I have spent the last fifteen years on the production floor and in the field, and I can tell you that most autosampler downtime I have seen did not come from a failed pump or a burned-out detector. It came from a two-dollar glass vial. A cracked neck, a septa coring problem, or a cap torqued wrong can take a whole sequence offline at 2 a.m. and cost you a re-run of forty samples. What follows is a field guide to the failures I see most often, what actually causes them, and the specific numb...