
There seems to be some convergence issues for reltol of 1e-5 or less.
For example:

test                        acBias_Ccb_EXMOD_2
temperature                 25 100
biases                      V(e)=0 V(c)=1.0
`include substrate_bias.inc
//biasSweep                   V(b)=0.7,1.0,0.02
biasSweep                   V(b)=0.7,0.9,0.02
freq                        lin 1 1.0e9 1.0e9
outputs                     C(c,b)
`include dtestpars.inc
modelParameters             EXMOD=2
`include ../test_specs/set_mult.inc

At the bias sweep value of 0.9 at 100C, the circuit doesn't converge.

After some WRspice mods, the issue is still present, but less commonly
observed.  At present, the issue does not appear with Sparse package,
double or long double, with reltol=1e-8.  It does appear with KLU,
reltol=1e-8, but not with reltol=1e-3.

XXXXXXX

Issue is no longer present after fixes to WRspice internal logic.

