Quadplane almost crashed in auto mode fw 4.3.1 & cube orange, please help!

ArduPlane Plane 4.3
1

Hello gentlemen,
I just upgrade my quadplane to Arduplane FW 4.3.1 from 4.2.3.
The drone flew perfectly using FW 4.2.3, and I have conducted several way point missions without any issue.
Just yesterday I did upgrade to FW 4.3.1, I have got some issues. First is all servos are showing erratic behavior, in FBWA mode elevator servos move very little, when I did test on the ground after arming. While aileron servos are normal. So servos shows very strange (erratic) behavior, compare with previous FW. (4.2.3). Secondly, I get error : ā€œCompass Inconsistenceā€ many times when I turn on the drone. I use dual GPS and Cube Orange.

After I rotate the drone, then I donā€™t see any Compass Inconsistence message. I can fly very well in FBWA mode without any issue. Then next day I did full AUTO MISSION from take off to landing. This is what happen during AUTO MODE:

1). Arm (no error)
2). AUTO mode (no error)
3). Start TAKE OFF (no error, setting TAKE OFF altitude is 45m)
4). Drone only reached 30 m, then heading to WP 2 (Setting WP2 is 80m)
5). Then I got :ā€Error Compass Varianceā€ messages, EKF become RED.
6). Drone was flying to WP2 at 30m altitude, and started to do Transition (supposed to be 45m heading to 80 m).
7). While Error Compass Variance keep going, speed reached 15m/s
8). Got message Transition done at 13 m/s
9). But drone was flying at about 30 m Altitude, and even down to 26m (almost crash to big trees).
10). Then I switch to RTL and tried to take over the control.
11). Drone seems to be confused and just hover at 30 m altitude
12). After a few second , drone start to fly a bit higher heading to home
13). Then I switch to FBWA to increase altitude, and drone can fly as normal.

So, first issue: Why drone can not achieve 45m take off height, but only 30m instead?
Second issue: why elevator servo donā€™t move as normal (so can not achieve target pitch in AUTO), but in FBWA mode it works as normal during flying.

Third issue: why all servos behaves so erratic using FW 4.3.1, while they behave normal in FW 4.2.3 ? Do I miss something in FW 4.3.1 ?

So I still donā€™t understand what is root cause of this problemā€¦I need help to analyse the log further to solve this problem. Maybe @tridge , @andyp1per
@Allister , @geofrancis or other experts could take a look the log.

This is the log files (bin log and tlog)
https://drive.google.com/drive/folders/1U3bNSnNaebpG5nUpsg1rrGgmGCfnD0tK?usp=sharing

Thank you very much for any help to solve my problem. I hope tomorrow I can do the test again.

Best regards,

2

Your EKF is definitely very unhappy early on in the flight which probably caused most of these issues. I can only suggest you check for interference and maybe recalibrate the compasses

3

Hi @andyp1per ,
Thank you for your quick response. Yes, I will do calibrate the Compass. I did suspect since beginingā€¦ But do you think Compass issue can cause the servos behave erratic ??
How about PID especially Pitch ?
Pitch behavior is not normal.
Any other advice ?
I appreciate any advice very muchā€¦
Best regards

4

Yeah not sure about the servo issue - would definitely check your hardware

5

Very strange that both Pitch servo (V Tail type elevator) donā€™t move at all when I push RC stick up and down, in FBWA mode . But Yaw control using the same servos is normalā€¦So the hardware servos are normal.
I did calbration the radio already. And input Rc signal is ok. But no output signal at allā€¦
What parameter that control this Pitch movement??
I did change the mixing gain from 0.5 to 1, still donā€™t solve the problem.
Any advice is highly appreciated.

1 Like
6

In the flight the V-tail servos are nearly maxed for the duration of the flight. Something is wrong with the setup. Could be C of G, or mechanical trim. But given how far apart those values are it looks like the FC is trying all it can just to keep the nose of the plane up.

Screenshot from 2022-10-27 23:05:35
Screenshot from 2022-10-27 23:05:351280Ɨ761 199 KB

There are also some sharp drops of the nose during flight. I donā€™t know enough about your plane to say for sure but the drops look like stalls to me, especially when they are at the same time as higher bank angles and higher pitch demands. The plane could simply be flying too slow, or it is too heavy, or the C of G is off. Maybe too far forward (but itā€™s late here so Iā€™m not sure about that.)

Plus what @andyp1per said about the compass. Iā€™d start by disabling the internal compass, then look to the wiring for better positions or better shielding for externals. If you havenā€™t seen it check out this post:

7

Hello @Allister ,
Thank you for your time to analyse my problem. But I find that in FBWA mode, the plane flew very well, without Compass error. I want to know how slow the plane can fly, so I put Throttle in 50%, and the plane flew well at 14-15 m/s speed. When I increase Throttle, it can get 20m/s without problem, but current will be much higher.
I have conducted AUTOTUNE in plane mode several time, and I think I flew quiet stable.
Do you think I have to do AUTOTUNE again, especially PITCH ?? I suspect that Pitch gain needs to be increased. Any advice?

Today I did test again, but I down grade to fw 4.2.3 to see the differences between these two . I disabled the Internal Compass. I did Calibration the Compass, and I got Compass 1 success Calibrated.
I found that using FW 4.2.3 I can control the pitch (Elevator V Tail) in FBWA mode (on the ground) by using RC. So all pitch servos looks normal.

But when using FW 4.3.1 in mode FBWA (still on the ground) the servo pitch donā€™t move at all using stick RC. I have tested both in ARMED and Disarmed condition. I checked on Mission Planner , there is no signal output on Pittch servo, eventhough input signal is ok. I had done RC Calibration and checked prior doing this testing. Any explanation about this ??
Thank you.

8

Thatā€™s very strange. It shouldnā€™t matter but obviously something is happening.

Can you post a log from the successful flight using 4.2.3?

9

@Allister ,
Yes it is strange. But unfortunately this new test was not as expected. It ended nearly same as previous test due to Compass Error.
In this fw 4.2.3 test flight, Compass error occured after the plane reached 32 m (It supposed to Take Off to 50m height) and will do transition to WP2. Then Compass Error keep going, and it did not fly higher to 80m height (as WP3) but just flew at 30m, then I toggle the RTL switch for safety reason.

For additional info, I have conducted about 10 times flight in AUTO Mission mode using this Vtol plane without any issue. Then I went to another Island and did this test flight. I believe this place has strong EM that cause Compass Error to my Vtol plane.
I never have this Compass error using exactly same configuration, when flying at my home town.
What do you think about my PID, especially PITCH?
I really like to hear your comment about current PID.
Here is the bin log and Tlog of today test flight using FW 4.2.3.
https://drive.google.com/drive/folders/1U3bNSnNaebpG5nUpsg1rrGgmGCfnD0tK?usp=sharing

10

Which FW do you think is ā€œbetterā€ based on this test flight?
FW 4.2.3 or FW 4.3.1 ?
I am not sure why the speed suddently drop during this fw 4.2.3 tes.
I am also not sure why altitude was down from 32m to about 16 m when I switch to RTLā€¦(It did NOT happen during test flight of FW 4.3.1).

11

I should clarify one thing: Iā€™ve never built a VTOL so I donā€™t know if thatā€™s a contributing factor here.

The fixed wing flight tuning looks better in 4.2.3 than 4.3.1. The V-tail outputs also look a lot better. They arenā€™t pinned like they were in that graph I posted previously. Comparing parameters I donā€™t have an explanation for whatā€™s causing that. Iā€™m going to have to defer to somebody smarter than me.

The compasses still look like theyā€™re having issues. You should consider running MAGFIT and setting up the COMPASS_MOT values.

https://ardupilot.org/plane/docs/common-magfit.html

12

I have the same question,

14

I use Plane-4.3.1 firmware .I switched the VTOL from QLoiter mode to manual mode on the ground, and then controlled the pitch servo, but there was no action; then switched to FWBA, and switched back to manual to control normally. This problem caused me to destroy a machineļ¼
I suspect itā€™s a firmware issue, Iā€™m using 4.3.0 firmware and everything works fine, but I havenā€™t done a flight test yet. so Iā€™m a bit worried. I donā€™t know if you have a stable version firmware recommendation, I hope to get an answer.

15

I was in the same situation.
When I changed the firmware to 4.3.0, the problem disappeared.
So I think itā€™s probably a firmware problem.

16

Recently, I want to make a VTOL Plane. Do you use the 4.2.3 firmware version for many flight tests? If can, I will choose 4.2.3. Thank you

17

Hi @Awesome
Based my experience until today, my Quadplane performs excellent using FW 4.2.3ā€¦ Perform long range mission waypoint without any signicant issues. I am not saying the later or latest FW is not goodā€¦ But I am not ready to take the risk of using later FWā€¦ But maybe next time I will try again using the latest FWā€¦

18

Thank you! :+1::+1::+1:

19

There arenā€™t any changes in 4.3.1 that could explain this. If you are able to reproduce this behaviour then please set LOG_DISARMED=1 to get a log while you reproduce the issue and I can take a look at it.