Nice app for it. The device itself is very flexible...... I would expect the thd to be flat with freq instead there is the rise beginning at 2Khz. What is that from... some phase shifting?
The AD844AN is at end of life..... I would not be surprised if the other variants of it go soon afterwards. But again they are expensive. And quantities on hand are low.
AD844 is a good current mode design example to discuss, IMO.
THx-RNMarsh
But such an amplifier (AFAIK) would typically have a small-value (ground-referenced) current-sensing resistor in series with the load to convert the output/load current into a voltage. Think of a linear deflection coil driver. "Current feedback" arguably doesn't fit here at all. I don't have a ref. immediately at hand but I can only recall ever seeing this described as "current-sensing feedback".
Last edited:
Not exactly.
The contestation of the CFA idea is born from the refusal to accept the inverting input as being a current input. Otherwise it would be the negation of the concept of feedback well established by Harold Black before 1930.
I am not aware of a valid contradiction to the fact that, whatever the sophistication of the surrounding circuit, the feedback operation in active circuits summarizes in the end to one active device sensing between two points, a potential difference which controls, according to physics laws well mastered by now, the current through it.
So, OK, the claim is NOT that the feedback signal is not returned as a current quantity, but the claim is that the inv input is not a current input? Do I have that right then? And if so, are the two statements really different??
It is late in the game, but understanding exactly what is or is not claimed or disputed seems long overdue.
BTW I cannot understand your last sentence, sorry.
Jan
Last edited:
Not necessarily; you can also make an amplifier with shunt feedback at the input and series feedback at the output and feedback by a transformer, for example. This results in a low input impedance, a high output impedance and an accurate current-to-current transfer.
When I was at the electronics research lab of the Delft University of Technology in the early to mid 1990's, they only used the term current feedback or its Dutch equivalent stroomtegenkoppeling when talking about amplifiers with a low-impedance current input and a high-impedance current output. Later, when I got interested in historical electronics, I learned that historically, the term was also used for amplifiers with series feedback at both the input and the output, like what you described.
Anyway, there are more terms like this: in 2015, after using computers for decades, I finally learned that in its original meaning a computer is a human being.
Last edited:
That much is obvious, which is why I just neglected error current and mentioned the CFA-specific frequency compensation behaviour earlier.
If the inverting input is not a current sensitive node then how to you otherwise explain the fact that the closed loop bandwidth is a function of the feedback network impedance (or in the simplest instance just the resistance of Rf) ?
I find the argument that the error signal is ideally a current a bit strange. Why would you call a case with nonzero error signal ideal?
As far as I'm concerned, in the ideal case, the transimpedance, transadmittance, voltage and current gain of the active part of the amplifier all become infinity. That is, the active part becomes a nullor. Both the error voltage and the error current are then zero and the input impedance of the active part is 0/0 or undefined.
As far as I know, the real process of feedback is substrating voltages and outputing a current in accordance to physics laws which govern the device behavior in charge of the subtraction.
However I am prudent here, because if all circuits I analyzed work that way, I do not exclude that some that I am not aware of may could work differently.
EDIT : I also may say that, in the above claims, I do not take in account the effects of the compensation and parasitic capacitors on the devices and on the circuits, I am just considering the behavior at the low frequencies domain where they can be considered as insignificant.
Last edited:
Well one example comes to mind, the I/V converter, where the feedback is current (the exact same current that is sourced from say a current output DAC).
However, the subtraction process to make it work is to subtract the voltages on the +input (normally ground or a servo) and the -input which both are high impedance.
So here you can say that the feedback current creates an error voltage used in the feedback voltage substraction. You can have both current feedback and voltage subtraction in the error amp.
You cannot make the statement that because feedback works by subtracting voltages, that therefor current feedback cannot exist.
Jan
I don't think I stated this, but anyway, I interprete the word 'ideal' here as ín 'principle' or 'mostly', indicating that the nature of the feedback signal is 'basically' current rather than voltage. Context should make it clear.
As seen from the low impedance of the inv input, any source with appreciable impedance would look like a current source. The current into the inv input would vary with the impedance of the feedback network. In my view this justifies the current feedback designation.
In a 'VFA' inverting input, because of its high impedance, anything feeding it would appear as a voltage source. The current into the VFA inv input does not vary appreciably with the feedback network impedance, justifying the designation voltage feedback.
Edit: I would like to note that the terms VFA and CFA are traditionally (last 30 years) used to make a distinction between two clearly differing input topologies. Seen from that vantage point, the terms are totally logical and immediately clear and useful. What is debated now here is whether strictly technically spoken, the terms are correct or not. I think they are (see above) but even if all debaters here by some stroke of magic would agree that yes, the terms are wrong, what do you think would happen? Nothing, nada, zilch. They would remain in use as they have for the last 30 yeas.
So, interesting as the debate may seem from a technical point, it is hardly a cause to loose sleep over.
Jan
Last edited:
Here are the feedback forms in an expanded view covering both CFA and VFA - might be a bit clearer.
Current Feedback Amplifiers, not only a semantic problem ?
Current Feedback Amplifiers, not only a semantic problem ?
Yes, as long as people start reusing names for other things, the context is needed to help (hopefully) to understand what is ment.
Feedback was allways named after it's starting point.If someone suddenly names it depending on it's destination, trouble starts.
Mona
In the context of this discussion, the meaning of current feedback should be substaction of two currents. Such a configuration may exist but I am not aware of any.
In virtual earth, there is only one current circulating in a resistive path from the input to the output of the circuit. A point on this path is connected to the inverting input which, having a very high impedance and voltage almost equal to the in+ 0V, determines the overall gain or the conversion of the input current to the output voltage in a DAC.
Speaking of DAC reminds an other interesting scheme which uses a current conveyor, it can be found in the Wadia CD player and Devialet DAC, I think.
An unavoidable subsidiary question of this thread is : why did the term CFA appear only 35 years ago with the diamond input op-amps and not before, where the current into the inv input already varies with the impedance of the feedback network ?
That's a tough one, I think Derek Bowers showed me a faded old copy once. In any case the Demrow instrumentation amplifier (1968) is certainly an early example of solid state CFA (in fact forr's pic is half of one).
