What does this have to do with it, we're not talking about long or short cables, we are talking about a feedback loop that connects the output and input of the amplifier ..
Here are inevitable or conscious RC network ( lead–lag compensator) which creates delay, and which necessarily limits the bandwidth and thus introduce delay in the feedback loop.
We fight constantly against instability, if delay too small, bandwidth too big, risking instability during fast input signals.
It is a balancing act for quick equal to instability, too slow too much "pulse distortion". either we must limit the input signal speed, or the amplifier's speed (with the errors it gives) to get stable results
please don't use the unqualified word "delay"
it really is annoying, the unqualified use of "delay" in a discussion that the distinctions between minimum phase lag functions vs non-minimum "essential delays" is centrally critical to understanding and explaining
phase shift, group delay can be shown for either type and by themselves don't make the distinction between minimum and non-minimum phase
basically if you don't know the distinction please quit "contributing" "explanations" based on semantics, naïve reasoning
it really is annoying, the unqualified use of "delay" in a discussion that the distinctions between minimum phase lag functions vs non-minimum "essential delays" is centrally critical to understanding and explaining
phase shift, group delay can be shown for either type and by themselves don't make the distinction between minimum and non-minimum phase
basically if you don't know the distinction please quit "contributing" "explanations" based on semantics, naïve reasoning
Sergio
Apologies for a time-loop!
One reason I do not see the effects of an input filter is that I use small value capacitors, not to limit audio inputs but to keep the input impedance low at high frequencies (for stability reasons). I agree if the input is band width limited then this transient distortion can be suppressed. Should have paid more attention to the replies.
But interestingly, transient Miller compensation does not show this effect.
I am working on a design now with the goal of minimising or avoiding feedback from the VAS collector since this contributes to crossover distortion in the sense that it applies a correction signal that slows down the response needed to minimise crossover distortion. This makes the difference between, say, 0.01% at 20kHz and 0.002%, but it is all a question of whether the unit is stable. When it is bench tested as working I'll report it.
Apologies for a time-loop!
One reason I do not see the effects of an input filter is that I use small value capacitors, not to limit audio inputs but to keep the input impedance low at high frequencies (for stability reasons). I agree if the input is band width limited then this transient distortion can be suppressed. Should have paid more attention to the replies.
But interestingly, transient Miller compensation does not show this effect.
I am working on a design now with the goal of minimising or avoiding feedback from the VAS collector since this contributes to crossover distortion in the sense that it applies a correction signal that slows down the response needed to minimise crossover distortion. This makes the difference between, say, 0.01% at 20kHz and 0.002%, but it is all a question of whether the unit is stable. When it is bench tested as working I'll report it.
Even a class A output stage may have non-linear transfer characteristics. Crossover is one reason but simple exponential dependency on Vbe is another ... high impedance (current) drive is more linear (especially with modern linear gain transistors). A class A OPS may eliminate crossover distortion but I suspect that THD will still be higher if the compensation capacitor is connected to the VAS collector than the output? (Next simulations?)
- Status
- Not open for further replies.