This is a comment it is not a question.
The older the mems sensor in acceleration the more likely it was to have a thermal effect below 1 Hz. On the old CX1 from Sensr it was pronounced on three axes, on the MKI109 -210 combination it was reduced, now it is only just present on the Z axis. This is a count of a couple of days of acceleration for the peaks below 35 Hz on a 16384 FFT at 1995 Hz.
To say the result is surprising is the understatement of the year.
The linear regression on the peaks spacing has a t Stat of 554, that is interesting.
less than 60 is Gaussian more than 60 is non-Gaussian. From the Gaussian I can tell the temperature.
Nice little unit.
| Coefficients | Standard Error | t Stat | P-value | Lower 95% | Upper 95% | Lower 95.0% | Upper 95.0% | |
| Intercept | -2.80533 | 0.040385 | -69.464 | 2.57E-07 | -2.91746 | -2.69321 | -2.91746 | -2.69321 |
| X Variable 1 | 5.750571 | 0.01037 | 554.5379 | 6.34E-11 | 5.72178 | 5.779363 | 5.72178 | 5.779363 |
