What’s not wrong with Chever’s thesis?
His advisor and thesis committee in an accredited institution issuing graduate EE degrees has no excuse for not knowing of negative feedback harmonic series expansions in the early literature and asking Cheever to at least reference them
They might have been expected to catch and insist on some corrections or rewording of his misstatements regarding negative feedback theory
They could have commented on his (nonexistent, undocumented, uncontrolled?) “experimental protocol”
They might have recognized that IMD renders any “harmonically consonant” distortion audibility theory moot for anything but flute solos – and people jumping on his bandwagon complain about engineers only testing with single sine waves!
http://www.diyaudio.com/forums/showthread.php?postid=700346#post700346
His advisor and thesis committee in an accredited institution issuing graduate EE degrees has no excuse for not knowing of negative feedback harmonic series expansions in the early literature and asking Cheever to at least reference them
They might have been expected to catch and insist on some corrections or rewording of his misstatements regarding negative feedback theory
They could have commented on his (nonexistent, undocumented, uncontrolled?) “experimental protocol”
They might have recognized that IMD renders any “harmonically consonant” distortion audibility theory moot for anything but flute solos – and people jumping on his bandwagon complain about engineers only testing with single sine waves!
http://www.diyaudio.com/forums/showthread.php?postid=700346#post700346
Sorry jcx, am I missing something? I honestly found nothing substantive, nothing even quoting Cheever, following that link. Did you mean the pdf files padamiecki uploaded. I'll read them at lunch tomorrow. 🙂
Agreed the protocols of Chever's listening tests were seriously lacking, though no more so than some IEEE demo accounts I've read, and the SET 'fanboism' counterproductive and embarrassing but I was hoping for something a little more concrete and technical than the usual boring and empty abuse. You provide references to what appear really interesting papers which I hesitate at dropping a hundred bucks to confirm they are directly applicable.
I'm not trying to be argumentative here, and am really interested where core aspects of Cheever's paper - distortion weighting in particular - fall down.
Agreed the protocols of Chever's listening tests were seriously lacking, though no more so than some IEEE demo accounts I've read, and the SET 'fanboism' counterproductive and embarrassing but I was hoping for something a little more concrete and technical than the usual boring and empty abuse. You provide references to what appear really interesting papers which I hesitate at dropping a hundred bucks to confirm they are directly applicable.
I'm not trying to be argumentative here, and am really interested where core aspects of Cheever's paper - distortion weighting in particular - fall down.
Feeding a systems output back to its input creates a regenerative system. The circuit amplifies a different signal. If the system is non-linear then the regeneration will alter the spectral content of its transfer function.
Isn't this obvious? Perhaps I didn't quite understand Rod's statement...it is somewhat rich in double negatives.
Think Inter-Modulation Distortion
http://www.diyaudio.com/forums/showthread.php?s=&postid=559330&highlight=#post559330
Czerwinski “Multitone Testing of Sound System Components – Some Results and Conclusions, Part 1; History and Theory” JAES V49 #11 Nov 2001 p 1011-1047
Part I is probably worth the $20 nonmember price
http://www.aes.org/journal/search.cfm
you may want to search: moore, brian for distortion perception articles
http://www.aes.org/e-lib/browse.cfm?elib=12197
us$75 for membership, then article downloads are $5/ea
+us$125 for 1yr online access to back issues – plenty of time to download them all
public libraries here can usually arrange interlibrary loans if your local university library doesn’t have what you want
Rod’s Article is really unfortunate, for all I agree that negative feedback is the primary tool in analog audio amplifier design he just gets too much wrong
Yet another silly error is his dismissing “square law” amplification as impossible with his zero bias example, by middle school he should have learned to expand (x + a)**2
to get: a**2 + 2*a*x + x**2
http://www.diyaudio.com/forums/showthread.php?s=&postid=559330&highlight=#post559330
Czerwinski “Multitone Testing of Sound System Components – Some Results and Conclusions, Part 1; History and Theory” JAES V49 #11 Nov 2001 p 1011-1047
Part I is probably worth the $20 nonmember price
http://www.aes.org/journal/search.cfm
you may want to search: moore, brian for distortion perception articles
http://www.aes.org/e-lib/browse.cfm?elib=12197
us$75 for membership, then article downloads are $5/ea
+us$125 for 1yr online access to back issues – plenty of time to download them all
public libraries here can usually arrange interlibrary loans if your local university library doesn’t have what you want
Rod’s Article is really unfortunate, for all I agree that negative feedback is the primary tool in analog audio amplifier design he just gets too much wrong
Yet another silly error is his dismissing “square law” amplification as impossible with his zero bias example, by middle school he should have learned to expand (x + a)**2
to get: a**2 + 2*a*x + x**2
jcx said:
Bought that one online, thanks for introducing me to more too-easy and too-convenient ways of spending money on this hobby!
Rod Elliott
---Feedback does not - repeat does not - cause the signal to travel from the output, back into the inverting input, and continue through the amplifier several (or multiple) times. ---
Sometimes, it does at high frequencies on fast square waves : there are unworrying and inaudible damped oscillations.
---Feedback does not - repeat does not - cause the signal to travel from the output, back into the inverting input, and continue through the amplifier several (or multiple) times. ---
Sometimes, it does at high frequencies on fast square waves : there are unworrying and inaudible damped oscillations.
It is one of the basics of circuit theory that feedback closed over non-linear circuit creates new harmonic components, that were not present in the signal before applying feedback.
I am very sorry that my friend Rod Elliott was not aware of this fact, in case he really was not.
I am very sorry that my friend Rod Elliott was not aware of this fact, in case he really was not.
Quoted
---A perfectly linear transistor doesn't exist.---
A transistor can be linearised by a fast local feedback loop and used a a building block. The most well known is the Sziklai pair which can be used in an amp differential LTP for example. Negative feedback has a lot less alleged defaults then.
---A perfectly linear transistor doesn't exist.---
A transistor can be linearised by a fast local feedback loop and used a a building block. The most well known is the Sziklai pair which can be used in an amp differential LTP for example. Negative feedback has a lot less alleged defaults then.
Now we have a parameter space to work with! I'll jump in here to follow up on my last simple question:forr said:
Let's assume for a second the speed of electrons within a copper wire. OK, everybody calibrated? What audible frequencies can survive a feedback loop or "add" to a "non-linear" circuit assuming those "lag" delays? These are the only new features that could possibly be added by negative feedback. Anyone care to do some of that advanced math you guys are clamoring about?
If your circuit is CREATING new harmonics via non-linearities in the circuit, negative feedback will work to damp these creations relative to the input signal. Not to be obtuse, traderbam, but I thought this was obvious as well. I really have no further desire to jump into a discussion about established electronic circuitry theory at this point, but it seems as if there is a fundamental misunderstanding somewhere about how electronic signals work.
Just to get an additional calibration point, how many people here have actually taken advanced electronics courses, or have advanced electronics degrees (heck, I'll even take physics or any other data acquisition field) before questioning Rod or Cheever? [disclaimer] I'm not necessarily defending Cheever's 8 year MS, but who else has an advanced degree? [/disclaimer] Let’s just make sure that EVERYONE’s qualifications are transparent in this discussion so that Joe Greenhorn can be sure to see where the opinions are coming from.
Are we talking about negative feedback here? I agree positive feedback can do this, but we are all aware that this discussion is centered around negative feedback, right??PMA said:
Speed of electrons in a copper wire has nothing to do with.
You will be surprised, but speed of propagation in copper is pretty slow.
for 50Hz it is 2.9 m/s
for 1kHz it is 13.1 m/s
for 10kHz it is 41.5 m/s
The information propagates as electromagnetic wave along the conductor, with velocity of 300 000 000 m/s in vacuum, a bit less in real world. Do not speak about speed of electrons in a wire.
You will be surprised, but speed of propagation in copper is pretty slow.
for 50Hz it is 2.9 m/s
for 1kHz it is 13.1 m/s
for 10kHz it is 41.5 m/s
The information propagates as electromagnetic wave along the conductor, with velocity of 300 000 000 m/s in vacuum, a bit less in real world. Do not speak about speed of electrons in a wire.
Quoted
---it takes a finite time for the feedback loop to operate.---
The delay (time) must be distinguished from the phase lag (angle).
In an amplifer, the delay between the arrival of a signal at the input and the arrival of the fedback voltage might be less than 100 nS.
An audio signal has a finite slew-rate (V/µS) and the integration time of the ear is longer than 10 µS (the minimum time during which the ear can discriminate two informations). So the delay in a feedback loop should not be a problem.
---it takes a finite time for the feedback loop to operate.---
The delay (time) must be distinguished from the phase lag (angle).
In an amplifer, the delay between the arrival of a signal at the input and the arrival of the fedback voltage might be less than 100 nS.
An audio signal has a finite slew-rate (V/µS) and the integration time of the ear is longer than 10 µS (the minimum time during which the ear can discriminate two informations). So the delay in a feedback loop should not be a problem.
dfdye said:
Negative feedback, of course. The case needs a bit of background. There is no simple answer for complex question 😉
Regarding new components created by feedback, one should take into account they would be usually quite small in amplitude. It depends on non-linearity shape and open-loop gain response.
Just to simplify /though I do not like to do it/, imagine non-linearity that creates 2nd harmonic without feedback. After applying feedback you reduce 2nd harmonic, but create 4th harmonic of smaller amplitude as well. You also create IM products in case of more complex input signal.
Just to simplify /though I do not like to do it/, imagine non-linearity that creates 2nd harmonic without feedback. After applying feedback you reduce 2nd harmonic, but create 4th harmonic of smaller amplitude as well. You also create IM products in case of more complex input signal.
Touche. SIGNAL propogation speeds were techincally the intent of the question. Of course physical electron propogation is painfully slow and has nothing to do with signal transmission speeds. 😀PMA said:
Um. . . Now you have lost me. . .PMA said:
Are we not assuming higher order dampening effects as well?
It is really not quite simple at first view. Some maths, simulations and measurements are very helpful.
Try to realize that in a feedback loop system you always subtract input signal X and divided output signal AX/K at the input of the amplifier, that is non-linear (A is not constant, but depends on X). The divided output signal came through non-linear amplifier, so it has distortion component included. This one subtracts at the input /to some extent/, and creates new distortion component when passes through non-linear amplifier.
Try to realize that in a feedback loop system you always subtract input signal X and divided output signal AX/K at the input of the amplifier, that is non-linear (A is not constant, but depends on X). The divided output signal came through non-linear amplifier, so it has distortion component included. This one subtracts at the input /to some extent/, and creates new distortion component when passes through non-linear amplifier.
can this be worked in to this topic?
http://www.physorg.com/news66582110.html
check the eyes of the 'man' in the picture
feedback is only a portion of the total sound of an amp
http://www.physorg.com/news66582110.html
check the eyes of the 'man' in the picture
feedback is only a portion of the total sound of an amp
- Status
- Not open for further replies.