Sip&Puff sensor spurious failing

[benjaminaigner]

@ma0027 had the problem of a failing sip&puff sensor.

Workaround: unplug & plug in for a few times.

Reason: unknown

Maybe we should add error handling for the sensor & a restart procedure of I2C.
In addition, it would be good to have error output and maybe a "storage" of error messages.

[ChrisVeigl]

i suspect that the problem is that some of the sip/puff sensors introduce significant noise.
Not all sensors (tested v3 boards) behave in the same way!
The attached screenshot shows the noise range of the (already pre-processed) sip/puff signal inside the web gui: no sip/puff activities were performed but the signal fluctuates with more than 30 (corr.: 50!) values. The signal processing default values for the sip/puff sensors have already been chosen sub-optimally (big deadzone around zero) - but even this is not enough for some of the tested boards: due to the high fluctuations it can happen that the zero value can't be adjusted correctly and the sensor value indicates permanted sip or puff - although there is no user interaction. This could be a reason of the malfunctioning although the i2C communcation is still working. (However this is still just an assumption, which should be verified).
This needs more testing and we should try to find the reason why some sensors behave so weakly!