2020-12-21 02:35 AM
Is there a number / measure for the linearity of the LSM9DS1 gyro across the 0-360deg range ?
I mean, how much 1deg measured at 0deg orientation is the same as 1deg measured at 90deg orientation ?
Thanks
Solved! Go to Solution.
2020-12-21 06:14 AM
The angular rate (the output of the accelerometer) is linear across the complete rotation, if the rotational speed is below the device full scale (for example, below +-245dps if this is the selected FS).
In principle, also the orientation angle is linear too. However, between the rate and the angle there is an integration procedure, and there are deviations from the theoretical behavior. The three most important factors are: the zero rate level (ZRL, offset), the noise level and the ODR parameters. The first accumulates during the integration procedure, so the indication is to perform at least a calibration of the offset at the beginning of the rotation. The noise should be less impacting, being on a variable around 0dps average. Last, the ODR determines the timebase for the integration operation, so you should calibrate it according to the speed of the movements you want to detect: for example, use low ODRs for slow movements.
-Eleon
2020-12-21 05:29 AM
Hi @PRebo.1 ,
the gyroscope linearity is usually measured applying a sinusoidal / rotational stimulus to the device (for example mounted on a rate table), and comparing the dps read by the sensor with the dps read by a reference sensor in the same position. This because the gyroscope measures the angular rate (in dps), instead of the angle.
If I'm well interpreting your question, you are asking if this gyro linearity is the same if the device is oriented at 0 degree (with respect to the floor?) or at 90 degrees. Supposing either to use a 3 axial rate table, in the sense that you can give a stimulus on all the sensor axis leaving the device in the same position, or to directly rotate the device, the answer is yes, since the MEMS internal driving and sensing structures are not affected by gravity.
-Eleon
2020-12-21 05:54 AM
My question was:
While maintaining the sensor parallel (0 deg) to the floor, and rotating it on its Z axis, does the angular rate measures linearly across a complete rotation on the X-Y plane ?
2020-12-21 06:14 AM
The angular rate (the output of the accelerometer) is linear across the complete rotation, if the rotational speed is below the device full scale (for example, below +-245dps if this is the selected FS).
In principle, also the orientation angle is linear too. However, between the rate and the angle there is an integration procedure, and there are deviations from the theoretical behavior. The three most important factors are: the zero rate level (ZRL, offset), the noise level and the ODR parameters. The first accumulates during the integration procedure, so the indication is to perform at least a calibration of the offset at the beginning of the rotation. The noise should be less impacting, being on a variable around 0dps average. Last, the ODR determines the timebase for the integration operation, so you should calibrate it according to the speed of the movements you want to detect: for example, use low ODRs for slow movements.
-Eleon