Posted on August 26, 2016 at 23:37Hello,
I am currently trying to do some initial noise analysis of
LSM6DS3 using
STEVAL-MKI160V1 adapter board and
STEVAL-MKI109V2 mother board. The software that I am using for obtaining data is
Unico GUI. The problem is that gyroscopes and accelerometers noise levels are higher than I expected by more than 2 orders of magnitude. Here is how I did:
I first changed a couple of registers for low noise operation;
CTRL1_XL was set to
10(hex),
CTRL2_G was set to
10 (hex) for the lowest bandwidth of the low pass filters.
High-performance bit was set to 0 for the high performance operation for both gyro and acc. This gives me very high noise outputs (see
fig.1 &
fig.2:
Y-axis unit is
deg/h and
mgfor gyros and accs, respectively). To plot these graphs, I saved LSB data into a file using Unico GUI save function and multiplied those LSB numbers by
8.75*1000/3600 for the gyros (full-scale range was
245dps), and by
0.061/1000 for the accs (full scale range was
2g).
I thought these results were very suspicious, so I checked on the register values directly from the registers
OUT_H_XL and
OUT_H_G, and they did not seem to be changing this much, but only by around
10-20 LSBs. So I looked into
FIFO data in
Unico GUI and saved those LSB values from
FIFO, and the results are shown in
fig 3 (only gyros are plotted for this one). They are much less noisy, and the noise level seems to be within the region, although the Z-axis shows a bit strange behavior.
Here is my question. Am I doing something wrong that I see very noisy outputs when I read data from those sensors using Unico GUI? Why are they so noisy, and why do they give me much less noisier results when I read them from FIFO? I am new to STmicro inertial sensors, and I hope anyone could help me with this. Thanks in advance!
Regards,
Max
