GPS loss ending in crash

#1

I have been flying my quad for some time now with no issues, until recently. I took the drone off and allowed it to hover. it was hovering well, until all of a sudden, it did a half toilet bowl and shot off about 200 ft horizontally. It landed in a tree and I was able to fly it out with minimal damage.

Can someone take a look at this log and let me know if you’ve had the same issues or have fixed an issue similar to this?

Some of the flight errors i received right before the crash:
failsafe enabled: no global position
EKF GPS fusion timeout-reset to gps
ALL DATA LINKS LOST
EKF GPS fusion timeout-reset to gps
EKF GPS fusion timeout-reset to gps
EKF GPS fusion timeout-reset to gps

Do you think this could be a bad GPS or flight controller?
I am using a Pixhawk2.1 and Here+

https://logs.px4.io/plot_app?log=4bf5fdf5-3645-430a-a5f2-7849bf4b598f

#2

I don’t understand what happened. @Paul_Riseborough does it make sense to you?

1 Like
#3

Looks like a combination of mag field interference - see field strength change when motors are running

and GPS interference - see elevated velocity error metric:

and inconsistency between baro and GPS indicating a large glitch

at the same time the GPS innovations spike

It is most likely an integration problem, not an autopilot problem.

1 Like
#4

I do have my GPS about 4.5 inches above my flight controller. Do you think raising it will help? Or do you think it will not make a difference?

@Paul_Riseborough

#5

I had a similar situation happen today. The drone was at a hover and drifted into a tree. Not very fast. Although, this was probably the 5th flight today.

Also, I did move the GPS 100mm higher, using 4 aluminum standoffs.

https://logs.px4.io/plot_app?log=c6304f8a-f08e-43f9-83ba-386a537c601c

#6

Are you also using the external mag with the GPS?

#7

I believe I am. Correct me if I am wrong, but on PX4 firmware, I don’t think you can specify using an external mag can you?

My basic setup is a Pixhawk 2.1 cube with a here+ gps. From what I see, it fuses the GPS data and mag data from the input of the GPS1 port on the flight controller. If you know any better, please correct me.

#8

If you are using a 3DR type radio for communication, in my experience (both drone and rover) they affect GPS reception, at least at 433 Mhz, even the frequencies are far apart. If close to GPS, they can make lock difficult or affect performance.

I discovered this time ago experiencing with two identical RTK GPS’s at the time on a rover: the one closer to the 3DR antenna got RTK_FLOAT lock; the other one RTK_FIXED. Other RTK instruments near were affected by the 3DR emission, working normally again turning it off.

So I mount the 3DR radio antenna as far as possible from the GPS, pointing down (both antennas parallel for best reception). If possible, I insert a copper surface between 3DR antenna and GPS. Much happier now.

#9

sorry for the late reply. I think I narrowed the issue down to vibration on the craft.