I am not sure to understand and/or agree with your definition of current or voltage feedback.
My personal understanding is the following :
A CFA is an amplifier where I can replace the feedback network by a current source controlled by output.
A VFA is an amplifier where I can replace the feedback network by a voltage source controlled by the output.
All the so called CFA I know are -according to my definition- VFA. This is not to say that so called CFA are identical to classical VFA, I see some differences like the impedance of the inverting input and the fact that the open loop behavior depends of the feedback network.
But I am pretty sure that all that can be done with so called CFA can be done with VFA.
This is an example of what is for me a CFA : feedback is done with only one resistor (Rfb) that I can replace by a VCCS ( a resistor is a particular case of VCCS)
Attachments
I am truly amazed at the inferences you draw from what I have written. You must get an incredible amount of exercise jumping to conclusions. First you accuse me of varying the values of gm in an earlier post when I did no such thing. Now you assume that I am unaware of the "undergraduate level" classical types of the four feedback topologies and their names and consequences. (I won't waste your time mentioning my post-graduate degrees.) You have spent an extraordinary amount of effort informing me of things of which I am quite aware.
By the way, you might ask yourself into which of these four categories of the traditional combinations of two types of sampling and two of summing a simple differential input, single-ended output amplifier falls. Doesn't fit, does it?
In another indictment of this nomenclature, we note that parallel summing at the inverting input is called current feedback. Now logically, this would suggest that if we change the amount of current flowing, we would change the amplifier operation. But if we simply multiply both impedances connected to the feedback input in a VFA by the same scale factor, we change the "feedback" "current", but, ignoring the effects of parasitic capacitance, there is no effect on the voltages at the amplifier output. Yes, yes, I know the "names" of this type of so-called "current" feedback. But it's pretty darned ineffective "current" feedback if changing the current has no effect on the output now, is it not? How utterly confusing...
Please refer to page 307 of the Radiotron Designer's Handbook, 4th Edition, published in the 1940's and available online for free. The circuit I described was mentioned therein and was called a Current Feedback Amplifier.
If we can get past the above, we can discuss the remainder of what you wrote if you'd like, but it would seem a pretty big bridge to cross at this point.
At some point in this thread someone mentioned that a good criterium for calling a topology 'CFA' would be that the output current of the input pair is in fact the current into the inverting input. In contrast, in a VFA the output current of the input pair comes from the tail current.
Not sure how/why we lost that on the way but to me it seemed sensible at the time. Still does, in fact.
Jan
Not sure how/why we lost that on the way but to me it seemed sensible at the time. Still does, in fact.
Jan
This is exactly the type of response I was expecting, which discourages any further discussions. Disregarding of your education level, you seem to have difficulties compiling what others are saying, a selective attention, and insist creating your own version of reality. How could anybody else that went through a feedback 101 course accept that
Which is claiming CFA defines a n-th feedback type, where n>4, exactly what it doesn't, since the amp it is STILL a two port.
or
FYI, feedback theory doesn't assume any inverting or non inverting inputs.
Oh, wait, that's not what you said, right? And then the show must go on 😀.
So I'm afraid I cannot get past this point; at least I tried, although I should take Reodor's advice, to start with. BTW, you are not the only reader here, so please don't think that everything it is said is addressed to you. There is a silent audience which knows how to distinguish good from trash.
Happy fighting your war, now I'm really done here. I would dare to advise everybody else, this is not a war that can be won, since it is not carried with conventional wisdom.
In fact what you are saying is that a CFA has a hybrid input.
The non inverting input has a very much higher impedance as the inverting input.
Isn't this a better way of saying it ?
Hans
No I don't think so Hans because then you immediately get the discussion about at what impedance level you draw the demarcation. I could imagine a high impedance starved emitter follower and a low impedance hot base input.
If you use the actual topology as the criterium as I mentioned above, there's no scope for confusion, at least not that I see.
If the -inp current is the output current, that works for CFAs formed with BJTs, FETs, even IGBTs if you like.
Edit: Come to think of it, if you want to give different names to different topologies, then you should take the topology as the deciding factor, no? Too much bandwidth has been wasted by trying to define things based on all sorts of things, even whether a BJT is current or voltage controlled*, which absolutely has nothing whatsoever to do with a topology. It is the topology that determines whether the -inp current is the input stage output current, or whether the tail current is the input stage output current. Clean, clear, unambiguous.
* it's both, of course. You pick one view depending on which one makes it easiest to analyze a particular circuit. But both the voltage-controlled and the current-controlled view can be used to correctly analyze any circuit, any topology.
Jan
Last edited:
But then, what exactly is the output current of the input pair ?
Aren't you saying that the inverting input draws practically all the current of the sum of both input currents.
In a VFA each input draws about the same amount of current.
Why is it a contrast that the output current now comes from a tail current.
Is it important where this current comes from ?
Hans
An excellent post. You exposed the fundamental points of the topic.
Thanks a lot, Syn08.
The collector current of the device that receives the current feedback on its emitter, the -input, IS the output current that goes to the current mirrors etc. The two currents are basically identical except for the base current. Say 99% identical. There is a direct relationship between the two, 1:1.
In a VFA input stage, the output current from the input device is a (fraction of) the tail current. That output current depends on the -inp of course, but there is no direct 1:1 relationship between that -in voltage and the tail current, it depends on circuit details. In a CFA, whatever the circuit details, the -inp current IS the input device output current.
The question is not whether this distinction is important, I'm just saying it is a convenient way to separate CFA topology from VFA topology. Beats anything I've seen here so far.
PS We talk about signal currents of course, disregard DC bias for this discussion, it is irrelevant for the discussion.
Jan
Last edited:
I could answer using the same provocative manner as yours but I won't. As a discussion with you, years ago, ended by the intervention of a moderator, I do not want to see this repeated.
P.S. Because you mentioned it, I spent a few minutes checking the 1940 reference you mentioned. I concluded this is yet another proof we live in parallel universes. At page 307 I can’t find any relationship with the “circuit I describe” and the only “current feedback” mentioning is in the context of output current sampling, which is the original Black naming convention. Oh wait, this is not what you were saying, right? 😀
Anybody can check it out http://www.tubebooks.org/books/rdh4.pdf pp. 307. Sorry, you have to do better than that in finding references, if you want to be taken seriously.
P.P.S. Perhaps you are confused by the +/- signs in the little generic schematics illustrating the feedback types? Those have obviously nothing to do with inverting and non inverting inputs 😀, but with the conventional signs for the circuits voltages. Something that should be obvious to any EE student that has ever dealt with a two-port analysis.
Last edited:
That’s the definition I proposed in 2014 - I put an article up on my website about it at the time as well.
What kind of ego assumes that my responses to you are indicative of those to others, especially when you have only recently re-joined this thread? Perhaps it's that you entered with an agenda crafted from listening too much to Reodor and foor, the former whom you've referenced and the latter who's referenced you?
Of course, it's simply impossible that you could be part of the problem, right? You, who accused me of asserting that the value of gm was varying when there was zero evidence of me having done so? Now just who is creating their own version of reality, huh?
Discourages further discussions? When I've asked questions that you've chosen not to answer? Who is it again that is discouraging discussion? There is no reply to the question about the feedback class of the differential input amplifier, and no response to the Radiotron reference to CFAs. Unless, of course, you want to count ad hominems, a type of response in which your associate Reodor also is well versed.
Where in the world are you getting this stuff from?
Yes, of course (sigh). I should have said "feedback input." Nitpicking minutia.
[/QUOTE]
Indeed. It seems that some of what passes for "conventional" wisdom is vindictive, agenda-driven, and challenge-averse. If you are "really done here," I will say that I will not miss your attitude. However, it does seem that you have some knowledge of this topic, and it's a shame that we will be losing that.
Yes, please do check it it out. The term is "current feedback" not "current sampling." I will grant you that the word "amplifier" is missing. Is that that important to you?
No confusion here. Wow, you and your friends really do specialize in ad hominems, don't you? Are you unable to make a point without getting personal?
When I say current or voltage feedback or abbreviate it as c.f or v.f., I am referring to to the Middlebrook Dual Insertion Test sense of this term. The article https://www.edn.com/design/analog/4458753/In-defense-of-the-current-feedback-amplifier is good reading and on page three under the section "Current Feedback or Voltage Feedback" you can find a good discussion of this technique. The type of the lesser of the two loop gains in a given circuit indicates the predominant type of feedback in that circuit.
My definition is that the signal current in a CFA's input stage flows through the feedback network, while that of a VFA does not.
Chris, please stop playing the victim game. You are not the subject of any campaign of ad hominem attacks, but only your understanding of feedback theory is under fire.
Please understand that by denying the evidence, shifting the targets, making unsubstantiated statements, challenging basic EE knowledge, denying you own statements made five posts up, making up quotes to defend the indefensible, etc... you are not helping the CFA cause (if anybody believes it needs to be defended, I don’t). I find that sad, but then it’s ultimately your own reputation on the line, not mine.
P.S. Check your facts. I left this forum over 8 years ago so I am new to this raging debate. There are enough people still around from then, to certify that I have no agenda or axe to grind against anything or anybody. Except Bybee devices, Shakti stones, green markers, magic pebbles, directional wires, and a few others, none of which apply here.
Please understand that by denying the evidence, shifting the targets, making unsubstantiated statements, challenging basic EE knowledge, denying you own statements made five posts up, making up quotes to defend the indefensible, etc... you are not helping the CFA cause (if anybody believes it needs to be defended, I don’t). I find that sad, but then it’s ultimately your own reputation on the line, not mine.
P.S. Check your facts. I left this forum over 8 years ago so I am new to this raging debate. There are enough people still around from then, to certify that I have no agenda or axe to grind against anything or anybody. Except Bybee devices, Shakti stones, green markers, magic pebbles, directional wires, and a few others, none of which apply here.
Last edited:
Fine. So we're done then. Wow! A thread that ends in a major agreement, that's a real accomplishment.
Except maybe for a some rear-guard skirmishes in a few isolated pockets, they'll die out when left to themselves ;-)
Jan
God has the last say, but the Dutchman has the last word. 🙂
Wishing you health and happiness for 2019, Jan,
Patrick
Wishing you health and happiness for 2019, Jan,
Patrick