g calculation correct for 2g range but wrong for other ranges

[mgrusin]

The floating point gains built into the library are handy, but are correct only for the 2g range; you'll get erroneous results when using the .get_Gxyz() function and a range other than 2. A solution in .get_Gxyz() would be to multiply each result by (range/2), e.g. 8 for the 16g range.

[ToniCorinne]

Want to set up a pull request @mgrusin ?

[mgrusin]

You mean if I find a bug I have to fix it!? This changes everything.

Coming right up.

[40943260]

I tested the ADXL345 on a GY-80 module in ±16g range and attempted to convert the raw readings to m/s² using the following code:

Serial.print(x * 0.00376390 * 8 * 9.80665); // convert to m/s²
Serial.print(", ");
Serial.print(y * 0.00376390 * 8 * 9.80665);
Serial.print(", ");
Serial.println(z * 0.00376390 * 8 * 9.80665);
delay(100);

In a stationary state, I obtained values such as:

-0.30, 0.30, 9.15
-0.59, 0.30, 9.15
-0.59, 0.59, 9.15
...

I notice that the Z-axis readings are slightly below 9.8 m/s², which could be due to the sensor not being perfectly level. I am not entirely sure if this conversion is fully correct, but multiplying by 8 for the ±16g range produces much more reasonable results compared to using the original factor..