I get a lot of time a simulation will run from orcad if I click the play button, but won't converge (bias point fails) if I do the advanced analysis or run it through a script. Is there something different with regards to initial conditions going on between doing a run from advanced analysis and doing a run from orcad capture? I'm using the latest version of 17.2. The log files look like they show the same simulation parameters. To be more clear, what I have done is observed a non-convergence in a monte-carlo analysis and then wrote a TCL script to set the parts in the schematic acording to the non-converging condition in capture, and then when I run that it converges fine and shows the plot. I have had the same issue when I do runs where I set the parts with a TCL script in PSpice and it does not converge, but when I set the same ones in capture and click play, then it does. There has to be something with initial conditions or .NODESET's for the bias point calculation that I'm missing?
I'm alsmost wondering if it's using bias points from a prevous run sometimes instead of starting from scratch, but I have it set to start form scratch I belive?