Enabling and disabling FDS bit shows unexpected result in the axis magnitude

elange · ‎2026-02-03

Hello STMicro Community,

I am working with the LIS2DH12 accelerometer, STEVAL-MKI109V3 and MEMS Studio environment. My application is the measurement of vibration on rotating equipment.

A vibration signal of 118Hz is being applied to the board and using ODR=400Hz, Scale +/-4g.

I am noticing a significant difference in magnitude measurements when switching the filter bit on vs off (FDS). See attached image of the time series data yellow (Z-axis) where:

- with FDS=0 I see -500 to -1500 for ~1000mg range

- with FDS=1 I see -800 to +800 for ~1600mg range

I DO expect to see gravity getting removed but NOT an increase in the magnitude. Is there an explanation for this?

-Edwin

Federica Bossi · ‎2026-02-09

Hi @elange ,

With FDS=0 you see the raw DC‑coupled signal (gravity + vibration), so Z is centered around −1 g and the AC amplitude is the “true” one.
When FDS=1, the internal high‑pass filter (set by HPCF bits at ODR=400 Hz) is applied before the outputs. This HPF does not have unit gain at 118 Hz: at that frequency its transfer function can slightly amplify the signal, so you observe a larger peak‑to‑peak value even though gravity is removed.
For accurate vibration amplitude it’s usually better to keep FDS=0 and remove gravity/offset in software (e.g. subtracting the mean or applying a digital HPF with known gain).

In order to give better visibility on the answered topics, please click on 'Accept as Solution' on the reply which solved your issue or answered your question.

View solution in original post

Federica Bossi · ‎2026-02-09

Hi @elange ,

With FDS=0 you see the raw DC‑coupled signal (gravity + vibration), so Z is centered around −1 g and the AC amplitude is the “true” one.
When FDS=1, the internal high‑pass filter (set by HPCF bits at ODR=400 Hz) is applied before the outputs. This HPF does not have unit gain at 118 Hz: at that frequency its transfer function can slightly amplify the signal, so you observe a larger peak‑to‑peak value even though gravity is removed.
For accurate vibration amplitude it’s usually better to keep FDS=0 and remove gravity/offset in software (e.g. subtracting the mean or applying a digital HPF with known gain).

In order to give better visibility on the answered topics, please click on 'Accept as Solution' on the reply which solved your issue or answered your question.

elange · ‎2026-02-09

Hello Federica and thanks for the feedback!

We tested some configurations today focussing on FDS=1 vs FDS=0 and also +/-2g vs +/-4g scale comparing our sensor output to the MEMS Studio outputs. We observe that:

1) our sensor is reporting raw values ~2x that of MEMS studio when using same FDS and scale settings

2) with FDS=1 (filter on, default mode '00') both sensor and MEMS studio remain consistent by slightly amplifying the raw data signal that you have indicated is expected

Questions:

a) What could account for the 2x factor above?

b) What is the recommended filter mode setting (HPM1,HPM0) for a vibration sampling use case?

c) Should Block Data Update (BDU) bit be set for High Resolution mode HR=1?