From: Alan G. <ag...@sp...> - 2009-09-19 22:53:39
|
Niels Egberts wrote: > This is still very easy to reproduce by simply creating a matrix > display. It will output "Invalid matrix!". And after adding some fixed > voltages and ground elements it crashes and outputs matrix A. This is > with the SVN-version. yeah, there are several underlying issues there. You can remove the assertion checks and relax the parameter for the matrix print-out on high error. But the real problem is how Diode is implemented. It routinely puts huge values into the matrix, sometimes even INFs, It is extremely glitchy. The best way around the glitch appears to be to put a resistor in parallel with each diode. I spent the whole day cleaning up the code for the matirx driver and the matrix itself. It still doesn't work though. =( All said, I've put hundreds of hours into trying to get nonlinear working right and ARGH!!!!! =(((((( -- New president: Here we go again... Chemistry.com: A total rip-off. Powers are not rights. |