Analysis

Q: How do I make .TRAN and .TRAN SWEEP results agree?

A: If you get different results from transient simulations than when doing a parameterized transient sweep (and you are using transmission line elements), you should explicitly set the option RISETIME. For example:

.OPTION risetime=<rise time of fastest component being simulated>

Solution: 28

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Q: Why do I get "Internal timestep too small" when I use UIC in the .TRAN statement?

A: If the UIC is used in the .TRAN statement, the DC operating point analysis is bypassed and a transient analysis is started using node voltages specified in an .IC statement.

During a simulation, DC operating point analysis is an important step. Most of the DC analysis algorithms, control options, initilizations, and convergence also apply to transient analysis. As a result, although a transient analysis might provide a convergent DC solution, the transient analysis itself can still fail to converge if UIC is used. In transient analysis, the error message "internal timestep too small" indicates that the circuit failed to converge. The convergence failure might be due to stated initial conditions that are not close enough to the actual DC operating point values.

Therefore, the best solution is to comment out the UIC from the .TRAN statement and rerun the simulation. At that point, the DC operating point analysis is performed and the correct values are used in the transient analysis.

Solution: 383

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *

Q: Is there a way to change the seed that pseudo-randomly generates the initial conditions in the Monte Carlo simulation?

A: Yes, you can set .OPTION seed=<value> where value is between 1 and 259200. This changes the random number generator starting seed.

Solution: 408

* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *