Followers, loops, and feedback

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Re: Re: Re: I agree with Charlie........

Charles Hansen said:
Steve, for a guy with encyclopedic knowledge of everything from Cooper pairs to altered states of conciousness, you sure seem to have a hard time reading. Let me help you out with this one. For an example of an open-loop follower, please refer to the data sheet for the Burr-Brown BUF634:

The BUF-634 being called an open-loop follower is in a completely different context from what you were referring to an open-loop follower. Here they're referring to a multi-stage circuit without a global feedback loop from the output of the last stage to the input of the first stage.

This is what you called an open-loop buffer:

An externally hosted image should be here but it was not working when we last tested it.


Which is incorrect as virtually 100% of its output is fed back to its input by way of the emitter resistor. That's negative feedback. That's closed-loop. Not open-loop.

But all of this is really silly. In case you haven't realized it, you've "hijacked" this thread. The original poster had a specific question regarding the construction of a specific circuit. None of this is helpful or illuminating to him (or anyone else, for that matter).

And your berating and belittling jcx was helpful or illuminating to the original poster how exactly? Who's the real hijacker here?

For you to drag in one of your old arguments from a thread on a completely different forum (Audio Asylum) is out of line and uncalled for. This argumentative behavior that generates heat and no light is exactly why you have been banned from some forums. It is exactly why new rules have been created on other forums. It is exactly why so many people on this forum wish the moderators would ban you.

I only mentioned a claim you had made in public to illustrate the sheer arrogance and hypocrisy of your berating and belitting someone, requesting that they not post at all unless they "really *know*" what they are talking about.

se
 
Re: Re: Re: Re: I agree with Charlie........

Steve Eddy said:


.

This is what you called an open-loop buffer:

An externally hosted image should be here but it was not working when we last tested it.


Which is incorrect as virtually 100% of its output is fed back to its input by way of the emitter resistor. That's negative feedback. That's closed-loop. Not open-loop.



se

Mr SE,

A emitter-follower like that in your image is in fact refered to as a open-loop-follwer. If you design a power supply regulator using a zener-diode connected to the base of a BJT, one refers to that regulator as an "open-loop-regulator"

:cool: :cool:

Best regards\Morello

BTW, nice to see you here. You might remember me from AA:)
 
Re: Re: Re: Re: Re: I agree with Charlie........

Morello said:


Mr SE,

A emitter-follower like that in your image is in fact refered to as a open-loop-follwer. If you design a power supply regulator using a zener-diode connected to the base of a BJT, one refers to that regulator as an "open-loop-regulator"


Morello,

the fact that somebody referred to this circuit as an "open loop follower" does not change the fact of almost 100% negative feedback in this circuit.

x-pro
 
Re: Re: Re: Re: Re: I agree with Charlie........

Morello said:
A emitter-follower like that in your image is in fact refered to as a open-loop-follwer. If you design a power supply regulator using a zener-diode connected to the base of a BJT, one refers to that regulator as an "open-loop-regulator"

Yes. Though it's not a term I've seen typically used to describe such a circuit. And a Google search for "open loop follower" brings up a whopping 17 hits. :)

The way I see it, negative feedback is an inherent trait of even the simplest follower. They're effectively a 100% voltage feeback circuit, which is why they have an ideal voltage gain of 1.

Also the way I see it, you can't have negative feedback without a closed loop. So to call an emitter follower an "open loop" circuit is ultimately a misnomer. An emitter follower functions as it does precisely because it's a closed-loop circuit employing 100% voltage feedback.

BTW, nice to see you here. You might remember me from AA:)

Thanks. And ditto. Though I remember you as Mr. Morello there. So I see you've decided to ditch the suit and tie, threw on some jeans and a shirt and became a more casual Morello here. :)

se
 
Re: Re: Re: Re: Re: Re: I agree with Charlie........

x-pro said:
the fact that somebody referred to this circuit as an "open loop follower" does not change the fact of almost 100% negative feedback in this circuit.

Yes. And it was ultimately the issue of negative feedback which brought about the "open loop follower."

The discussion concerned two emitter follower topologies, the Darlington and the Sziklai (or "complimentary darlington" or "complimentary feedback pair").

It was said that the Sziklai was inferior to the Darlington because it employed vastly more negative feedback than the Darlington. The reason the Darlington had so much less negative feedback being due to the Darlington comprising two "open loop" followers.

However the way I see it, both the Darlington and the Sziklai have fundamentally the same amount of negative feedback. They both function as followers, they both have virtually the same gain multiplication and therefore employ virtually the same amount of negative feedback.

se
 
diyAudio Senior Member
Joined 2002
Hi,

However the way I see it, both the Darlington and the Sziklai have fundamentally the same amount of negative feedback. They both function as followers, they both have virtually the same gain multiplication and therefore employ virtually the same amount of negative feedback.

That part I'd agree on..anything else is just a matter of how you'd tag the kid, really.

If you allow me, in tube electronics I've heard many bold claims stating this or that all triode circuit does not use any feedback.

Well just as the follower above uses 100% inherent feedback so do CFs in tube circuits...and surprise surprise, all triodes have feedback built in...chuckling...

Quite often both circuits will be marketed as no FB designs, probably because they meant no additional global NFB loop from output to input is used?

Just don't let appearances fool you, everything else is just BS no one ever cared to define.

And guess what, those "clever" marketing guys just exploit that grey area...can you blame them?

Add to that the fact that most people don't have a clue as to how the AC current flows in circuits and voila, you once again have yourself a new niche to drop your product into...

NFB isn't necessarily a bad thing, just don't use it as an excuse for incompetence.

Cheers,;)
 
fdegrove said:
If you allow me, in tube electronics I've heard many bold claims stating this or that all triode circuit does not use any feedback.

Well just as the follower above uses 100% inherent feedback so do CFs in tube circuits...and surprise surprise, all triodes have feedback built in...chuckling...

Yup. In a BJT I believe the same is true due to its equivalent emitter resistance, re. So even in a common-emitter configuration without a literal emitter resistor, you've always got some small amount of negative feedback.

Quite often both circuits will be marketed as no FB designs, probably because they meant no additional global NFB loop from output to input is used?

Yes, that's how the term's typically used. Some manufacturers actually make the distinction.

NFB isn't necessarily a bad thing, just don't use it as an excuse for incompetence.

Yup. Whatever gives you the greatest satisfaction in the end is a "good thing" as far as I'm concerned. Well, except when it's Martha Stewart saying "It's a good thing." :)

se
 
Disabled Account
Joined 2002
Re: Re: Re: Re: Re: Re: I agree with Charlie........

Steve Eddy said:


Also the way I see it, you can't have negative feedback without a closed loop. So to call an emitter follower an "open loop" circuit is ultimately a misnomer. An emitter follower functions as it does precisely because it's a closed-loop circuit employing 100% voltage feedback.


se


Correct....another 'follower' of interest perhaps is the common-base stage.....which possesses 100% current (series) derived, current (shunt) applied, negative feedback....ergo...a current buffer.
 
Eddy,

I think the keyword is the LOOP. You can't creat loop with one stage only. The feedback You mentioned is LOCAL feedback.
The biggest problem with the loop feedback is the delay between the input and the output of the amplifier. This means that the feedback signal allways delayed compare to the input.
In case of local feedback this problem not exist. The feedback caused the input signal itself, so there is no delay.

Sajti
 
sajti said:
I think the keyword is the LOOP. You can't creat loop with one stage only. The feedback You mentioned is LOCAL feedback.

You seem to be contradicting yourself here. You say the feedback I'm talking about is local feedback and that you can't create a loop with one stage only. But you can't have feedback without a loop. So clearly there is a loop in even this single stage. Yes, it's local feedback. But then I didn't say it wasn't.

The biggest problem with the loop feedback is the delay between the input and the output of the amplifier. This means that the feedback signal allways delayed compare to the input.

Delay's only a problem in that it effectively reduces the amount of feedback as frequency increases. Some would argue that it's the feedback itself that's the problem, reducing low order distortion at the expense of increasing higher order distortion.

In case of local feedback this problem not exist. The feedback caused the input signal itself, so there is no delay.

As long as there's a loop of any non-zero size there will be some delay. And any reactance within that loop will also add to the delay. Whether the feedback is global or local doesn't change that.

se
 
"You seem to be contradicting yourself here. You say the feedback I'm talking about is local feedback and that you can't create a loop with one stage only. But you can't have feedback without a loop"

There is feedback without loop. In this stage the feedback caused by the emitter current itself. As the emitter current increase, the emitter voltage increase too. The ouput point and the feedback point is same. I can inmagine loop between two points. But in this case we have only one.

"Delay's only a problem in that it effectively reduces the amount of feedback as frequency increases"

Delay can be problem at any frequency. You can measure this problem even with 1kHz signal.
In well designed amplifier the results is very-very small, of course.

"And any reactance within that loop will also add to the delay. Whether the feedback is global or local doesn't change that."

Would You explain the parts of the loop, with this single transistor emitter follower?

Sajti
 
sajti said:
"You seem to be contradicting yourself here. You say the feedback I'm talking about is local feedback and that you can't create a loop with one stage only. But you can't have feedback without a loop"

There is feedback without loop. In this stage the feedback caused by the emitter current itself. As the emitter current increase, the emitter voltage increase too. The ouput point and the feedback point is same. I can inmagine loop between two points. But in this case we have only one.

I'm sorry but I've never heard of feedback without a loop.

In this case, the loop is the base/emitter loop with the emitter being fed back to the base via the load.

"Delay's only a problem in that it effectively reduces the amount of feedback as frequency increases"

Delay can be problem at any frequency. You can measure this problem even with 1kHz signal.
In well designed amplifier the results is very-very small, of course.

What exactly is it that you're measuring? As I said previously, the delay effectively reduces the amount of feedback due to the phase difference between the input and output signal and when you reduce the feedback, you increase nonlinearity. You'd get the same result by simply reducing the feedback though the reduction in feedback wouldn't be frequency dependent the way it is when there's a delay involved.

"And any reactance within that loop will also add to the delay. Whether the feedback is global or local doesn't change that."

Would You explain the parts of the loop, with this single transistor emitter follower?

Sure. The loop is formed by the base, the emitter, the load, and then back to base.

se
 
"emitter being fed back to the base via the load"

Sorry, but I don't understand it. How the emitter feed back the base, via the load?
The load is not connected to the base.

"What exactly is it that you're measuring? "

Apply 1kHz square wave on the input, and measure the collector of the first stage. You will see overshot on the squarewave. That is the delay.

Sajti
 
I don't see the negative feedback there. Even though you have a resistor from base to emitter, the emitter's going straight to ground so no emitter degeneration.

The emitter terminal goes straight to ground, but the emitter resistance, it seems to me, should be degenerating things, otherwise you'd have infinite gain. If I think of a T model for a bipolar, there's an equivalent resistance re in series with the emitter lead.
 
sajti said:
Eddy,

You can't creat loop with one stage only.

What about a common emitter gain stage with little or no degeneration, the gain of which is set by a resistor or capacitor from collector to base?

In this you can conceive an open loop configuration (i.e. to impedance between collector and base), hence there is an open loop voltage gain. The feedback impedance then takes the voltage signal and takes it back to the input where it is current-added with the input signal.

Speed and distortion wise, this is probabyl similar to emitter degeneration, which is in fact a current derived feedback. But I have more trouble defining what the open loop case is here...
 
Re: Re: Re: Re: Re: Re: Re: I agree with Charlie........

mikek said:



Correct....another 'follower' of interest perhaps is the common-base stage.....which possesses 100% current (series) derived, current (shunt) applied, negative feedback....ergo...a current buffer.

mikek,

sorry, but common-base stage does not amplify the current .......

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