Under normal circumstances, an oil separator's efficiency increases as entering velocity decreases, as would be the case with the single variable of condensing pressure lowering. (I expect arguments here )

I think something else is changing that you are missing. Possible starvation of the evaporators which could mitigate the oil return from the suction lines camouflaging itself as an inefficient oil separator?

Okay, if that doesn't pan out, how about liguid overfeed!? That could dilute the oil in the separator, perhaps providing the same misindication of the loss of oil separator efficiency? Hmmm. .... possibly caused by ambient subcooling?

Good question.