Steve, you nailed it.
The picture below shows the two semiconductor pads (arcs). The one on the right I took care of with an eraser. The one on the left is worn completely thru to the metal substrate; apparently the idle zone.
There is one thing I cannot visualize.
If I leave the original sweep arms in the air meter housing and just replace the cover containing the fixed circuit board (a new version of the part pictured below), the only calibration I can see is the 1-2mm of play when seating the cover.
And if that was enough to throw the calibration off, don't the odds favor miscalibration after using an eraser and reinstalling it?
If I understand this correctly, the resistance is a function of the distance along the arc from the sweep arm to the end of the arc.
What am I missing here?