Mchambin, you have shown an excellent example of the difference between SOFT TIM and HARD TIM. You have soft Tim here, and that is why the distortion rises well before the hard TIM limit is reached. This is why we usually factor a multiplier of 5-10 (10 preferred) to estimate the
WORST CASE SLEW RATE necessary to avoid TIM in every case.
WORST CASE SLEW RATE necessary to avoid TIM in every case.
But it all depends on the source. I can slew limit ANY amplifier with a 1MHz square wave.
How can anybody put a requirement for slew rate if you don't know what the slew rate of the source is? You guys discuss slew rate capabilities of opamps without having a clue of what is needed.
Traderbam put the relevant question and it is ignored.
You guys all put the cart behind the horse and consequently all this discussion does is waste bandwidth by going nowhere.
As one of my mentors once said: "If you don't know where you're going, any road will take you there".
Jan
How can anybody put a requirement for slew rate if you don't know what the slew rate of the source is? You guys discuss slew rate capabilities of opamps without having a clue of what is needed.
Traderbam put the relevant question and it is ignored.
You guys all put the cart behind the horse and consequently all this discussion does is waste bandwidth by going nowhere.
As one of my mentors once said: "If you don't know where you're going, any road will take you there".
Jan
Not the Threshold NS-10. Its slew rate is 100 volts/microsecond BUT they fitted a 300 kHz lowpass filter at the input. At the maximum possible output (±10V peak to peak), when you put in a 300 kHz square wave, you get out an exponential-rise, exponential-fall wave whose rise time is 1.16 microseconds. 20 volts in 1.16 usec is 17 volts/usec. Much lower than 100 V/usec which the NS-10 can do when the filter is bypassed. So, you can't get a factory made NS-10 (with filter intact) to slew rate limit.
Last edited:
No you can't, at least not always.
Try slew limiting a current-on-demand amplifier (CFAs could be a good start to find one). You'll note most of those amplifiers are never slewing, but always enter into bandwidth limitations first.
What I think you are missing is that there's essentially no general relationship between bandwidth and slew limitation. However, in particular cases, like the standard Miller compensated Blameless, these two phenomena ARE coupled; both depend on the Miller compensation cap, that a) sets the ULGF and hence the closed loop bandwidth and b) defines the slewing behavior, since the long tail pair input stage bias current charges it at a 2*I/C rate, when the input stage enters nonlinearities.
Yes.
And all discussions that ignore this are a waste of time.
You cannot compare SR which is about a large signal non linear behavior with dV/dt of signals that stay within the linear behavior.
They both tell about speed, they both use the same unit, however there is no general relationship in between. The former belongs to non linear behavior, the later belongs to linear behavior.
No, I am very well aware of this, having made the point several times here myself.
The context of my post was that I was missing a clear indication how much slew rate we can expect from a source, so that we can sensibly develop the slew rate requirements for our amps.
I note that even you Waly quietly sidestep this basic question.
So we remain on the road to nowhere.
Jan
No, I did not. You made (and I quoted) a wrong statement (about being able to slew any amplifier) and I thought I should correct it.
P.S. Besides, the frequency of the input signal is irrelevant for slewing (or not) an amplifier. The amplitude and the rise/fall times are.
Last edited:
I captured two short videos to show the difference between VFA and CFA amp slew-rate. (rise-time vs amplitude)
VFA
https://youtu.be/MBuQ5E-ruyw
CFA
https://youtu.be/7C-1RJLsu2w
VFA
https://youtu.be/MBuQ5E-ruyw
CFA
https://youtu.be/7C-1RJLsu2w
No? By your calculation, if I input a +/-100V 300kHz square wave with very fast rise time it rises 200V in 1.16uS. That should be about 172V/uS which is higher than the preamnps' 100V/uS.
So from this it appears I can slew rate limit any amplifier you throw at me. Or am I missing anything?
As to the CFA example quoted, I understand that they slew very fast. But in the limit there is always a limit to the current available to change node capacitances so the difference with a voltage amp in this respects seems quantitative rather than qualitative.
Jan
Nice videos! Especially the two were put side by side to show the difference. Thanks a lot!
The VFA apparently was over driven by the input signal's high rate of transition and its output was slewing, we can tell that from the constant rate of transition (V/uS) in the waveform that was independent of the amplitude, meaning the feedback loop is out of control.
I may have missed something, but the CFA video, on the other hand, does not show an apparent evidence of slewing, as the rate of transition in the waveform appears to be following the amplitude in a tight proportion, indicating the feedback loop is in control. This CFA may have a slew rate that is greater than even the maximum rate of transition in the waveform shown in the video. In another words, we are not seeing the slew rate of the CFA from the video because it was not slewing.
As a side note, in my old book signals do not slew. Amps do. I felt sometimes signal's rate of transition got mixed up with slew rate might have caused confusion.
again - a signal always has a slew rate - its time derivative
circuits, amps may have slew rate limits
use some intelligence to determine the context rather than waste bandwidth 'correcting' people who aren't trained in the same tradition you are but who's posts make sense when you allow for their usage, context
circuits, amps may have slew rate limits
use some intelligence to determine the context rather than waste bandwidth 'correcting' people who aren't trained in the same tradition you are but who's posts make sense when you allow for their usage, context
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.