It would be of great help if the most common and very cryptic message “Fatal error: Timestep too small” could be somehow clarified with a hint. If I would understand where the problem lies, I could save a lot of time not digging through the whole schematic.
3 Likes
I dont know the inner working of Spice, and whether its possible for it to pinpoint the trouble maker.
However, I am sure determining the error rootcause is not trivial. And can even give you wrong suggestion.
Thus, if possible…its always better to always check if the simulation stays converging whenever you add anything, before you ended up with big mess
1 Like
@physicboy Thanks Arief for the tip, appreciate it. Paying attention to what changed helps most of the time, the trouble is with dormant issues that are popping up when the complexity increases. I have seen in LTspice a note after this error, something like “trouble with the node Nxxx”. Maybe we can get something like that in QSPICE.
1 Like
I completely agree. LTspice indicated a “bad” node in which the rate of voltage change was excessively high. If this node was formed by a voltage source, then I replaced it with an equivalent circuit consisting of a current generator and a parallel resistor. To prevent a very sharp front, I added a small capacitor. Oto limited the rate of voltage change at the “bad” node. In Qspice it’s even easier. You just need to set the internal resistance to the voltage source and add a capacitorю
.
1 Like