Why qspice have these results in case of this circuit?


test3.qsch (4.7 KB)
Why V(u1) and V(u2) show like that?

It should show more like below:

For the Switch model, if Ron=1n, you need to disable Fast Math (Edit > Preferences: Deselect Fast (less accurate) Math). Alternatively, if you want to keep Fast Math, change Ron=1u.

This issue may be related to precision of math.

Good, thank you ! ! !

I have another question:

test3.qsch (7.4 KB)

Is it normal that voltage V(u1_a, u1_b) to be like that when V(SW) have pulses:
Should not be also pulses on V(u1_a, u1_b)?

Check V(u1_a) and V(u1_b) at steady state. They are charged to the same potential. Therefore, regardless of whether you turn ON or OFF the switch, the potential difference is 0.
This is not a Spice simulation problem but rather a basic electronic analysis.

OK, my question was if this is the expected behaviour or not, as in this case to be no pulses on V(u1_a, u1_b) when having pulses on V(SW)?

If you increase R1 to for example 1000 ohms, you can see potential different for few cycles until C2 is fully charged.

Good, thank you, this I was looking for to know.

I simulated the circuit with Ron=1u (SW) and Fast math disabled, the result seems ok. But if I simulate the circuit with a simulation time equal to 10000000 s and Fast math disabled, I obtain a strange result as in the figure. The result seems to depend on the simulation time.

Stefano, what are you asking for is over 115 days of simulated time. You can get the result you want if you force max step size for example.