The most general definition I know is 'using a sample of the plant output fed back to the plant input to modify the plant behavior'..
Not sure it clarifies anything
Jan
Thanks, Jan, I recall that now and I think it's consistent with "my" definition. We'll see what others say.
Jan,The most general definition I know is 'using a sample of the plant output fed back to the plant input to modify the plant behavior'..
Not sure it clarifies anything
Jan
I also agree in case it is possible.
But suppose there is no direct connection between output and input like voltage out and current in, just as an example.
Then you have a conversion to make.
The voltage could well be be transformed into light, the light into a RF signal to a relay station, Rf back converted into heat and heat into current and there you are.
The last in the chain determines whether the Plant behaviour has to be modified, after a feedback chain in 5 steps.
Take one away and the loop is open.
Feel free to see this all together in one go as "Feedback", but it is completely hiding the actual road that was taken.
Hans
Yes, true, but the definition at the highest level says what is happening, not how to do it, not 'which road to take'. Early systems even had a man in the loop, and a fly-by-wire jetfighter has a man in the loop and actually is unstable without continuous feedback control of the flight characteristics.
The control systems people have been at this much longer than we audio types.
Jan
The control systems people have been at this much longer than we audio types.
Jan
Yes, In a VFA the error current is out of the LTP (into the 2nd stage) and not in the feedback network - that's the point I was trying to make - apologies I should have been clearer.
Here is the infographic about 'maths difficulty' I mentioned a few weeks ago.
Control theory is a little way up from the 'genius gap', but way deeper than tensor calculus.
This is not to say feedback in audio amplifiers is difficult, but simply to make the point that what we do in audio wrt control theory is extremely elementary.
Mathematics trench. - Imgur
Control theory is a little way up from the 'genius gap', but way deeper than tensor calculus.
This is not to say feedback in audio amplifiers is difficult, but simply to make the point that what we do in audio wrt control theory is extremely elementary.
Mathematics trench. - Imgur
I showed that in my damned indisputed simulations
Forr, We already told you many posts back that at LF the error current into a CFA -in is very low and you can calculate that. The Vbe deltas are concomitantly very small and a function of the DC loop gain and Ro. This compares somewhat to the Vbe deltas you would see in a a VFA.
What high frequencies ? Remember, this topic on CFA is oriented towards analogic audio.
Question of the day : is it not strange that the currents coming from the input stage to the next stage are very similar in both configurations, provided that the working conditions are the same of course ?
Sorry, I won't please you. I could ask you the same but I think such a request is perfectly incorrect. You do not personify the electronic science and I am free to express ideas which are not specifically mine (and cause, in other places, the same similar hostile reactions rather than indifference; this a very weak point of the CFA defenders).
It is already done. Know that, until now, the arguments of CPaul and you are the best advocates to show that CFAs are VFA in disguise.
Here is the infographic about 'maths difficulty' I mentioned a few weeks ago.
Control theory is a little way up from the 'genius gap', but way deeper than tensor calculus.
This is not to say feedback in audio amplifiers is difficult, but simply to make the point that what we do in audio wrt control theory is extremely elementary.
Mathematics trench. - Imgur
Green's functions (Linear transformations) were my downfall, so I never even made it to "serious math".
Goodness help me if I tried anything that in-depth now.I also couldn't wrap my head around spatial Fourier transforms either (k-space), even though I could do the math(s); but that may have had more to do with condensed matter physics part of it.
Anyways, this whole thing has gone utterly circular. :/
I rest my case.
Indeed. My arguments are such a good support of the claim that CFAs are VFAs in disguise that forr won't respond to them!
???
Are actual solid state devices fast enough to provide flat open loop at 20KHz in a high gain power amp ?
Of course, you are free to think it does not matters.
Even if the feedback amp is ruler flat to 100kHz, still the error voltage or error current is a major indicator of what goes on inside it. Remember, the error voltage is really Vout/Aol (for a VFA at least) so that error voltage documents the complete amplifier open loop behaviour.
Most people are not consciously aware that the opamp in a feedback loop still works in open loop - it's just that the feedback manipulates the effective input level, but the opamp itself is not changed of course and ALWAYS works open loop...
Jan
Most people are not consciously aware that the opamp in a feedback loop still works in open loop - it's just that the feedback manipulates the effective input level, but the opamp itself is not changed of course and ALWAYS works open loop...
Jan
Last edited:
Well DPH, you’re a whole lot further down the road than I am. I’ve pulled out my old math books and just bought s few new ones. I’m going to try to re-educate myself in 2019. But I’ll leave the k-space stuff for now
My wife's PhD thesis was on Krylov subspace techniques and she ended up at an internship under Kevin Kundert at Cadence but she decided not to enter academia. It's all Greek to me.
Try anything > DC. Nothing useful is gained by solving feedback equations at DC only, no AC response, no stability criterion, virtually nothing of any use.
Exactly
It also shows that one have to use a Ccomp inside a CFA as you have to in a VFA.
Some here seems to think that it's unnecessary with a Ccomp in a CFA and one person thinks that it's not necessary to use a Ccomp in a H-bridge amp because it is based on a CFA.