--- In LTspice@..., Philip Bellingham <rmhc78a@...> wrote:
Emma,
?
I would also suggest changing your output load from a constant current to a resistor or some thing that more closely approximates a real world load. There are times in this simulation when both IGBTs are off. Since the current source still tries to maintain the specified current, the output voltage will have a magnitude of several thousand volts due to the current source forcing the current through the high impedance presented by the cut-off transistors.
?
I have uploaded a version of your schematic file modified as I have suggested (igbt_infineon_modified.asc). It ran for me using the Normal solver with modified trap integration.
?
Regards,
?
?? - Philip
Typically, the load on IGBTs is highly inductive so it is appropriate to use a current source as a load. Therefore, you need reverse diodes across the IGBTs.
There is obviously something wrong with the IGBT model, since the simulation takes so long.
See faq_11.txt under the Files=>FAQ section of this group for information on timestep and convergence problems.
Rick
