home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.physics
- Path: sparky!uunet!decwrl!world!jcf
- From: jcf@world.std.com (Joseph C Fineman)
- Subject: Re: Order-of-Magnitude problems
- Message-ID: <BtD0BI.9u3@world.std.com>
- Organization: The World Public Access UNIX, Brookline, MA
- References: <1992Aug15.105735.19350@midway.uchicago.edu> <1992Aug17.162216.9332@amhux2.amherst.edu> <18AUG199200260639@zeus.tamu.edu>
- Date: Sat, 22 Aug 1992 00:35:40 GMT
- Lines: 40
-
- Rob:
-
- I wrote this reply several days ago, but am new to Internet, & have taken
- this long to figure out how to post it. In the meantime a good deal of it
- has appeared in other followups, but some of the details may still be of
- interest.
-
- Consider first the *classical* problem of a pencil balanced on its
- point (at 0 K, in vacuo). This is just an upside-down pendulum, and
- has the same differential equation as an ordinary pendulum with the
- sign of the force term reversed. Once again we may make the
- approximation sin(theta) = theta, this time not because theta will
- remain small, but because once it becomes large we may deem the pencil
- to have fallen over. Then the motion, instead of being A sin(OMEGA t)
- + B cos(OMEGA t), is A exp(OMEGA t) + B exp(- OMEGA t) (same OMEGA),
- where only the first term is of interest. Let the pencil be given a
- tiny initial displacement theta(0) and/or angular velocity omega(0) at
- t = 0; plugging in, we find A = omega(0)/(2 OMEGA) + theta(0)/2.
- Hence the time it takes for theta to become of order 1 is about
- - (1/OMEGA) ln{(1/2)[theta(0) + omega(0)/OMEGA]}.
-
- Now the *handwaving* step. It seems plausible that the amplification
- of the quantum-mechanical fluctuations takes place almost entirely in
- the classical regime, so that the latter can be incorporated entirely
- in the constraint they place on the initial conditions.
-
- Finally the *q.m.* Let I be the moment of inertia. Then theta and
- I omega are canonically conjugate, so theta(0) and omega(0) cannot
- both be zero; at best theta(0) omega(0) is of order h/I. Choosing
- theta(0) subject to this constraint to minimize the quantity in { } &
- thus maximize the lifetime, we obtain t = - (1/OMEGA) ln sqrt[h/
- (I OMEGA)]. Plugging in reasonable values, we find that it takes
- about a minute for the pencil to fall over.
-
- If I were Von Neumann, I would supply a rigorous justification of the
- handwaving step. If I were Fermi, I would have intuited the logarithm
- & supplied its argument on dimensional grounds. Perhaps you can do
- something of the sort.
-
- Yours &c, Joe
-