The schematic is as below.
ExtOsc, sweep Fmult, measure the CLK frequency
Single run is OK and accurate.
When running .step, the measurements are wrong and non-accurate.
Checking the output waveform, the CLK frequency is OK and accurate.
qsch file attached
ExtOsc.qsch (5.8 KB)
Example:
VFmult=2.5 → 2.5MHz
.step
VFmul from 0.5 to 2.5 step 0.5 → 500K, 1MHz, 1.5MHz, 2.0MHz, 2.5MHz
@bordodynov
scroll up post process tab
meas number 38 should be 4.4e+06, rather than 4.38596e+06 (VFMUL=4.4)
Single run is correct
It may be related to the algorithm that decides the time step size.
In .step simulation, the time step size might not be adaptively calculated for every step.
@bordodynov Thanks for your information.
Changing MAXSTEP to 0.1n will yield more accurate measurement results but will sacrifice simulation speed.
@bordodynov still not accurate enough
