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: cavendish expt



Tungsten or DU (depleted uranium) is even better.

Go to the tire places, as they may give you old weights. I have about
10# of them from picking them up from the lead bushes on my walks for
the last 40 years and lead from old wine bottles. When liquid, one may
remove most of the impurities (the steel clips for example). An easy
form (mold) is a tin can. Cylindrical geometry is not much less
efficient than spherical (Do you recall a quiz a few months ago on the
best geometry?), and is less sensitive to vertical misalignment.

Recently I discussed the advantage of a rectangular cross section for
suspensions. I assume the amplitude used is only a few degrees, and,
therefore, suspect the torsion constant will be constant. [JD! Why do
you think (know) otherwise?] The Leybold uses a gold suspension with a
rectangular cross section. Their new other apparatus uses a metal wire;
I think tungsten.

Strong (Sci. Am. Sept. '63) reports Sam Epstein of LA used # 36 (55"
long) music wire for his apparatus. I suspect it was a PSSC apparatus
that used 1/4" OR (open reel) recording tape.

I suspect (with out much thought) that a large massive apparatus will be
less sensitive to draughts, still must be covered. The commercial ones
are also sensitive to light (IR).

Finally, and I think, most important, one will achieve considerably
greater accuracy by plotting the oscillation and fitting to a sinusoid
with an exponential decay. Then shift the external masses and plot
again. At UCSC the data was collected manually (now they have an analog
output with a capacitor detector.). One student was the time keeper
(1/2 min. intervals, another scribe, and the third called out the
position of the LASER spot on a two meter stick. I've already reported
my use of a camcorder and a VCP (video cassette player) that steps
frames (or is it fields?). The fits reveal the equilibria.

[Another method is to treat it as a ballistic pendulum. When the
pendulum has stopped, quickly shift the masses to the other side and
collect data. The first ~ third of the max. amplitude will be
sufficiently linear (do a fit) to use F=m*a. Note the force is double
that of the g force alone.]

bc



"John S. Denker" wrote:

thomas pfeiffer wrote:

I'm using two joint compound buckets filled with sand
as the larger masses.

Lead would be better.

1) It gets the center of mass closer to where
the action is ... 1/r^2 and all that.

2) All the sand I've ever seen is full of iron
particles. I would be worried about magnetic
systematic errors. It is easy to get Pb free
of impurities.

3) Although lead is more costly than sand, you
don't need huge amounts of it.

I have a meterstick horizontally
oriented with bottles of sand on each end. This is
suspended from the ceiling with video tape.

Videotape seems odd. Monofilament fishing line
is pretty much standard for this application.
Videotape will do bad things to the sensitivity.

I'm finding that it takes an incredibly long time for
the suspended meterstick to find its equilibrium
position.

So add some hydraulic damping.
http://www.fourmilab.ch/gravitation/foobar/

My questions are:

1)Do you think the incredibly small Fg would produce
any effect?

Yes. There shouldn't be a problem due to
lack of signal. The apparatus doesn't have
to be super-tall or super-massive.

> 3) Any suggestions?

It's not hard to find good information on this topic.
http://www.google.com/search?q=cavendish+experiment+air-currents

Note that I didn't simply google "cavendish experiment"
... I added the term air-currents in order to select
sites that discussed the real issues.

The difference between hackers and scholars
is that the scholar checks the literature
before hacking.

=================

It would be amusing to see how small you can
make the apparatus. I'll bet it wouldn't be
too hard to make it smaller than a breadbox.
You can get some very small fibers cheaply
and easily as follows: Go down to the boat
store and get some good Dacron kernmantel
rope. Cut through the outer braid with a razor
knife. Pull. Inside you will find angel
hair, so fine that a single strand can't
be seen unless the light hits it just right.
One strand can only support a few grams. I
haven't measured or calculated the torsional
spring constant, but it must be reeeeally
small.

=========================

Over 350 years ago there appeared a book called
"Discourses on Two New Sciences". One of the new
sciences was the laws of motion. The other one
was scaling laws.

Scaling laws are important. Never pass up
the opportunity to make a scaling-law argument.

It is amusing to see how the sensitivity of the
Cavendish apparatus scales with size. Work it
out. The back of my envelope says if you make
the apparatus 10 times smaller, the raw
sensitivity goes down by only a factor of
sqrt(10). And I don't think raw sensitivity
is the issue -- what you care about is signal
to noise ratio, and your control over noise
sources (air currents etc.) gets better when
you keep it small.

You easily win back that factor of sqrt(10)
by converting from sand to lead, as soon as
the thing is small enough that the mass can
be affordably implemented in lead.