Hello everyone,
I'm new on the forum, and in the autopilot world as well. I need some help to understand what happened with my plane.
I was flying in FBWA mode tuning the PID as described in the manual.
Suddenly, the motor stopped, the plane got unresponsive and crash on the land.
When we found the plane, luckly the battery was still plugged. Here are some thing we were able to see:
1- The multi color led was Green, even with the radio turned off.
2- Turning on the radio, there was not servo response
3- There was no telemetry signal. The RDF900 only has a Green light (no red flashing light)
4- The radio Rx was working fine, we plugged a servo on it and worked fine.
5- The orange B/E led was blinking.
After recycling power, it all seems to work well.
The setup was:
Plane: Canary 40 ARF
Motor: EVO 10GX gas
Radio: Spektrum DX6 with Thomas Scherrer LRS
Pixhawk Autopilot with latest firmware at the time (Nov, 15th)
Mission Planner 1.3.10 build 1.1.5369.11976
I have read that noise in the I2C bus can lockup the Pixhawk. I have extended the wires for the airspeed sensor as well as the wires for the power supply.
Any suggestion on how to avoid this Pixhawk lockup in the future will be usefull.
Thanks!
Cheers,
Carlos
Replies
Thank you all guys for your answers. I will try with the new 3.2.0 release!
i have recently installed apmplane 3.2. During an auto flight we lost altitude which almost resulted in a crash. When looking at the data flash log i saw we were getting negative spikes on the I2C log. I checked earlier logs and did not see a similar issue. I have the 3DR external compass and the pixhawk pitot tube connected to the I2C port.
Michael,
How did you recover the plane from losing altitude? did you pass to manual control ?
Rgds.
we first went into FBWA mode and then manual. Managed to recover 5 m from the ground. Having now also looked at the I2CErr log i see no problem with this signal. Bit of a mystery. Should fly again in the next few days and will give an update if I find anything
I2CErr isnt logging anything, it's simply unused in current firmware.
do you have very long I2C cables ?
post the .bin file (log) - that will help.
Hi Carlos,
based on your description it is quite likely that this was the I2C bug that was fixed in the 3.2.0 release. The ignition system from a gas plane can give quite a lot of RF noise, and can cause I2C errors. I fly gas planes myself, and notice a number of I2C errors (for example, you see occasional downward spikes in the the airspeed logs, which are caused by I2C errors if using a digital airspeed sensor).
Cheers, Tridge
yes, but it's hard to know how "extended" cables here were performing, the I2C bug needed poor signals or much traffic to kick in, so it's hard to say for sure if it was the reason.
Andre,
based on your experience, would you say what other reasons could cause a lockup?
except of power issues (power loss, brownout) I have no known reasons. - and yes, I too think that might have crashed your plane, just pointed out it's hard to *know*.
You could check your logs for PM/I2CErr , (I've never seen anything >0 there), even on a plane with a little bit longer airspeed and GPS wires, (custom crimped) But I would expect a UAV that is prone to this I2C bug, to log some bad packets.
-
1
-
2
of 2 Next