Updated PolledBno055

[bparks13]

This update refactors the PolledBno055 device so that it no longer uses the HighResolutionTimer Juce class. Now, it runs on a regular Thread class and utilizes std::chrono::steady_clock to count intervals to ensure that the timer is not shared across multiple sources.

This PR also takes greater advantage of the new multi-byte read functionality to minimize the number of register reads while polling. Instead of reading two bytes at a time to get the full word for each value, now it reads four bytes at a time and then separates the higher and lower bits to create the individual values.

Testing shows that setting the sample rate at 100 Hz was able to achieve this rate, on average. There is some jitter in the acquisition, but the mean value is very close to 10 ms between successive timestamps.

[aacuevas]

Since you are adding locks to many things now, I'd say to be consistent.
You are adding an optionLock to read options but not to write.

Also, in truth, optionLock and registerLock should be the same, they both access the same stream

[bparks13]

@aacuevas Thanks for the review, I have consolidated the option and register lock into one object. The lock for writing options was already present in the setOption method, but now it uses the register lock for consistency with the stream that is used for communication.

[jonnew]

Brandon and I looked at the behavior starting from scratch on a different dev machine with OEPS-supplied hardware and the behavior looked good.

Changes were committed and tested.