Skip to main content
Associate
May 15, 2025
Solved

IIS2DULPX slope filter and reference value for wake-up threshold

  • May 15, 2025
  • 1 reply
  • 433 views

Hi,

my company is currently evaluating usage of the IIS2DULPX accelerometer in one of the products we are developing. Our use case requires that a wake-up interrupt is generated whenever acceleration sample exceeds a set threshold value. However, we would like to generate this interrupt with a threshold that takes into account either:

  • an absolute value of a data sample, or a difference between a sample and zero-g offset (interrupt generated when sample > threshold - zero-g offset)
  • a difference between a data sample and a configured, constant reference value (interrupt generated when acceleration sample > threshold +/- constant reference value).

We would like to do so while keeping the ODR at 400-800 Hz and the ability to configure threshold as a value in a full +/- 16g range. 

From the data sheet it seems that the wake-up feature is fed by an internal slope filter, making it impossible to use wake-up feature as I have specified above. Therefore, is it possible to either disable the slope filter or achieve the aforementioned functionality with use of one of the advanced features, such as finite state machine or machine learning core modules?

Kind regards,
Robert

Best answer by Federica Bossi

Hi @td-robhyn ,

I confirm that wake up is not suitable for your case, but you can make your own MLC or FSM algorithm, keep in mind, however, that each device has its own offset, each of them will have to be configured with FSM and/or MLC that take into account its specific offset.

1 reply

Federica Bossi
Federica BossiBest answer
Technical Moderator
June 4, 2025

Hi @td-robhyn ,

I confirm that wake up is not suitable for your case, but you can make your own MLC or FSM algorithm, keep in mind, however, that each device has its own offset, each of them will have to be configured with FSM and/or MLC that take into account its specific offset.

In order to give better visibility on the answered topics, please click on 'Accept as Solution' on the reply which solved your issue or answered your question.
td-robhynAuthor
Associate
June 6, 2025

Hi @Federica Bossi,

thank you for your reply. Indeed, we have managed to create an FSM algorithm that would work and evaluated it on the MKI109D eval-board. Unfortunately though, it turned out that the power consumption of the sensor with FSM enabled and running at 400 Hz is too high for our use case (>32 uA at 1.8V).

Kind regards,
Robert