Hello together,
I see this strange behavior of Mosfet leakage current:
The current seems to increase indefinitely when reducing reltol!
I could not resolve it by tightening other convergence parameters. Also couldn’t remove the apparent noise!
Indrit
mos_leakage_report.qsch (2.1 KB)
How about trying method=gear?
Noise looks different, but same current levels!
I don’t know which part in the NMOS model, but the simulation results suggest a drain-source element depends on reltol setting.
However, as you generally set reltol=1e-3 (0.1%) as default, and leakage generally model with a resistor. So, for example, you expect 10V with 1nA leakage, you can add a Rds=10/1n into .model, which will give you practical what you need.
The ringing is actually trapezoidal ringing in resolving the circuit. There is capacitor model in the mosfet model and solving that in this current magnitude you notice that very very small fluctuation. If capacitor model removed from the model, you see the trap ringing eliminated. Changing to gear integration added damper to the circuit, but it also change the solver from trap to gear.
