Anomalous acceleration values from some LSM6DSLs: are they damaged?

[ADi S.2]

I've used a number of times LSM6DSL IMU sensors and I never had any problem with them: they are in general very precise and straightforward to use. However, in a recernt pre-production run, some accelerometers provided strange values on one or two axes compared to the usual values. In these conditions (stationary board in horizontal position) the expected values on two axes should be close to 0 g, i.e. raw values in the range of about +-200 for a full scale of 16 g. This happen with most units, but on some others (identical firmware, identical boards) we get much greater values, something like >5000 or even >16000, that, by the way, are normally sensitive to the self test stimuli. The problem seems to be only in the accelerometer values: other functions work correclty.

My question is: are these values something "normal", that can/must be corrected through calibration (with the offset registers or in firmware)? Or have I to assume that these devices have been simply damaged during board assembly? In this last case, are there some thermal or mechanical reccomandation to follow during assembly? (I was not able to find them on the product page).

Thank you!

[Eleon BORLINI · Accepted solution]

Hi @ADi S.2​ ,

please consider that the typical linear acceleration zero-g level offset accuracy is ±40mg (datasheet, p.21). In the 16g FS case, it corresponds to 100LSB, so a value of +-200 is a little higher than the typical. In this case, however, the offset can be compensated by properly tuning the X_OFS_USR (73h), Y_OFS_USR (74h) and Z_OFS_USR (75h) registers (datasheet, p.95).

Values >5000 or even >16000 are surely too high, and in this case -provided that the software, the soldering process and the application boards are the always the same- a damage of the sensor cannot be excluded. Are you facing other issue, for example anomalous current consumption?

As general recommendations, please consider this Surface mounting guidelines for MEMS sensors in an LGA package application note for further details on the soldering rules.

-Eleon

[ADi S.2]

Ok, thank you for the answer. Yes, normally for 0 g the value is less than 100 (slighly more if the horizontal alignement is not perfect, as happen in our case). But I agree, >5000 is most probably indicative of a damage. I will check out with the assembly company.

As for the electrical consumptions, I have to check out. Thank you for the suggestion!