persistent currents

["John S. Denker" <jsd@MONMOUTH.COM>]

>  Does anybody know how such currents are
> started and how large they are, typically?

You start them using a circuit like this:

          _______/___
         |           +++++++++++
         O           +         +
         O           X         +
         |           +         +
         |_______/___+++++++++++
                 s

Where "+" indicates superconducting wire,
X denotes a piece of wire that would like
to be superconducting, but is initially held
above Tc by a heater, the switch "s" is
initially closed, and "OO" indicates a
regulated constant-current source.  There
is a magnetic field in the vicinity of the
three superconducting legs.

Then the heater is shut off.  There will
be some trapped magnetic flux in the
superconducting rectangle, which now has
four fully-superconducting legs.  There
will be no current flowing in the western
leg at this point.

Then the power supply is switched off.  There
is now a persistent current flowing around
the superconducting rectangle.  The power
supply wires can be disconnected at "s" if
desired.

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

Another way to do it would be to make the
superconducing loop be the 1-turn secondary of
a transformer.  Drive the persistent-current
switch "X" above Tc, and use the primary
to drive some flux through the non-closed
superconducting loop.  Then let the
persistent-current switch "X" cool below
Tc, trapping flux in the now fully-superconducting
loop.  Then switch off and withdraw the
primary.

This works at the conceptual level only;  the
previous method is what is used for real
superconducting magnets, because it works
better if the superconducting magnet has
more than 1 turn.

===

>  Does anybody know how such currents are
> started and how large they are, typically?

You can buy superconducting wire that has a
critical current density in excess of 50,000
amps per square cm.  Jc in excess of 300,000
A/cm^2 has been claimed for small samples.

A typical off-the-shelf lab magnet (10 Tesla
with a bore of 5 cm diameter 45 cm long) might
require 50 or 100 amps.

Homework:  Calculate from first principles an
estimate of the inductance of that critter.

> What kind of power supply is used in commercial
> superconducting electromagnets?

The output stage is just a bunch of ordinary bipolar
transistors.

The tricky part is that you can't change the current
very fast, because of all that inductance.  If you
try to change it quickly, especially when the thing is
almost fully charged up (or should I say fluxed up?),
you'll just drive everything normal and a whole lot
of B^2 energy will be converted to heat all at once.
Wheeee!!!

So you need a controller that synthesizes a nice slow
linear ramp up to the desired current.  A microprocessor
plus D-to-A is the only game in town.  If somebody
tries to sell you an analog controller, tell him
you'll call back later.  Much later.