Retry focus position when solution is pending, is before pass1 minimum, and HFR got worse

Small update to Linear Focus algorithm. One the 2nd pass, when approaching the pass1 minimum, and having already achieved a within-tolderance HFR value, previously the algorithm stopped as soon as a measured HFR was worse than the previous measurement. This update retries at the current position a few times in that situation, in case the worsening was just HFR noise.

Updated the unit test.

Tested successfully for several nights.