I did figure it out. To me, "real speakers behaviors adding complex signals in the feedback loop" indicates a feed back amp is in action in correcting the misbehavior of the speaker, hence is normal, natural by design, and desirable. My question was why such property was necessarily bad and needed to be "taken care of" as viewed by some including you.
What about that waterfall plot? Have you compared two waterfall plots of a same speaker driven with amps w/ and w/o global f/b and examined which was cleaner? I would love to learn the result and that would be far more interesting to me.
Where did i pretended "necessarily bad" ?
Just, i said the resulting signal can have higher slew rate than the same input signal on resistive load. Nothing else, nothing strange. Right ?
What i feel strange is how, in this forum, lot of inputs are looked-at in a suspicious way (knee-jerk reaction) and misinterpreted.
Now, what is the slew-rate margin we have to ensure ? Same answer than signal/distortion or signal/noise: can't say, the higher, the better ?
I have nothing against GNFb. It would be stupid: global or local, it is just a scale difference. We are talking about CFAs, in this topic. Don't this IMPLY GNFb ?
If I have misread you, you have my apologies. Your posting of that waterfall plot and the context around confused me.
Nothing against CFA here, in fact I'm in the middle of projecting member dadod's 200W CFA.
No offense, no need to apologize ;-)
My life: Lot of what i write is, in fact, questions i ask to myself. You know, measurements and their correlations with what we hear, why two topologies so close in results can have such a different character when we listen to ? How to stay balanced between what we know that works without total explanation, and a pure scientific approach (comfortable but not as productive as expected) ...
No, natural sound of loudspeaker in negative op impedance is coming from loudspeaker is free ringing, just like natural instrument ringing.
I am asking the schematic available to built or amp available to buy, of your friend's amp, so I could compare with my natural amp. I like good sounded amps, thats all. Please any link?
Seems like a very important issue you have pointed out. This is where we left off pages back... what OPS is the best to have the minimum signal back at the VAS input - using a real speaker load?
How do we reduce it to the lowest amount? With current-mirror VAS stage? With FET OPS or triple bjt OPS? Combo?
Accepting wild Guesses, multi-SIM, And, measured with different OPS topologies using various VAS stage and CFA topology over-all.
I posted a similar question to Bob C over on his thread.
Thx-RNMarsh
Last edited:
I'm not sure to understand your question.
We know the signal in the VAS input is the result of the difference between the input signal and [the signal at the output divided by the feedback ratio bridge].
The OPS distortion will be added in phase opposition.
So the best OPS to minimize the signal at the VAS input is the OPS with the less distortion. Hence, the interest, may-be, to use local feedback/error correction in the OPS stage. To be tried.
Now, for the 'parasitic' signals generated by the speaker's resonances and witch enter back in the feedback loop, there is nothing we can do: they are part of the GNFb principle, and what give low impedance at the output: The amp try to fight against them in order to damp the speakers. The only way to reduce them would be to add a serial resistance, reducing damping and efficiency, it is not what what we want, right ?
Or to add compensating networks in the speakers, in order they present a flat impedance curve (and damp themselves better). That i tried to explain in an other thread, before to be exhausted by the usual opponent's noise.
Did-we want to listen to the 'natural' sound of loudspeakers, or force them to reproduce closer the music ?
Last edited:
Dick,
I would think that varying loads and/or back-EMF at the amp output dynamically change the load at the output, which dynamically changes the load the driver sees, which dynamically changes the load the Vas sees, which dynamically changes the Vas input Z, etc....
Jan
The waveform at the VAS simply shows that the NFB is doing its job! I fear you don't really understand feedback theory.
I'm in two minds with this one:
1) GNFB takes care of it
or
2) You want to minimise these effects which would appear as varying loads at the VAS. This either means either increasing the current gain of the OPS to minimise the amplitude of these variances or having a VAS that "rejects" these signals.
Is this not an argument for a triple OPS? and/or possibly a Baxandall super pair based VAS which was suggested by another member a while back in another thread. IIRC, Baxandall based VASs are supposed to be good at rejecting current distortion? Or am I talking rubbish?
I used real components. It seems to me that almost purely reactive components are more difficult loads for an amplifier, particularly capacitors, than anything else.
Yes, a series resistance is not the most elegant solution but I remember that
John Linsley Hood recommended exactly that (0.22 Ohms) and there are some
Naim amplifiers with such a resistor too.
Absolutely agreed, definitely yes !
Originally Posted by forr
The main effect of negative output impedance is to lower the Qt in box of the speaker.
Robert Munnig Schmidt in Linear Audio, volume 6, pp 33-46, provides some data : it did not found that it does much for distortion reduction and better control of the cone movement.
If I understand it is about the application of a negative output resistance by known methods in the subwoofer, for example:
Variable Amplifier Impedance
Effects Of Source Impedance on Loudspeakers
These methods are applicable only in a frequency range not higher than 300 Hz, and have serious drawbacks.
Furthermore the author Effects Of Source Impedance on Loudspeakers
conducted research at output impedance minus 3 ohms!
Getting a negative impedance with a positive operating current output shifted relative to the input voltage is not the best way.
According to my information to make the output impedance more negative than minus 0.3 Ohm undesirable, there are problems at frequencies of 100 ... 200 Hz - to "depression" impedance of the subwoofer.
Therefore, I think that put the stamp on the negative output impedance of the amplifier only on the basis of failed experiments is extremely cautiously, especially because the subwoofer is not responsible for the reproduction of mid-range, where the largest concentration range signals.
Some even suggest the use of reverberation for the re-entry of music.
Volume 5
System design
How do you increase your enjoyment of reproduced audio if your whole system has already been optimized? Use High Frequency Reverberation for finer sound reproduction says Richard Burwen.
Originally Posted by ontoaba
I am asking the schematic available to built or amp available to buy, of your friend's amp, so I could compare with my natural amp. I like good sounded amps, thats all. Please any link?
Description amplifier with negative impedance and without general NFB at the end of the magazine log at:
Ðàäèîõîááè ¹4 (àâãóñò 2013)
best regards
Petr
The main effect of negative output impedance is to lower the Qt in box of the speaker.
Robert Munnig Schmidt in Linear Audio, volume 6, pp 33-46, provides some data : it did not found that it does much for distortion reduction and better control of the cone movement.
If I understand it is about the application of a negative output resistance by known methods in the subwoofer, for example:
Variable Amplifier Impedance
Effects Of Source Impedance on Loudspeakers
These methods are applicable only in a frequency range not higher than 300 Hz, and have serious drawbacks.
Furthermore the author Effects Of Source Impedance on Loudspeakers
conducted research at output impedance minus 3 ohms!
Getting a negative impedance with a positive operating current output shifted relative to the input voltage is not the best way.
According to my information to make the output impedance more negative than minus 0.3 Ohm undesirable, there are problems at frequencies of 100 ... 200 Hz - to "depression" impedance of the subwoofer.
Therefore, I think that put the stamp on the negative output impedance of the amplifier only on the basis of failed experiments is extremely cautiously, especially because the subwoofer is not responsible for the reproduction of mid-range, where the largest concentration range signals.
Some even suggest the use of reverberation for the re-entry of music.
Volume 5
System design
How do you increase your enjoyment of reproduced audio if your whole system has already been optimized? Use High Frequency Reverberation for finer sound reproduction says Richard Burwen.
Originally Posted by ontoaba
I am asking the schematic available to built or amp available to buy, of your friend's amp, so I could compare with my natural amp. I like good sounded amps, thats all. Please any link?
Description amplifier with negative impedance and without general NFB at the end of the magazine log at:
Ðàäèîõîááè ¹4 (àâãóñò 2013)
best regards
Petr
Then you missed the point of what I was saying about distortion when driving a real speaker. It behaves differently from real components which are, by and large linear.
If I miss the point, you coud be specific in what.
I already know Ottala's and Self's papers on the subject.
It seems there was also a paper by Keith Howard which provides facts about currents in loudspeakers with passive crossovers; they appear to be even less demanding to the power amp than Douglas Self considers. I did not find this paper, maybe someone knows it.
Perhaps you're confusing 'reactive' with 'non-linear' - I wasn't referring to reactive loads. A speaker is non-ohmic in the sense that when driven with a zero-distortion sine, it draws a non-zero distortion current. Imposed across any impedance in the output circuit this distorted current gives rise to voltage distortion. Those aiming for multiple zeros after the point in THD will do well to bear this in mind.
Perhaps the paper you're referring to is this one? Loudspeakers: Effects of amplifiers and cables - Part 1 | EE Times
Perhaps the paper you're referring to is this one? Loudspeakers: Effects of amplifiers and cables - Part 1 | EE Times
Of course, but only due to the input stage distortion ?
We can consider with a MOSFET output device that what happens in the source/drain currents has no influence in the gate(near infinite current gain).
Well, not totally true at high frequencies, due to source and drain/gate parasitic capacitances...
With a bipolar, your question is even more interesting. We can see the effect on distortion of an emitter follower inserted between VAS and OPS.
I'm sure you have an idea behind.
There are some complaints that we are not dealing much about CFA topology. Most of what has already been said by CFA afficionados (do not take this term as pejorative) belongs, or is related, to the general theory (and practice) of negative feedback amplification. So I estimate we do not leave the topic of the thread
I would like to quote excerpts of what Douglas Self wrote in Electronics World about
1. Delayed feedback
2. VAS input signals
which are topics by which many people intervening here are concerned.
1. Delayed feedback
Douglas Self
Letters "Self reacting"
Electronics World and Wireless Wolrd
September 1995, pp 774-775
Self comments on the claim :
"limited open-loop bandwidth prevents the feedback signal from immediately following the system input"
Douglas SELF
"Power amplifier input currents and their troubles"
Electronics World, May 2003, p53-56
to this remarkable article.
I would like to quote excerpts of what Douglas Self wrote in Electronics World about
1. Delayed feedback
2. VAS input signals
which are topics by which many people intervening here are concerned.
1. Delayed feedback
Douglas Self
Letters "Self reacting"
Electronics World and Wireless Wolrd
September 1995, pp 774-775
Self comments on the claim :
"limited open-loop bandwidth prevents the feedback signal from immediately following the system input"
About sinewaves claimed inaccurate for test-signals :
2. VAS input signals
Douglas SELF
"Power amplifier input currents and their troubles"
Electronics World, May 2003, p53-56
Curiously, the sixth edition of "Audio Power Amplifier Design" seems not to refer
to this remarkable article.