If I read your post correctly, the plate count abruptly dropped from 13k to 3k at injection #50, right?
Take a look at the peak shape. If it abruptly changed at the same time, and especially if several peaks in your samples show the same problem, something is wrong at your column inlet. Could be an air bubble (in which case it will go away all on its own), a partially plugged frit, or a void space at the head of the column.
Next question: over what period of time did your run those 50 injections (a day, a week, a month . . .)? At pH 7.5 you're near the upper end of the stability range for that packing, so if the column were in that mobile phase for an extended period (several weeks out to several months), development of a void space would not be surprising. If the problem occurred in a short time (say a week or so), then a partially-plugged frit is the most likely culprit. You should be able to backflush that column (check with Thermo first to be sure!!!), and that may fix the problem.
