Hello all, I encountered a weird problem with xschem and ngspice. In the tran analysis of this triangular wave generator circuit, if I change the transient analysis step from 5ns to 4ns, the simulaiton result of the output triangular wave would have a large jump to an irregular point at a certain time, as shown below. This is not the first time I run into this problem, it happens pretty randomly when I change the timestep of the transient simulation. Sometimes it also happens if I change some values in the circuit even though the timestep is kept the same. Does anyone know what is happening?

It looks there is some deadlock or stable point in the circuit. But i can not be more precise since i don't have the OTA implementation. My suggestion is to plot all other nodes (also OTA internals) to see what is changing at ~12usec.