mikeks said:
What 'doesn't exist' is in fact a variation on voltage feedback....Get it?
No, you are missing the point, although now you are starting to contradict yourserlf, which is confusing. Are you now saying that CFB in the sense used in this thread does after all exist, although you prefer to view it as a variation on VFB? Either it exists or it doesn't exist. You may be undecided about which is the case, but I hope you agree that only one of these alternatives is true.
Anyway, the point of my original post was not to debate whether CFB exists or not. I got tired of the non-constructive disussions about this in previous threads and had no intention of starting all that over again. Rather my post was a comment on the more abstract level of argumentation. Forget about VFB and CFB. Let's just say that Jack made a comment about some property of X. Then you disagreed with that claim while at the same time you confirmed that you don't think X exists, which not sensible. To bring it back to a more concrete level, consider the following example. Person A who believes in god claims that god has a certain property. Then person B who does not believe in god insists that god does not have the property A claimed. That is not very reasonable, don't you agree?
mikeks said:
Oh no, are you back to that old pointless discussion again? The term current feedback is used in the literature for at least two different concepts. That is unfortunate, but something we have to live with. Such clashes of terminology are not uncommon, as you should very well be aware of if you have actually ever managed to get a university degree. This thread is discussing CFB in one of these senses so it is totally pointless if you insist on interpreting it in the other sense in this context.
Of course there is nothing to prevent using also other, non-clashing, terms for these phenomenae, but until we have universally agreed-on such terms, it would probably just cause even more confusion than we already have.
Christer said:
Ill-informed literature....The truth and nothing but:
http://www.diyaudio.com/forums/showthread.php?postid=947000#post947000
Christer said:
mikeks said:
Truth? What har truth got to do with this? We are talking about definitions.
I agree that those diagrams cover one of the concepts often referred to as current feedback, although it is funny that you refer to it without further comment since the term current feedback cannot be found in the text there.
They do not, as far as I can see, cover the other concept referred to as CFB, as far as I can see. I still cannot second guess you if you are only protesting about the term CFB in that case or of deny the existence of the concept? I agree that the term CFB is unfortunate, but for the lack of a better and accepted alternative, I think we have to stick to it. If you insist the concept doesn't even exist, then I must be very stupid since I have managed to derive equations proving that it does indeed behave differently from ordinary VFB. Well, that doesn't bother me then, since it only means I am just as stupid as the engineers at TI, AD and other semiconductor companies who have come to the same conclusion and arrived at the same equations.
This discussion still has not given a clue about what your original comment meant, since you have just avoided the question by constantly changing topic. Oh well, I guess I knew all the time that was what would happen, so why did I bother to ask.
Christer said:
I agree that those diagrams cover one of the concepts often referred to as current feedback, although it is funny that you refer to it without further comment since the term current feedback cannot be found in the text there.
I just cannot be bothered to go over this territory again:
http://www.diyaudio.com/forums/showthread.php?postid=423522#post423522
http://www.diyaudio.com/forums/showthread.php?postid=859757#post859757
It gets even more unfortunate when you look at the terms used in switch mode power supply literature. There are two general feedback concepts, one is called “Voltage mode” and the other is called current mode control.
The “voltage mode” refers to a system that measures the output voltage and uses this as input to a compensation circuit whereas the “current mode” refers to a system that uses both information about the switching current in the main switch and information about the output voltage.
Generally we need to agree on terms – why most books on the subject spend a great deal of time explaining what different terms cover. I guess this could be needed in this discussion also.....
Peace out
\Jens
The “voltage mode” refers to a system that measures the output voltage and uses this as input to a compensation circuit whereas the “current mode” refers to a system that uses both information about the switching current in the main switch and information about the output voltage.
Generally we need to agree on terms – why most books on the subject spend a great deal of time explaining what different terms cover. I guess this could be needed in this discussion also.....
Peace out
\Jens
mikeks said:
I agree with that, except that there is also another CFB concept in the literature, which requires that you consider also the properties of the voltage-transfer function. I am not sure what your problem with this is. Of course, what the op amp manufacturers call VFB and CFB are just two theoretical models with fundamentally different properties for the voltage transfer function, and all real (op) amps will be somewhere in between these two extremes. However, they are usually designed to approximate either of these two ideal models.
I could even argue that feedback doesn't exist. It is just a conceptual model we use make it easier to design and analyse amps and other systems. The feedback is only visible if you write the mathematical equations on certain forms, and it disappears if you rewrite the equations. Hence, it is somewhat arbitrary to claim that a certain kind of feedback does or does not exist. We simply define the conceptual models we find useful.
Of course, all this has been a sidetrack since you insisted on dragging up the old discussion again instead of trying to answer my original question, although you have managed to make the post I originally commented on even more confusing on the way.
mikeks said:
Of course I can. I just did. Didn't you read the rest of my post?
However, I must admit I haven't investigated this matter in depth, so it may be that the equations describing a system always can be transformed into a certain canonical form if and only if they are what we usually classify as feedback systems. It is however not generally possible to decide if a system has feedback or not from just looking at an arbitrary collection of equations that describes the system correctly.
http://people.ee.ethz.ch/~hps/publications/2000cas.pdf
presents a classification scheme that shows the "cfa/cfb op amps" as one of 9 "op amps"
Of course you have to be prepared for clarifications that read like this:
"All amplifiers [in Table I] have already been named in the literature, with the exception of the H-I amplifier, which we call current-feedback OTA (CFB OTA) because its relation to the OTA is the same as the CFB op amp’s relation to the op amp. Both names are misleading, since the CFB OTA is actually a current amplifier and not a transconductance amplifier, just as the CFB op amp is a transresistance amplifier and not a voltage amplifier."
elsewhere in this thread someone mentioned "cfb" op amps not running out of gain - not true, they have a open loop voltage gain defined by Rt/r_in, the transresistance and the input resistance as well as an ultimate "gbw" defined by the internal Ct (which is in parallel with Rt) and which cannot be charged/discharged faster than the r_in input resistance allows (yes there are some "cfa/cfb op amps" using multiplying current mirrors so the r_in or the Ct could be seen as being divided by the current mirror gain)
I must agree that as much as I like the specs of the tpa6120 as a audio amp I couldn’t just use it by itself – it begs to be put in a multiloop amplifier (in fact I used 6 of the tpa amps per channel)
http://www6.head-fi.org/forums/showthread.php?t=175815&page=28
my amp pics #542,4 explanation post #549, ~ halfway down the page
presents a classification scheme that shows the "cfa/cfb op amps" as one of 9 "op amps"
Of course you have to be prepared for clarifications that read like this:
"All amplifiers [in Table I] have already been named in the literature, with the exception of the H-I amplifier, which we call current-feedback OTA (CFB OTA) because its relation to the OTA is the same as the CFB op amp’s relation to the op amp. Both names are misleading, since the CFB OTA is actually a current amplifier and not a transconductance amplifier, just as the CFB op amp is a transresistance amplifier and not a voltage amplifier."
elsewhere in this thread someone mentioned "cfb" op amps not running out of gain - not true, they have a open loop voltage gain defined by Rt/r_in, the transresistance and the input resistance as well as an ultimate "gbw" defined by the internal Ct (which is in parallel with Rt) and which cannot be charged/discharged faster than the r_in input resistance allows (yes there are some "cfa/cfb op amps" using multiplying current mirrors so the r_in or the Ct could be seen as being divided by the current mirror gain)
I must agree that as much as I like the specs of the tpa6120 as a audio amp I couldn’t just use it by itself – it begs to be put in a multiloop amplifier (in fact I used 6 of the tpa amps per channel)
http://www6.head-fi.org/forums/showthread.php?t=175815&page=28
my amp pics #542,4 explanation post #549, ~ halfway down the page
Christer and everybody else, I'll suggest that you take a peek at Walt Jung's book. Mike, I think you too should take a look. 
Page 26 and 115
http://www.analog.com/library/analogDialogue/archives/39-05/Web_Ch1_final_R.pdf
Page 26 and 115
http://www.analog.com/library/analogDialogue/archives/39-05/Web_Ch1_final_R.pdf
peranders said:Christer and everybody else, I'll suggest that you take a peek at Walt Jung's book. Mike, I think you too should take a look.
Page 26 and 115
http://www.analog.com/library/analogDialogue/archives/39-05/Web_Ch1_final_R.pdf
Oh my, it seem that Walt is just as stupid as I am.
Feedback of any kind, is really not any good in audio. The problem is basically, that the good THD, S/N ans other spec, are achieved by comparing the input and the output, and then correcting the difference.
But... you can't fix the problem before it's already there. So... when you correct the signal, it is already damaged.
THD-measurements will not suffer from this, but sound will.
Therefor non-feedback amps, although having 1000 times more distortion, normally sounds better.
If you have to use feedback, make sure to have a very fast loop. The faster the loop, the less damage to the sound.
Current Feedback is great for high speed, and therefor also for audio.
See.: http://www.analog.com/UploadedFiles/Application_Notes/492001115525484056221917334AN211.pdf
But... you can't fix the problem before it's already there. So... when you correct the signal, it is already damaged.
THD-measurements will not suffer from this, but sound will.
Therefor non-feedback amps, although having 1000 times more distortion, normally sounds better.
If you have to use feedback, make sure to have a very fast loop. The faster the loop, the less damage to the sound.
Current Feedback is great for high speed, and therefor also for audio.
See.: http://www.analog.com/UploadedFiles/Application_Notes/492001115525484056221917334AN211.pdf
- Status
- Not open for further replies.