My point is that "back EMF" issues can be investigated by proper modelling (including cross over and more than one speaker driver), there is no mysterious back EMF related to speaker.
Regarding "first cycle distortion", sudden sine turn-on at peak amplitude etc., these reflect time response of the system and can be derived in case we know impulse response of the system. BTW, music will never bring sudden changes in derivative, as they are related to infinite bandwidth. Musical signals are BW limited by intruments itself, microphone BW and response time and by media BW and response time.
To conclude - I can see a bit of paper dragon in these issues.
Regarding "first cycle distortion", sudden sine turn-on at peak amplitude etc., these reflect time response of the system and can be derived in case we know impulse response of the system. BTW, music will never bring sudden changes in derivative, as they are related to infinite bandwidth. Musical signals are BW limited by intruments itself, microphone BW and response time and by media BW and response time.
To conclude - I can see a bit of paper dragon in these issues.
Hi Rodolpho,
Oh I fully agree that listening is the final judge, but that actually is how I started out, for conventional investigations were not giving the right explanations due to proven 'appropriate' study methods not looking at output terminal relations from t=0 at input, ie. assumptions that amplifier and component (lead) group delays are too short to have audible effects upon sound reproduction.
Also, isn't it strange how conventional square wave circuit evaluation which has been with us for decades can be used for valid investigations and even NFB set-up, yet a suddenly starting 10kHz sine which tests circuit bandwidth limits in a similar way can be deemed inappropriate.
I have long simulated citcuits from input to output with a view to minimising distortion not just for continuous sines, but for a low 'first cycle' sine distortion from t=0, which includes not only amplitude non-linearities but group delay and realistic load induced effects.
Better 'first cycle' results do actually follow through into reproduction as more stable imagery, especially minor presences which are not disturbed by more powerful instrumental transients as loudspeaker drive is substantially increased. Everyone knows that you must run several cycles to obtain a 'best' thd figure, but hey, that is not what happens in the music world !!! You can have a good thd figure yet still have a compromised transient performance, but you cannot have a good first cycle performance without the thd being even better to start with (unless there is an asymmetry to circuit activity which makes one first cycle polarity better than opposite). And yes bandwidth does come into this, but amplifiers must be stable and so often it is the application of stabilising components that affects the untested (suddenly starting / transient) first cycle accuracy.
Every crossover section, driver, cabinet etc has its own natural reactivity and thus independent back-EMF generating capabilities, but steady sine examination does NOT illustrate all possible combined loading effects upon an amplifier, cables etc. especially with suddenly changing amplifier input, so more strings to the testing bow should not be deemed superfluous to progress.
Hi Hugh,
And if that upper mid prominence is not apparent with a tube amp ........ then where are we ?????
Hi PMA,
But that D2S load is no more than a single fullrange, and not as complex a loudspeaker as many amps will have to work with.
I don't know what kind of output characteristics that output stage has when used as part of a (presumably global NFB loop) amplifier.
Oh I fully agree that listening is the final judge, but that actually is how I started out, for conventional investigations were not giving the right explanations due to proven 'appropriate' study methods not looking at output terminal relations from t=0 at input, ie. assumptions that amplifier and component (lead) group delays are too short to have audible effects upon sound reproduction.
Also, isn't it strange how conventional square wave circuit evaluation which has been with us for decades can be used for valid investigations and even NFB set-up, yet a suddenly starting 10kHz sine which tests circuit bandwidth limits in a similar way can be deemed inappropriate.
I have long simulated citcuits from input to output with a view to minimising distortion not just for continuous sines, but for a low 'first cycle' sine distortion from t=0, which includes not only amplitude non-linearities but group delay and realistic load induced effects.
Better 'first cycle' results do actually follow through into reproduction as more stable imagery, especially minor presences which are not disturbed by more powerful instrumental transients as loudspeaker drive is substantially increased. Everyone knows that you must run several cycles to obtain a 'best' thd figure, but hey, that is not what happens in the music world !!! You can have a good thd figure yet still have a compromised transient performance, but you cannot have a good first cycle performance without the thd being even better to start with (unless there is an asymmetry to circuit activity which makes one first cycle polarity better than opposite). And yes bandwidth does come into this, but amplifiers must be stable and so often it is the application of stabilising components that affects the untested (suddenly starting / transient) first cycle accuracy.
Every crossover section, driver, cabinet etc has its own natural reactivity and thus independent back-EMF generating capabilities, but steady sine examination does NOT illustrate all possible combined loading effects upon an amplifier, cables etc. especially with suddenly changing amplifier input, so more strings to the testing bow should not be deemed superfluous to progress.
Hi Hugh,
And if that upper mid prominence is not apparent with a tube amp ........ then where are we ?????
Hi PMA,
But that D2S load is no more than a single fullrange, and not as complex a loudspeaker as many amps will have to work with.
I don't know what kind of output characteristics that output stage has when used as part of a (presumably global NFB loop) amplifier.
Hi PMA,
Yes music sources are bandwidth limited, but they still start suddenly, esp percussion, thus I fail to see your reasoning here.
And AF amp input might be medium limited but of course the amp itself must be must faster for internal control to generate faithful output.
Yes music sources are bandwidth limited, but they still start suddenly, esp percussion, thus I fail to see your reasoning here.
And AF amp input might be medium limited but of course the amp itself must be must faster for internal control to generate faithful output.
Graham Maynard said:Hi PMA,
Yes music sources are bandwidth limited, but they still start suddenly, esp percussion, thus I fail to see your reasoning here.
No.
20kHz BW signal has rise time of 17.5 microseconds (Tr 10%-90%). You can see nothing faster.
40kHz BW signal has rise time of 8.75 microseconds (Tr 10%-90%). You can see nothing faster. These are fastest "sudden changes" for BW mentioned.
These are elementary basics, as basic as Ohm law is.
Properly designed amp (even with NFB) will be much faster than any captured and reproduced real musical signal.
Pro amp manufacturer like Crown recognize the term "Back EMF".
http://www.crownaudio.com/pdf/amps/136224.pdf
(first) if "back EMF" existed and (second) if we agree to take output node as another input for the feedback system, then it can affect sound.
http://www.crownaudio.com/pdf/amps/136224.pdf
(first) if "back EMF" existed and (second) if we agree to take output node as another input for the feedback system, then it can affect sound.
Graham Maynard said:......
I have long simulated citcuits from input to output with a view to minimising distortion not just for continuous sines, but for a low 'first cycle' sine distortion from t=0, which includes not only amplitude non-linearities but group delay and realistic load induced effects.
......
Every crossover section, driver, cabinet etc has its own natural reactivity and thus independent back-EMF generating capabilities, but steady sine examination does NOT illustrate all possible combined loading effects upon an amplifier, cables etc. especially with suddenly changing amplifier input, so more strings to the testing bow should not be deemed superfluous to progress.
.......
Graham,
You are addressing several issues here.
Nonlinear transfer in amplifiers is one thing, linear circuit nonlinear phase response is quite another, both making for incorrect rendering of the sound field. For both NFB provides a powerfull tool to deal with, and the compound operation may be confusing to some.
The first issue, active amplification nonlinearity, is probably the best addressed one with currently available device and circuit design expertise, being 0.01% THD20 and lower a mostly routinely achieved specification.
Phase nonlinearity, resulting from even ideal linear reactive circuit components or, more precisely, complex polynomial fraction transfer functions, are an issue depending on context.
For the amplifier part, just like the case for nonlinear active transfer it is fairly easy to achieve practical 0 group delay within the relevant frequency range even well beyond the proverbial 20 KHz upper limit, and NFB lends a helping hand here too.
Transducer performance is quite another thing, being the effects of single full range driver limitations, or multi way systems combined performance a far cry from the desired flat, constant group delay, uniform radiation one should like. Interaction with the amplifier is most frequently overlooked and a source of nasty surprises.
Some time back I made a quick simulation taking into account published driver data and box design for the LDC6 2 way system here in the "The many faces of distortion thread". Being a particular case, it nonetheless highlights the challenges at the transducer end of the chain.
I guess a review of that long thread sould be a worthwhile refresher in view of the current discussion, and something recent comers to the forum may want to check.
As noted in the referenced simulation, the back EMF due to actual mechanical effects is negligible, tribute to the very poor transduction efficiency of current loudspeakers.
Most listeners will probably agree with me in that the vast majority cannot be fooled, behind a curtain, about distinguishing a real sound source from its reproduced version, and this goes a long way to recognize we are still quite far from adecuate technological means.
Rodolfo
lumanauw (back EMF): Yes, but it is caused simply by output stage driving RLC resonant circuit.
Hi PMA
---So called "back EMF" issue can be completely described by speaker electrical equivalent circuit.---
Self has shown that, on asymmetrical signals, currents can be a bit different from those described by the pure electrical equivalent circuit, about twice greater, I think.
Before him, Ottala said about the same thing, but he gave a value of six times greater and prooved it on... apparently irrealistic signals.
---So called "back EMF" issue can be completely described by speaker electrical equivalent circuit.---
Self has shown that, on asymmetrical signals, currents can be a bit different from those described by the pure electrical equivalent circuit, about twice greater, I think.
Before him, Ottala said about the same thing, but he gave a value of six times greater and prooved it on... apparently irrealistic signals.
Hi forr,
the weak point of the equivalent circuit is that it consists of linear components, though some of them are non-linear in the speaker (voice coil inductance, e.g.).
Nevertheless, equivalent circuit shows that there is no mystery in back EMF. It is only about precision of the model used.
the weak point of the equivalent circuit is that it consists of linear components, though some of them are non-linear in the speaker (voice coil inductance, e.g.).
Nevertheless, equivalent circuit shows that there is no mystery in back EMF. It is only about precision of the model used.
AKSA said:The electrical inductance of the speakers is probably trivial at 1-2mH, and would not influence proceedings until 30KHz and beyond. But the back emf is created by inertia of the cone reverse energising the voice coil; this can and does affect sonics because the damping factor of the amp is not infinite. Propagation delay through the amp pales into insignificance due to this effect, which is measured in the milliseconds.
This might mandate more care in crossover design of speakers than any special attention to the amp. Of concern would be controlling phase shift at crossover and resonance. I have noticed small anomalies in passive crossover design can create marked phase shift around the crossover point, typically high in the midrange on a two way, and this is unquestionably audible.
Cheers,
Hugh
Yes, and this is a good argument for a reasonably high damping factor and good current-drive capability on the part of the amplifier. It is also an indictment of tube and other amplifiers that have low damping factor (certainly anything less than 20) in respect to accuracy and lack of coloration. But in the end, there is no unknown mystery here.
Cheers,
Bob
Graham Maynard said:Hi Rodolpho,
Oh I fully agree that listening is the final judge, but that actually is how I started out, for conventional investigations were not giving the right explanations due to proven 'appropriate' study methods not looking at output terminal relations from t=0 at input, ie. assumptions that amplifier and component (lead) group delays are too short to have audible effects upon sound reproduction.
Also, isn't it strange how conventional square wave circuit evaluation which has been with us for decades can be used for valid investigations and even NFB set-up, yet a suddenly starting 10kHz sine which tests circuit bandwidth limits in a similar way can be deemed inappropriate.
I have long simulated citcuits from input to output with a view to minimising distortion not just for continuous sines, but for a low 'first cycle' sine distortion from t=0, which includes not only amplitude non-linearities but group delay and realistic load induced effects.
Better 'first cycle' results do actually follow through into reproduction as more stable imagery, especially minor presences which are not disturbed by more powerful instrumental transients as loudspeaker drive is substantially increased. Everyone knows that you must run several cycles to obtain a 'best' thd figure, but hey, that is not what happens in the music world !!! You can have a good thd figure yet still have a compromised transient performance, but you cannot have a good first cycle performance without the thd being even better to start with (unless there is an asymmetry to circuit activity which makes one first cycle polarity better than opposite). And yes bandwidth does come into this, but amplifiers must be stable and so often it is the application of stabilising components that affects the untested (suddenly starting / transient) first cycle accuracy.
Every crossover section, driver, cabinet etc has its own natural reactivity and thus independent back-EMF generating capabilities, but steady sine examination does NOT illustrate all possible combined loading effects upon an amplifier, cables etc. especially with suddenly changing amplifier input, so more strings to the testing bow should not be deemed superfluous to progress.
Hi Hugh,
And if that upper mid prominence is not apparent with a tube amp ........ then where are we ?????
Hi PMA,
But that D2S load is no more than a single fullrange, and not as complex a loudspeaker as many amps will have to work with.
I don't know what kind of output characteristics that output stage has when used as part of a (presumably global NFB loop) amplifier.
As I mentioned earlier, there is nothing wrong with first cycle analysis or squarewave analysis as long as it is somewhat reasonably bandlimited to, say, 100 kHz or so. Even Otala understood the need to bandlimit the squarewave he used in his DIM test to 100 khz or less. To not do so can lead to mis-leading results.
On the other hand, if you are testing for evaluation of feedback stability, then it is appropriate to use a very wideband squarewave.
Bob
lumanauw said:Pro amp manufacturer like Crown recognize the term "Back EMF".
http://www.crownaudio.com/pdf/amps/136224.pdf
(first) if "back EMF" existed and (second) if we agree to take output node as another input for the feedback system, then it can affect sound.
No one here is denying the existence of back EMF. For a useful discussion of some of its effects, see my paper on Interface Intermodulation Distortion (IIM) on my web site at www.cordellaudio.com.
Bob
PMA said:I have a puzzle concerning output stages, NFB and EC as well. I am attaching 2 images. Who would guess what is it?
A guess: An x-y plot of the square wave response vs input wave....?
Edit: No. It is the response versus input for a sine wave. Question is which response. The large error peaks suggest the response from either the EC/FB network or the input signal at the input differential stage.
Jan Didden
PMA said:Here is the 2nd image:
Should be the initial response to an input signal. Output from the EC or FB network to a suddenly starting signal? X = output voltage or current, Y is EC or FB error signal?
Jan Didden
PMA said:I have a puzzle concerning output stages, NFB and EC as well. I am attaching 2 images. Who would guess what is it?
My guess: class B output stage distortion vs output current at higher frequency?
- Home
- Amplifiers
- Solid State
- Bob Cordell Interview: Negative Feedback