x 6 Goodmans speaker 28 Gns or £29.40 ) . Instead of telling me off he explained how it worked . He said that is the part you must be careful of pointing to the transformer . When he was at work I would used his hi fi . I even made needles for 78's on a little machine he had . He caught me and was furious . My mother defended me and said I had been doing it for ages . That night he gave me his hi fi .
One day dad was overlooking a valve amp my brother made and asked about how it worked . We explained and he said where did you boys learn this ? From you dad ! He was a radar instructor .
Sorry to reprint it all . I know that as it has been said before , but I feel uneasy with it . To me a loop must have echos , or loops . If the energy is dissipated , OK it is gone . Whatever the reason when high order harmonics appear I say I got that wrong didn't I . Then I relax the feedback loop and it's gone . Funny even if my maths let me down my spectrum analyzer didn't .
If you have time this is worth a read . Worth printing . It is a total package , nearly all topologies are covered .
http://www.its.caltech.edu/~musiclab/feedback-paper-acrobat.pdf
Last edited:
- Status
- Not open for further replies.