Auto-pilot clutch slips under load to portside

My old Raymarine auto pilot had Alzheimer's in that it would suddenly let go with 'no drive' warning, I took the motor off to test the clutch, all good and yes its either engaged with pins or its not, so it can't slip. Thinking that it may be an issue between the fluxgate and the compass within the unit I tested the fluxgate. Deciding I didn't really know whether my findings would tell me anything I could understand and having upgraded the E series plotters to Axioms I opted to replace the auto too.  The ACU 200 was fitted and off I went only for it to fishtail the boat until we nearly gybed. I undertook several tests and once again it opted to, under engine (little wind), follow a search pattern without being asked to.

It was a struggle to get Raymarine to accept anything other than a limpware (user) issue, but eventually they sent over an Engineer who came and inspected with the Engineer who fitted it. When the auto pilot was asked to steer +10degrees and the helm was held to simulate weather helm, they were surprised that 15amps was called for by the Jeffa motor, bearing in mind it is protected with a 40amp fuse. They've since changed the unit for an ACU 400 which I assume handles the amperage better and once installed and the boats back in the water we'll be testing again.

Check the rudder sensor, probably by disconnecting it and see if it's the same both sides or cures the problem. You could have a similar issue in that the amperage being asked by the motor exceeds that of the unit fitted.