temperature impact over gyroscope zero-level offset

[MSham.2]

Hi,

I'm using the LSM6DS33 MEMS and found that in different temperatures the zero-level offset of the device gyroscope changes significantly.

I recorded several recordings while the device is static, and found that consistently when the temperature increases the gyroscope measurements in the Z-axis increase, in the X-axis decrease and in the Y-axis don't change.

See for example the plot attached of the gyroscope and temperature measurements over 20 minutes recording. The temperature initial value is 27C and final is 45C, the gyroscope Z value starts with -2.8845 and ends with -1.6794, and X value starts with 4.2124 and ends with 2.9963.

The gyroscope values are not the same for each temperature in different trials, but as mentioned - the changes seems correlated to the temperature change.

I'm using 250dps range in gyroscope dps and 8g range in the accelerometer.

The temperature was taken from the device (not measured using any external device).

My questions: Are these findings make sense to you? If so, is there any additional way to correct the gyroscope offset based on the temperature (didn't find one in the device manual)?

Thank you,

Mor.

[Eleon BORLINI]

Hi Mor. @MSham.2​ ,

please note that, on the datasheet, only the typical value of 0.05dps/DegC, that is specified at p.15. This means about 1 dps every 20 degrees, which is more or less the value you are seeing.

 

I can also share you an example of characterization lab measures for a better specification of the Gyro ZRL vs temperature variation across a broader thermal range for this product.

 

Please note however that the LSM6DS33 is an obsolete product, and that the more recent iNemos such as the LSM6DSO are better performing vs temperature.

If my reply answered your question, please click on Select as Best at the bottom of this post. This will help other users with the same issue to find the answer faster.

-Eleon

[MSham.2]

Hi Eleon, thank you for your immediate response!

What are the colours in the plot represents?

I'm trying to understand rather there is a way to know the gyroscope ZRL, or to predict the change by time / temperature.

For example: have you seen in the lab that a specific temperature is corelated with specific ZRL values? If so, I can decide the ZRL to remove from each gyroscope sample based on the measured temperature.

Please let me know if you came across any other way to predict the ZRL values in real time (except from measuring it during a static period before).

Thank you,

Mor.

[MSham.2]

Following my last question: if the colors in the plot represent different trials (N=48), it seems like the temperature is not correlated with the ZRL offset, as you can see that in some trials the ZRL offset increases when the temperature increases and in other it decreases when the the temperature increases.

Another question: does the range of the gyroscope (250 or 2000 dps) effects the gyroscope sensitivity to temperature?

I appreciate your help,

Mor.

[Eleon BORLINI · Accepted solution]

Hi @MSham.2​ ,

the different colors refer to different devices. The values are normalized with the 25°C offset, that is typically different from 0 before the calibration process, but that is close to 0 after it (it is a software calibration, since the LSM6DS33 has no internal register for it, just the reserved registers for the calibration at inline tester level).

There is no clear correlation between the offset level and the offset percentage variation in temperature, at least from what emerges from the characterization.

It is better to refer to the general range of 0.05dps/DegC. You might understand the direction of this gain measuring the temperature drift in a small range as you just did.

About the dependence of the offset from the FS, there is no appreciable scale factor, so you can consider the same value shared among all the FSs.

-Eleon