Charles writes:
<<I would have bad feelings about this solution if, as likely, it is due to field excitation current effecting the compass.>>

I have an older B&C standby alternator and it uses permanent magnets for the field. If the iron path in the magnetic circuit is saturating then you will have a stray field that can effect the compass. When the alternator (generator?) is turning the field is alternating and the average field goes to zero, as does the compass error. To track down the source of the (DC) magnetic interference, use a hand compass and triangulate to the source.

You can tell if your SB alternator uses permanent magnets by turning it by hand and seeing if it has significant cogging torque. No cogging, no magnets.

BTW, the same may be true for an electromagnetic field if the field coil has magnetized the pole pieces.

Regards
Brent Regan