I love this topic! I would like to build a wind vane/wind speed transducer to use to track performance of small (but human sized) sailing boats. I plan to mount the vane/anemometer on a bow sprit elevated off the bow (front) of the boat for easy install/removal since typical mast head instruments are exposed to the elements 24/7.
Question 1: I have an old APM 2.5 from my drone days, will that board be sufficient for this, or should I look into a pixhawk or other?
APM is too old to run the current code, there are loads of good hardware options thanks to ChibiOS. They vary in price, quality and features.
Regarding the Wind Vane. Reading above, and the Wiki I want to understand this correctly, as it may influence the flight controller I choose.
-A challenge is the Hall sensor outputs up to 6 volts, but that some of the flight controller Analogue to Digital Converter (ADC) only accept up to 3.3v input. This requires a voltage divider out the on the Hall sensor output to bring the output of the sensor, below the max input of the ADC. I am clear there.
The pixhawk (for example) has a 6.6v ADC input. Are people using other boards that donāt have a 6.6v ADC and only 3.3v ADC inputs? or is it that the 6.6v ADC pin is being used by other sensors? Of is there some other reason that I am missing? Iām not afraid to solder a voltage divider, I just want to understand fully.
edit: Links to other comments about this
The 6.6v ADC is just a 3.3v one with a voltage divider on the pcb on the pixhawk. On all other boards you will need to solder your own voltage divider. The other consideration is the number of ADC inputs. On smaller boards you only get one, often labeled as RSSI. This means you cant use an analogue airspeed sensor or the rev P hot wire sensor for wind speed. This basically leaves you with the only option of a cups type wind sensor using the RPM lib.
Because this will be on larger vessels (minimum 3m and more likely 6m to 8m), size of a flight controller is a non issue. It seems to me the regular pixhawk may be the best and most cost effective option as it has a 6.6v ADC and then two more 3.3V ADC slots. Unless there is a stronger recommendation from anyone here.
I will start with the rev-p hot wire sensor (it said backordered but they already shipped it!) since the cup anemometers are out of stock.
On big and I assume expensive craft i would go for a CUBE. But you say to track the performance? If ArduPilot is not driving then it doesnāt matter if it fails and a standard pixhawk or anything else will be fine.
Yes, just to track boat speed vs wind speed, heading and apparent wind angle. All the steering and sail controls will be done via real life FPVā¦eg, people on the boat 
And this is purely for amusement. Basically Mission Planner will be attached to the flight controller via Wifi telemetry or just a USB cord, making a tablet or small netbook into a chartplotter/data logger.
-P
Second sailing session today 
I spend some 3 hours fiddling with settings before sorted most out and could get some time in the water before the wind totally died out after 6 pm.
No video this time. These are smaller things and likely I will figure them out but comments are welcome.
I noticed often when starting up the boat after some days the flightcontroller is reset to default setting and have forgotten all my settings. I really do not know what causes this but will pay more attention.
Then I try loading settings file from disk that I saved last time all was fine. However, that brings most parameters back to what I set but not all. One parameter I specially notice is not set is the WINDVN_SPEED_TYPE, this is usually back to 0, but it should be 3 RPM library. I am pretty sure I have saved it many times to disk but it will be reset to 0 when loaded.
Many time the Wind direction is bouncing around, is off and needs recalibration when I start rigging the boat. Today I noticed that unless I also assemble the wind speed cup instrument, the wind direction will not show a stable value. I think its because these two instruments share a ground wire, so I could make one less wire from the rigg all the way to under deck where the flight controller is located. When both are assembled, both works okay. This lesson will save lots time in future.
I checked and confirmed that in Acro mode, flightcontroller compensation is correct if a sudden shift to the right boat tries to give left rudder. Also sheeting is correct when wind direction changes, as is the sheeting out when keeling is over limits. In Manual mode when I control the rudder directly, it steers correctly to right and left. BUT the odd thing, when in Acro mode the rudder is opposite, so if I give go right rudder, the boat will turn left. A bit unexpected
I need to correct this, but keep the Manual rudder behavior as is and also keep the compensation from flight controller when in Acro mode.
Thanks
#4. try reversing the steering in your transmitter and RC in for the same channel.
There was a issue on master that would cause this, it has now been fixed, so to get rid of it you can load the latest firmware. That issue has been fixed but there maybe some new ones.
This is due to setting enable parameters. So it has to set WINDVN_TYPE to something other than zero and reload the parameters before it will see the rest of the WINDVN_ params. The trick is to load from disk, reload parameters and load in from disk again. Then reboot and use the compare tool to double check everything has been set.
As David says for number 4 it is mostly likely transmitter and RC output reversed. You can check the RC inputs are correct in the mission planner RC input tab, all the green bars should move the same way as the sticks (except pitch but you donāt need that on rover)
All items solved
Thank you both.
I am pleased to report I have been given a GSoC project this year to carry on my work on sailboats. So hopefully there will be some improvements coming this summer. More info here:
So if you have any ideas or feature requests, now is the time, I will add them to my list.
That is so great Peter!
At the moment the sailboat do not support Smart RTL Mode right?
I am sailing in to small bays and if RTL is entered as failsafe it will be bad to going home just using straight lineā¦
I think Smart RTL should work although I havenāt tested it, i will test soon and report back.
I did an Autonomous mission around a pond 12 or 13 times IIRC. Total distance was like 6 KM according to mission planner. Only problem I had was failed tacks due to wind shifts. I also had a max cross track set at 10 meters so I would not run into shore. Not bad for the early sailboat code.
I have been thinking about how to best deal with this, I think a turning rate to boat speed at start of tack parameter. So if your going slowly it will turn slowly. Maybe also disable the turning integrator. And for bearing away I think the solution is to set the sail for the target heading rather than the current one. So it would let the sail out first then turn rather than trying to do it all at once, maybe have a wind speed threshold where it does this, no point in light wind.
setting the sail to the target heading first is smart, it will reduce weather helm and make the steering easier. For boats with two sails, differential sheeting between the forward and aft sail can help bring the boat around for tacking and jybing. In sailing school they would do an exercise on light wind days where the instructors fix the rudder in the middle, and then students have to use sail and boat heel only to drive. Frustrating and fun day.
Iām tempted to get a yacht and follow along: a couple of decades ago, as a young midshipman, I learned āblind pilotageā on a twin masted Ketch. Two masts, a chart, a compass, an echo sounder, and a speed log. And a curtain across the cockpit/bridge.
A challenging, but achievable, navigation exercise for a 19 yr old. It would be great to see ArduPilot do the same.
(caveat: these days Iām an Aerospace Engineer. Itās been a long timeā¦)
Peter, for the point #1, the support for commercially available wind instruments that can be installed at mast top, I have two thoughts (you probably got them covered):
. some of these wind instruments also have a compass included, it would be great to get access to that data in a plug and play way, since in a limited water proof box on board, there is so many items disturbing the standard compass in the flight controller.
. When implementing the (true) wind speed usage,
(currently it is mostly used to shut down trust of wind direction data when speed is too low? but for using Polars it will have much more impact on auto sail operations),
just consider that a boat can move faster than the wind speed, especially for an ice sailer that normally travels at 3-4 times the true wind speed.
I hope to build an ice sailer and use the same fine software
The only problem with commercial wind instruments is the weight. They will be good for a custom built hull ballasted properly. Even my EC 12 behaves differently with just and extra 3 oz. of equipment up on the mast. I do love the idea of hardsails with an elevator and the ability to motorsail.