BMI270 Measurement Range

[RW199]

Hi,

I have purchsed an inertial measurment unit from MbientLab which incorporates the BMI270. 

I am using this sensor unit placed within a concussion model to detect the angular velocities reached during a concussive whiplash motion.

I am using the maximum range quoted of 2000 degrees per second, which the model comfortably reaches during the first movement (+2,000dps) of the head forward. The issue is, in the reverse movement of the head backwards, the angular velocity only goes to zero / slightly negative, when I expect it to reach close to -2,000dps (shown in Image attached, blue line is data from sensor, red line is the plot I would expect if working properly).

Is this due to the BMI270 only being able to measure -2,000 dps starting from the positive 2,000 dps peak? (Positive peak to negative peak occurs in 0.05 seconds). Or, should the BMI270 be capable of detecting the +2,000 peak quickly followed by the -2,000 peak, suggesting something else in my experiment design is at fault for these faulty results?

Thanks,

Robert 

[FAE_CA1]

Hi,

Thanks for your inquiry.

Maybe the concussive whiplash motion is supposed to be like the graph in your attachment. Here is how to verify if BMI270 gyro can detect +/-2000dps angular velocity or not. You can put your device on a table with BMI270 z axis up or down. And then you can use your two hands to move your device quickly back and forth. After you plot BMI270 gyro data you should be able to see sine waveform with magnitude of +/-2000dps. You can do the same thing with x axis up or down and y axis up or down.

Thanks.

[RW199]

Thanks for your reply @FAE_CA1 ,

I tried this out and achieved a sine-like graph with peaks at +/- 2000 dps as you described... However, it seemed consistent that everytime I repeated this, it took ~ 0.2 seconds for the angular velocity to go from positive peak to negative peak (reading @ 800Hz). Again, quoting I need the unit capable of reading peak to peak within 0.05 seconds, and as I'm unfamiliar with specs of these units, is there any other parameter which could affect how quickly it can read from peak to peak? Of course I understand in my test, I may not have been rotating the unit fast enough but just seemed coincidental to me how similar the peak to peak times were.

[FAE_CA1]

Hi,

With 800Hz sampling rate on BMI270 gyro data you should be able to capture the dynamic angular motion of your unit nicely. It always takes time from positive peak to negative peak or vise versa, because there is a 0dps point when turning back. I don't think it is possible in reality to make positive peak 2000dps to negative peak -2000dps in 0.05s.

Thanks.

[RW199]

@FAE_CA1 "It always takes time from positive peak to negative peak or vise versa, because there is a 0dps point when turning back" - true, like in the image shown for y-axis data; peak to peak with a 0 dps in the middle. If I was to do an error analysis regarding this, what spec of BMI270/reason would you say makes this plot unattainable for a 0.05 s duration. 

EDIT: Just wanted to ask about the effects of poor RSSI, was hovering around 80dbm as sensor is placed within a solid skull and maybe 1-2m from connected phone. Could this affect data, accuracy etc.?

Thanks.