Gaston - There are rudder position sensors / ru...
There are rudder position sensors / rudder angle transducers from Raymarine and other brands. Most of them are based upon a potentiometer. You have to install them next to your rudder connecting them with a stud. Sometimes there isn't much space to mount it, and when your rudder is hanging on your transom it's probably fragile.
Did anyone ever think of mounting a (3 axis) compass sensor on the rudder and, calculate the rudder angle by comparing it to another compass sensor mounted on the boat? These sensors aren't expensive, small, easy to mount, and there are no moving parts that can get stuck. Or is the potentiometer still more reliable and accurate?
4 Replies
based on observation of my rudder angle indicator (simrad transducer) .. using a couple of dec places in signalk to see smaller variations,
i can definitely move my rudder in increments of 0.5 degree or less… and the reading is very stable .. when rudder set to a particular angle.
i suspect that “heading” sensor using magnetic compass or similar … will be a lot less accurate and stable … given that even very expensive heading sensors
still only have accuracies of 1 degree ? or so.
the autopilot will do weird stuff if it sees the rudder angle jump around ….. eg for my (planing hull) boat, 1 degree of rudder produces a very distinct turn!
On a planing hull rudder changes for sure have a much bigger impact compared to slow sailing vessels.
I do agree that heading accuracy is not much better than 1° for cheap chips (which I think means that bad readings still are within +/-1°, so in general readings are much more accurate). And in my suggestion you need two of them, and that might mean that accuracy could become 2° (but again, that's when both sensors have the maximum inaccurate reading in opposite direction). I don't know what accuracy is needed, because frankly I don't even know how the rudder position is used by the autopilot. Autopilots also work without rudder angle sensors. Maybe the software is self learning to make it sail smoother using the rudder angle and other variables.
Then a rudder angle sensor based upon a potentiometer might have a very good precision, but I am not sure if the sensor knows the real angle, because that would mean that you have to calibrate it using an angle scale.
for certain the APs have a bunch of algorithms to cater to various situations.
i recentky had a problem where my rudder “center position “ was out by a small ammount.
that is, when AP was being told it was 0, it actually had around 1deg of left rudder.
result was the AP would input rudder adjustment to correct the course, but when it returns ruderr to (what it thinks) neutral,
the boat is still tracking wrong…. so it was a constant fight.. with lots of rudder movement.
once i realised this and reset its neutral rudder calibration… things worked way better.
.. hence i know from recent experience… what happens if actual vs “reported” position are incorrect…
… i had some really ugly behaviour with a big following sea, and the boat surfing (badly ) … with AP not working correct…
I asked around and for what I have understood about the use of the rudder position sensor:
* end stops (most important function)
* centering the rudder (but I don't know if that's a pilot function or just a handy function)
* to set the absolute position of the rudder. If the absolute position can be set, you can perhaps improve algorithms if you can predict the result of a certain rudder angle in specific cases (as you said). That should be a machine learning function, because it completely depends on boat/rudder characteristics and environmental conditions. I am wondering if these algorithms already exist (in particular for sailing vessels).
And a follow up question would be what accuracy the feedback sensor should be for the algorithms to deliver proper results.