DC Correction Fails (Functional Form Method)
The Simulation Does Not Converge to the Correct DC Value
The following are the most commonly observed issues that occur when using the DC Correction algorithm. Details are provided in each to help troubleshoot the problem.
An unstable time domain simulation could be the result of a variety of issues. As it pertains to DC Correction however, the following are suggested:
Solution:
Increase the shunt conductance (say to 1e-9).
Reduce the lower frequency bound (to say 1e-3 Hz). This should increase the fitting accuracy at low frequencies.
Decrease the maximum curve fitting error. This will help increase the curve fitting accuracy.
This problem may occur when the magnitude of the propagation function is significantly higher than zero at the upper frequency bound and the optimization algorithm has not converged.
Solution:
Increase the upper frequency bound (ex. 1e8 Hz instead of the default 1e6 Hz). The log file will indicate:
|
Applying DC Correction --------------------------------------------------------------------
Unable to apply dc correction (Functional Form method) -increase the upper bound for fitting |
If all else fails, try using the Add Pole/Residue method.
This is most likely due to the presence of diminishing low frequency transients.
Solution:
Follow the same solutions given above under Simulation is Unstable.