Hi,
We are using LSM6DSL for one of our products and in that we have noticed that the Z-axis Accelerometer deviates due to change in temperature for a stable orientation as shown below which is affecting pitch of the product by -1 to -2 deg.
The IMU is placed near a LED driver which changes its brightness from 20% to 100%. The layout image shows the distance (7.5mm shown in thin silver line) between the thermal vias of the LED driver(which is in bottom layer) and IMU (in top layer). The entire board is potted with a thermally conductive epoxy material.
As shown in the graph,
1. Initially from 0 to 20 sec it is in 20% brightness (LED driver dissipates less heat) at 130 milliG.
2. At 20sec, we switch the driver to 100%. We see that the due to sudden change in temperature by 1 deg the Z-axis accelerometer immediately shifts down by 30 milliGs (130mG to 100mG) and then slowly tend towards to its original value.
3. At 225sec we switch the driver back to 20% and the Z-axis accelerometer immediately shifts up by 30 milliGs (115mG to 145mG) and then slowly tend towards its original value.
From temperature graph we can see that the max change in temperature, that too over the entire period not immediately is 3.5 deg (25 deg to 28.5 deg) and according to datasheet
There is another scenario where just 60mV noise on my 3.3V was fluctuating the Z axis Gyroscope from FA to 05 in just couple of minutes and the fluctuation would increase as time went on. This was resulting in high fluctuation in my roll calculation. we solved it for time being by using Gyro in-built low pass filter but this behavior is not expected from such low noise level and multiple filter caps.
If someone can help me with what could be the reason and how to resolve it.
