Neil, if I understand you correctly, I don't think the rest position of the airflow meter should be significant. Even at idle, airflow should cause a significant deflection of the plate due to pressure differential between the top and bottom surface. Of course, there could still be a mechanical misadjustment of the plate position relative to the plunger into the fuel distributor - or the connection to the feedback pot. In my case, though, the key datum is that turning the engine off then back on usually restores normal performance - or in any event it seems to be the only thing that will. I don't see that affecting the mechanics of the sensor plate, nor - in the case of Duke's suggestion - the TPS.
I returned to the issue last night after a weekend away camping, and I removed the new OVP relay. I disassembled it and checked for problems. The relay contacts are hidden inside an internal case, but I would think they should not yet have problems. Several solder connections looked suspicious, though - problems on the assembly line. I reflowed these with fresh solder, repackaged, and reinstalled. I will report back in a week or so on my conclusion, but a short errand in the evening and the drive to work this morning proceeded with completely normal operation.