Pete Smith spoke:
>Since both numbers are comfortably within spec, this rotator should have no
>trouble turning or stopping this array, and in fact it doesn't. but I'm
>uncertain how far one really can push it, and many people will have more
>aggressive antennas than I do.
Therein the rub... I wonder if the science exists to really manufacture
gears to destructive tolerances against absolute maximums, at least for the
kind of money we are willing to spend.
One can readily take a group of rotor parts, stress them all until they
break, and say that for a given confidence factor, do not exceed x stress.
This works fine for the statisticians, but is suicide for marketing. The
problem is that *one* failure that hits the list-serves is bad sales, and
*every* customer wants his rotor to last forever.
The statistician can fairly well predict where 30% will fail, where 60% will
fail, but anyone who has worked with these kinds of figures will tell you
that one gets into voodoo when you are trying to predict where 0.5% will
fail. Especially when parts are cast.
The best thing is to measure where 1% or so break down, and then don't get
*anywhere* close to that, like one third or one half of that. This is what
the "K" factor smells like to me.
A rotor goes up in the air at a place where people don't want to go, and
they get really peeved when the rotor breaks. So some engineer will make a
marketing vs. profit vs. probability choice and publish a spec that doesn't
have a 30% failure return rate at one extreme, or
"so-expensive-that-we-couldn't-sell-one-for-that-capacity-at-that-price" on
the other extreme.
Bet you there isn't an engineer that will even talk to you about what the
true 95% capacity is, based on whatever acceptable failure rate you might
agree on.
They don't want you cutting it that close. They can't guarantee your
success.
--
FAQ on WWW: http://www.contesting.com/towertalkfaq.html
Submissions: towertalk@contesting.com
Administrative requests: towertalk-REQUEST@contesting.com
Problems: owner-towertalk@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|