XPost: or.politics, alt.politics.trump, alt.fan.rush-limbaugh   
   XPost: alt.astronomy, rec.aviation.military   
   From: muratlanne@gmail.com   
      
   "R Kym Horsell"  wrote in message   
   news:v62l3f$1r7h$4@nnrp.usenet.blueworldhosting.com...   
      
   According to the contract they had to test them in a certain way the   
   company officials considered a big old waste of time and money.   
   If a fauly was detected the whole test rig was SUPPOSED to be shut down   
   and colled off, then re-started and the controller had to the get up   
   to speed then pass the test. The company method was to just keep going   
   until the controller passed the tests and ignore any failures along the way.   
      
   ------------------------------   
      
   I was in the industrial testing business for a while. "Test until pass"   
   wasn't an uncommon practice, there's enough random variation in repeated   
   test results to somewhat justify it, and the fail limit may be arbitrarily   
   overcautious. It may have been set by Marketing to look good on a spec   
   sheet, and you won't go to jail for setting it too high but you might for   
   setting it too low. When weight doesn't matter much the safety factor is   
   typically at least 3 to 5, while in aviation it may be 1.5. For the steel   
   chain I just bought it's either 3 or 4 depending on grade, I don't know why   
   the difference. Chain rated to hold 6600 Lbs in use is tested during   
   production at 13,200 and expected to break at 26,400.   
      
   In one case I found a timing bug in the test station's computer microcode   
   that caused occasional failures which didn't repeat when the part was   
   retested, so passing it through a second time was the procedure. The test   
   was the problem, not the component. My reward was being assigned to a new   
   product's development team. Unfortunately such small highly specialized   
   companies go broke or at least lay off the engineering staff when capital   
   investment in new equipment vanishes during a recession.   
      
   The experience I gained there in high speed computer controlled measurement   
   positioned me well when the increasing speed of analog to digital converters   
   made digital radio practical. By then I had a stronger background in   
   computer hardware design than the radio engineers.   
      
   The designer of the Japanese Zero fighter couldn't achieve the specified   
   performance without reducing the strength margin to 1.5x the expected load   
   to save weight, and others followed. The pilots were ordered not to dive   
   faster than 350 kts because the lightly built wings would fail at higher   
   speed. He knew because a diving plane had ripped apart and killed the test   
   pilot while he was watching. Improving high speed performance would have   
   added weight and cost some low speed maneuverability, which the authorities   
   considered more important to win a dogfight. British pilots learned the hard   
   way that even a Spitfire couldn't dogfight a Zero, though they were   
   vulnerable at high speed. I read that an F-15 can sometimes outmaneuver an   
   F-16 because the computer controlled F-16 has an angle of attack limiter   
   while the F-15 doesn't, and can be pushed closer to a stall. The pilot is   
   expected to handle the danger.   
      
   --- SoupGate-DOS v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)