Nelson Pass said:
Ok, let's back up here a moment.
Take a given transistor configured as a single-ended, single supply, grounded-emitter common-emitter amplifier with a given value of resistance for the collector load RC. And in that configuration, you end up with a voltage gain of let's say 20dB.
Now take that same transistor and the same resistance and configure it as single-ended common-collector amplifier (emitter follower).
Question: When configured as the latter, do we have an amplifier circuit operating with 0dB of negative feedback or 20dB of negative feedback?
se
Nelson Pass said:
Well when we say "global feedback" we mean something other than when we say "local feedback" but they're both feedback and fundamentally they behave exactly the same. If it's feedback it's feedback. I fail to see how one form of feedback makes it fundamentally any more or less feedback than a different form of feedback.
Is blue paint any more or less paint than red paint?
If emitter degeneration is a form of feedback, then it's feedback. Period. And since it's degenerative rather than regenerative, that makes it negative feedback.
se
I agree with Nelson Pass and Charles Hansen. We must differentiate between local feedback, followers, and loop or global feedback. While they have some similarities, they also have significant differences.
It is interesting to look at an ideal follower with R(load) being very high. In this case you have TOTAL degeneration, with the interesting result of the follower being close to ideal, because there is virtually no significant change in Gm, therefore no distortion is generated. Yet, a follower with an 8 ohm load would behave much like a common emitter (source, cathode) with 8 ohms of resistive degeneration, so far as distortion is concerned. In this case, there is almost no 'feedback', but also no voltage gain.
What do you think?
It is interesting to look at an ideal follower with R(load) being very high. In this case you have TOTAL degeneration, with the interesting result of the follower being close to ideal, because there is virtually no significant change in Gm, therefore no distortion is generated. Yet, a follower with an 8 ohm load would behave much like a common emitter (source, cathode) with 8 ohms of resistive degeneration, so far as distortion is concerned. In this case, there is almost no 'feedback', but also no voltage gain.
What do you think?
john curl said:
On the whole the differentiation is irrelevant as far as I'm concerned. There's either the existence of feedback or the absence of feedback and whether it's regenerative or degenerative. If feedback exists, then it's just a question of how much.
Yeah? What's the difference between negative feedback and negative feedback?
Why wouldn't you have "TOTAL degeneration" simply with no resistance on the collector?
Is it the result of no significant change in gm or is it because of the effects of negative feedback?
A common-source with what sort of collector load?
If emitter degeneration is indeed negative feedback (and I've so far seen no conclusive evidence that it is not) then the amount of feedback and therefore the amount of linearization will depend on the ratio of the collector load and the 8 ohm emitter resistor.
se
"It is interesting to look at an ideal follower with R(load) being very high. In this case you have TOTAL degeneration, with the interesting result of the follower being close to ideal, because there is virtually no significant change in Gm, therefore no distortion is generated."
"Objection. Assumes facts not in evidence."
"Objection sustained. Strike the statement from the record."
"Yet, a follower with an 8 ohm load would behave much like a common emitter (source, cathode) with 8 ohms of resistive degeneration, so far as distortion is concerned. In this case, there is almost no 'feedback', but also no voltage gain."
"Objection. Assumes facts not in evdience."
"Objection sustained. Strike the statement from the record."
se
"Objection. Assumes facts not in evidence."
"Objection sustained. Strike the statement from the record."
"Yet, a follower with an 8 ohm load would behave much like a common emitter (source, cathode) with 8 ohms of resistive degeneration, so far as distortion is concerned. In this case, there is almost no 'feedback', but also no voltage gain."
"Objection. Assumes facts not in evdience."
"Objection sustained. Strike the statement from the record."
se
Objection. There is no significant change in gm because there IS feedback. From the transistor's point of view as a sort of differential amplifier (it's really the differential voltage between emitter and base that's the variable in the Ebers-Moll equation), the feedback causes the majority of the input to be common mode.
Y'know Steve, I think on reflection you should get a real job.
You really do spend far too much time here, tilting at windmills, arguing the unwinnable and generally plying your winning ways on other denizens. And I'm sure you're a lovely bloke, just misunderstood, that's all.....
Cheers,
Hugh
You really do spend far too much time here, tilting at windmills, arguing the unwinnable and generally plying your winning ways on other denizens. And I'm sure you're a lovely bloke, just misunderstood, that's all.....
Cheers,
Hugh
Charles Hansen said:
Why would we go around in circles?
Do you agree with this pesky moderator's statement or not?
There is no significant change in gm because there IS feedback. From the transistor's point of view as a sort of differential amplifier (it's really the differential voltage between emitter and base that's the variable in the Ebers-Moll equation), the feedback causes the majority of the input to be common mode.
If you agree with it, then we'll not be going around in circles.
se
Here we go 'round the mulberry bush...
http://www.diyaudio.com/forums/showthread.php?postid=331141#post331141
http://www.diyaudio.com/forums/showthread.php?postid=331141#post331141
insert suitable quote here!
as someone who is about as dangerous to himself with a soldering iron as a monkey is with dynamite and matches...
I have one semi useful observation that I have never heard an amplifier with darlington output devices that I liked the sound of, not that one decent designer couldn't make a decent one, as I said just an observation.
anyway I remember in a former life at working for an audio dealer being terrified by the managing director of a company building very famous Scottish turntable throwing a tantrum when I even suggested that his "baby" wasn't as good as he made it out to be and he ought to listen to a Roksan!
his ultimate answer was "everyone has the perfect right to be wrong"......I always remember that, as it was by him in a sly way an admission that even he get's it wrong sometimes....
very entertaining thread by the way.
Kevin
as someone who is about as dangerous to himself with a soldering iron as a monkey is with dynamite and matches...
I have one semi useful observation that I have never heard an amplifier with darlington output devices that I liked the sound of, not that one decent designer couldn't make a decent one, as I said just an observation.
anyway I remember in a former life at working for an audio dealer being terrified by the managing director of a company building very famous Scottish turntable throwing a tantrum when I even suggested that his "baby" wasn't as good as he made it out to be and he ought to listen to a Roksan!
his ultimate answer was "everyone has the perfect right to be wrong"......I always remember that, as it was by him in a sly way an admission that even he get's it wrong sometimes....
very entertaining thread by the way.
Kevin
Andy, you forgot to go back to the start of the circle. Here's how it works:
1) Lukio asks for information on a "no-feedback" amp.
2) Many people tell him that that's a terrible thing because "no-feedback" amps have all kinds of problems.
3) Someone posts a schematic of an amp that has several feedback loops, but no loop than encompasses the entire amplifier.
4) I assert that there are more kinds of feedback in amplifiers than we have terms for, and that it is meaningful to distinguish between these different types of feedback by using different terms and futhermore suggest terms that we should use, and give reasons for making those distinctions.
5) People argue that all feedback is feedback, period.
6) Someone else points out why some feedback is different from other feedback.
7) People argue that all feedback is feedback, period.
8) Someone else points out that in the realm of audio amplifiers that the word "feedback" has come to mean a specific thing, and that to differentiate the local degeneration that occurs in an emitter follower from the word "feedback" that is normally applied to audio amplifiers is, in fact, a useful thing.
8) People argue that all feedback is feedback, period.
9) Go back to step #4. Repeat as often as necessary.
Now, isn't that a fun game?
1) Lukio asks for information on a "no-feedback" amp.
2) Many people tell him that that's a terrible thing because "no-feedback" amps have all kinds of problems.
3) Someone posts a schematic of an amp that has several feedback loops, but no loop than encompasses the entire amplifier.
4) I assert that there are more kinds of feedback in amplifiers than we have terms for, and that it is meaningful to distinguish between these different types of feedback by using different terms and futhermore suggest terms that we should use, and give reasons for making those distinctions.
5) People argue that all feedback is feedback, period.
6) Someone else points out why some feedback is different from other feedback.
7) People argue that all feedback is feedback, period.
8) Someone else points out that in the realm of audio amplifiers that the word "feedback" has come to mean a specific thing, and that to differentiate the local degeneration that occurs in an emitter follower from the word "feedback" that is normally applied to audio amplifiers is, in fact, a useful thing.
8) People argue that all feedback is feedback, period.
9) Go back to step #4. Repeat as often as necessary.
Now, isn't that a fun game?
Charles Hansen said:
Ok. So to avoid the figure-of-eight dilemma, I'll reply here.
Well, while the emitter is at the same node as the output, it's not actually the output. The output is the voltage across the emitter resistor as a consequence of the current flowing through it.
Look at the transistor's equivalent circuit (as a transconductance device) rather than its basic schematic symbol and I think it becomes more clear why the emitter isn't the output.
In the equivalent circuit, you have a current source whose current is controlled by the base/emitter voltage in the form of gmVbe.
As for feedback, I just remembered that Kurt Strain did a rather comprehensive analysis of this over on Prop Heads back in July of 2003, where he shows that a transistor configured as a common-collector/emitter-follower meets all the criteria for a negative feedback circuit.
Unfortunately the graph he posted with the equivalent circuit and the equations is gone, but the post can be found here:
http://www.audioasylum.com/forums/prophead/messages/4742.html
No you wouldn't. It might be problematic for the marketing department. But that's about it.
Yes, they would claim that there is feedback. Because there would be feedback. There is no semantic problem because however much or little feedback there is, that doesn't change the definition of feedback.
Is one drop of paint any more or less paint than a gallon of paint?
No it doesn't. Feedback still means what it means regardless.
Has distortion lost all meaning simply because there are amplifiers with distortion significantly below what many would consider negligible? Of course not. No one's touting "zero distortion" amplifiers are they?
Just use the word feedback in a more honest, quantitiative sense, such as decibels.
Yes. But pretending something doesn't exist when it does doesn't help us define and understand the world. It may help us define and understand "marketing" however.
The key phrase being "one of the input ports."
But the current, which is what flows through the emitter resistor to give us the output voltage, is controlled by the voltage BETWEEN the base and emitter. And since the input signal exists between the base and ground, how can there NOT be a time delay?
Well it certainly wouldn't be useful for someone who is making their living selling what they claim to be zero feedback amplifiers would it?
To call something other than what it is is confusing the issue.
It's not misleading to call feedback feedback.
I find it interesting that the two who have pounded their fists the loudest, and who have more than once resorted to personal attacks on this issue, are both in the business of selling amplifiers.
se
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.