diamondaviators.net

When I first got my plane many years ago, if I was flying straight on HDG mode, then turned the bug anything substantial to change course, the FD would intercept the new heading consistently with amazing precision, no wind, big crosswind, whatever. It was always dead on. Now, when I do the same, the FD usually overshoots by 5-10 degrees or so, then quickly corrects back to the set course. It seems it is the FD making the overshoot, not the autopilot, as the pink and yellow chevrons are locked throughout the whole maneuver. Any thoughts on why, or if others see this behavior as well? Intercepting a localizer however, is still dead on.

Winds have no effect on changing heading, only on the resultant course. I use that function frequently, and my wife was using it a lot yesterday in some training. It's very reliable about turning to the selected heading.

I suppose it could be an issue with the magnetometer?

Russ, you are correct, winds should have no effect on that. I guess I noticed it as well last time flying on AP NAV when a waypoint was passed instigating a turn, which also then overshot the new desired course by 5-10 degrees before correcting back. That time, I suppose winds would play an effect in the new calculated heading for the desired track, but I had no significant winds at the time, and prior to a year or so ago, those always pegged the new track dead on as well, irregardless of winds.

Steve, since the FD appears to be having difficulty (as opposed to the autopilot itself) it could be that Russ is onto something with the magnetometer. Hopefully there is a Garmin Aviation service center in the Vegas area that can investigate.
Good luck and keep us updated,
Wayne

BTW -- likely not related... I had the Red X HDG problem and it did indeed turn out to be the magnetometer. In aviation terms, not all that pricey to replace with new: parts ~US$750 plus a few hours labor.

Magnetometer otherwise seems to work great, no problems. It my be worth while though to try doing the recalibration process (described elsewhere in the forum). Thanks.