The unstable Sampling rate for LSM6DSL

[JLee.9]

Hello all,

Our company is trying to switch from BMI160 to LSM6DSL due to it low power nature.

For our product, accurate time representation in the level of ms is important.

However, when we were experimenting with LSM6DSL, we experienced that when we are sampling data, it doesn't represent the sampling we set.

For example, if we set it to 416Hz, we get around 419hz.

We would have to do more experiements but we also see different sampling rates as well(419, 425,423,etc)

Is there something that we are not understanding or is this a given characteristic of this sensor?

Thanks in advance.

[Uwe Bonnes]

Sampling rate is probably determined by an internal oscillator. Some % deviation and jitter are not unexpected for such an oscillator. Can you perhaps trigger sampling from outside? Otherwise, did the BMI160 do really better in timing?

[JLee.9]

Thanks for the quick response.

I never characterized BMI160 because we didn't run in any issues but can do.

Is there a way where I can get the expected deviation?

[Eleon BORLINI]

Hi @JLee.9​ , the behavior you are facing is quite strange... Are you using I2C or SPI? And at which clock frequency of data reading? You could also test the FIFO ODR instead of the device ODR, and then read the data from the FIFO (not continuous mode). You should get a better synchronization. Regards