Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-l] Another uncertainties question...



At 11:58 -0800 1/25/08, Bernard Cleyet wrote:

This reminds me of specification in mechanical drawings I've seen (and
written), e.g. 2.010 " +0, -.002 Why not 2.009 +/- 0.001? No
idea. In my case it musta been having written 2.010, not wanting to
cross out after discovering what I later wanted, or 2.010 is ideal but
no more and 2.008 is OK?????

A specification tolerance is different from a measurement uncertainty. For example, if you have a rod that must fit into a hole, you'd better make sure that the rod will be at least no larger than the hole, so you specify a maximum size for the rod, but allow it to be smaller than that, and a minimum size for the hole, and let it be larger than that. The requirement is obviously that you can get the rod into the hole ant that it won't wobble once fitted.

There's the famous story of when the first atomic bomb was being assembled at the Trinity site in New Mexico in 1945, and the plutonium pit wouldn't fit through the hole in the implosion casing that was left for the final assembly. At first they thought it had been improperly machined (due, presumably, to being improperly specified), but later they realized that the two pieces were at different temperatures and once they were allowed to get to the same temperature, the fit was perfect. If the pit had actually been machined too large, the test would have been delayed, possibly by as much as a month, because of the dicey weather situation, the test wouldn't have been completed in time for the Potsdam conference at which Truman was planning to use it to intimidate Stalin, the bomb wouldn't have been able to be dropped in August (at least the Nagasaki bomb, but if we had used the Hiroshima bomb on August 6th we would not have had any others in reserve, since we had only enough enriched uranium at the time for one bomb), and the whole course of history might have been changed. (This last part is obviously idle speculation, because the course of history could just as well have not been changed by all this.)

Hugh
--

************************************************************
Hugh Haskell
<mailto:haskell@ncssm.edu>
<mailto:hhaskell@mindspring.com>

(919) 467-7610

Hard work often pays off after time. But Laziness always pays off now.

February tagline on 2007 Demotivator's Calendar