2020-07-30, 09:00 AM
I've gotten it to work yesterday, following the built in route capabilities of OpenCPN to test the basic functionality. Calibration of magnetic compass could be better (my fault, some interference), but following a GPS route worked. It followed the route "on the spot" and did not wonder off of it's track. I am very happy it finally worked. The configuration I posted earlier is working.
When testing on land it is honestly very hard to say if it will work in real life. Interpreting the GUI and how course is handled by the autopilot, when it chooses it's waypoints is a bit tricky. Sometimes it will point directly to the first waypoint, but I have found out that in real life when the waypoint is a bit far, it will first steer for the track rather than directly to the waypoint. Also, a GPS simulator does not take into account a magnetic compass, so this will be another factor when testing on the water.
Probably common knowledge for others, but I think for people who are trying to set this up for the first time need to know this:
- Test on land, or with a gps simulator if the autopilot at least will react when activating the autopilot (look at the blinking waypoints and if using GUI overlay of pypilot, look if the green line goes to the "general direction" of the heading it should go, does not need to be the waypoint directly.
- If using pypilot, see if automatically switches to GPS
This should be enough for you to know that it will probably will try and follow the route and that it will work.
Then, if the following issues are there:
- Pypilot is holding magnetic heading correctly, but does not react to course changes when pressing button: change the opencpn connection priority
- If the tiller is reacting to quick and oversteers all the time, just decrease the P and D gains bit by bit until you hit the sweet spot.
Only if you get it working, you get a general idea of how opencpn communicates heading changes so it will follow a route and how direct / aggressive this is.
Next step, test Sean's autopilot route plugin.
When testing on land it is honestly very hard to say if it will work in real life. Interpreting the GUI and how course is handled by the autopilot, when it chooses it's waypoints is a bit tricky. Sometimes it will point directly to the first waypoint, but I have found out that in real life when the waypoint is a bit far, it will first steer for the track rather than directly to the waypoint. Also, a GPS simulator does not take into account a magnetic compass, so this will be another factor when testing on the water.
Probably common knowledge for others, but I think for people who are trying to set this up for the first time need to know this:
- Test on land, or with a gps simulator if the autopilot at least will react when activating the autopilot (look at the blinking waypoints and if using GUI overlay of pypilot, look if the green line goes to the "general direction" of the heading it should go, does not need to be the waypoint directly.
- If using pypilot, see if automatically switches to GPS
This should be enough for you to know that it will probably will try and follow the route and that it will work.
Then, if the following issues are there:
- Pypilot is holding magnetic heading correctly, but does not react to course changes when pressing button: change the opencpn connection priority
- If the tiller is reacting to quick and oversteers all the time, just decrease the P and D gains bit by bit until you hit the sweet spot.
Only if you get it working, you get a general idea of how opencpn communicates heading changes so it will follow a route and how direct / aggressive this is.
Next step, test Sean's autopilot route plugin.