The low impedance of the (inverting) input (let's say 20 Ohm) has the strange property that it does not load Rg (let's 200 Ohm). This has been approached in the thread.
Insert the small resistor as I suggested and you will see if the feedback current is arbitrarily entering the inverting input or getting out of it.
This is not to support the legitimacy of devices being necessarily called CFA's or VFA's, however the current mirrors in chip VFA's and CFA's are current sources that can only source current into the Tz node resistor. This is for reasons they cannot sink currents back into their ideally infinite source impedances.
Hence, the inverting terminal can only source signal currents, hence with an outward pointing arrow. This seems ought to be agreeable as the starting point of further analysis.
Without elaborating, and contrary to my earlier opinion, VFA's and CFA's are accurately portrayed under normal circumstances. This is based upon the Ro of the buffered inverting terminal being compared in relation to the Thevenin equivalency of the feedback network.
In the brevity, CFA source resistances can be typically 50 Ohms, while VFA source resistances can be beyond 20K Ohm. It is considered that values in the parallel combination of Rf and Rg normally fall in between these two values, hence can differentiate a CFA from a VFA. This also leads to the conclusion that Thevenin equivalent resistances (the parallel combination of Rg and Rf) that fall below 50 Ohms turns a CFA into a VFA, and values exceeding 20K Ohm turns a VFA into a CFA.
Hi Hans,
I disagree.
1) the map is not the land
2) I can write V(s) = A(s)*s0, as the output stage is a voltage controlled source.
This is wrong.
Of course, a voltage controlled transfer function doesn't need input impedance. A commmon source or emitter transistor is a voltage controlled source and you don't know "a priori" its input impedance, you have to calculate it.
Do you mean Z(s) is known and fixed ?
I disagree.
1) the map is not the land
2) I can write V(s) = A(s)*s0, as the output stage is a voltage controlled source.
This is wrong.
Of course, a voltage controlled transfer function doesn't need input impedance. A commmon source or emitter transistor is a voltage controlled source and you don't know "a priori" its input impedance, you have to calculate it.
Excellent.
Do you mean Z(s) is known and fixed ?
Attachments
Currents, like voltages, can be positive or negative. Arrow directions mean little on a schematic unless the associated meanings are further defined, for instance by naming the arrows and relating the names with equations.
I think I recall you arguing that since the current through Rf does not flow through the CFA's in (-), this can't be current feedback. If this is incorrect, please say so and disregard the rest of this post.
Consider the case in which in (-) is being fed by the Thevenin equivalent of the standard Vout, Rg, Rf circuit: a voltage source equal to Vout * Rg / (Rg + Rf) in series with a resistor Rfg = Rg || Rf. (The open loop output impedance is neglected here for simplicity, but could be accounted for.)
Surely the same current from the Thevenin voltage source flows through both Rfg and in (-). Can we agree that this can be called current feedback? If not, please say so and disregard the following.
in (-) cannot distinguish between these two circuits. The difference between them is the difference between the outputs and the feedback networks. Am I correct in assuming that this difference is enough for you to declare that the CFA does not employ current feedback?
Hi Forr,
Do you think Tryphon calculate the inverting input closed loop impedance before claiming it is low wrt Rg ?
Jan,
I see some contradiction in your definition.
When the "how to do" or "which road to take" for feedback is unimportant, this whole discussion about Current Feedback and Voltage Feedback both clearly describing the "how to" and "which road was taken" wouldn't make sense.
When following this line, what remains is just and only FA isn't it ??
What should be mentioned when feeding back is the product (like current, voltage, temp, light or whatever) that ultimately controls and dominates an optimal production at the output.
Then we also have to discriminate between positive and negative feedback, where usually only the second leads to a stable system.
My attempt to a definition for feedback could be:
Bring the product created at the output, eventually after a possible conversion to a size and dimension, back to the input where it can take the dominant steering function of making the end product as close as possible after comparing it to what was originally envisioned at the start of the production.
With this wider description there is room for:
A VFA, voltage A out, Voltage -X*A in.
But also for a CFA, Voltage A out and Current -Y*A in.
For Car spraying quality feedback, measured paint thickness A out, spraying time -Z*A in, etc, etc.
Hans
Are you stating then that the inverting terminal can source or sink current as independent of the direction of the current flow in the current mirrors?
Of course it does make sense. You can define 'travel' as going from A to B. But it makes a hell of a lot of difference whether you go by car, mule cart, train or bike. You wouldn't say that bikes and mule carts are the same thing because they both facilitate travel. They're two different ways of doing 'travel'.
Jan
There are those here obviously determined to not listen to any reasoning along other than their predetermined path even to the point of silly comments.
Attachments
