Our actual diode implementation has the problem of discontinuous behavior in case of activated recombination current. Test case and picture is attached.
Recombination current test case
v1 a 0 0
d1 a 0 dio
.model dio D Is=1p N=1 Rs=1m Cjo=1p M=.5 Vj=1 bv=6 Ibv=10m nbv=1.4 ikf=100m
+ Isr=10p Nr=2
.control
dc v1 -6.3 1 0.001
plot -i(v1)
plot -i(v1) xlimit -5.5 -4.5 ylimit -50e-12 10e-12
plot abs(i(v1)) xlimit -1 1 ylog ylimit 1e-13 10
rusage totiter
.endc
.end
Even this effect is only in forward direction applicable the model has to be modified that also the base recombination current appears in reverse and breakdown region too.
Problem fixed by commit c1acc44 in pre-master-46.