analysis pss terminated prematurely due to error

I am trying to simulate a down mixer using pss in spectre. The RF and LO frequencies require a small beat frequency of 35MHz. I have verified the operation of the mixer using transient. However, when I tried to find the 1dB compression point with swept pss, the simulation is extremely slow. And the simulation would stop prematurely after about 3~4 swept pss iterations. The simulation log window did not show any error message and the analog environment would show the status as "Ready" after the simulation terminated. The pss simulation can finish only if I swept no more than 3 points, which is not useful. I even reduce the "number of harmonics" parameter to 1, but it did not help the situation. Using a larger beat frequency would help a lot, but my assignment requires a 35MHz beat frequency.

Has anyone run into similar problem before?

There should be a topology problems that influences the convergence of the simulation such as high impedance nodes with ver small capacitors, differential floating lines, branch loops etc.

It should be increase the harmonic number instead of decreasing to simulator be converged.

There may be numbers of problem that distirb convergence efficiency including bugs of cadence, so to say something is almost impossible for that..

But check the topology of the circuit from the point of view redundancy ...

it is better to check ur transient analysis first.
sometimes u r making an osc but u r not aware.

hi,

its better to do qpss analysis for mixer especially for icp and iip3 as its fast even with small beat frequencies. see the link below for the mixer simulation using cadence.

https://www.edaboard.com/viewtopic.php?p=300285#300285