Re: Input & feedback signals
Congratulations, you have discovered phase shift in square waves after they pass through a system with a pole. 🙂
Unfortunately, your figures show no delay, there is only phase shift since changes in the slope of the output signal are inmediately reflected as changes in the slope of the feedback signal.
Jorge said:
Congratulations, you have discovered phase shift in square waves after they pass through a system with a pole. 🙂
Unfortunately, your figures show no delay, there is only phase shift since changes in the slope of the output signal are inmediately reflected as changes in the slope of the feedback signal.
Here is an interesting artcile with a logical thinking :
http://tkhifi.homepage.dk/div/The_ Negative_ Effects_ of_ Feedback.html
But a detail nullifies all its conclusions :
there is no mention of what amount of delay could be in an amplifier.
As far as I remember, about thirty years ago, Nelson Pass, in the documentation of his Treshold cascode amp, gives a delay of 70 nS between a voltage at the direct input and a voltage appearing at the feedback input.
The time resolution of the ear is at best, about 10 mS.
Audio signals have a finite rise time (my data indicate that the fastest audio signal coming from a microphone is about 10 mS) and a finite slew-rate : they are rather slow. Most, if not all, audio amplifiers can easily deal with them (more than twenty five ago, Peter Baxandal and Neslon Pass have given numeric values for the minimum of slew-rate amps should have).
Conclusion :
Many discussions about audio are about existing effects,
but so small that they are absolutely irrelevant in audio :
the delay in the feedback signal is one of them.
~~~~~~~~ Forr
§§§
http://tkhifi.homepage.dk/div/The_ Negative_ Effects_ of_ Feedback.html
But a detail nullifies all its conclusions :
there is no mention of what amount of delay could be in an amplifier.
As far as I remember, about thirty years ago, Nelson Pass, in the documentation of his Treshold cascode amp, gives a delay of 70 nS between a voltage at the direct input and a voltage appearing at the feedback input.
The time resolution of the ear is at best, about 10 mS.
Audio signals have a finite rise time (my data indicate that the fastest audio signal coming from a microphone is about 10 mS) and a finite slew-rate : they are rather slow. Most, if not all, audio amplifiers can easily deal with them (more than twenty five ago, Peter Baxandal and Neslon Pass have given numeric values for the minimum of slew-rate amps should have).
Conclusion :
Many discussions about audio are about existing effects,
but so small that they are absolutely irrelevant in audio :
the delay in the feedback signal is one of them.
~~~~~~~~ Forr
§§§
1] time delay in audio is BS
2] you mean you though of this yourself.. your own bright original idea..
3] I think you're ready to school now... and try to learn a bit this time.. you messing up concepts you don't seem to understand.. you don't seem to be hindered by any understanding.. .. phaseshift is NOT time delay...
Actually group delay IS a -derivative of phase
And delay is mainly caused by frequency compensation circuit and other 'slow' parts of an amplifier. So a goal is to have minimum phase shift non-linearity within audio range and group delay is a good measure of phase shift non-linearity. A nice thing is to have dominant pole above audio frequency, like 25-30kHz Building a stable amp with 30-40dB of feedback factor and dominant pole at 25kHz is a key to make feedback work as advertised.
A DELAY IN AUDIO HAS NOTHING TO DO WITH NIETHER SPEED OF CURRENT NOR A SPEED OF A SINGLE ELECTRON
Understood?
Again a potentially interesting discussion, but.....
forr said:
Well, this is just one of those examples where a perfectly valid theory just has zero practical application in this particular niche of electronics called audio.
Other examples include back EMF and first cycle distortion.
darkfenriz said:
it depends on what devices we are talking about. in wires, you are 100%. In vaccum tubes or carrier-based semiconductor devices, you are incorrect.
forr said:
10ms ? Would make a max of 50hz ???
BTW, the ear can detect phasehifts below 100us, that's how the
stereohearing works, the brain measures phasehifts between left
and right "channel", this gives a very exact information for a 180°
position.
tschrama said:
So when one uses a RC to produce a time delay it's only imagination, since all there is is phase shift?
I think you should take a closer look at time-frequency behaviour of circuits...
Jorge said:
Exactly. Any change in the input signal is instantaneously reflected by the output signal but with phase of their frequency components shifted.
Ok, since it seems I'm wrong, could someone pls explain:
The delay between the sin signal and fbin signal as posted;
The overshoot produced at the very initial rise of the input signal (and with a band limited input), also as posted.
Taking in account this is an ideal amplifier, with 90 deg phase shift over the band.
Thanks for enlightening me.
The delay between the sin signal and fbin signal as posted;
The overshoot produced at the very initial rise of the input signal (and with a band limited input), also as posted.
Taking in account this is an ideal amplifier, with 90 deg phase shift over the band.
Thanks for enlightening me.
An externally hosted image should be here but it was not working when we last tested it.
First figure shows a square wave after and before passing through a complex phase shifter. Believe it or not, both waveforms contain the same frequency components and sound exactly the same.
Second figure shows a square wave and a delayed version of it.
Thanks, Eva
I'm an EE and I've studied this (albeit a long time ago).
But my questions are still unanswered...
I'm an EE and I've studied this (albeit a long time ago).
But my questions are still unanswered...
This figure shows a square wave after passing through two different first order low-pass filters. Red filter has a pole at 2340Hz, blue filter has a pole at 3386Hz. That's what you call delay, it's just a lower pole. Note that there is no delay since both waveforms react to input transients at the same time.
PD: I have not studied EE and judging by what I routinely see I think I'm hardly going to study it in the future. 🙂
I should have realized that. Add a leading zero to my calculations even if they are not germaine. Any way you look at it, the best description of feedback in audio is : "instantaneous".
Which is the point of my exersize. In a long ago thread I fantasized (with a hint of humor, I hope) about an amp where the output devices and the NFB take-off point were in geosynchronous orbit while the IS ans VAS were sitting on the equator. That should be a long enough path to make the delay worth considering.
There are some advantages to entering a field without the impediment of the standard education. If nothing else, you quickly learn not to take the jargon at face value. "Fast" and "slow" are terms that bug me even if I know (I think) what is meant. As far as I'm concerned fast amplifiers are the ones in Ferrarri's an slow ones are in Hyunai's.
Hi, Jorge,
You will have difficulties discussing this "delay" with guys who build kind of amps that handle this very well. JLH (Tschrama/Graham Maynard), AKSA, NP's are some of them.
The delay does not go very deep into electron speed or current path, it is simply because the capacitance around base/gate. There are 2 of, B-C and B-E capacitance.
If an amp having signal passing more of this B junction, you will get more delay, because these capacitance have to be charged/discharged.
In short, amps with less stages will be better if feedback is used, because less B capacitance have tobe charged/discharged.
JLH and NP's are very obvious, they are 2 stages power amp.
AKSA uses bootstrapped VAS. Bootstrapped VAS has very good AC relationship from VAS to output node. Even if the output stage is triple darlington, AKSA kind of amp have 2 stages (AC'ly).
But if you build amp with 3 stages (without any special trick) or 4 stages, or 5 stages, with global feedback, you willl notice the sound difference due to more of what you call "feedback delay".
You will have difficulties discussing this "delay" with guys who build kind of amps that handle this very well. JLH (Tschrama/Graham Maynard), AKSA, NP's are some of them.
The delay does not go very deep into electron speed or current path, it is simply because the capacitance around base/gate. There are 2 of, B-C and B-E capacitance.
If an amp having signal passing more of this B junction, you will get more delay, because these capacitance have to be charged/discharged.
In short, amps with less stages will be better if feedback is used, because less B capacitance have tobe charged/discharged.
JLH and NP's are very obvious, they are 2 stages power amp.
AKSA uses bootstrapped VAS. Bootstrapped VAS has very good AC relationship from VAS to output node. Even if the output stage is triple darlington, AKSA kind of amp have 2 stages (AC'ly).
But if you build amp with 3 stages (without any special trick) or 4 stages, or 5 stages, with global feedback, you willl notice the sound difference due to more of what you call "feedback delay".
Hello, lumanauw
It sems this posting was not a good idea...
Nobody is understanding the spirit of it.
The idea was just to show what happens inside an feedback amp when submmited to a transient (much more like a music signal than a senoid).
How many here had seen it before?
This is in no way related to this or that amp. The sole reason I've based it on the Leach is because it's a quite well known DIY project, and very well documented
Now, the delay is not related to the number of stages - this 'electronic time' is in the order of nanoseconds, and not significant in an amp.
The delay is due to the amp's finite BW (on the Leach's case about 350 kHz - 8MHz total BW divided by the 23 times gain) and determined by the main compensation pole of the amp (for a single pole circuit).
A single pole rise time is defined by 0.35/BW, so in the Leach's case 0.35/350000 = 1 microsecond...
Simple, basic electronics here. Get an amp with a wider bandwidth and this rise time will be less (I posted this statement before).
And as I've posted also, I believe this delay is not significant audio wise.
And I believe the names you've cited know enough basic electronics to agree!
It sems this posting was not a good idea...
Nobody is understanding the spirit of it.
The idea was just to show what happens inside an feedback amp when submmited to a transient (much more like a music signal than a senoid).
How many here had seen it before?
This is in no way related to this or that amp. The sole reason I've based it on the Leach is because it's a quite well known DIY project, and very well documented
Now, the delay is not related to the number of stages - this 'electronic time' is in the order of nanoseconds, and not significant in an amp.
The delay is due to the amp's finite BW (on the Leach's case about 350 kHz - 8MHz total BW divided by the 23 times gain) and determined by the main compensation pole of the amp (for a single pole circuit).
A single pole rise time is defined by 0.35/BW, so in the Leach's case 0.35/350000 = 1 microsecond...
Simple, basic electronics here. Get an amp with a wider bandwidth and this rise time will be less (I posted this statement before).
And as I've posted also, I believe this delay is not significant audio wise.
And I believe the names you've cited know enough basic electronics to agree!
- Status
- Not open for further replies.