Well, perhaps not fully, but I agree that it is misleading.
I'm sure that you will understand when I decline to even consider commenting on what an indistinct group of others have said about the CFA.
Thank you for commenting.
A conclusion of post 2548 is that gain blocks with finite, non-zero input impedances have finite, non-zero voltage and current loop gains and finite, non-zero voltage and current feedbacks. The dividing line for loop gain predominance is the point at which zf = rd.
In what are called VFAs, practically speaking, rd is a lot larger than zf, and so there is virtually no current feedback and a lot (in some cases approaching infinity) of current loop gain. There is substantial but much less voltage loop gain and significant voltage feedback.
In what are called CFAs, rd is generally (but not always) smaller than zf. There are always significant amounts ( but none approaching infinity) of all four parameters here. CFAs employ significant voltage and current feedback and enjoy significant voltage and current loop gains.
One of the reasons, perhaps the main reason, that this thread goes nowhere is that the one, ground-basic analysis method based on the superposition principle is unknown to the most vocal members here.
It's matter of seconds to set the independent variable like Vb(ac) to zero and then see what the inv input looks like. Then you see the emitter of a common base transistor and the input characteristics are in clear sight and well known.
Then again, if you are ignorant about this basic method, you get what you see here - hundreds of posts going into circles and people groping in the dark. Hilarious if it wasn't so sad.
Jan
It's matter of seconds to set the independent variable like Vb(ac) to zero and then see what the inv input looks like. Then you see the emitter of a common base transistor and the input characteristics are in clear sight and well known.
Then again, if you are ignorant about this basic method, you get what you see here - hundreds of posts going into circles and people groping in the dark. Hilarious if it wasn't so sad.
Jan
Chris, it is similar as the analysis of a feedback loop with a Middlebrook probe for instance. You set the independent variable (Vin) to zero and that lets you 'look into' the loop without the complication of an input signal.
After you have analysed and characterized the loop, you use the superposition to determine the output of the circuit: you multiply (because it is a gain) the loop with the input signal to get the output signal.
Most simple example: two voltage sources V1 and V2 in series and a load R. Set each voltage source to zero in turn, determine the current through the load due to the other source, and then add the two currents to get the load current from the two voltage sources.
So when you calculate the load current due to V1, you set the other independent variable V2 to zero, and vice versa.
Jan
After you have analysed and characterized the loop, you use the superposition to determine the output of the circuit: you multiply (because it is a gain) the loop with the input signal to get the output signal.
Most simple example: two voltage sources V1 and V2 in series and a load R. Set each voltage source to zero in turn, determine the current through the load due to the other source, and then add the two currents to get the load current from the two voltage sources.
So when you calculate the load current due to V1, you set the other independent variable V2 to zero, and vice versa.
Jan
I think you've mistaken me for someone with a misperception. Yes, superposition, Freshman 101 electronics, yes, your analysis, yes...
I responded to forr as I did because
(1) in references to common base mode, the base is generally AC and often DC grounded, and of course this is not the case with typical CFA operation.
(2) I didn't want to get caught up in his particular viewpoint about common base operation - that's not the "fight" I wanted to have with him.
The inference you drew from my comment was not warranted.
Chris I was not trying to bash you, sorry if it came across like that.
I think what was missed is that I nor anybody else ever said that that transistor works in common base, because it doesn't. The statement that started this discussion would be based on either misconception about the difference between analysis and actual operation, or deliberately twisting words, or sloppy reading/thinking. In either case, discussion will not be fruitful.
Jan
I think what was missed is that I nor anybody else ever said that that transistor works in common base, because it doesn't. The statement that started this discussion would be based on either misconception about the difference between analysis and actual operation, or deliberately twisting words, or sloppy reading/thinking. In either case, discussion will not be fruitful.
Jan
Last edited:
Common base
For long, I thought that the input device of a CFA, and even more of a VFA, worked in an hybrid mode. With the help of some posts of the discussion here (yes, it is constructive despite the unsuccessful opposition it meets). I now have a clearer view on the question. Even better, during the last weeks, I discovered that I am not alone to have pondered on it :
Can we see the negative feedback principle in the operation of the common-base stage? Can we think of it as of a disturbed common-collector stage?
What are actually the input and output of a negative feedback system? Can we consider the input quantity as a disturbance and the output quantity - as a reaction?
For long, I thought that the input device of a CFA, and even more of a VFA, worked in an hybrid mode. With the help of some posts of the discussion here (yes, it is constructive despite the unsuccessful opposition it meets). I now have a clearer view on the question. Even better, during the last weeks, I discovered that I am not alone to have pondered on it :
Can we see the negative feedback principle in the operation of the common-base stage? Can we think of it as of a disturbed common-collector stage?
What are actually the input and output of a negative feedback system? Can we consider the input quantity as a disturbance and the output quantity - as a reaction?
We know that signals are applied to both the non-inverting input base and the inverting input emitter (see the Thevenin equivalent in Post 2548) of the CFA diamond input stage. Generally, a transistor "common X" stage is one in which signals are not directly applied to terminal X. Accordingly, the CFA input stage is neither common base nor common emitter nor a combination of the two. However...
Jan's suggestion of applying the Superposition principle is helpful here if we wish to understand the diamond input stage on a stand-alone basis. Ironically, this would involve summing a common base and a common emitter circuit. Unfortunately, we can hardly understand the role of the input stage in a closed-loop CFA by setting to zero the signal directly exciting the emitter.
Personally, I did not find the references you provided to be particularly enlightening, but perhaps that's just me.
May I return your attention to Post 2563 which elaborates on Post 2548?
No need to call abstract concepts, ignored by so many people, for help to analyze such a simple circuit.
I involed the common base topic because I saw this sentence repeated many times :
the input devices are in common base, seen from the feedback network.
A very puzzling sight.
The feedback network is the output load of the buffer, and behaves like the load of a voltage regulator : it sees a voltage source and reacts according the Ohm law.
And not the emitter of a common base transistor which, if true, should only be considered as belonging to a CB only from a higher altitude.
From a higher point view, the one which shows the ciruit as a whole, the feedback network is nothing else than the load of a buffer which dictates its voltage.
By the way, note that there is also no need to refer to a diamond circuit for the analysis of the CFA concept. Referencing to Op-amp applications
page section 1.2, by W. Kester, W. Jung and J. Bryant page 1.25
The basic concept of current feedback can be traced all the way back to early vacuum tubes feedback circuitry, which uses negative feedback to the input tube cathode. This use of the cathode for feedback would be analogous to the CFB op-amp's low impedance input.
I involed the common base topic because I saw this sentence repeated many times :
the input devices are in common base, seen from the feedback network.
A very puzzling sight.
The feedback network is the output load of the buffer, and behaves like the load of a voltage regulator : it sees a voltage source and reacts according the Ohm law.
And not the emitter of a common base transistor which, if true, should only be considered as belonging to a CB only from a higher altitude.
From a higher point view, the one which shows the ciruit as a whole, the feedback network is nothing else than the load of a buffer which dictates its voltage.
By the way, note that there is also no need to refer to a diamond circuit for the analysis of the CFA concept. Referencing to Op-amp applications
page section 1.2, by W. Kester, W. Jung and J. Bryant page 1.25
The basic concept of current feedback can be traced all the way back to early vacuum tubes feedback circuitry, which uses negative feedback to the input tube cathode. This use of the cathode for feedback would be analogous to the CFB op-amp's low impedance input.
I don't question that you have seen it. I do not support the claim.
... and a means for feeding a signal to the emitter. Please review Posts 2548 and 2563.
We have to be careful with words here. Current feedback (what I think you mean by "the CFA concept") can be applied to an input stage consisting of a single transistor, triode, etc. But "CFA" is the term that the industry employs when a diamond differential stage's emitters are connected to a gain block's inverting input. (In a related, less frequently seen variant, one transistor pair of the two is replaced with diodes.)
Agreed.
Kindly address Post 2563.
Chris,
Also, your model seems to show that the current is flowing between the 2 inputs, that is not the case in real life.
I don't understand what allow you to say that Z is the product of the differential input impedance by the differential voltage gain.
Also, your model seems to show that the current is flowing between the 2 inputs, that is not the case in real life.
forr, in the model,
vo = A · (vd) = A · (rd · if) = (A · rd) · if = (Z) · if
So, Z = A · rd
So, Z = A · rd
It is true that you cannot generalize that current flows between the inputs. I shouldn't have called rd a differential input impedance.
However, there are generally impedances between the inverting input and ground and between the inverting and non-inverting inputs. In the model I presented, these can be combined into a single impedance, rd.
With these understandings, please comment again on the Posts.
I don't dispute the calculation, but the model.
So, with the exception of the sign the two inputs are equivalent, I think you have to have to refine your model.
I discovered them a few days ago and Cyril Mechkow a month ago with this discussionhttps://www.researchgate.net/profile/Cyril_Mechkovhttps://www.researchgate.net/profile/Cyril_Mechkov What is the truth about the exotic current feedback amplifier? Is it something new or just well known old? Is it really a current feedback device?
I found interesting to show that similar ideas have come to the mind of very different persons whose education and formation completely differ and who did not know each other at all. Many times when reading Mechkow, I felt consolidated in my thinking.
I think you'll also find a wide array of folks who think pretty much as I do about CFA's. I know, for instance, that one such well-known and well-respected individual is self-taught and to the best of my knowledge does not have a technical degree. Others have a Ph.D. Yet others write application notes for the semiconductor industry.
Whichever group is misunderstanding how the circuit actually works, I don't think it is telling that it is made up of a significant number of people with different backgrounds. Unless a group has a hidden or ideological agenda, it just shows that the matter is sufficiently complex that reasonable people can draw the wrong conclusions. What would be notable is if that group consisted of a really small number of people or a single person. Then it would probably be populated by morons - or geniuses.
I have never claimed, nor can I recall anyone on this thread claiming, that the industry's CFA embodies a new concept. In fact, I have been clear that feedback to a triode's cathode involves current feedback. The main notable CFA innovation that I can think of is the diamond differential stage, which probably dates back to the 1980's.
Would you be so kind as to reply to Posts 2548 and 2563?
The concept of misunderstanding does not apply. If something is misunderstood it would imply you could change it and advance to a better place. This has not been the goal of this thread for weeks.
Long before.