No need to wait after Wire.requestFrom() about rapid HOT 5 CLOSED

Thanks for spotting that! I will remove the line. Awesome!

from rapid.

Thanks. That was fast.

You use the I2Cdev code, which is using the Wire library in a wrong way ( jrowberg/i2cdevlib#265 ). That library is no longer maintained and the issues are piling up. The library is also a little outdated. Your copy of "init_mpu6050" is therefor also a little outdated. The FastWire option is not really needed and setting the TWBR for the speed is no longer needed, since there is a Wire.setClock() function now.

Changing the I2Cdev.cpp is risky. There is not much you or me can do about the library at Github. Keep in mind that the library is slower than needed, with extra unnecessary bus activity that slows it down. With the current version of the AVR Wire library the I2Cdev library works (despite the wrong use), so for now it is okay.

from rapid.

I guess it would be possible to use the i2c_t3 library in I2Cdev, however as you said it could be a lot of work.

I have sufficient power on my microcontroller to be able to do my own sensor fusion, so this library by Kris Winer is worth looking at for me.
In the long run, I want to use a different sensor anyway, as InvenSense has presented really good new options as recent as CES.

Thank you for pointing out the issues!

from rapid.

You might consider to upgrade to the MPU-9250 and try the Kris Winer AHRS fusion. Sometimes a MPU-6050 module is shipped with a smaller capacitor which causes extra noise. The MPU-9250 doesn't have that capacitor.

from rapid.

Ok, that sounds like it will improve the IMU quite a bit. I am wondering about the tradeoffs of using on-chip sensor fusion vs. doing it on the main board. I guess I will try both!

BTW, great job on helping all those other people use the Wire library correctly.

from rapid.