LPS33HW temp compensation performance

[DManc.139]

We are experiencing big pressure variations with respect to 3/4 degrees of temp change. Is this normal? We are seeing up to half a millibar (2000 raw count) of difference changing the temp of the product by taking it in the hand. For our application this is not really manageable. Please let us know what are the compensation specs (time constants, error, etc...)

Thanks in advance!

[Eleon BORLINI]

Hi @DManc.139​ ,

are you facing this issue on different samples (so that it is a systematic deviation), or just on one sample?

In the first case, you should check if you are correctly implementing the pressure reading and data conversion (lps33hw_reg.c:(

float_t lps33hw_from_lsb_to_hpa(int32_t lsb)
{
 return ( (float_t)lsb / 4096.0f );
}
 
float_t lps33hw_from_lsb_to_degc(int16_t lsb)
{
 return ( (float_t)lsb / 100.0f );
}

In the second case, you should check if the pressure is not increasing because you are somehow pressing the air while handling the device in your hand to increase the temperature.

If it is not the case, it could be a sample-related issue: you could try to apply the one-point-calibration technique to mitigate the issue, as described in the datasheet, p. 41:

If, after the soldering of the component, a residual offset is still present, it can be removed with a one-point calibration. After soldering, the measured offset can be stored in the RPDS_H (19h) and RPDS_L (18h) registers and automatically subtracted from the pressure output registers: the output pressure register PRESS_OUT (28h, 29h and 2Ah) is provided as the difference between the measured pressure and the content of the register 256*RPDS (18h, 19h). 

In normal conditions, the full temperature range (0°C to 65°C) accuracy should improve from 2.5hPa to 1hPa.

-Eleon