Odd Autopilot Behavior

[NorthernFocus]

We installed a Raymarine EVO AP last year. We put over 200 hours on it last season and it worked flawlesly. And continues to do so this season... for the most part. But every once in a while it does something really weird. I haven't noticed a common set of conditions when it occurs other than most of the time it occurs within five minutes of switching to AP control. But it has happened once or twice after the AP has been in control for an hour or so. The symptoms are as follow:

- AP operating in "course keeping" mode (i.e. not integrated with GPS so no route following)
- boat drifts off course and AP starts to correct as indicated by display
- rudder turns in opposite direction of indicated course correction
- AP continues to attempt to correct with harder rudder command and rudder goes harder over in opposite direction
- this behavior continues until AP command indicates hard over in one direction and actual rudder goes hard over in opposite direction
- during this event compass heading appears accurate and AP appears to be trying to do what it should be doing but the rudder is responding in reverse of the AP command

This is really a weird situation when it occurs. It seems to be a hydraulic problem rather than electronics. But I can't imagine what could make it behave this way. Haven't had any issues with manual steering and even during this event the manual steering will function normally if you turn the wheel.

Any ideas/suggestions/etc that anyone can offer would be greatly appreciated.

 

[knotflying]

Well, you may want to call Raymarine support and see what they say. I am definitely not an autopilot guru, but could the hydraulic lines be reversed?

 

[Brunski]

Does it always happen in one direction? That is, always to port or always to starboard? Did you update the AP software at any point between last year and this year? Does the AP actuate the wheel using a mechanical servo, or does it use a hydraulic pump and a separate ram directly connected to the tiller, or does the AP hydraulic pump sit in between the wheel pump and the tiller ram? Was the steering system serviced over the winter? Are there any hydraulic leaks in evidence? Is the hydraulic fluid reservoir full?

It would be very strange for the hydraulic portion of a system to function normally most of the time, then occasionally reverse. Could be a pump issue, or maybe something is going "over center" mechanically somewhere in the linkage to the tiller, or some wiring open/short/corrosion between the AP computer and the pump/servo is the culprit.

Assuming no software updates and no hydraulic leaks or servicing, I would check the mechanics of the system to be sure nothing is loose or broken and that the geometry is OK. Then I would check that all wiring is good, connectors clean and corrosion free. If all looks OK then I would start suspecting the pump or mechanical actuator.

Sorry to be so vague, but without specifics it's really hard to be of any real help...

 

[NorthernFocus]

Does it always happen in one direction? That is, always to port or always to starboard?

Goes both ways.

Did you update the AP software at any point between last year and this year?

No updates.

Does the AP actuate the wheel using a mechanical servo, or does it use a hydraulic pump and a separate ram directly connected to the tiller, or does the AP hydraulic pump sit in between the wheel pump and the tiller ram? Was the steering system serviced over the winter? Are there any hydraulic leaks in evidence? Is the hydraulic fluid reservoir full?

Hydraulically actuated via the pre-existing actuator. Lines off of the pump tee into the lines from the wheel. No leaks, reservoir is full.

It would be very strange for the hydraulic portion of a system to function normally most of the time, then occasionally reverse. Could be a pump issue, or maybe something is going "over center" mechanically somewhere in the linkage to the tiller, or some wiring open/short/corrosion between the AP computer and the pump/servo is the culprit.
...

My thoughts are that there is maybe an internal check valve sticking in the pump. But even that is odd because it has to be something that is failing in both directions. The wiring has only been in place for a year and all control signal connections are in the cabin and not subject to corrosive environment.

However... your mentioning the wiring reminded me of something. There is a way to correct backward wiring via programming the controller to operate in reverse. So the function does exist within the electronics to make the actuator move opposite of the direction indicated on the controller UI. It's still a mystery why it only occurs ocassionally
but it's a straw to grasp at. I think I'll do a factory reset of the system and set it back up from scratch. Not all that hard to do compared to previous generation systems. The compass auto calibrates so no need to swing it etc.

Thanks for taking the time to reply. Often times talking through things with a third party sparks a thought even if indirectly. Will give the reset a try. Unfortunately we won't be back on the water for 10 days or so and will have to wait till then to see if it works.

 

[Brunski]

My pleasure, hope the reset solves your issue!

 

[NorthernFocus]

Well I think I figured this one out.

These new, whizbang APs are so smart that they auto-tune/calibrate. That's great in that it makes for easy set up and gets you on the water with it quickly and without the need for service/support. But as I've now discovered there is a down side.

This past week the AP worked flawlessly for three and a half days. On day four it had another episode of "confusion". However this time I recognized what appears to have been the initiating event. We were cruising along on AP control when I decided to stop and better secure the dingy. I forgot to switch the AP out of auto prior to stopping the boat. While I was aft securing things the AP off-course alarm sounded(duh). When we got back underway and I switched the AP back into auto it had one of its episodes, indicating turning in one direction but rudder going opposite. I recalled that last time when I was having this issue we had been fishing and I'd done the same thing, stopped the boat and forgot to switch out of AP control. So armed now with the knowledge that there is an electronic function in the AP to reverse rudder command/response it occurred to me that when I stopped the boat the AP must have "recalibrated" itself by reversing rudder direction commands when it was unable to bring the boat back on course. As if it assumed the rudder control unit was connected backward.

Hmmm. Well if the problem is in fact caused by the AP recalibrating itself during that brief period of drifting then it should be self correcting if given a chance. To test my theory I switched back into AP control and just let it go. Sure enough rudder command indicated more and more correction to starboard while the rudder went over harder and harder to port. But after swinging through an arc of 120 degrees or so the AP figured it out and reversed itself again and came back under control. I watched it carefully for the next few minutes and it showed a clear bias in one direction for five minutes or so then was right as rain.

Amazing stuff those ones and zeroes....