Strange code in MPU6000 module

Archive
1

Hi!

I’ve discovered strange code in MPU6000 module

github.com/diydrones/ardupilot/ … 0.cpp#L377

Strange thing is X and Y values are read in reverse order.
Bytes 0,1 goes to Y accel value
Bytes 2,3 goes to X accel value

Same for gyro values.

According to specification X-axis-bytes comes first.

Is it a hardware-specific hack? Looks like IMU turned in 90 degrees.

Some background:
I’m porting arducopter on custom hardware with MPU6050 as IMU.
i2c is the only supported bus for MPU6050 so I’m hacking MPU6000 code to support i2c transport.

2

Could be that we did that because the APM2 has the chip rotated 90degrees, or maybe we have a difference definition of what X vs Y is from the MPU6k datasheet.

You’ll find that SirAlex who I think works on the Crius board and/or MegaPirates has already ported our MPU6k library to work on the 6050. The problem with using that is that it’s much slower than using an SPI connection. about 20x slower I think. 8Mhz vs 400khz.

3

Thanks for information!

MPU6050 supports i2c only so in my case it is 400kHz VS Zero :wink:

4

Looks like I found the confirmation for the answer (through looking QGroundControl source code)

The strange calculation just ‘fixes’ HW axes to match logical aircraft axes:

X axis goes forward
Y goes to the right
Z goes down.

Could someone confirm my findings?