Normally the dominant pole is the VAS and gate stoppers bring the OPS pole too close to it so that there is too much phase lag above 0dB loop gain, not enough phase margin, hence instability. So, simulation likes zero gate stoppers, but simulation does not account for local instabilities. It may help to space the gate wire closer to the drain than the source, (or vs versa) since the current in the drain and source go in the opposite directions. Vendors should look for a better lead arrangement like maybe GDS instead of DGS, but they have probably been more concerned about the voltage difference and not RF coupling. I did a simulation of coupled inductors on MOSFET, but I can't find it just now. Yes, it oscillated.
Last edited: