Fall 2016 Mainsheet Article: ST4000 Mk I Autopilots

Previous topic - Next topic

0 Members and 2 Guests are viewing this topic.

Breakin Away

I just looked at the Fall 2016 Mainsheet issue, and enjoyed John Nixon's excellent article on the Raymarine ST4000 series autopilots. It came about a month too late for me, as I had already experienced an autopilot failure on my new (to me) Catalina 34 about 2 hours into my maiden voyage. It turned out to be a shredded/dryrotted drive belt, which was easily fixed as described here:

http://www.sailnet.com/forums/general-discussion-sailing-related/256778-need-help-diagnosing-raymarine-st4000-autopilot-post3575569.html#post3575569

I found a service manual online that helped to guide me through the repair, but John's article would have gotten me there faster, especially since it pointed me to the source for the aftermarket replacement belts (whom I found myself, but not until after much searching).

Although I am brand new to the black-ring Mk I version of the ST4000, my prior boat (C250WK) had the gray-ring Mk II version, so I've learned a number of things that apply to both. I'll share a few here:

  • Drive tension adjustment is critically important IMO. Too loose and the wheel will slip in heavy conditions, too tight and the clutch may fail to disengage. I believe that, for some people, this adjustment may be at the heart of the ratcheting noises that John describes (although the drive ring distortion that he describes is also an interesting observation the I'll have to look out for). The service manual describes a very detailed procedure for adjusting tension which involves measuring motor amperage under load, etc. I did not have the right electrical connectors for that, so instead I followed a different procedure described in the Sailnet link above.
  • A Google search revealed several complaints of clutches disengaging when the AP turns to port. I believe that root cause is the cam rollers (for clutch and also for tension adjustment) binding up and not rolling properly. Cleaning those rollers is very important, especially since they are at the bottom of the drive ring where organic matter can build up.
  • In his article (and several posts on this site and elsewhere), John and others describe the very common problem of autopilots oscillating port to starboard, and the oscillations getting gradually larger. Some knowledge of PID control theory can help diagnose this. Although I am not sure that the ST4000 is a full PID controller, the rudder gain adjustment is basically the "P" in PID, and general control theory says to use the minimum gain required to maintain a stable course. Too low and the boat will fall off in one direction; too high and the boat will experience growing oscillations in both directions like John described. I made a regular practice of adjusting the rudder gain to fit the sea state, and had set it up as one of the data pages on the control head to bring it up for quick adjustment by simply pushing the "Display" button. My personal experience (with my prior boat and with charter boats) is that high-power-draw settings that increase the AP's sensitivity and speed of adjustment can often make things worse. It's often better to reduce sensitivity so that the AP will ignore little disturbances and adjust for the big ones.
  • I also see a lot of issues with using the AP in following seas, just like many others. On my prior boat, I could never get the AP to work well in following seas. Once again, PID theory says that regular, slow period perturbations can be more difficult to any controller to overcome than random, quick ones. Most controllers do not know what is a random perturbation and what is a regularly repeating one. It just sees a change and adjusts, and when the perturbations happen slowly, it will get confused into thinking that it's a permanent change (like shift in wind or current) and attempt to follow it when it should just ignore it. Sometimes you just have to hand steer, because your brain can figure out which perturbations to ignore better than any AP can. It can also do a "feed forward loop" based on what your eyes see coming, whereas an AP can only do a "feedback loop" what it has already detected.

I have not yet tried the "Catalina cocktail" settings, but some of them look surprising to me. Do most of these C34 installs include a rudder position sensor? (I'm embarrassed to say that I don't know of my new boat has one - will check next time I go down there.) Some of the cocktail settings only have an effect if there is a rudder sensor.

2001 MkII Breakin' Away, #1535, TR/WK, M35BC, Mantus 35# (at Rock Hall Landing Marina)

Stu Jackson

#1
    Quote from: Breakin Away on July 22, 2016, 08:57:39 PM
    • I also see a lot of issues with using the AP in following seas, just like many others. On my prior boat, I could never get the AP to work well in following seas. Once again, PID theory says that regular, slow period perturbations can be more difficult to any controller to overcome than random, quick ones. Most controllers do not know what is a random perturbation and what is a regularly repeating one. It just sees a change and adjusts, and when the perturbations happen slowly, it will get confused into thinking that it's a permanent change (like shift in wind or current) and attempt to follow it when it should just ignore it. Sometimes you just have to hand steer, because your brain can figure out which perturbations to ignore better than any AP can. It can also do a "feed forward loop" based on what your eyes see coming, whereas an AP can only do a "feedback loop" what it has already detected.

    I have not yet tried the "Catalina cocktail" settings, but some of them look surprising to me. Do most of these C34 installs include a rudder position sensor? (I'm embarrassed to say that I don't know of my new boat has one - will check next time I go down there.) Some of the cocktail settings only have an effect if there is a rudder sensor.

    Great report, thanks.

    Following seas are always the bane of recreational boat autopilots unless one moves to the more robust below decks units.

    John contributed to this early FAQ:

    http://www.c34.org/faq-pages/faq-autohelmwacky.html

    IIRC, the Catalina Cocktail was developed well before separate rudder sensors were required by Raymarine.  The really old units had simple heading sensors, then they had them built in, then they made them without them and then required them to be added.  Go figure.  :D[/list]
    Stu Jackson, C34 IA Secretary, #224 1986, "Aquavite"  Cowichan Bay, BC  Maple Bay Marina  SR/FK, M25, Rocna 10 (22#) (NZ model)

    "There is no problem so great that it can't be solved."

    Breakin Away

    The cocktail settings that I found include a setting for rudder damping. But that has no effect unless you have a rudder sensor installed. Hence my question about how common this feature is on these boats.

    2001 MkII Breakin' Away, #1535, TR/WK, M35BC, Mantus 35# (at Rock Hall Landing Marina)