AP heading setting changes radically, why?

[danstanford]

Sorry for another question as the answer is likely to replace the unit! ST6002 Raymarine wheel pilot, rudder sensor accurate always.
So, heading on Autopilot screen is say 70 degrees off accurate so I go through the calibration steps and set it to match the real heading from the Chartplotter. Next day the heading is off again by a similar factor and the unit swings wildly to a new heading after pushing the auto button. It never occurred to me to let it go to see if it would work properly once it got to the heading it thought I was on...duh. Tied to the dock the 10 degree change button worked properly both ways.
So the question is, where is this setting stored and why would it change? Could it be the flux gate compass?


Dan

[danstanford]

Is this likely to be in the Autopilot unit itself, or the flux gate compass? Any input welcome

[transmitterdan]

Here are the most likely problems in no certain order:

Some electrical device (inverter, alternator, battery charger, fridge, etc.) is interfering with the flux gate compass. When the interfering device turns on or off this moves the magnetic field and the flux gate changes accordingly. This can also happen to a regular card compass. We had a boat that when bilge blower turned on the compass rotated 10 degrees. Try turning off everything except the AP and instruments.

The wiring from the flux gate compass to the computer is corroded or loose.

The flux gate compass is sticky. Sometimes they get that way after many years.

The calibration data in the AP computer is getting lost after power is turned off. Depending on age the nonvolatile memory may be going bad.

When you "calibrated" did you do the long 360 degree turns to accomplish the auto calibration? It's not enough just to enter the compass offset. You have to do the lazy 360 turns as described in the AP manual. If the engine alternator wires to the battery run near the flux gate that can prevent a good calibration.

[danstanford]

TDan, thanks so much! This is the first information I have been able to correlate to my understanding of the systems and the issues. I did just enter the compass offset because despite trying 3 times, I could never get the calibration to work. Inevitably, no matter how big a circle I made, the too fast response would come back. I think this means the turn was too sharp because I was literally running at the lowest possible speed even going into neutral when I turned down wind.
The boat is now out for the winter so I cannot test these things but I feel a ton more confident that the problem lies in the compass or related wiring than the computer.
Still understanding that it is a risk, I will buy the other instruments I want plus a new foresail and leave the autopilot for some other time. The sticky compass fits the circumstances perfectly. The compass is only a couple of feet from the instrument cluster in the cockpit coaming but I think it has all been there since new in 2007.


Any input welcome, Dan