|
Post by tedp on Sept 10, 2013 12:39:05 GMT
Despite last year's repair of my tiller drive motor unit described on this forum, I'm still dissatisfied with the performance of my ST4000+ autopilot. The response of the system to wind and waves is too slow, and there is no reliable end-of-stroke shut off. The only time I can use it without constant attention is under engine, or in very light weather.
No amount of fiddling the settings of the control head will make it steer a steady course in windy conditions, or even in moderate conditions with the wind abeam. It will start to yaw around the course, often as much as 25 degrees each way. Response to a course change will come too late which compounds the yawing. In windy conditions the available control range (10 inches of thrust stroke) is too small - the end of stroke is reached in no time and the motor will heat up as the shut-off is ineffective.
Trimming the sails so they will spill a little when the boat luffs will help, but the result is unpredictable and you will lose performance.
My question is how much better the SPX range of tiller pilots will perform. Is course holding better (as Raymarine claims it is), and is the end-of-stroke shutdown better than that of the ST4000+?
Anyone prepared to share experience? Note that it's tiller pilots I'm interested in - a wheel pilot may perform better due to the greater admissible rudder angle.
|
|
|
Post by MartyB on Sept 12, 2013 4:52:47 GMT
I wonder if you could get a tiller option to connect to the rudder post below deck, so as you do not need to use the connection from the cockpit seat to the tiller! That would be my prefered way. Not sure if I could find one as of yet, with out going to one that is a min 20K disp boat. Pretty pricey and over built to a large degree.
Marty
|
|
|
Post by tedp on Sept 12, 2013 8:07:14 GMT
I could live with having to check the motor drive isn't buzzing against its end stops, or even install a rudder sensor, but what I don't like is the inadequate response of the unit to wind and waves.
I have tweaked every possible setting, and remember I know a bit about autopilots as we were taught about them in nautical college. It simply doesn't do the job, and from what I remember about the ST1000 I had on my old boat that wasn't much better with the wind abeam.
I wonder how much better the gyro sensor on the SPX-5 is before I invest €1100 in a brand new unit.
The under deck option you suggest, Marty, would mean the drive unit is connected continuously. There must be systems that offer that functionality (I think Flightdeck has installed something like that), but I fear it would add up to even more expense.
|
|
|
Post by MartyB on Sept 13, 2013 4:45:03 GMT
Not that I have seen one, nor is one close, but what does an SF3200 use? seems to me it is below deck attached. So that might be the first place to look. I am sure the unit in working order is more than a st2000/4000 or equal from any brand!
Marty
|
|
|
Post by tedp on Sept 13, 2013 10:12:47 GMT
It seems we are talking about two issues: drive motors and course keeping. drive motorsThere are below deck drives ('direct drive' units) available from several producers. Prices are far beyond a complete Raymarine tiller pilot - think of €1500 for the drive alone. The Danish firm of Jefa produces a complete direct drive pilot with compass and control panel for around €2000. These are highly thought of but I cannot write from experience. I think Flightdeck of this forum fitted one to his boat. course keepingThe longer I think of it, the more I become convinced the main control problem of my ST4000+ is inadequate rate-of-turn information given by a flux gate compass. A human helmsman notices a minor heading change way before a flux gate compass does. When I steer the boat myself, I give tiny helm corrections that catch any tendency to turn before they go out of hand. The ST4000+ just sits and waits until the boat goes, then starts an inadequate helm correction (too little, too late). Fiddling the rudder gain and rudder damping controls on the unit only results in over-compensating. As I said before, I tried every combination of autopilot settings conceivable and the one I have now is the best I can do. rate-of-turn dataSo what remains is feeding rate-of-turn data to the autopilot so it senses a heading change before the flux gate compass does. It seems that this is what Raymarine tries to do with their SPX-5, but no one seems to have reliable info of improvements over the ST4000+. I read of someone who replaced the Raymarine flux gate compass by an Airmar gyro compass (the H2183 network gyro) which uses NMEA2000 protocol to communicate with his SPX-5 tiller pilot. This suggests even the SPX-5 needs improvement. Price of the Airmar unit is $530 to $600. Someone else even managed to use a model helicopter flight attitude gyro and a home-brewed circuit to feed NMEA info to his autopilot. Obviously you need to be an electronic wizard to pull this off.
|
|
|
Post by MartyB on Sept 13, 2013 13:18:22 GMT
Now I think........OUCH........need to quit thinking.....hmmmmm.....too early in the morning on the left coast here of NA......anyway, I see the issue. You could be correct in that some compass styles are slow in working. Probably like my hiking days of an air vs water emulsed compass. The water style was way more accurate and did not move as much. If the two styles of compass's in the steering system sorta of operate the same, yes, it could be in that part that is not getting info to the computer, which is not getting info to the hydraulic rod in a quick enough fashion to truly keep you on course.
Off to work......
Marty
|
|
|
Post by rene460 on Sept 14, 2013 11:29:49 GMT
Hi Ted,
Another aspect of course keeping is the inherent course keeping stability (or lack of it) of the boat. I know you put a good main sheet traveller and achieved much better sail shape, but does the boat maintain its course well through gusts or does it require relatively large tiller movements to keep to your course? Upwind in reasonable conditions, can you let go of the tiller for a reasonable time?
Certainly beam seas or aft of beam are a severe test of any boat and the auto pilot will need to work pretty hard to maintain a course. Normally the helmsman will allow it to go with the flow a bit, not so easy to achieve on auto, and in bad conditions your 25 degrees sounds quite good. But if your sail cloth is a bit soft and the draft moves with wind strength you will do a lot of work manually controlling course with the tiller and the autopilot will have a much more difficult job if it copes at all. Very hard on the batteries if the tiller is being driven all the time.
Regarding the SPX series, I believe the question is whether they have a derivative or rate function in the control algorithm, in addition to the normal proportional plus integral. More tricky to tune, but without it I would suspect it will be no better than your existing unit and upgrade of sails may be more satisfying and productive. If a more powerful unit is all that is required, I doubt it matters which series.
rene460
|
|
|
Post by tedp on Sept 14, 2013 18:55:42 GMT
Thanks, rene460. The SO32 is not a very steady boat regarding course stability - the tiller needs to be watched rather like that of a dinghy. The old SO36 I sailed in Turkey last month is far more steady, but then it had the same sail area for a bigger boat. Despite the sensitive steering, I find manual steering on the SO32 is quite positive, even in gusts, and it is far better than when I use the autopilot. So the rig and the sails are good enough as they are. The course instability may amplify the course holding problems of the autopilot, but I think it should be far better than it is. The main observation is that small heading changes are picked up far too late by the pilot controller. When I see the boat's heading change, the pilot should intervene immediately, but it picks up far too slowly. These are the basic control settings of the ST4000+ (apart from several others concerned with driving the rudder motor, magnetic compass error etc.). - rudder gain - proportional control of the rudder angle given a course offset
- response - sets a 'dead course offset' inside which the pilot won't intervene. This can only be set on and off.
- turn rate limit - sets a maximum rate-of-turn, i.e. this is the differentiator.
- auto trim - reduces the course offset of proportional control, i.e. this is the integrator
For positive control however the pilot needs sensitive input from the compass which I think isn't good enough. Raymarine claim the 'gyro' of the SPX-5 (or possibly this is a piezo-electric inertia sensor) takes care of this. I have a spare flux gate compass - I might try and substitute that first to see how it performs.
|
|
|
Post by rene460 on Sept 15, 2013 10:24:35 GMT
Hi Ted,
Changing the flux gate unit should at least eliminate any faults with the unit.
It is a pity that you can't borrow a new series control unit for a trial.
That gyro unit will be the key to whether the new series will perform better for you. It would be interesting to know what it really does in the control algorithm especially in the light of the boat characteristics. In principle, it should have more possibilities than a turn rate limiter.
rene460
|
|
|
Post by tedp on Oct 8, 2013 12:15:28 GMT
During a short trip in fine autumn weather I checked several properties of my ST4000+. Wave height during the trip didn't exceed a few inches due to the wind being offshore. The wind was 10-12 knots slowly increasing and gusting to 16. I sailed on a reach with full hoist.
First, I checked on flux gate compass readout response for tiny course changes when steering by hand. There proves to be nothing wrong with the compass, at least not when sailing in a calm sea. The readout changed very promptly, about as quickly as I saw the heading change by eye. So I think there is no gain in a gyro sensor.
So what remained was testing the algorithm response of the computing unit. I sailed on a reach, steering 330 degrees out and 150 degrees on the way back. Provided the sails were trimmed well, the autopilot kept the course within about 10 degrees of the setting, even in gusts. The setting for rudder gain was 5 (maximum is 9) and the response setting was 1 (leaving some dead band around the course). Increasing the rudder gain to 6 had no significant effect, 7 or more resulted in a corkscrew course. Changing the response to 2 (no dead band) increased rudder activity and seemed to improve course keeping marginally.
The rudder drive at no time ran to its end stops, which I think was due to the flat sea conditions.
Obviously the system works well in moderate conditions. The problems occur in a rough sea and strong gusting wind, when the boat is pushed off course much more, or possibly the compass sensor may start to swing of its own in its gimbals. This will have to be tested next, although I fear the advanced season may need it to be postponed till next year.
|
|