Iris,
yes, i have seen this too.
It seems to be a bug, because the simulation results look ok.
Lets report it to Protel.
GMIN is one of the spice variables. Actually its the conductance (1/R) of
the resistors that spice places in parallel with any semiconductor p-n
junction during the simulation
I do a lot of simulation with large currents through MOSFET's and
particularly with reverse recovery issues there is a lot going on in a
small amount of time compared to the wavelength. When I get gmin
problems a crude way is just to increase the Points per cycle on the
Transient page. I