A wonderful MP feature is the tracker (MP->SETUP->Optional Hardware->Antenna Tracker): knowing vehicle and tracker GPS positions, send on a COM port pulse widths for pan/tilt servos, and have pan/tilt angles shown as AZtoMAV/ELtoMAV.
But AZtoMAV/ELtoMAV start late and while flying (copter) or running (rover), which implies:
- Tracker preparation has to be done blindly.
- Gimbal servos move abruptly, as can be seen here (real thing) and here and here (SITL rover simulations).
Even the camera used for the real thing detects fast panning:
This stresses the servos.
Knowing MP the vehicle position (after GPS lock) and the tracker position (home or being set manually or with a GPS on the laptop (which BTW I couldn’t make it work)) it seems (unless I miss something) it has all the information to output servo pulse widths, and fully prepare the gimbal. In addition, placing the camera on the gimbal after pulse widths appear would avoid stressing the servos.
Why do servo pulse widths appear so late? Seems long after movement starts.
At the end of this other copter SITL simulation AZtoMAV/ELtoMAV go to 0 abruptly, possibly around landing:
In the real case it would also cause servo stress, so angles should be maintained.
BTW1, these AZtoMAV/ELtoMAV discontinuities on the SITL rover simulations mentioned above are mysterious:
This was a capture with the TrackerLocation.Alt workaround mentioned applied, so may be it is incorrect or not innocuous.
BTW2, for the rover case with tracker on podium (higher than rover), which needs the workaround, it would be useful to inform MP of tracker altitude relative to rover, and not need that workaround.