Bonsai, you must be a pretty young fellow to have a memory that stops at the 80-ies
jan
You can have outputs both from the emitter and the collector.
If the resistor in the emitter is equal to the resistor in the collector, it's a phase splitter. This scheme often seen in tube amps has been used with bipolars by Norman Thagard for a bridge configuration of power amps.
In my opinion, it is sometimes preferable not to be too strictly tied to the terminology EC, BC, CC. In an input differential long-tail pair, I think it is correct to consider each bipolar as working in two or three modes simultaneously.
The feedback network is a passive one, typically 2 resistors and maybe a capacitor for CFA. For VFA it can be more elaboate. The "transfer function" is ambiguous. There are 4 that come to mind, Vout/Vin, Vout/Iin, Iout/Vin, Iout/Iin, orIin/Iout, etc. The transfer function is simply an output/input ratio, it can be V/V, I/V, etc.
I explained 8 years ago how the error signal is indeed a current. If the gain is 2 noninverting, with a 10V rail and ground rail, and the Vin is 6 volts, of course the output cannot reach 12V. The 2 input terminals are at the same potential due to the emitter follower buffer. But the current at the inverting input node is not zero since the feedback current does not equal the current in the grounded resistor. If the input is reduced to 4V, then the grounded resistor current is reduced and the output slews to 8V to make sure the feedback current matches the shunt current. The current at the inv input drops to zero, this being the error signal. See my schematic from 2006.
The VFA under same conditions will exhibit a large voltage difference across the input terminals. The differential input Vd, would be 1V minimum if 6V is at the noninv input. The output can only slew to 10V (rail value), the feedback network drops half this 10V value, and the inv input is at 5V. The error signal is Vd. Easy enough. Any app note fron Natl Semi, Tex Instr, Analog Devices, etc. explains it in detail.
Also, the schematic you erferred to does not resemble commercial CFA available today. Again, my 2006 schematic and posts explain in detail. Regarding Dr. Cherry, we covered that in 2004-06 exhaustively. I ask readers to scroll back and I will elaborate if needed. The schematic you gave has no current mirrors, which are important to operation of CFA.
Please refer to my 2006 schematic and raise your differences based on that and what I've posted to date. Thank you.
Claude
No it isn't.
1) For a voltage feedback amplifier, the transfer function of the foward path is Voltage/Voltage.
2) For a transimpedance amplifier, the transfer function of the forward path is Voltage/Current.
3) For a transadmittance amplifier, the transfer function of the forward path is Current/Voltage.
4) For a current amplifier, the transfer function of the forward path is Current/Current.
There is no ambiguity. Period.
See Table 1 here:
http://users.ece.gatech.edu/mleach/ece3050/notes/feedback/fdbkamps.pdf
Current mirrors are NOT essential to the operation of so-called "CFAs". There are plenty of so-called "CFAs" manufactured by the IC industry that do not use current mirrors.
This has already been rebutted. See:
http://www.diyaudio.com/forums/soli...ck-how-do-i-see-difference-34.html#post438891
See also fig 15 on pg 8 here:
http://www.ti.com/lit/an/snaa004/snaa004.pdf
Last edited:
The first patent I am aware of is from 1985: "Reduced settling time for wideband amplifiers". That was befor marketing got involved.
It's really interesting to look at the various circuits, and indeed current mirrors are not essential at all.
jan
Last edited:
The first (SS complementary current-feedback ampilifier) CFB (CFA) amp that I remember having seen is from 1966 or 1967.
The first patent (I've seen) is from 1974.
BTW: if some here take the trouble to analyze a CFB (CFA) circuit in inverting mode, it is certainly easier to see that everything is about current and not voltage. Remember that in a CFB (CFA) the input Z on the Inverting input is close to zero (ideally it should be zero.)
After analyzing the inverting mode, you can analyze the non inverting mode.
Stein
The first patent (I've seen) is from 1974.
BTW: if some here take the trouble to analyze a CFB (CFA) circuit in inverting mode, it is certainly easier to see that everything is about current and not voltage. Remember that in a CFB (CFA) the input Z on the Inverting input is close to zero (ideally it should be zero.)
After analyzing the inverting mode, you can analyze the non inverting mode.
Stein
Last edited:
Yes ideally the inverting input Z is zero. Yet, if you analyse, you will see that the current into it is not infinite, what you might expect. In reality, that input Z DOES NOT determine the current into it....
BTW Do you have a number for that 1974 patent?
jan
Of course it is not infinite, who has stated that?
U.S. Patent 3852678
BTW: the 1967 circuit is designed by Eckes.
Another great resource is Roy Gosser (ADI)
Stein
A voltage across a zero impedance gives infinite current....
My point was that the (low) input impedance of the inverting input does not determine the current into that node.
jan
I see current feed back as current injected into a node between a CCS and the active device. If the device is feed by a resistor, things become more defuse as the current swing will also alter the voltage at that node. and you have a mixture. If feed back is applied through a BJT or af Fet, it'll be purely voltage feedback. To me current feedback is an additive process into the line of the driving current.
Roy is a good buddy for the last 25 or so yr. He has his own style an interestingly different spin on this. His extra stage on the classic CFA has made a great difference.
"Series feedback may be provided for gain stabilisation and input impedance improvement".
Exactly!
jan
These are a real stretch as mainstream products, you should do better. This is a joke yes?
Last edited:
Michael Kiwanuka wrote
"1) For a voltage feedback amplifier, the transfer function of the foward path is Voltage/Voltage.
2) For a transimpedance amplifier, the transfer function of the forward path is Voltage/Current.
3) For a transadmittance amplifier, the transfer function of the forward path is Current/Voltage.
4) For a current amplifier, the transfer function of the forward path is Current/Current."
Michael, you have COMPLETELY missed the point!
50 pages of posts and you are still banging on about the forward path!
"1) For a voltage feedback amplifier, the transfer function of the foward path is Voltage/Voltage.
2) For a transimpedance amplifier, the transfer function of the forward path is Voltage/Current.
3) For a transadmittance amplifier, the transfer function of the forward path is Current/Voltage.
4) For a current amplifier, the transfer function of the forward path is Current/Current."
Michael, you have COMPLETELY missed the point!
50 pages of posts and you are still banging on about the forward path!
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.