Hi @Tim Edwards @Leo Moser, a simulation in cace is not giving the expected result. The design is here. The xschem testbench works fine, it generates the signal that conmutates between 0 and 1.8V. The cace testbench is giving a "logarithmic" curve which has no relation to the design. This is the entry on

SDC.txt

name:		mean_130
status:		skip
description:	Mean output value at 130 degrees C
display:	At 130 °C
unit:		V
spec {
	typical:	any
}
simulate {
	tool:		ngspice
	template:	transient.spice
	format:		ascii .data null result
}
measure {
	tool: python
	filename: analyze_temps.py
}
conditions {
	name:		temperature
	typical:	-40

	+
	name:		corner
	typical:	tt
}

The `analyze_temps.py`script is executed after the simulation results, and it works as expected.

I changed settings to "keep simulation results", then ran CACE, then removed the "quit" from the control block and re-ran ngspice and did "plot V(out)" and I get the result from your first screenshot, which is correct. My best guess is that the pin order of the standalone subcircuit is not the same as the pin order of the call to that subcircuit. This could be due to xschem if your version of xschem is too out of date, as I think this problem was fixed a few months ago. What is your xschem version?

The xschem version was the problem but just updating xschem was not the solution. It work only after re-making the SDC symbol. Thanks Tim.