P1db (linearity/power handling capacity) simulation problem for a t/r switch

Hi! i need your kind help to solve my simulation problem.

I am simulating a design of series shunt T/R switch. As we know insertion loss, isolation and p1db point are the main performance parameters of a switch. (), i am doing simulation to evaluate the parameter values for my design. I am using 0.13um CmOS in mentor graphics software.

I have no problem when i simulate for isolation and insertion loss. But for p1db measurement every time after 4db transmission power, the simulation got stuck. the simulation log shows that there is convergence and divergence occurred.

Hence my query is:

what are the reasons for this? and what can i do to overcome the problem?

Any suggestion and clue will be very helpful for me. Thank you in advance for your time.

A common-enough problem is simply models that blow up
when fed overrange voltage or current. Or even non-
overrange, but difficult to digest situations like an
intrinsic-body (VT=0) BSIMSOI FET given a gate voltage
that swings above and below source, making body doping
(numerically) pass through zero and making "interesting"
things happen to the S/D diodes' depletion regions. This
kind of thing, you can wait for years on Berkeley and the
CAD tool vendor to settle their finger-pointing. So while
the cause may be interesting somewhat, you probably
need to move on to expedient approaches.

Maybe you want to creep up on 4dB (what dB? Usually
RF power is in dBm, and 4dBm is not much at all) and get
an idea of whether the switch is set up right for simulation,
or whether perhaps it's in a state where your stimuli are
making some internal voltage / current crazy. Maybe
charge pumping with a lack of restoring leakage is taking
things beyond where models are fitted and sane.

Maybe you want to start at higher power and back it off
instead, approach the singularity from the other side and
maybe you get to the sought result before you hit the
landmine.