Arduino, Wire, and I2C Part 3: gyro

The SEN-10724 board includes three distinct sensors on it. The linear accelerometer and magnetometer were covered in previous posts. The experience gained in working with those two sensors made it a relatively trivial matter to implement an Arduino sketch to get measurements from the 3-axis gyro.

Set-up of the IC is relatively straight-forward. The defaults are fine in most cases. Because my intent is to eventually couple the measurements from each sensor (the first being the magnetometer and gyro), I configured the data rate to be the same as the magnetometer's default. There is a single register that sets that, the sample rate divisor (SMPLRT_DIV). The datasheet provides a formula (in section 8.2) to derive the desired output rate:

F_sample=F_internal/(divider+1)

F is the frequency (sample rate) in Hz. The internal sample frequency can be either 1kHz or 8kHz. I chose an internal sample frequency of 1kHz to start with, and with the desired 15Hz output rate, that resulted in:

15Hz=1000Hz/(divider+1)
divider+1=1000Hz/15Hz
divider=65

One other piece of initialization needs to be done to take measurements, and that is to set the DLPF_FS (digital low-pass filter/full-scale) register. The "full-scale" part of the register has only one valid value, which is to set the full sale range of the gyros to ±2000°/s. The DLPF portion of the register sets the bandwidth of the filter in conjunction with the internal sample rate. For this test, I picked the value that had the most bandwidth in the 1kHz sample rate. I honestly don't understand what the low-pass filter is being used for, but that's not important right now.

I've made the Arduino sketch available. There's no self-test available, but I can get an idea as to whether the data makes any sense by picking up the board and rotating it around. It seems to work. Also, I can look at the temperature sensor measurements and do a quick sanity check.

According to the table in section 3.1, the reference point for the temperature sensor is -13200LSB=35°C, with 1°C=280LSB (i.e. each bit in the temperature measurement is 1/280°C). The measurements looked fairly stable so I just picked one at random: -15106:

(-15106+13200)/280=-6.8°C + 35°C = 28.2°C = 82.75°F

This seems a tad warm to me, but it's not completely unbelievable.

Now that I've gotten measurements (or at least test measurements) out of each device, it's time to start aggregating the data...