Thanks to all who responded (29 in all). Most of them agreed that a
failure of the seals at the top and/or bottom of the magnet was the
cause, and that warming up the outside and possibly tightening the
screws MIGHT allow me to pump it down and recover, if no ice bridges
had formed.
I was in fact able to save the magnet. I continued pumping and
(after a conversation with a Varian engineer) also tried carefully
tightening the screws holding the upper and lower seals. We do
have indium seals on this magnet; it is one of the older ones that
we got to replace a broken R2D2 a few years ago. It was my impression
at the time that the indium seals were considered more reliable,
at least as far as not requiring the periodic pumping of the magnet
that we had been used to with the newer model; perhaps this experience
is evidence of why they were abandoned even if this was so.
On Friday only the upper inner seal screws seemed to move at all
with the very modest pressure that I was willing to apply. Still,
a few minutes after the very slight adjustment the boiloff went
down substantially. I pumped over the weekend, and it continued
to improve. On Monday it was not much different from Sunday, and
I tried once again adjusting the screws; this time the outer lower
ones seemed to move a little. There was no immediate effect, but
later in the day the boiloff was about normal. Today I took it off
the pump, and so far it looks good, though I'm holding on to the
pump for a few days.
I have a new question: what should I do in the future? This was
the result of about 3 hours of operation at -90C (I think). This
is hardly an extreme use, and not that different from what we have
done without problems in the past. We do use a little different
system for cooling than the standard one: instead of a heat exchanger
we use boiloff of LN2 from a vessel with a submerged heater (it's a
Bruker piece, actually). Thus the actual temperature of the incoming
gas is probably lower than when running N2 gas through a heat
exchanger. I don't know whether the failure was at the top or
bottom of the magnet or both, or whether any of the drop in boiloff
during the pumping was due to my tightening the screws.
For the moment we are doing NO low temp work until we have a more
satisfactory answer. I can see several possiblities.
Maybe the probe was defective, and got colder than usual. I don't
think so, but I will check this out. I believe the probe gets
pretty cold at the top even if its dewar is intact, though, so I
don't trust fixing the probe to be a complete solution even it is
bad.
Some people have apparently tried to actively warm parts of the
system during low temp operation, e.g., by wrapping exposed part
of the upper stack in heater tape. I'm not sure that would do the
trick, if the problem is created by the probe cooling the bore tube
through the RT shims. If the probe were longer we might be able
to heat its outer case during low temp operation, but there is only
about 1/2 inch exposed below the RT shims. Insulating the probe
better somehow seems difficult, since there is so little space
between the probe and shims, and between shims and bore tube. Maybe
really high flow of probe "cooling" air, possibly pre-heated, would
be in order.
If this problem is especially severe on R2D2 magnets with indium
seals, then we could move our VT operations to our other 300 MHz
magnet. It is an Oxford magnet, of 1996 vintage, about the same
height as the R2D2 but fatter. Are these Oxford magnets less
vulnerable to low temperature effects? Or do people have to take
special care when doing low temp experiments with Oxford magnets
too?
Again, thanks for the all the advice so far
Feeling much better now,
Steve
--
Steve Philson philson_at_nmr.chem.umn.edu
Director NMR Lab 612-626-0297
Chemistry Dept. University of Minnesota
Received on Tue Nov 18 2003 - 17:22:48 MST