I checked for the 'sticky' problem on an Etrex 30 running version 2.70 software, by comparing it with a Vista Cx, holding one device in each hand. I marked a location next to a lamp post as a waypoint on both devices, and then set both to navigate to it. Just as other users have reported, the Etrex 30 did not immediately start updating its position if I walked very slowly towards the waypoint (under 2 km/h), but I did not get such large errors as other people have found (never as much as 10m), and when I arrived at the waypoint, the Etrex 30 slowly updated its position until it was within the normal accuracy of hand-held GPS devices. If I walked briskly (5-6 km/h) both devices showed essentially the same distance from the waypoint at all times. Initially I had GPS, GLONASS and WAAS/EGNOS enabled on the Etrex 30, but turning of GLONASS made no apparent difference, and neither did turning off WAAS/EGNOS. I did the test on flat open ground, with no significant obstructions other than the lamp post that I was using as a waypoint. It seemed to be impossible to fool the Vista Cx, no matter whether I walked slowly, walked quickly or stayed in one place.
As regards compass calibration, we can't really blame Garmin for the need to recalibrate after changing the batteries. The casing of NiMH batteries is made of a magnetic material (I think it's steel), so different batteries will need different compass compensation.
