LPY plots choosing y-limits poorly

The lpy plots seem to be choosing their y-limits so that they don't always include the data on that channel. Maybe the y-limit generator is looking at data too long before the lockloss.

https://ldas-jobs.ligo-wa.caltech.edu/~lockloss/events/all/1238520278/lpy_WIDE.png

https://ldas-jobs.ligo-wa.caltech.edu/~lockloss/events/all/1238520278/lpy_ZOOM.png

I don't know how relevant seeing this data will be to commissioners, but it seems like something we should fix regardless.