transient simulation of the same circuit--different results?

Now I am using transient simulation in cadence to simulate the circuit. If I take a look into the small spike of the circuit output current in the time scale of several hundred nano seconds. I find the amplitude of this spike fluctuates. The amplitude of the spike can vary from 1,5uA up to 20uA, as far as I see.

1, the amplitude of the spike is different in each periode of the same simulation. For example, if the periode of the output signal is 1ms, then the amplitudes of this spike at 3ms and 4ms are different.....

2, the amplitude of the spike is different in different simulation. For example, each time I run the simulation, the results are a little different.

3, the amplitude of the spike is different on different computer. I have tested the simulation on 32bit computer and on 64bit computer, and cadence is installed in both of them. The results are also quite different.

And with all these three conditions, the spikes seem to vary randomly, there is no rules for the variations at all.

I am wondering what are the reasons. Is it because that the transient simulation is not accurate enough for such a small amount of current during several hundred nano seconds?

Re: transient simulation of the same circuit--different resu

Increase the accuracy of the simulator: reltol, abstol, etc

Re: transient simulation of the same circuit--different resu

in fact, I was using ULTRASIM in cadence to perform the simulation
because for large circuit, spectre is a little too slow

If you use different versions, it may be the difference of algorithm.
If the same, indeed amazing me.

Re: transient simulation of the same circuit--different resu

which version do you mean? the version of cadence?

Hunter_BJUT said:

If you use different versions, it may be the difference of algorithm.
If the same, indeed amazing me.

Click to expand...

yeah, version of cadence.
in the same condition, simulation results should be consistent, unless configuration difference.