[LIS3DH] Axis data jumps at some temperature when temperature changes

[Chelsea]

I did the temperature test and get the below graph. The Axis data always jumps at some temperature between 10-30 ℃.

Settings: 2g scale, high resolution mode and ODR is 100Hz.

Did anyone else have encounter such problem? The sensitivity stated on datasheet is 0x01%/℃.

[Eleon BORLINI]

Hi @Chelsea​ , which device (part number) are you testing? Regards

[Chelsea]

I am testing LIS3DH sensor ​

[Chelsea]

Hi, @Eleon BORLINI​ , I am testing LIS3DH sensor ​

[Eleon BORLINI]

To better understand the issue... do you perform your test starting from low temperature (-5°C) and slowly increasing the temperature itself, or did you start from 25°C, where the jump is more evident, going up and down in different times? How many data did you acquire for each graphical point, and do you have the temperature vs time graph, synchronized with that data graph? And, last question, which is your DUT Vdd and VddIO?

Btw, please consider that the 0.01%/°C Sensitivity change vs temperature value declared in the datasheet is a typical value valid for FS 2g, high performance mode selected (proper ODR) and Vdd = 2.5V, because the factory calibration is performed in these conditions. Regards

[Chelsea]

The testing process is listed below:

1. We set the sensor to temperature mode in room temperature, and put into oven and heat to more than 70°C
2. In the code,
   1.  it will enter recording mode after the sensor temperature reaches 63 °C
   2. After all testing samples start recording, we slowly decrease the temperature to -20°C in 3 to 4 hours
   3. The recording stops when the sensor temperature below -10°C.
3. For each data points [-10, -5, 0, ..., 55, 60], it only recorded once when the sensor temperature first reaches the target temperature.
   1. Axis data sampling is performed every 50ms, and we buffered 32 samples to get averaging axis value.
   2. Every 200ms, we will read temperature once. And as long as the temperature reaches recording points , it will count down for 12 seconds and store the axis data.
4. After the recording finishes, we move the samples out of oven and read out axis data from memory.
5. We do not have a real-time recording for temperature vs time.
6. DUT VDD and VddIO are all fixed at 3.3V by a booster.
7. Sensor configurature: FS 2g, high-resolution mode (12-bit), 100Hz ODR, I2C connection with 10K pull up resisters.

Actually, we just wonder why there is a so big gap within the "common" temperature range - usually happens within 5°C change between 0 - 30°C. The axis data shows the same trend (incremental or decremental) when temperature increases.

We have tested for 20 times, and all DUT have exactly one gap at individual temperature range.