Is the output value of the H3LIS331DL 12 bit long or 16 bit long? On page 9 of datasheet (sensitivity), 12 bit is assumed. On page 1 of datasheet, it says 16 bit output-data...Thanks a lot!

[Tesla DeLorean]

12 isn't a multiple of 8, whereas 16 is

[AKais.1]

True, still this doesn't mean anything... did you take a look on page 9, section sensitivity, of the datasheet? that's why I am confused... there it says 12-bit representation

[Peter BENSCH · Accepted solution]

Table 3 of the data sheet confirms the width of the measuring value with 12bit = 0...2047.

For example, with FS1,FS0=0b01 in CTRL-REG4 the measuring range FS is set to 200g, the sensitivity So is 98mg/digit, resulting in a necessary number of values of 200'000/98 = 2040.8. For this 12bit are exactly sufficient, ditto for the other measuring ranges.

As @Community member​ mentioned, the 16bit data results from the necessity that the higher-level controller typically wants to have multiples of bytes. Nevertheless, only 12 bits of significant data are contained in this 16-bit data packet for the measured value.

Is that a little clearer now?

Regards

/Peter

[AKais.1]

Hi Peter!

Thanks a lot for your answer! So, the 12 less significant bits of the output data do always contain the full information, right?

[AKais.1]

to be honest, the results I get make more sense if I consider all 16Bits to be valid... that would mean for range FS = 200g -> 6,1 mg/digit ((2*200g*1000)/(2^16-1) digit)...are you sure that table3 on page 9 is correct?

[Peter BENSCH]

Well, you could call it centred, because the measured values are output in two's complement.

Yes, table 3 is correct and refers to the actual measurement result.

Regards

/Peter