Bosch Sensortec | BMP388 Datasheet

14 | 57

Modifications reserved |Data subject not change without notice | Printed in Germany

Document number: BST-BMEP388-DS001-01 Revision_1.1_032018

3.4.3.

IIR filter

The environmental pressure is subject to many short-term changes, caused e.g. by slamming of a door or window, or wind

blowing into the sensor. To suppress these disturbances in the output data without causing additional interface traffic and

The output of a next measurement step is filter using the following formula:

������������������������_�������������������������������������������� + 1

where data_filtered_old is the data coming from the previous acquisition, and data_ADC is the data coming from the ADC

before IIR filtering.

The IIR filter can be configured using setting iir_filter in “CONFIG” register (see 4.3.18).

When writing to the register filter, the filter is reset. The next value will pass through the filter and be the initial memory value

for the filter. IIR filter is reset if the temperature or pressure measurement is disabled (temp_en or press_en registers changed

from ‘1’ to ‘0’) or when a transition from sleep mode to normal mode occurs. After enabling of pressure or temperature

measurement, the filtering will start, thus the next incoming value will pass unfiltered and be the initial value of the IIR filter.

The step response (e.g. response to in sudden change in height) of different filter settings is displayed in Figure 6.

Figure 6: Step response at different IIR filter settings

3.4.4.

Oversampling

Noise depends on the oversampling and filter settings selected. The stated values were determined in a controlled pressure

environment and are based on the average standard deviation of 32 consecutive measurement points taken at highest

sampling rate (for details please refer to Table 21). This is required to exclude long term drifts from the noise measurement.

E.g. environmental fluctuations caused by windows being opened and closed might have a frequency <5 Hz. Consequently, a sampling rate of ODR = 10

Hz is sufficient to obey the Nyquist theorem.