Before starting a Auto mission I ensured that the following stuff works
Calibrated IMU and compass, ensured that the facing direction in APM Planner is the actual direction
Tested the copter in stabilize mode
Tested the copter in loiter mode
Ensured that GPS has enough satellites (11 on startup)
Ensured that throttle is around 50% for hovering
With my navio / APM setup I experienced the following problem today when doing the first auto testflight:
I did a small mission with a takeoff commando letting the copter ascent to 10 meters, 3 waypoints and a Return to launch command at the end. The first time it worked quite well EXPECT it definitely climbed not 10 meters but less (around 3-5m).
The second time I launched the same mission the the copter immediately flipped forward onto it’s back and killed all it’s propellers. Since the first waypoint was in front of the copter it looked like it immetiately wanted to “fly” to the first waypoint even when standing being on ground.
Since all the other modes worked well the “typical” flip on start problems like wrong probs or wrong motor order can’t be the problem. I’m new to ArduPilot, but have at least some experience with multicopters in general (MultiWii based).
Does anybody know about this behavior and know how to resolve it (before I kill the next 4 propellers)? Are 10 meters too less for a “safe” height? Since it did not climb to the expected altitude I’m suspecting the barometer. Is there something I missed about it? I already put some soft foam onto it to prevent wind and sun from disturbing it. Or can something else be the problem?
Here are some additional infos for the quad copter (yes it’s quite big…):
DIY frame made from aluminium rods and glass fiber reinforced plastic for center plates
[EDIT] Another Info: Before buliding this custom copter I tested my setup with a small warthox frame (8" propellers, 1100 KV motors). Here I experieced exactly the same problem. Since the small frame was a bit underpowered for lifting all the stuff (2 accus, raspi, navio…) I deceided to build this custom frame.
It sounds like its in your mission setup / PIDS me.
10m is plenty high, if you airframe is well balanced and pids are dialed you should be able to lock your unit into althld a foot off the ground and have it keep its position.
Hard to tell without knowing these things:
commands you are giving
auto/waypoint nav PIDS
default way-point diameter
default altitude
etc
To eliminate variables and to tune PIDS I like to drop the auto/waypoint nav PIDS to 1 MPS and do controlled low level (1-4 meter) flights with small movements up down left right. Always try to switch into auto mode from the air first till you have everything dialed then add in auto take off and landing later.
Thanks alot for your help,
So it seems that I have to spend some effort at least into PID tuning. I thought I could leave them default since it flew quite stable in stabilize mode…
The other values you mentioned are definitely also the default “out of the box” values. I’ll make some screenshot today from my mission. and try to export my parameters.
Before playing with the PID, I would advise you to check how much vibrations you have on each axis.
Another common problem is unwinding of the propellers.
Could be that one became lose and the copter flipped suddenly.
I 2nd this, vibration can have a huge effect, I had an air frame that had a mid throttle at 740, after I balanced motors and props it dropped to 680 without reducing any weight, just a good balance. It was just that much more efficient.
If you ask me solid flight in stabilize doesn’t mean a thing, beyond your props are facing the correct direction
Yes you definitely need to PID tune, but as dgrat mentioned, if you haven’t already, a balance will clean things up a lot and help you get the right pids. Balance 1st, pids 2nd, then reduce your way point radius to 1m and way point navigation to 1mps, then bring them back up slowly till you have the desired performance. Setting a fixed yaw orientation for auto missions helps too. When the copter tries to orient it’s self towards the next way point it can add more inconsistencies.
With balancing the frame you mean that all arms are exactly the same weight and the frame is in balance when it is hinged in the center?
Then I’ll check my vibrations, balance my props and then i’ll retry if it behaves better. How do you measure vibrations? With on board sensors and the apm logs?
