Safety workaround for hardware interlock design

[ryan-summers]

This PR fixes #3 by adding a check during channel enable to ensure that the interlock thresholds are properly set before enabling the channel.

This ensures that any over-power condition properly trips the interlock.

I've also renamed the input_overdrive to reflected_overdrive in the status structure so that it's clear this is not the input signal, but rather the reflected output signal.

[jordens]

I agree that this should be in finalize_enable() and not in start_enable().
With #45 this should probably not be a fixed value but the measurement of the current power. OK for now.