Bitcraze Forums

Sorry for the english! I use a translator!
Recently won CrazyFlie10DOF!

Before the advent of firmware Crazyflie 2014.01-rc1 was virtually impossible to fly, copter pulled in different directions and stabilize it impossible to me.
Used to control Android tablet.

Noticed in the client for Windows (cpclient), values ​​Pitch, Roll, Yaw, not zero.

Copter is at rest on a table, and the values ​​are as follows:

Pitch 1.48
Roll -0.46
Yaw -8.18
ASL 68.88
And the skyline from the indicator declined from 68.88 Digits (ASL).

It is a failure?
Is this normal?

Thank you ...

If you're absolutely sure your table is level, the IMU chip probably got mounted a tiny bit askew.
You can probably trim it out.

If you're bothered by the readout, that's another story.
It'd probably make sense to add sensor trim as a configurable parameter to cflie firmware.

Thanks for the answer!

I tried trim in bootloader.
I appropriated the return on a sign (+/-) Roll and Pitch values.
But at the following loading in the bootloader, the appropriated values were interchanged the position.

Initial values appropriated by me (for example):
roll: 0.5
Pitch: 1.0

At the following reset in a butloader (for example):
roll: 1
Pitch: 0.5

It is a failure?

And why not to make function of independent trimming at inclusion?
What the copter himself trim also exposed all values in zero?

Trimming and calibration is definitely something we would like to improve more. As soon as we have time for it...need more time!

The Pitch and Roll values looks really good so I wouldn't bother trying to get them close to zero. I don't think you will be able to find a reference good enough either.

The trim in the cfclient is a "high" level of trim and takes into account motors that are a bit angled, propeller differences etc. The calibration parameter in the config block (bootloader) is meant to adjust for the chip for not being mounted flat. This isn't working that well though and a better calibration method is on our wish list.