FIxes for improved timeout handling

Spec sheet says 37ms echo pulse maximum. My logic analyzer says it is actually 200ms or on some faulty sensors- infinite. Code now waits for the echo to drop before proceeding.

Update readme

I forgot to update the readme for v1.1.0

Improved out of range handling

It appears that many of the HC-SR04 sensors being sold take significantly longer than the datasheet 37ms to return after and out of range incident (OOR). Some I have take 200ms - some I have lock up forever. This is now handled by:

• pausing the rotation until the interrupt pin goes low
• ignoring one OOR, returning the previous value. This can be configured to more for noisy environments. The old behavior returned the last value forever.

Initial Release

This is my first pass at this, and has been bench tested by not yet used in anger.

Things I plan to do:

1. Eliminate the floating point arithmetic
2. allow a more nuanced control of the polling frequency of the individual sensors.

Ideas and suggestions to [email protected] - put "SonarI2C" somewhere in the subject line.