Is it possible to create any active circuit with zero feedback?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The one and only
Joined 2001
Paid Member
Yawn. This can only end up being a semantic argument.

The internal mechanism of a gain device creates feedback,
therefore you can't technically have a no-feedback circuit.

But as a practical matter, we are really asking the question
of whether the designer has created a feedback loop for the
purpose of controlling gain or reducing distortion. If not,
these circuits are commonly called "no feedback".
 
PMA,

Greetings from the USA. The answer to your question is as follows. In an absolute sense no, but in a practical sense, yes. Let me clarify. "Active devices" include op-amps, SCR's, transistors, comparators etc. An op-amp operating open loop (no feedback at all, or a comparator), could be considered as an active circuit without feedback if you view the op-amp as a single device. Internally, however, the op-amp consists of gain stages, each of which has inherent local feedback. Capacitance and resistance inherent to the active devices result in a portion of the output of that stage being coupled to the stage input, even to a small degree. This is "local" feedback, as opposed to "global" feedback, where the op-amp output is fed clear back to the op-amp input. The end result may be barely detectable, so that considering the circuit as having "no feedback" may be accurate enough for all practical purposes.

Where this issue becomes controversial is when I, or others suggest that the inherent emitter resistance or emitter-to-base inherent capacitance constitutes "internal feedback". Some will not accept "internal" or "inherent" feedback. The following example should demonstrate that internal feedback is indeed present with semiconductor junctions.

Have you ever done work with "SCR" devices (silicon controlled rectifier)? An SCR is a four-layer "p-n-p-n" device. A drive voltage is applied to the gate-cathode junction, and current flows from anode to cathode. If the gate to cathode drive voltage is then removed, the current will continue to flow from anode to cathode until it is externally set to zero. What keeps the current flowing without gate drive is internal feedback. The p-n-p-n junctions form an npn-pnp bipolar junction transistor pair. The emitters and bases are interconnected in such a way, that one transistor's emitter current drives the other transistor's base, and vice-versa. This is "regenerative" feedback, aka positive feedback. The mechanism responsible for this feedback is entirely internal. In control system theory, a system with positive feedback is unstable, because it is uncontrollable. An SCR gets turned on by applying gate drive, but the same gate drive signal cannot turn it off (I know, gate-turn-off SCR devices were developed in the '70s, but that's beside the point).

A common emitter, or "CE" amp stage without an emitter resistor, or with an emitter resistor bypassed by a large capacitor can be considered an active circuit without feedback to a large degree of accuracy. If you're splitting hairs, and you wish to consider the inherent emitter resistance present in the transistor, then there is a very small amount of negative feedback. The effect of such is quite minimal. To all reading this, let's please not argue over this. Also, there is stray capacitance across all junctions as well as circuit board traces. There is an ac path from the collector, which is the output of the CE stage, back to the base, which is input to the same. This is feedback, but not very much.

To summarize, active circuits intentionally designed without feedback still posess a small degree of feedback due to inherent R and C, but in most cases can be neglected. My stereo amplifier at home, as well as my CD player in the cellar claim "zero feedback" designs. As far as I'm concerned, this claim is legitimate. By the same token, circuits which claim to be "balanced", are legitimate as well, even though no circuit is perfectly balanced. I hope I've answered your question. Best regards to all.
 
Since everything in this world suffers from feedback [not to talk about the usually forgoten thermal, mechanical, capacitive and inductive coupling phenomena present in every circuit], I think it would be more sensible to talk about 'low feedback' [feedback is itentionally avoided] and 'high feedback' [feedback is intentionally added] circuits. 'Zero feedback' is just a purely theoretical concept
 
correcting my error

A common emitter, or "CE" amp stage without an emitter resistor, or with an emitter resistor bypassed by a large capacitor can be considered an active circuit without feedback to a large degree of accuracy.

What I just said in that preceding post is correct at ac, but not at dc. With a large enough emitter resistor in a CE stage, there is dc feedback, but no (actually very little) ac feedback. With zero ohm emitter resistance in a CE stage, ac and dc feedback are practically zero. I should have made that clear. Interesting responses. It looks like universal concensus. Best regards.
 
I have allways considered a emitter degeneration resistor to be a linearising component, not feedback. What happens when a emitter degeneration resistor is added is that the nonlinear voltage variation over base emitter with collector emitter current is given less influence. You do not actually feed a current from the emitter to the base with this resistor and you do not measure the collector voltage with this resistor. Whait a second, it does. The collector current will pass through the emitter down through this degeneration resistor thereby generating a negative feedback signal on the base emitter voltage.

I surrender.
 
AX tech editor
Joined 2002
Paid Member
Pabo said:
I have allways considered a emitter degeneration resistor to be a linearising component, not feedback. What happens when a emitter degeneration resistor is added is that the nonlinear voltage variation over base emitter with collector emitter current is given less influence. You do not actually feed a current from the emitter to the base with this resistor and you do not measure the collector voltage with this resistor. Whait a second, it does. The collector current will pass through the emitter down through this degeneration resistor thereby generating a negative feedback signal on the base emitter voltage.

I surrender.

If I may stir the pot a little more, what Pabo correctly describes is voltage feedback: the output signal modifies the input voltage, even if it is caused by the output current.

Jan Didden
 
AX tech editor
Joined 2002
Paid Member
PMA said:
Of course, emitter resistor means voltage feedback. In case of emitter follower this is a 100% voltage feedback.


Indeed. And if we look into the low-impedance input of a so-called 'current feedback' opamp, what do we see? Basically the same emitter follower, except that now there's two of them, sharing a common Re (the feedback resistor). The feedback signal modifies the input voltage, which is the difference between the two input voltages. That's why I maintain that the term 'current feedback' opamp is a complete misnomer, invented for marketing reasons to be able to sell something 'completely new'. Yeah, my foot!

Sorry, I know we have beaten this to death already, but just couln't resist making the point again.


Jan Didden
 
It's easier for a marketer to sell amplifiers based on the myth that feedback is evil, rather than a different myth that global feedback is evil and local feedback is not. So the solution for the marketer is to completely avoid using the term "feedback" to describe local feedback, while characterizing those who disagree with this view as engaging in arguments of semantics. That's what I've observed anyway :)
 
janneman said:
........ That's why I maintain that the term 'current feedback' opamp is a complete misnomer, invented for marketing reasons to be able to sell something 'completely new'. Yeah, my foot!

Sorry, I know we have beaten this to death already, but just couln't resist making the point again.


Jan Didden

As you wish Jan ;) but the input impedance at the emitter side is in the ideal case zero ohms and in reality quite low. Sorry, but you can’t drive a zero impedance with voltage.

Cheers ;)
 
I really hate these marketeers promoting evils of global negative feedback.

Any amplifier needs proper design with simulation. Sadly a lot of designs don't seem to have this. Douglas Self explained this decades ago and people still don't get it.

Global feedback can work just fine if the delay through the system is not too long. If this isn't the case, then it has to be done more locally.

For audio amps using modern components, the delays are generally short enough, except maybe the reconstruction filter used in the output of Class D amps.

Using negative feedback corrects errors. Amps that reduce this end up with more distortion and higher output impedances (the latter makes them more sensitive to the type of speaker load).

Now if we could replace 10 marketeers with 1 good designer, we'd be onto a winner!
 
AX tech editor
Joined 2002
Paid Member
Pjotr said:


As you wish Jan ;) but the input impedance at the emitter side is in the ideal case zero ohms and in reality quite low. Sorry, but you can’t drive a zero impedance with voltage.

Cheers ;)

Yes, you are right. But who said we want to drive zero ohms with voltage? What does this feedback modify? The input VOLTAGE, so it's voltage feedback. The fact that the feedback node happens to be quite low impedance doesn't make a conceptual difference. The important point is that it modifies the input VOLTAGE, that makes it voltage feedback, independent of how it does this.

And yes, current feedback does exist, as in when you modify the input current by the feedback signal (which may itself look like a voltage...).

Current feedback opamps aren't, except for the Marketing Dept.

Jan Didden
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.