You would expect one peak to be quite a different from the other, because it is basically nonlinear. Consider a step input going from low resistance to high; the loop gain will be high and the bandwidth low. Step change in the opposite direction is to a higher bandwidth regime with low feedback resistance and low loop gain. So, rise and fall times could be quite different (so long as neither test exceeds the op-amp slew rate).
Thus, when driven with a sine, depending on the dR/dt, the waveform COULD be quite different at the positive and negative peaks.
Jim Wagner
Oregon Research Electronics
toggle quoted message
Show quoted text
On Apr 14, 2013, at 7:33 PM, Andy wrote:
MOHAMMAD A MAKTOOMI <amaktoomamu@...> wrote:
Andy, how could you guess so quickly that waveform was missing some
peaks and the step-size be reduced in the setup
< ()?
I ask this because this will help me start thinking the way an experienced
user thinks.
When I looked at the waveforms, I noticed that the positive peaks of
the 1 MHz output were irregular, even when zoomed in. It looked to me
as if it suffered from aliasing, being sampled but not with a high
enough sampling rate. That led me to check for the plotwinsize=0
option, and then to add the max timestep parameter.
Andy
