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: Bar magnets

• From: "John S. Denker" <jsd@MONMOUTH.COM>
• Date: Mon, 19 Feb 2001 10:25:37 -0500

At 09:29 AM 2/19/01 -0500, Michael Edmiston wrote:
Students ask (each year) if they can measure attraction
rather than repulsion. I say, "Sure, give it a try." Of course no one is
successful. As far as I can tell (both theoretically and experimentally) it
is not doable.

What theory is that?

If the torsion wire is stiff enough to allow a stable
equilibrium for attractive forces, then it is not sensitive enough to yield
a meaningful measurement. If it is weak enough to allow sufficient
sensitivity for a meaningful measurement, you simply cannot attain a stable
equilibrium.

Using only modest resources, this is quite fixable.

Consider the following apparatus:

. TTTTTTTTTTT turntable
. |
. |
. |
. |
. _ | _
. / \===m===/ \-----[-]--
. \_/ \_/ needle in cage

That is, it is a pretty conventional torsion balance except that
a) The torsion fiber is hung from a turntable, and
b) A needle is added to the end of one arm. Add one to the other arm
for symmetry. The needle goes through a tiny hole. This prevents any
large movements.

In operation, apply whatever electrical force you intend to
measure. Perform a _null measurement_ by rotating the turntable to create
a torsion that just nullifies the force and pulls the needle off the wall
of the cage.

This may cause the needle to snap from one wall to the other, forcing you
to move the turntable back part way to unstick it again. In the worst
case, there is hysteresis, but even then you get an upper and lower bound
on the force -- and when you couple that with the appropriate near-case and
far-case distances, things get even better. In a professional lab (as
opposed to a student lab) you would just put electronic feedback around the
whole thing and get rid of the hysteresis entirely.