2025-02-14 03:32 AM - last edited on 2025-02-14 04:37 AM by Peter BENSCH
Hello community,
I am an electronics engineer at the CNRS in Montpellier (France).
I am currently working with the IIS3DWB vibration sensor
When I use the LFP2 low-pass filter, the action of gravity is added to the measurements.
I use the following data registers:
- CTRL1_XL
- CTRL3_XL
- CTRL8_XL
This result is as unexpected as it is disconcerting.
Has anyone ever encountered this type of problem?
Thanks in advance to anyone who can help me.
Olivier Maudens
2025-02-14 09:43 AM
IIS3DWB is a wideband accelerometer therefore suited for vibrations monitoring.
This does not mean that you will only see vibration. The device is an accelerometer and will therefore measure every acceleration, including the gravity acceleration.
You can however filter out the gravity because it is seen as a DC/low-frequency component, and keep only the vibrations/high-frequency components.
Check section 3.1 of application note AN5444: there is an high-pass filter that you can enable and configure, and you can route its output to the FIFO and output registers. This will give you the high-frequency information, removing low-frequency.
Alternatively, if you want to keep low-frequencies as much as possible, you can choose to subtract a given DC component (user offset) from the output of the low-pass filter, and then route this output to the FIFO and output registers. The DC component, aka user offset, to be subtracted must be computed by other means. When the device orientation is given, one usually takes the average in steady conditions and uses that as DC component to be subtracted. This will work until the orientation changes of course.