Just perhaps bringing this topic back to 'top' as I've also recently developed this problem. Attempting to follow through some of the suggestions here hasn't really helped and I've got the even more unusual scenario of the problem seems to correct itself when the wind speeds are more noticeably higher. I can't say there's a specific value when the problem stops, but the quick veering to a N'ly direction does seem to happen when winds are lighter and it's not because the winds are simply variable.
Today has been a good example where the dominant direction has been N'ly, even though it hasn't with wind speeds around 5 to 10 KT, however, wind speeds in association with the passage of a front have increased and are persistently in the 10-20KT range from a general W'ly direction, this is now being correctly recorded by the Davis VP2 wireless anemometer.
I would hope that a completely new unit would eradicate the problem, but as someone posted months ago, that process was carried out without the problem being solved, so sort of stuck between a rock and hard place here.
Has anyone else had any further experiences with this issue since this topic was created?
Many thanks.