I have been following the discussion concerning damage to magnet cryostats
by flying tools and other objects. Below, I have crafted some comments for
general consideration. This was going to be a short response, but a
description of the internals of a cryomagnet is tough to convey concisely
via text.
After many years of operating these beasts, I have yet to see a hole --
large or small -- punched through the outer cryostat. We have a mix of
Oxfords and Varian R2D2 magnets. Of the two types, the R2D2 type is
fabricated from a lighter gauge metal (probably spun aluminium). I would
expect it to be more likely to be punctured by a driven edge or point --
such as a screwdriver or crowbar end-on.
Modern cryomagnets are suspended to minimize heat transfer, thus the longer
hold times to that we have all prefer. The suspension points (in tension)
are at the circular welds in the helium stacks on the top of the cryostat.
All internal parts hang from these points, including the various radiation
shields, the helium and nitrogen cans, and the magnet itself. The bottom
of this complex assembly is aligned coaxially about the room temperature
bore tube by a set of fiberglass locating rods (Oxford) or monofilament
lines (Varian). These some of these parts are in compression others in
tension. The alignment assembly is at helium temperature and quite brittle.
The more subtle type of damage is due to Newtonian reaction forces. When a
magnetic object (say a steel hoist, cryogen dewar or crowbar) is near the
side of the magnet, a significant counter force is required to hold the
object at a static distance. There is an equal and opposite force on the
magnet pulling it toward the object. This force is applied directly to the
alignment assembly. If the external object is close enough to the magnet,
the force to pull it away is large enough to overstress the alignment
assembly. That assembly can fail under load. The helium can now swigs to
one side and touches a radiation shield or the bore tube. The result is
high helium boil off due to the resulting thermal short circuit, but no
external evidence mechanical damage. Often this effect produces an
unusually low nitrogen boil off rate. This is because the helium can is
cooling the nitrogen can.
The take-home lesson is that the nearby magnetic object does not have to
collide with the magnet to cause significant trouble. It is sufficient
that it be close enough that the force required to pull it away is more
than the alignment assembly can sustain. A relatively small object under
high acceleration away from the magnet can also exert the necessary force.
The helium can may tilt far enough to the side that one of the welds above
takes a "set" or even cracks. Occasionally, the weld does not set, the
helium can swings back into position, and spectra can be obtained for a
while with noticeable audio side bands present. Inevitably -- usually the
next time a nitrogen fill is done -- the magnet swings toward the liquid
nitrogen dewar, and a sudden increase in helium boil off is seen. In
either case, the magnet must be decommissioned to replace the alignment
assembly. Broken rods will be found after field disassembly of an Oxford
unit. See the PS concerning field repair of an R2D2 type.
--Bob Santini
PS -- Concerning the original posting that described an R2D2 magnet that
was hit hard enough to produce dents and a puncture in the outer cryostat,
I would be concerned that the alignment assembly was stressed when the
object was pulled away suddenly. It is possible that the failure modes
described above will occur later if the alignment assembly was damaged but
did not fail outright. If strange things are observed during a subsequent
nitrogen fill, this is the probable reason. If this is the case, I believe
that the magnet cannot be easily repaired in the field. The R2D2 dewar
must be cut open to access its internal parts.
Robert E. Santini, Director of Instrumentation
Department of Chemistry, BRWN-1393
Purdue University
W. Lafayette, IN 47907-1393
E-mail santini@chem.purdue.edu
Voice: (765) 494-5227
Facsimile: (765) 494-9388