Hi all,
I found that with the latest arduplane code, the attitude of the plande levels out real quick... In other words, if I bank or tilt the plane, the attitude readings go to 0 (level) in a few seconds. This is not supposed to happen right?
I am using Arduplane 2.65
Thanks for any suggestions!
Here is a movie of the behaviour.
Replies
try
open terminal
setup
reset
y
then after board reboot setup hardware, radio
Hey,
I have also tested with an older IMU shield (with which I have flow quite some succesful flights) which produced the same results. I have also tried with and without magneto; same result.
What is strange is that the gyro and accel readings seem to be alright. (accel readings remain static if the plane is static, gyro readings go to 0 if the plane is static and excite when the plane moves). Also the initial estimation when moving the plane is as expected. When tilting forward, the attitude estimation moves along well (maybe a little less then expected), but then starts running back to 0.
A summary of what happens in terms of a and g readings:
bank right 45 degrees : ay = -30, gx excites to 1500 and returns to 0 (hmm... shouldnt the ax value change when banking instead of the ay?)
bank left 45 degrees : ay = 30, gx excites to -1500 and returns to 0
The other values remain roughly stationary; az = approx -1000 in any position.
Tilt forward 45 degrees : ax = -20, gy excites to -1500 and returns to 0
Tilt backward 45 degrees : ax = 20, gy excites to 1500 and returns to 0
The other values remain roughly stationary; az = approx -1000 in any position.
I have attached the log files of a short bench test where I did roughly the same as in the video. Tilt the plane 45 degrees forward, wait (attitude estimation levels out) and tilt back; same thing happens. Then I did the same for roll.
Ok while typing I think the problem may be that the accelerometer x value corresponds to the tilt movement and the gyro y value; which would mean that the gyros y measurement is wrongly corrected for drift by the accels x measurement... or am I babbeling? =]
If it would be an error in the code I would assume others had the same problem; but I don't know what else it could be... Any ideas would be very much appreciated; I am puzzled =]
Thanks!
Serge
2012-10-01 19-43-04.rlog
2012-10-01 19-43-04.tlog
No that definitely does not look right.
Check your accelerometer and gyro status in the CLI.