KeplerCBVCorrector breaks when the module fails

[barentsen]

When you try to CBV-correct the lightcurve derived from a targetpixelfile observed during Quarter 4 on the channel that failed during that quarter, it fails:

https://archive.stsci.edu/missions/kepler/target_pixel_files/0119/011904151/kplr011904151-2010009091648_lpd-targ.fits.gz

This is the lowest-priority issue ever, just recording it here for posterity.