Mikeks,
Instead of repeatedly referring to a quote from Cherry, taken out of context and without the figures referred to, could you instead point out what is wrong in the mathematics of all those applications note that we have linked to? They agree with each other, and I arrive at the same result when I try to repeat the exercise. The quote from Cherry make me suspect he is not really talking about the same type of circuit.
Instead of repeatedly referring to a quote from Cherry, taken out of context and without the figures referred to, could you instead point out what is wrong in the mathematics of all those applications note that we have linked to? They agree with each other, and I arrive at the same result when I try to repeat the exercise. The quote from Cherry make me suspect he is not really talking about the same type of circuit.
OK, I have had a brief look at that section in Cherrys paper where he discusses CFB amps. He obviously does accept the concept itself and its properties, as described in various other papers. What he does not accept is the name, and that it is given a special status distinguishing it from VFB. He also beliefs it is better to use a VFB amp with what he calls a more optimal choice of resistors.
So basically it is just the philosophical and terminological dispute that we have had also here on the forum zillions of times before. It is just different ways of looking at it and defining it. I could even say that Cherry, which I presume to be in academia, is surprisingly pragmatic, and do not view things as abstractly as the engineers at the semiconductor companies do, and who would be expected to be the less abstract and more pragmatic ones.
Cherry, you and some others, think that everything fitting the first of the four feedback schemata posted in this thread, and in Cherrys paper, should be labelled VFB, whatever properties the amplifier four-pole in the schema has. The semiconductor companies and many of us others think that VFB and CFB are theoretical models, which are both subclasses if this schema, and no real amp is truly only VFB or CFB. VFB is in our opinion basically the usual model of the ideal op amp, while CFB diverts from this ideal model, thus forming a different but also ideal model. Both of them form extreme cases of that singe schema.
So basically it is just the philosophical and terminological dispute that we have had also here on the forum zillions of times before. It is just different ways of looking at it and defining it. I could even say that Cherry, which I presume to be in academia, is surprisingly pragmatic, and do not view things as abstractly as the engineers at the semiconductor companies do, and who would be expected to be the less abstract and more pragmatic ones.
Cherry, you and some others, think that everything fitting the first of the four feedback schemata posted in this thread, and in Cherrys paper, should be labelled VFB, whatever properties the amplifier four-pole in the schema has. The semiconductor companies and many of us others think that VFB and CFB are theoretical models, which are both subclasses if this schema, and no real amp is truly only VFB or CFB. VFB is in our opinion basically the usual model of the ideal op amp, while CFB diverts from this ideal model, thus forming a different but also ideal model. Both of them form extreme cases of that singe schema.
Christer,
I can see your point(s), but I also think that the term CFB coming from the semi corporations is not necessarily the result of their conviction that indeed it is a different class. I am convinced that an important reason behind it was for marketing purposes.
It is difficult to convince customers that your umpteenth opamp is really different, unless you give it a different name. Hence CFB.
And possibly the guy who coined the term wasn't aware of the previous use of the term current feedback in another context.
These are historical things, and the best we can do is to make clear what we mean whenever we use the terms.
Jan Didden
I can see your point(s), but I also think that the term CFB coming from the semi corporations is not necessarily the result of their conviction that indeed it is a different class. I am convinced that an important reason behind it was for marketing purposes.
It is difficult to convince customers that your umpteenth opamp is really different, unless you give it a different name. Hence CFB.
And possibly the guy who coined the term wasn't aware of the previous use of the term current feedback in another context.
These are historical things, and the best we can do is to make clear what we mean whenever we use the terms.
Jan Didden
mikeks said:
I am not sure I buy his simplified circuit as an equivalent, so in there I agree with you, but would make his case weaker.
He does agree that an amplifier with unity gain buffer between the inputs have the same properties as other papers claim. I admit I didn't take time to check his equations, but trusted his own interpretatin of them.
Hello Jan,
I am very happy to see you posting again. Having seen your most recent previous posting I was afraid you had left us for good, which would have been very sad.
Of course there could be a lot of truth in your suggestion too, and they decided they had to overemphasise the differences for marketing reasons. I am not really claiming that either way of looking at this is more right than the other. Further, as I have stated many times before, although I think it is useful to distinguish the two different (sub)types of amplifiers, I am in no way defending any particular names. I agree that the term 'current-feedback amplifier' is perhaps not the most fortunate choice of name in this case, but I use it since it used a lot and good alternatives seem to be lacking. We might consider calling in an H-V amplifier, though, according to the classification in the paper somebody linked to earlier in this thread. We could call them apple amplifiers and pear amplifiers and say that both these are of voltage feedback type, if you prefer that. A name is just a name, but I think there is a point in distinguishing the concepts, whatever we choose to call them. That said, perhaps there is some even better way of doing the classification. I do admit that CFB amps disguised as VFB amps, such as LM6171, raise certain questions about how to best categozise these things. However, I am not a dictator of the IEC, ISO, ANSI or similar, so for the moment I prefer to stick to the VFB/CFB distinction that after all most people seem to agree with.
I am very happy to see you posting again. Having seen your most recent previous posting I was afraid you had left us for good, which would have been very sad.
Of course there could be a lot of truth in your suggestion too, and they decided they had to overemphasise the differences for marketing reasons. I am not really claiming that either way of looking at this is more right than the other. Further, as I have stated many times before, although I think it is useful to distinguish the two different (sub)types of amplifiers, I am in no way defending any particular names. I agree that the term 'current-feedback amplifier' is perhaps not the most fortunate choice of name in this case, but I use it since it used a lot and good alternatives seem to be lacking. We might consider calling in an H-V amplifier, though, according to the classification in the paper somebody linked to earlier in this thread. We could call them apple amplifiers and pear amplifiers and say that both these are of voltage feedback type, if you prefer that. A name is just a name, but I think there is a point in distinguishing the concepts, whatever we choose to call them. That said, perhaps there is some even better way of doing the classification. I do admit that CFB amps disguised as VFB amps, such as LM6171, raise certain questions about how to best categozise these things. However, I am not a dictator of the IEC, ISO, ANSI or similar, so for the moment I prefer to stick to the VFB/CFB distinction that after all most people seem to agree with.
mikeks said:
No, it is a special case of that, whatever we call it. Then, if VFB should also refer to another special case or the whole of this class (thus including CFB or whatever we call it) seems to be a dilemma. The actual characteristics of the amplifier four pole in the class you refer to actually matter for the overall properties of the circuit.
Being a reasearcher I don't like that either, but as I said above, they do not really rename that whole class, but rather give a name to a subcase. Cherry seems to avoid the problem by acknowleding this special class but considering it of questionable use, thus not needing a name of its own. Many others seem to find it useful, which means we do need a name for it.
What do you think of that more fine grained classification in the paper linked to earlier, where the "CFB amp" was classified as one of nine different cases under the name H-V amp?
BTW, being a computer scientist I have seen the industry reinventing new names for the same old well known concepts over and over again every five years of so, that I have almost gotten immune to such things. I see new terms and abbreviations all the time, not even bothering to find out what it is, because in 99 % of the cases when I ask somebody about it I realize, aha it is just a new name for this old concept.
I don't mind if we do an attempt to standardize terminology on this forum, though, to understand each other better, but will that work out in the end?
mikeks said:
Agreed, but that's life. I don't want to start again the naming discussion, but you can't always say: aha, the semi corporations call it XWZ so that's what it is; they have other ereasons than engineering and logic to name their products.
Take the AD844, which really is a 2nd generation current conveyer with a buffered output. I once overheard a discussion on the AD844 with an AD rep asking why they called a CCII an opamp. Answer: "Are you crazy? Nobody would buy it unless we call it an opamp".
Jan Didden
....so... ash_dac asked about an opamp (or a discrete design) with a certain content and properties which the market calls "current feedback" was or was not suitable for audio.
I think we all can agree on that a VFB looks in a certain way with it's properties and a CFB is different and can't be used everywhere and when you do can use it you must know how to connect it.
.... so _is_ CFB suitbale for audio? What do you say Mike? (I have already said yes)
I think we all can agree on that a VFB looks in a certain way with it's properties and a CFB is different and can't be used everywhere and when you do can use it you must know how to connect it.
.... so _is_ CFB suitbale for audio? What do you say Mike? (I have already said yes)
peranders said:
I agree. We should not a priory limit ourselves to traditional solutions. If you have a specific application, there are generally many different options, CFB, VFB, discrete, hybrid, tubes, etc. Part of the fun in audio for me is trying out radical different things.
Look at chipamps. You can't say that they are or are not suitable. They have there own set of advantages and disadvantages.
VFB generally is better in precision and low offset, CFB generally better in bandwidth. But either can be used in a high quality audio design.
Jan Didden
which was...
So far, when I listen, I like the sound quality. If you listen, what is opinion?
.... and the proof of that is.....mikeks said:
PM...there is nothing to recommend so-called 'current feedback' in audio applications, unless such a topology is used as part of a compound output stage with gain in excess of unity.
So far, when I listen, I like the sound quality. If you listen, what is opinion?
- Status
- Not open for further replies.