View Single Post
  #1  
Old February 17th 05, 05:53 PM
ray
external usenet poster
 
Posts: n/a
Default

Steve wrote:

>
> People just do NOT understand production tolerance problems, it seems.
> Of course, many THOUSANDS of them ran perfectly! The problem was that an
> unacceptably large percentage of them were assembled out-of-tolerance,
> and had piston slap and/or swilled a quart of oil every 600 miles. The
> fact that thousands of others run over 200,000 miles without a hiccup
> doesn't mean that the problem didn't exist! It just means that its a
> PRODUCTION problem, not an inherent DESIGN problem.
>


It's both. If you're designing parts and don't take into account
production tolerances.... you're gonna end up with crappy results.

Just because your new design makes 10 more horsepower or weighs 10
pounds less means nothing if it costs the company $100 extra to assemble
it and it's impossible to repair.

(Even robots on assembly lines can go out of spec.)

I guess it comes down to the same stuff like in computers:
is 99% good enough? 99.9? 99.999?
1 million engines with 95% problem free: 50,000 "bad" engines.
1 million engines with 99% problem free: 10,000 "bad" engines.
1 million engines with 99.9% problem free: 1,000 "bad" engines.
It all depends who's in charge - the beancounters - and how much money
it costs to go that extra mile... because I'm sure the cost to go from
99% to 99.9% isn't a nice linear graph...
Ads