Ampman
u talked about scratch/pitch things etc demanding extremly high slewing amps, imagine if we would have a sample on the CD with one samplepoint at -32768 and the next sample is +32768(which is a maximum signal swing on CD), if this would be translated into a full swing output at your amp with for instance hefty +- 200 Volts rails, this would tranlsate to a slewrate at only 17,64 V/uS for a very unlikely signal output from a CD!
U C... a 400 Vpp signal swing which would be around ~2500 Watts into 8 Ohm if we assume sinus waveform, is very far from ur slewrate fantasy figures needed for an amplifier producing this rate of wattage.....
But asides from this I myself believe that an amplifier should be fast, or more exactely have a very small passthrough delay of signals through the amplifier, but this is not necessary the same as high slewrate.
Cheers!
u talked about scratch/pitch things etc demanding extremly high slewing amps, imagine if we would have a sample on the CD with one samplepoint at -32768 and the next sample is +32768(which is a maximum signal swing on CD), if this would be translated into a full swing output at your amp with for instance hefty +- 200 Volts rails, this would tranlsate to a slewrate at only 17,64 V/uS for a very unlikely signal output from a CD!
U C... a 400 Vpp signal swing which would be around ~2500 Watts into 8 Ohm if we assume sinus waveform, is very far from ur slewrate fantasy figures needed for an amplifier producing this rate of wattage.....
But asides from this I myself believe that an amplifier should be fast, or more exactely have a very small passthrough delay of signals through the amplifier, but this is not necessary the same as high slewrate.
Cheers!
Regarding the max slew rate requirement,:
Do the higher sample rate formats, SACD, DTS, DVD-A potentially get converted to analog signals quite a bit beyond the CD (and nominal audible) limit of ~20kHz ? Or does the process cut off lower?
Having inserted an RTA in my signal path a couple of years back and watched it with wide range source material I am convinced that excepting the very ocassional gamelon or electronica bits there is virually nothing going on source-wise in that top 1/3 octave and darn little in the 1/3 octave below that, too. Given that it would seem that the question reduces to what is neccessary to preclude audible IM (or TIM) artifacts generated by 20k+ nonlinearities. Oh yes, include anything that affect amplifier stability.
Do the higher sample rate formats, SACD, DTS, DVD-A potentially get converted to analog signals quite a bit beyond the CD (and nominal audible) limit of ~20kHz ? Or does the process cut off lower?
Having inserted an RTA in my signal path a couple of years back and watched it with wide range source material I am convinced that excepting the very ocassional gamelon or electronica bits there is virually nothing going on source-wise in that top 1/3 octave and darn little in the 1/3 octave below that, too. Given that it would seem that the question reduces to what is neccessary to preclude audible IM (or TIM) artifacts generated by 20k+ nonlinearities. Oh yes, include anything that affect amplifier stability.
wimms said:
I think it depends on where the slew rate limiting occurs. Probably it occurs, when it does, in the Vas stage. If you have an amp with strong feedback, the feedback tries to keep the output in line with the input, the error signal gets much greater resulting in gross input overload. After the slew limiting signal is gone, it may still take the amp many milliseconds to return to normal operation. In a common emitter output stage , where low output impedance depends on the feedback, the Zout would be very much higher than normal. In a common collector output stage, which has already a relatively low Zout without feedback, the effect may be less, on Zout, that is.
Jan Didden
To add to my previous post: the effect on Zout will be minor compared to the gross distortion resulting from effectively a very non-linear amplification. That is with full slew rate limiting. Appreciable distortion may start even when approaching limiting. That is why most people want to take a safety factor for the slew rate capability. But a competently designed amplifier will have no problem with 20V/usec or more which should be OK for most applications.
Jan Didden
Jan Didden
sam9 said:
Casual measurement on my part indicates that power-wise, you
see what appears to be a 12 dB/octave decline above 5 KHz.
This is reasonably consistent with slew rate measurements of
audio material. Drum rim shots will take you up there, but even
they don't seem to make use of the full 20 KHz power bandwidth.
I think Ampman's assumption comes from ordinary audience's perception. We can hear that synthesizer effect or turntable DJ scratch are fast tones. At a glance anyone who hears it will feel that it will need a fast amp to reproduce that sound.
People like Nelson Pass do not make simple (and possibly wrong) perception like that. To be sure, he MEASURE those tones.
We can still hear the effect and scratch that DJ makes. That means it is still under 20khz, where any descent audio amp should be able to reproduce it, not just an amp with special slewrate.
If the effects and scratches are made by Doggy DJ and Batman DJ, maybe they will come up with 30khz or 40khz tone effect. This tones are fast, but can we hear it?
People like Nelson Pass do not make simple (and possibly wrong) perception like that. To be sure, he MEASURE those tones.
We can still hear the effect and scratch that DJ makes. That means it is still under 20khz, where any descent audio amp should be able to reproduce it, not just an amp with special slewrate.
If the effects and scratches are made by Doggy DJ and Batman DJ, maybe they will come up with 30khz or 40khz tone effect. This tones are fast, but can we hear it?
Then you start opening garage doors and driving dogs nuts. I'm sure there is someone who will say they can hear 30KHz +.
The zobel network is now at risk if it exists. I think you are further ahead building a wide bandwidth amp within reason, and filtering the input. You thus guarantie the slew rate limiting with music will not occur. Building a DC to light amp is an exaggeration that will be at the expense of something else. All designs are a balance of different factors.
-Chris
lumanauw said:
Not frequency response, but full, or preferably, rated power bandwidth of at least 160KHz (according to Jung), but i prefer 200KHz, together with a curtailed 'small-signal' frequency response ,(circa 3dB ~200KHz or less), solely to reduce the possibility of RF driving the unit into slew overload...
I have (some) friends that likes to comment audio system. They talk, saying this system is like this, this system is like that, lack this frequency, the frequency is over in ....khz, etc-etc. They say it without RTA!! They also adjust EQ without RTA. When it is monitored with pinknoise and RTA, the spectrum is like chaos theory. Far from Fletcher Munchen curve or ordinary step-down response.
There are people like this. When they hear a high trebles, they say it is 20khz. When monitored in eq, it is only 12khz or 16khz top.
Sound can give different perception for different people. That's why measuring is the only way to be sure.
There are people like this. When they hear a high trebles, they say it is 20khz. When monitored in eq, it is only 12khz or 16khz top.
Sound can give different perception for different people. That's why measuring is the only way to be sure.
Maybe it would be advisible and also easier to keep RF from entering the amp in the first place, than building the amp in a way that it can process it cleanly !?
Regards
Charles
Edit: Forgot to mention the fact that an amp might process a signal cleanly if it enters through its input (i.e. it has a high slew-rate capability) but could still be affected if it enters somewhere else (i.e. output, PSU etc). So what is gained then ?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.