LSM6DSL, LSM303AGR, LSM9DS1 temperature dependence

[sergeyvslnko]

Posted on July 07, 2018 at 09:55

I have

conducted little research of temperature influence on accelerometr and gyroscope readings.

The sensors were cooled to a negative temperature on celsius, and then heated under the IR lamp to approximately 50 degrees in complete rest without movement.

And I have noticed some strange thing. On the graph of the dependence of acceleration and angular velocity on temperature for LSM6SL present nonlinear jump. I want to get explanation of this effect from ST developers.

PS LSM9DS1 has such effect in in a more less degree. LS303AGR hasn't such effect.

#lsm303agr #lsm9ds1 #lsm6dsl #accelerometer-gyroscope-temperature-dependence

[Miroslav BATEK]

Posted on July 13, 2018 at 08:10

Can you please provide more details, how did you perform the test (equipment, timing, ...)? Especially how did you cool down the device? 

The behavior could be caused by water accumulation during the test.

[sergeyvslnko]

Posted on July 13, 2018 at 10:59

The board was cooled down in freezer compartment to approximatly -5 degrees Celsius, after that it was mounted at heavy metal substrate on by means of racks with magnetic fasteners under IR lamp with reflector. Also, the board was fenced on all sides by walls to prevent the movement of air masses.

Heating lasted about 15 minutes (see graph) At the same temperature at which the drop occurred, braking is observed in the growth of temperature

Condensation of water droplets actually occurred, but how does this affect the difference in the readings of the accelerometer and the gyroscope?

[Miroslav BATEK]

In this case it is confirmed that the problem is induced by the water (humidity). The package humidity absorption and the consequent humidity reduction due to T increase induces a mechanical stress that modify the package forces equilibrium, thus inducing stresses that affect the MEMS.

This change is transduced as a change of digital output.