I am simulating a fairly complex parallel buck converter power system with a discrete time controller which dictates the output voltage (1kHz sampling rate).
I am trying to compute the loop gain of the system (PLECS, very similar to MATLAB Simulink), but one issue I am running up against is that even if I perturb the system with the largest perturbation the linear behavior region of the system can accommodate, the change on the quantized controller output is only a handful of bits.
The ultimate effect of this seems to be that the simulated loop gain can vary, sometimes quite significantly, depending on the size of the perturbation and how it interacts with the discretization of the controller output.
Can anyone point me in the direction of some resources for me to understand better what the effect of controller output quantization is on closed-loop stability? At present I am computing the loop gain with the output quantization inactive, but I'm not sure if this is appropriate when trying to give a definitive metric for loop stability (gain/phase margin).