Thanks for looking! Can you tell me what you look for in the bin file, or what Mission Planner is flagging? When the message came up there were tons of satellites and hdop was small.
Here’s a video from that day of sailing
Its a mission planner warning, based on the sensor status bit-masks I would guess. You could look in the T-Log.
This is where they are set.
1 Like
Not sure if this is the right place to post - maybe we should have a separate sailboat category.
So, I am almost ready to take my plunge into an autonomous sailboat, but just have a couple of questions;
My plan it to have the wind speed and direction sensor on the front of the wing sail to balance the wing sail elevator. My question is how the wind direction is measured relative to heading if that sensor is rotating with the wing. Am I better off installing a compass in the wing, doing the math, and just sending the wind direction via NMEA to the Pixhawk?
We don’t currently support such offsets. But it would be nice too. How are you planning to measure the relative position of the wing?
The quick approach would be to read some sensor in scripting and use its reading to change the offset angle param from scripting.
I have thought it would be nice to support windvanes on AP_Periph, you could then wind speed and direction along with the heading from a CAN node, you could also connect your elevator servo to it. The we would only need to run power and CAN between the boat and wing.
1 Like
thanks for getting back to me.
I was planning to have an absolute rotary encoder in the mast step or maybe even a compass in the sail.
The whole miniature wind direction think appears to be difficult. I’ve read the thread about wind vanes - no definitive answers.
I’m building a test wing and elevator this week to test my theories.
Again, would be great to have a sailboat section.
I am also interested in what Joe is looking for, to balance the weigh of the wing sail elevator with some useful ballast. With our current limitation the best I could do was to place a motor and LiPo there. Not ideal but was a good stepping stone.
Your idea sounds very interesting, can you expand on it a bit more please @iampete
AP_Periph is a ardupilot CAN peripheral firmware.
https://ardupilot.org/dev/docs/ap-peripheral-landing-page.html
This is shipped on several off the shelf CAN based GPS units, but can also be flashed to any AP flight controller. It then reports back its sensor readings, and outputs to servos (possibly that is not in a release yet) over UAVCAN. If we were to add a UAVCAN message for windvanes we could connect them to the AP Periph node. We could then have the GPS, windvane and servo out all done from one controller mounted in the wing, then run the CAN wiring down to the main flight controller on the vehicle, much easier than having to wire each individually.
I have been thinking about this a little, if the wing is fully rotating it will always be pointed roughly into the wind. This would allow us the sense the wind speed with a pito-tube airspeed sensor. We might be able to also workout the wind direction from the angle of the wing and elevator servo position. If not we can get by a AoA style sensor that need only do ±45 deg rather than the full 360 needed for a fixed wind vane. TLDR: a all rotating wing with elevator is basically a big wind vane.
could that controller mounted in the wing also connect directly to the antenna you think? (rc control and telemetry)
Since it’s good to have the antenna mounted in the mast top it would be nice not having to bring that cable under deck as well.
maybe just turn things around and put the peripheral unit under deck to just handle rudder servo and the controller in the wing do all the rest, gps, elevator servo, wind vane, compass, RC and telemetry …
RC and MAVLink telemetry are two things missing from the UAVCAN spec unfortunately. It also cant do IMU’s so the main flight controller still has to be mounted on the hull.
Hopefully all of that will be possible in the future.
1 Like
The Burns analog wind vane potentiometer seems to be obsolete. I found one that appears good in my initial tests.
Digikey Part Number: 987-1379-ND
I am designing a 3D printable wind vane for it if anyone is interested. It is based off of this paper:
https://hal.archives-ouvertes.fr/hal-02141286/document
New design on the printer has a balanced center of mass. Hopefully all goes well!
2 Likes
Wow what an interesting paper! Thanks for the reference.
I really like the P3022 sensor I’m using on my boat, it has very low friction, a nice bearing, and magnetic position sensing with analog output. Check it out here.
I have built a sailboat with a wing using the following link. However, when in Acro mode, the elevator trim does not move in response to the direction of the weather. The FC (F405 wing) is armed and I have tried moving the throttle to get the ACRO mode to work, but the problem persists. I have selected “128 - Wing Sail Elevator” as the output to which the elevator servo is connected, and in manual mode, I can properly trim the elevator.
I’ve attached the param file.
I need your help.
Hard to tell without knowing what the physical boat looks like and how the servos are connected. I would guess you need to reverse the servo output for elevator.
Thank you for your responce.
I took a video. As can be seen in the video, when in Acro mode, the elevator trim does not move in response to the direction of the windvane.
The servo for elevator is connected to servo CH4
I see you have the motoring switch setup on RC 8, since the motor is going I guess it is high. With RC 8 low it should only use the sail and disable the motor.
I turned the motor assist switch to Low and tried again. However, the sail only works in manual mode.
When I try to move the weather vane from port to starboard and vice versa in Acro mode (ARMed), I don’t see any elevator trim movement.
After this, I disabled the RC8 motor assist, but the situation is the same. I am currently using Ardurover FW 4.2 (for F405 wing).
Do you have any good ideas for the solution?
Thanks again.
I attached latest param file.
A log would be more useful that just a param file.
I’m not sure what is going on, I’m not aware of any changes that would have broken this functionality.
I hadn’t used my boat in a while, so today I sent N4N1 on a course that circumnavigated Saildrone Surveyor (on her mooring). I’ll edit the video soon. Worked well enough to go twice around but not a great driver in the gusts (on a path with lots of reaching) – I think I will try enabling the heel control PID. The integrator on rudder eventually gets it but I had some big transient wiggles.
7 Likes
Nice! Really gives a idea of the scale.
Can we claim: Ardupilot sails rings around Saildrone ?
5 Likes