my RC plane installed pixhawk is out of control after take off ,the good news is the entire process has been record included camera and osd video except telemetry log ,please find the detail crash below:
i switched to manual mode and increase throttle for take off ,then switched to FWBA mode after the wheel off ground .I try to make a left turn ,but the plane do not response to my controller input .then I realize the problem and switched RTL mode immediately ,but the plane keep flying without come back home .the GPS locator help me to get back the plane .
the dataflash log can be find in the below link:
data flash log
please go to YouTube visit the camera video
Hope some one can help to find out the issue this time,thank you for all your attention.
I am looking at your log. Did you verify the directions for your ailerons in both Manual and FBWA mode before flight? This is a very important step, and if your ailerons were reversed, I think that explains your crash.
Your flight travels over a densely populated area in Dongguan, China. You are very lucky that no one got hurt, and no property was damaged.
Your flight was irresponsible, and you are an unsafe operator. Do not continue to fly unsafely.
In order to fly safely, EVERY PILOT MUST BE ABLE TO CONTROL THEIR AIRCRAFT IN MANUAL MODE. Do not fly any aircraft you cannot manually control, exactly for the reasons demonstrated by this flight. We are all fortunate to learn this lesson with no one getting hurt.
thanks,i had check the ailerons in manual and FBWA mode before flight, it works fine.
and I review the osd video today and find out the airspeed and ground speed are big difference, there are an impact in airspeed tube at last landing, but I fixed the tube and blow to the pitot tube , the airspeed increase, so I thought the airspeed sensor works fine. imagine the airspeed value are wrong, it sends the wrong data to FC , the bank angel limit to 20 degree in FBWA MODE because of stall prevention, so it can’t make left turn because of wind affect even I push the stick to end of left. the plane keep nose down to increase airspeed afer switching to RTL mode, that should be. the reason of crash, are you agree with that? or do you think there’s any other reason?
I very much agree with your advice on safety. and will not flight unless find the reason of crash.this really good lesson about safty. I didn’t change to manual mode because I can’t see the ATT of plane at that time.
final question how to make sure the airspeed sensor works fine?we can’t make sure it works fine even the airspeed increase by blow to the pitot tube as we don’t know the exact airspeed that we blow out.
Safety Recommendation: Never fly further than your ability to visually see and control the aircraft in manual mode. (For folks in the USA, this is essentially a law.)
For diagnosing the issue: Good idea about the airspeed being incorrect. I have attached a plot showing that the airspeed thought it was flying too slowly. I looked more into that. Did you perform an airspeed calibration on the ground before flight? If you did an airspeed calibration while the pitot tube was uncovered and wind was blowing, that may have caused the problem.
Looking through the log, I see that the airspeed was re-calibrated at 272 [sec], changing the offset from 65.5722 to 73.8288. I also see ARSPD_AUTOCAL is set to 0, so the sensor was unable to fix the problem mid-flight. This is consistent with forgetting to cover the airspeed sensor during the pre-flight calibration.
Not that it helps now but in the next major release we should be introducing a basic airspeed health check. Its in master now just waiting for it to make its way to the plane release.
Looking at the log it would have helped you if you had a AHRS_WIND_MAX of 15 or less. This is good as some doubted this issue was occurring and that the fix put in place was going to help anyone, I actually had trouble getting logs with this problem.
As i said not that it helps you now, it’ll be released when it is. It will be turned off by default…
I’d recommend you turn ARSPD_USE = 0 and do another FBWA or manual flight. Ensure its calibrated (covered) and then fly. Compare the log results afterwards to see if its faulty.
Thanks zbrk, one more questions for this parameter, For example, level four winds（5.5-7.9m/s） should not continue to fly for foam aircraft. so we set the AHRS_WIND_MAX of 7.9m/s, and One day I felt the wind was very low when I was flying. It must be below the fourth level，but the air speed sensor fail in the air for some reason, the airspeed is 15m/s while the ground speed is 25m/s, do flight controller take ground speed of 15m/s to use as airspeed？the most important things for a pilot is to find out the wrong value and take correction, so does flight controller.
the problem is whitch value choose to believe .
the controller use ground speed as airspeed if it find the two speed value greater than AHRS_WIND_MAX, I don’t thinks the controller do the right things.
think about plane nose down or nose up in the air, the two speed would be much different. but the airspeed sensor works fine in that case.
In the example where you have winds up to 7.9 m/s and you have the AHRS_WIND_MAX set to say 10 m/s to give some leeway on gusts. Also the ARSPD_OPTIONS bitmask for the failover would need to be on for this to work.
In the failure it seemed your airspeed was around the 15 m/s mark, it didn’t really get any higher. The groundspeed however surged to almost 35m/s.
If the ground speed is 30 m/s it would take the 30 m/s and using AHRS_WIND_MAX it would check if the airspeed was in the range of expected values (in this case 20 m/s and 40 m/s). If it was outside of that it would be treated as suspect, this would decrease a confidence value going through a low pass filter. It will issue a warning at a certain level and then if it continues to fall then it will switch the plane to use the synthetic airspeed instead (switches ARSPD_USE to 0). If the failback option is on, if the reading gets better it will increase the confidence level (at a much slower rate). This process of failing would typically trigger in approx 3 to 5 secs in the logged flight.
If confidence is restored it will resume using the arspeed sensor but that is an option that is separate to the failover option.
Nose attitude has no bearing in this simple check we simply take the 2d gps ground speed and apply the AHRS_WIND_MAX to generate a range of expected airspeed. In your question its one of the reasons for the confidence levels to filter out short duration’s or strong wind gust where it will temporary be less. If the flight profile is one where you’ll likely be stationary (ground speed) then this airspeed option is not the one to use.
Once I finish two other PR’s I’m revisiting this with a much more robust and through solution. I was seeing reports of planes crashing (and I have had a tube come lose midflight but luckily I was only testing the airspeed sensor and had ARSP_USE=0) so put a fix in place that matched the current documentation.
Hope that helps.