Originally posted by sholtzma Wouldn't that approach have caught the sensor "stain"? The mirror flop?
Not necessarily. Cameras for testers would have come out before the line was running for production, so in the case of the stain the sensors would likely have come from evaluation lots rather than the actual production lot. So in that particular case everything might have looked fine until they received the lot ( or the second lot or whatever) that had the stain on it.
Mirror flop? Maybe, but again using pre-production cameras that used sample or evaluation parts might not give quite the same results as the next lot of parts used when the line starts running. Would not surprise me that part suppliers would hand select parts for evaluation that were rated higher than the "run of production" parts that are actually shipped. My k-5 had about 30,000 exposures on it before I saw the first mirror flop issue, so not something the average beta tester is going to find.
I'm sure beta testing catches a lot of issues but I would think more of the firmware bug variety than actual hardware problems, at least ones that do not manifest until later.