Accelerometer Calibration

PatR · May 9, 2019

I was indeed referring to the user manual as it is the manufacturer created manual released to the buyers. The only other documentation available from Yuneec is the Quick Start Guide and firmware update PDF’s.

i won’t debate further over calibrating accelerometers when they are in motion, i’ll just let those that do so live with the results. There’s ample online resources available describing accelerometers, their make up, and industry standard calibration practices.

Zugflug · May 13, 2019

While reading the debates whether the manual is wrong and what is the industry standard, i started to wonder about the other thing - the aircraft is capable of performing calibration (of something) both in the air and on the ground. So the same mistake must have been made in the software

. And while i perfectly understand the mechanics behind stationary calibration, my subjective feeling has been so far that the aircraft is most stable after performing both of them - stationary followed by no-wind in flight (it actually fails in-flight calibration in unsuitable conditions or if you jerk it around too much). So if anyone has any ideas about what (and why) is being calibrated in flight it would be enlightening to know.

P.S. I thought pixhawk was used in Plus, not original H.

jandjfrench · May 13, 2019

I don't know what mistake in the software you suspect. I believe the procedure that is given in the RealSense version of the Yuneec owner's manual is correct. I believe that unless differences in the center of lift and center of gravity are not taken into account and the gps is striving to stay at the same latitude and longitude this could result in something like a dog chasing its tail. As a thought experiment, I first imagined perfect conditions where the centers of lift and gravity were coincident in the Z axis and there was no wind. Increasing thrust would cause the H to rise without any x or y change. I next considered strapping a weight on one of the arms with the resultant shift in the C of G. There would have to be a change in thrust to compensate for this. I believe this thrust variation would be the new standard from which to work to achieve altitude changes without changing the x, y, values. This new baseline is what I suspect is being established by the procedure in the Yuneec manual.
And as we don't live in a wind free perfect world it seems likely that, unless there were large differences in the centers of lift and gravity, simply doing the calibration with the H level would be fine.

Zugflug · May 14, 2019

I should have written quotes around the "mistake". My point was that there is actual functionality for calibrating _something_ in the bird that is available under "accelerometer calibration" and performed in flight and is thus consistent with the procedure written in manual. What and how is actually calibrated this way is another question entirely. And this makes me doubt that it is just a mistake in the manual even if it contradicts the usual practice of accelerometer calibration.