I discovered that quad, that is calibrated well to fly fine in auto mode, is prone to enter e high speed toilet bowl effect if , while flying with high ground-speed, sticks are released, and loiter mode engaged.
A small compass error, will then cause the multirotor to do a huge high-speed turn instead of “break” and return to the desired position.
This is clearly dangerous if you should find your multirotor being taken by a strong wind, and try to solve it by going into loiter mode.
Please test and tell me if you can confirm my findings.
1.- in stabilize, fly at high groundspeed
2.- release sticks and switch to loiter.
you would expect the machine to overshoot, stop and return to the point.
you may see it take huge high-speed turn, failing to return.
Please note, that at normal speeds,Loiter, 20 WP auto missions, RTL , autoland, and any other “normal” uses - this quad behaves perfectly.
Loiter projects a point in space in front of the copter based on the speed and direction that the copter is travelling at and how fast the vehicle can decelerate, and locks in on that point at the moment you release the sticks. It will then smoothly decelerate to that point in space without any overshoot.
AFAIK, that’s only so when you fly in loiter mode = “push the copter around”. Andre was writing about using loiter as “panic button” when he is about to lose control in STAB.
IIRC, the developers have warned against using loiter as some kind of “panic button” several times on DIYD. So you simply shouldn’t do it .
Yes, I understand that loiter depends on many things, and is not a “save-the-newbie” solution. But it should be reliable within the normal flight envelope.
For a polished AP, the goal is to do the job as good as possible in any scenario it can handle.
…Think if somebody is flying hard and fast, and battery, throttle or gs failsafe is trigged, and set to land - (afaik,Land holds GPS posistion it had when trigged, just like Loiter ) - that could result in a huge high speed turn, while reducing altitude…