Hi Jan,
The induced transient voltage was not what I was worried about. Rather, it was the time-response of the current. The current from the power supply is, or becomes, the signal that we hear. Voltage transients on the power rails are mere artifacts.
The question was, or should have been, can the inductance of the power rails significantly affect the rise time of the current that is demanded by the power amplification devices?
If it can, then local decoupling could provide a significant improvement in the accuracy of the amplification, especially for fast transient components of the signal.
With a feedback system, this would have implications in the areas of overshoot and stability.
It's even very easy to imagine a slightly-delayed leading edge rise for a quasi square wave causing a feedback system to try harder to correct for the difference, just as the current finally catches up, causing overshoot.
Tom
Hi Hugh,
My only recent cogent arguments were about the definition of "transient" and about whether or not feedback was the main cause of instability and overshoot.
My post about the violinist without feedback was not a cogent argument. It was more or less just a provocative rant.
I have to admit that one of my character flaws is that I enjoy doing that, on occasion. But I do apologize for its overall tone.
Just so you know that feedback is not like a religion for me, I will also postulate that the best system would be one that could not benefit from any feedback.
To answer your question, no I have not listened to any amps like the one you described. As to why no feedback would be preferred, it's difficult for me to say much because I am not familiar with those amps but I think it would have to be because either a) the system does not need feedback and performs better without it, or, b) the feedback was implemented incorrectly.
Tom
Yes, intuitively it was also expectable for me, but at the same time I saw many statements from respectable designers, that they recorded musical signal and did analyze it, and have not seen more than near 1V/uS.
Maybe during the recording something has been truncated?
One also should remember, that spectrum limitation does not mean limitation in number of sine constituents being in game, there is an infinite number of them, and increasing the slew-rate threshold, we only decrease the probability when they sum up in such a manner that the threshold is exceeded.
Jan,
After all, every halfway-good amplifier should be able to perform extremely well on the relatively-slow-changing parts of any signal. So I tend to believe that the fast-transient response, and the timing and phase accuracies, are the most important things to worry about, in trying to achieve the ultimate in high reproduction fidelity.
After all, the Fourier harmonic components of any signal each have a required phase angle as well as a required amplitude. If we can't get both of those correct, far enough out in the infinite sequence of Fourier components, then we will not have accurate-enough reproduction.
Note, too, that the effect of the power rails' inductances on response time would be frequency-dependent (and slew-rate-dependent). Nasty.
If we think of the square wave, and the Fourier sinusoid components in the summation required to construct it, we could see "overshoot" as merely the result of some of the Fourier components not having the correct phase or amplitude, resulting in a differently-shaped waveform.
As is often the case, it all boils down to whether or not the causes and effects being discussed are significant-enough to worry about.
Thirty years ago I had all of the math and system theory down cold and could have come up with some actual numbers within a few minutes. Maybe someone here can do that. Otherwise, it will take me "a while". (Or maybe I can get LTSpice to help me.)
Tom
Last edited:
VladimirK,
Just to corroborate the range of your numbers and give a little better basis for audio slew rate discussion:
-----
slew rate max of sine (in volts per microsecond) =
[(2 x Pi) x (freq in Hz) x (amplitude in volts)] / 1,000,000
So, for example, for a 10V 0-to-Peak 20 kHz sine, the maximum slew rate is about 1.257 V/us. For 20V 0-P it would be about 2.513 V/us.
-----
So it depends on the amplitude. When those designers you mentioned recorded and analyzed the signals, did they use a voltage amplitude that might be seen in a high power amplifier? If not, you would need to scale the maximum slew rates upward.
Tom
forr,
The ear can easily spot both time errors and transient distortion, however, has a hard time detecting harmonic distortion, nevertheless, performs a harmonic structure analysis with incredible sensitivity. THD says nothing about harmonic structure, therefore, strictly speaking it´s worthless.
Regrettably, yes, with today´s sound recording, storage and reproduction technique and attitude, the sound stage suffers badly, but that´s not a good reason to make things even worse. (Sound stage is a lost feature, vanished a long time ago. Who needs it anyway? What matters is low THD now).
Those are the cause of dynamic transient distortions (TIM, SID, DIM, whatever). Such short-lived events can`t be seen with modest apparatus, only limitedly even with highly sophisticated (and pricey) apparatus.
The ear can easily spot both time errors and transient distortion, however, has a hard time detecting harmonic distortion, nevertheless, performs a harmonic structure analysis with incredible sensitivity. THD says nothing about harmonic structure, therefore, strictly speaking it´s worthless.
Small phase errors are an indispensable condition for a sound stage to have spatial dimensions, resolution and acuity; harmonic components must have a fairly accurate representation in time. But certainly, there`s more to it, phase distortion occurs without any global feedback, other distortions play a role as well.
lets not use even minimal common sense - if we can trash "conventional engineering"?
remember that the the audio power maplifier's task is to reproduce recorded musical signals
"infinite" SPL slew rates in air are incompatible with human hearing, or even continued life - not music
the fastest microphones sold for recording studio use roll off at 50KHz, the vast majority less than 25KHz
high feedback amps, like Bob Cordell's 1984 Mosfet error correction amp can have feedback bandwith corner frequency ~ 5MHz (around the output Mosfets - measured properly inside of the error correction loop)
many amps put ~ 200 KHz low pass filters at the inputs to reject EMI like AM radio
all commercial mass market music recording playback technologies have power bandwidth (1st order) or higher order low pass filtering of < ~ 100 KHz
remember that the the audio power maplifier's task is to reproduce recorded musical signals
"infinite" SPL slew rates in air are incompatible with human hearing, or even continued life - not music
the fastest microphones sold for recording studio use roll off at 50KHz, the vast majority less than 25KHz
high feedback amps, like Bob Cordell's 1984 Mosfet error correction amp can have feedback bandwith corner frequency ~ 5MHz (around the output Mosfets - measured properly inside of the error correction loop)
many amps put ~ 200 KHz low pass filters at the inputs to reject EMI like AM radio
all commercial mass market music recording playback technologies have power bandwidth (1st order) or higher order low pass filtering of < ~ 100 KHz
Baxandall's paper "Audio power amplifier design" which seems currently not to be available on Jan's site :
http://www.google.fr/url?sa=t&sourc...sg=AFQjCNHmQv6qAwdMmjV0GztYsHv2c4toQQ&cad=rja
WuYit,
Such short-lived events can`t be seen with modest apparatus, only limitedly even with highly sophisticated (and pricey) apparatus.
Then I can only suppose you have this kind of sophisticated apparatus to show the facts you alleged. Many of us would be happy to see you prooving them.
http://www.google.fr/url?sa=t&sourc...sg=AFQjCNHmQv6qAwdMmjV0GztYsHv2c4toQQ&cad=rja
WuYit,
Such short-lived events can`t be seen with modest apparatus, only limitedly even with highly sophisticated (and pricey) apparatus.
Then I can only suppose you have this kind of sophisticated apparatus to show the facts you alleged. Many of us would be happy to see you prooving them.
It is typical, that persons speaking about tiny audible effects, are being attacked and proposed to demo the effects they speak about.
Finally, Ed Simon, has created super low-noise setup, and has demonstrated differences in measurements of resistors, cables, connectors, etc.
And what has changed, the cable discussions have stopped?
Many of us are interested just dispute, not to explore the subject in more details.
God has not make any claims, that standard present-days measurement instruments covers all specifity of his creation - people's hearing.
Finally, Ed Simon, has created super low-noise setup, and has demonstrated differences in measurements of resistors, cables, connectors, etc.
And what has changed, the cable discussions have stopped?
Many of us are interested just dispute, not to explore the subject in more details.
God has not make any claims, that standard present-days measurement instruments covers all specifity of his creation - people's hearing.
Last edited:
Sure, but if the current ' cannot follow' the demand because of some impedance, it will necessarily show up as voltage deviations from a steady supply voltage, i.e. ripple. So, if there is no ripple on the supply, there's no lack of current. Ohms law is non-negotiable.
jan didden
About absence of the ripples, one could ask, what is the criterion of whether they are present or not.
As showed Ed Simon, resistor effects are very very small, but we hear them.
It could be, that ripples are of uV amplitude, but produce listenable effects.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.