> Is it theoretically possible
> to have a spherical magnet that has the north pole
> on the outside and south on the inside?
> I wanted to make make a toy with a ball that
>hovers over a magnet that has north up.

Theoretically, No
Actually, almost.

Long ago it was proved that it is impossible to achieve stable levitation
with static magnets - it can be shown that you need a dynamically variable
magnetic field to achieve this.

Fortunately, for practical purposes, the laws of Physics are wrong :-) (but,
y' canna break the laws of Physics).
I have also intended to try exactly what you have described but haven't
tried it. Since I thought of it I have talked to someone who says that they
did essentially that, and that it worked OK. He generally tends to have done
what he claims and magnetics are his special area so I guess that means its
possible .
It certainly makes sense that it can be done but the levitation height will
not be vast.

The Levitron top (search for Levitron - some very interesting stuff out
there) achieves very long term stability (tens of minutes) under very
specific and carefully arranged conditions.
In this case mechanical rotation with gyroscopic affects helps.

I have also thought that a mechanical arrangement where motion moved
material with variable magnetic susceptibility MAY be able to assist
stability (no, not perpetual motion :-)).

I have seen an extremely effective demonstration that achieves most of what
you describe.
A superb and provocative "toy"

Dimensions are arbitrary for the purpose of this description.
Take 2 circular ring magnets - say 2cm dia and 6mm dia hole.
Magnetise these so the outer circumference is all one pole (eg North) and
the inner circumference is eg South. You will NEVER find ring magnets
magnetised this way unless they were aiming to do what I'm about to describe
or similar.
Place ring magnets on a 6mm rod about 50mm apart.
Glue a dead PIC to the base (this is the PICLIst after all).
Arrange a number of short bar magnets with all North poles pointing upwards
in curved rows on a base - see ugly ASCII diagram below.
Ring magnets sit in "wells" between the two rows of upwards pointing North
poles - sink too low and the repulsion becomes excessive.
"Walk sideways" and they must climb out of the wells so it tends to be end
stable.

It ALMOST works.
The model I saw had to have mechanical assistance.
One end of the rod was sharpened to a point and rested against a mirror.
The rod levitated statically but had a net lengthways force against the
mirror.
Still extremely effective.
The rod and ring magnets could be spun up by hand and then remained spinning
for many minutes while resting with its point against the mirror.
Spinning is not required for static stability (with tip against mirror)
I imagine that a fully successful free floating unit could be built this way
with more work BUT 1) Theory says NO 2) The fact that they didn't;t suggests
it is too hard.


Side view

    |        N                            N             <- ring magnets
    |        S                            S
    |<XXXXXXXXXXXXXXXXX       <- Rod
    |       S                            S
    |       N                            N
    |
       N       N                  N       N        <- base magnets
       S       S                  S       S
___________________________


End view

                N
                S
          NS O SN
                S
                N

          N          N
              N   N


A TRULY stable levitator can be achieved using an elctromagnet plus a hall
sensor which works to keep the field zero as a magnet approaches. I won't
describe it in detaiul here but the electronics are simple. This essentially
simulates the classic Messnier (spelling?) effect levitation demonstration
which occurs when a magnet approaches a superconductor.



RM



_____________________________
What can one man do?
Help the hungry at no cost to yourself!
at  http://www.thehungersite.com/

-----Original Message-----
From: Brian Kraut <engalt@EARTHLINK.NET>
To: PICLIST@MITVMA.MIT.EDU <PICLIST@MITVMA.MIT.EDU>
Date: Sunday, 26 September 1999 15:24
Subject: <OT> MAGNETIC FIELDS


>This bring up an idea I had a while back.  Is it theoretically possible
>to have a spherical magnet that has the north pole on the outside and
>south on the inside?  I wanted to make make a toy with a ball that
>hovers over a magnet that has north up.  The magnet for the base would
>be dish shaped to keep the ball floating in the center.  I tried to cut
>up a bunch of small magnets a long time ago and glue them to a ping pong
>ball with the norths out, but it became a project that was never
>finished.
>