@Tim Edwards there is this line in the combined model set (

libs.tech/combined/continuous/models_global.spice

, line 90) that looks like the result of a substitution gone too far:

.param mismatch_factor*MC_MM_SWITCH*AGAUSS(0,1.0,1) = 0 mismatch_factor*MC_MM_SWITCH*AGAUSS(0,1.0,1) = 0

Surprisingly ngspice does not seem to care about, while Xyce dies on that. https://open-source-silicon.slack.com/archives/C017P3RAD42/p1710343650668069?thread_ts=1710339391.855419&cid=C017P3RAD42

Agreed. The equivalent line in the original models was setting

mm_z1

and

mm_z2

which have been replaced throughout the models with calls to

GAUSS()

. So the line should be eliminated because there is nothing to set. If ngspice ignores the line then that was (inadvertently) the correct behavior. I will fix the source.