Flew my hex x a few days ago and it flew great. Over 50 mph and very stable. Pixhawk came out of an Iris+. DJI e310 kit.

The following day I was getting compass errors. Calibrated it several times and it was okay. Took off and it was very shakey. Then it just went up and would not come down. The battery was low so it only got to 300 feet all before it came down.

I reconfigured it as a quad because two motors burned out.

I put on a new pixhawk mini and had issues with the motors not spinning right on px4 firmware. Motor spun according to the pitch of the uav. Put on ardu 3.4 latest stable and motors were working good.

Flew around for a while at 40 feet. Quad was very shakes again.

Suddenly it started going up. This time it had more battery remaining. Went into land mode. Still going up. Rtl still going up. Althold same thing.

Looking at the thr in it was at minimum. Thr out was at max.

Battery gave out around 12000 ft.

Any ideas what could be causing this or what to check for?


Views: 738

Reply to This

Replies to This Discussion

posting your log files would really help to see what was going on.

Crazy vibrations cause climb, but log is needed if you want help

I have attached the log file. The flight starts at around 50% in the log.

When it started going up from 40' it was out of control. Looks like that is when some vibration started.



Can you put up the .log or .bin file?

Cannot read the .mavlink and is it from the flight controller or the ground station?

It is from the ground station. Written by qgroundcontrol running on Linux. Mission planner will open it, you just need to change the file type mission planner is looking for.

Type * in the File name list box in the open file dialog.

I do not have any other files. The battery went dead enough to cut the telemetry radio at 11,500 feet and rc radio went dead a few seconds later.

I did not find it yet. The wind up there was probably 70 mph.


The copter had real problems with compass variances and EKF in the red.

A lot of vibration as well which either contributed or raised the issues.

When it comes to EKF problems it is out of my area.

You will probably get better answers by posting on the Ardupilot Forums

I didn't look at your log, it is most very likely a vibration issue. Let me explain. Most of the vibes are in the Z axis (up & down) on a multi-rotor, for what ever reason, it just is. The Z axis is different than the X or Y because it fells the force of gravity. The accelerometers have a max value of +/- 16 g. They are constantly vibrating +/- a certain amount and this vibration is filtered out to get an average reading. The amount of the vibration is usually more much more than 1 g. If the vibration surpasses 15 g you will get clipping. This is when the accel reading is more than+/- 16 g. Since we are normally at 1 g, when the vibs are greater than 15 g we exceed 16 g on one end and this data is lost. The - g is 14 when the + g is 16 and so the altitude calculator is getting bad data on one end where it exceeded + 16g. This makes the copter's EKF to think that it is falling down quickly. It chases the false altitude solution and climbs. This revs up the motors and makes the vibration worse, and well you saw the result. 

Did you have stabilize mode set up? In stab mode altitude doesn't mater, you control the throttle directly and you could have got it down. Learn to fly in stab mode. Check critical thing live vibration levels and clipping (exceeding 16 g). Mission planner does this for you. Think about were that machine fell to from 1200 ft. and consider yourself lucky. Read the wiki again, please.

Thanks for all the information. I am certain something happened to the frame or one of the motors when it crashed the first time and that time it was in stabilize mode but for some reason the throttle went wide open at takeoff. I was only testing the hover time at that point and there was white out conditions so I really could not see it after a couple of seconds. The day before it flew perfectly.

It does look like once I opened the throttle past a certain amount it gradually started giving itself more and more lift as the vibration increased.

The wind was around 25 on the ground but as it got up around 10,000 feet it seemed like the wind got it so there is no telling where it came down. Probably blew at least 1000 feet. Where ever it is its covered in snow.


Mike Boland said:

The copter had real problems with compass variances and EKF in the red.

A lot of vibration as well which either contributed or raised the issues.

When it comes to EKF problems it is out of my area.

You will probably get better answers by posting on the Ardupilot Forums

Thanks for the info. That's what I was thinking. Once I gave it enough throttle to create the vibration it started increasing itself.

Just a thought, the baro is a mechanical diaphragm and it is definitely sensitive to vibration.

Basically the vibration was throwing off all your sensors, at leas baro, gyro and accel and that means the FC was reading invalid conditions on almost all channels.

Basically the FC tries to protect you from performing a throttle off and crash maneuver by ensuring the motors have enough throttle to hover or descend slowly.

But if all the sensors are inputting nonsense, it just keeps trying to compensate.

And compensating for nonsense produces - nonsense!

Lots of things can cause compass errors, motors or iron or high amperage circuits too near the compass IC for instance, but it is the one sensor that probably didn't pay too much attention to the vibrations.

You really need to fix the vibration problem(s) before even bothering with the rest of it.

Bad motor shaft or bearing maybe.



Thanks for the input. It was apparently a damaged motor.

How about cold weather? Is -40 too cold for autonomous flying? I was flying in -40 for a couple of days but most of the time it has been above 0 F.


Hi David,

In short yes, -40 can easily be really excessive.

Most of the IC's used in the FC's are not rated for that low a temperature and their performance is likely to be incorrect.

In particular, the sensor IC's - all of them can be drastically affected by too low a temperature.

Also you're battery really hates being that cold, it is only the self heating from high discharge rates that lets it work at all.

Also, any silicone or rubber shock mounts you use for gimbal or flight control stabilization cease to work effectively at that low a temperature.

There are special silicones and plastics that will continue to work, but they are not commonly available or used.

Rob Lefebvre (number one DIYD heli guy) actually operates his high end helicopter in this temperature range a lot, but he is a true expert and fully understands and compensates for the problems encountered.

Basically, -40 for most of us is just a recipe for a crash.



Reply to Discussion



Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service