AnsweredAssumed Answered

ADXL362 unexpected behaviour

Question asked by Keplen on Dec 12, 2013
Latest reply on Dec 1, 2016 by dhoow

Hello everyone, I would like to use an accelerometer ADXL362 with a MSP430 uprocessor as a low-power pedometer, but even I tried almost all possible settings, I am still having problem with reading data from the accelerometer. The problem is, that I am receiving"corrupted" data, when I start to move more or I place the accelerometer to some positions. In the picture you can see an example. When I start to move with my board, where is the accelerometer, some of the data I receive are corrupted. But when is the board just laying on table, I can read the values of acceleration without any problems. This behaviour is the same for all axis.

I also discovered, that higher measurement range I set the more is the accelerometer sending the corrupted data.

For reading out the values I am using FIFO in stream mode and reading it every 50 samples.Data_error.PNG

The setting of my accelerometer is:

    Write to register FIFO_CONTROL value 0x02 - FIFO in stream mode

    Write to register FIFO_SAMPLES 0x29,0x96 - Number of samples when FIFO Watermark is set is 150

    Write to register INTMAP2 value 0x84 - FIFO watermark to INT2, active LOW

    Write to register FILTER_CTL value 0x12 -  +-2g, 50Hz ODR, Antialiasing filter is set to 1/4 of ODR

    Write to register POWER_CTL value 0x02 - Begin measurement


Reading and writing to/from registers is without problems.

I also have another board with ADXL362 (only board connected with cables to a uP) but the behaviour is quiet similar.

In first case is the acc powered directly from an Ucc, in the second case is powered from uP pin. In both cases is the Uacc ~3V to 3,2V.


Anybody any idea, what might be wrong? What else I can try to check/change?