Of the seventeen respondents, five said that a long stinger could certainly be
problematic. Two of these warned that the stinger must be "in the can" or
within the "helium chamber" of the Dewar. Four respondents had documentable
bad experiences with long stingers. To quote a colorful one:
>I think your hypothesis about your stinger being too long is
>probably correct; I once encountered a similar situation when
>I was trying to configure an old system so that I could just
>wheel the storage dewar on a ramp up onto a raised platform.
>To minimize the height of the platform I tried increasing the
>length of the stinger I was using (haven't used one of those in
>a long time, either!) and the transfer efficiency just went to
>crap.
Unbeknownst to me at the time, a Varian Service Bulletin dated 3 March 1987
states:
>Determine the maximum length of stinger allowed for the storage Dewar side of
the
>transfer tube. ... [ which should be less than ] the full fill height on the
>height/capacity chart attached to the Dewar.
Varian's new 6.1B System Administration manual, page 320 states:
>Caution: The stinger on the storage container side of the transfer tube must
not be
>longer than the height of the LHe chamber (distance C). Exceeding this
maximum
>will cause filling difficulties and can cause a quench.
One spinlander has seen no evidence for a long stinger being a problem when
filling from a large Dewar. To quote:
>I don't have any idea what the problem is, but I don't think it is
>the stinger getting warm. I fill 4 magnets from a 100L dewar.
>Three are in one room, and the fourth is in a different room, on a
>different floor, in a different wing of the building. That magnet
>is sometimes filled last, sometimes first. I have not noticed any
>real difference in the fill time. I usually almost empty the dewar
>(10 L or so left) so there is plenty of exposed tubing.
These apparently contradictory statements are not disparate; all parties are
correct. A 100 liter Dewar has a helium chamber depth of about 63 cm, and
using a stinger as long as that would not be problematic. 30 and 60 liter
Dewars have depths of about 33 cm and 40 cm respectively (Cryofab). Clearly
our error was using the 48 cm stinger and not raising the Dewar more.
Three respondents suspected a problem with vacuum in the transfer line. Two
suggested reevacuating the line. One had an experience with a leaking tube:
>We had a similar problem for a while... in fact, except for the length of
>stinger issue, it was identical. We believed that we had a microscopic
>leak somewhere in our transfer line which only became evident after a few
>minutes of transferring. We speculated that the microscopic hole was
>maximized when the cryogenic cooling (due to liq He flowing through the tube)
>allowed the external sheath to warm up. Thus, it could hold enough of a
>vacuum to allow transfer and to appear warm to the touch, but still prevented
>a truly good vacuum. WHile we figured this out, and tried to find out where
>in our area we could have it repaired, we simply performed our fills a little
>slower and more controlled.
>
>After a few fills, our problem "went away" by itself at about the time we
>prepared to send the transfer line to get repaired. We don't know what
>changed, but we no longer need to fill slowly. Perhaps our filling technique
>is slightly altered? (If so, I don't know how.) Perhaps something physical
>happened the transfer line? (Perhaps something as "unbent" or perhaps dirt
>got in the microscopic hole?)
With respect to our 0.25 Torr evacuation, Marc Stchedroff writes:
>I would be looking for 1x10-2 or 1x10-3 Torr. 0.25 Torr is soft!
Jerry Hirschinger adds this sage advice:
>We monitor a transfer line's condition not only by feeling for cold, but
>also by how long it takes to precool. The old Varian fixed lines should
>precool in only 15 or 20 seconds. Larger flexible lines for modern magnets
>should probably take less than 90 seconds to achieve liquid transfer at a 6
>lpm rate. Due to cryopumping, a soft line might not feel cold depending on
>when you feel it. Another way to watch for a soft line is to see if the
>line sweats after the fill as it is warming up and the cryopumping stops.
>We use a turbo-molecular pump for both our transfer lines and our dewars.
Ice blockage was the concern of two respondents. To all indications here,
however, ice is not the problem.
Insightful Valerie Robinson suggested that I consider the gas-tightness of the
fitting between the transfer tube and stinger. If that fitting is not snug,
then gas could be shunted into the tube.
For correcting the problem, four respondents said that they do or have raised
Dewars to accommodate filling. Heights for raising ranged from 9 inches to 5
feet.
...
Today, we finished constructing a 36 cm platform and filled WITH SUCCESS!
We added 6.9 liters using 13.0 liters which is 1.88 liters used for each liter
added, which is our best ever. The time to fill was 8 min 40 sec which is our
fastest rate yet at 1.50 liters per minute ... with our gas flow rate never
exceeding 5 l/min and pressure of 1.3 psig.
Please don't inform the safety police. Two of us lifted the 122 pound Dewar
onto the platform without kidney belts.
Now that we can fill with confidence, we hope to extend the time between
fillings to save both the cost of helium and the possibility of contamination
from frequently opening the magnet.
Thanks (in chronological order) to:
Gary Strahan <strahan@mmiris.umaryland.edu>
Jeff Walton <jhwalton@ucdavis.edu>
Steven K. Silber <ssilber@nmrsun2.chem.tamu.edu>
Tom Pratum <pratum@u.washington.edu>
Bill Thurmes <thurmes@displaytech.com>
John Wright <jmw@chem.ucsd.edu>
Gregory Leo <GLEO@prius.jnj.com>
Charles Amass <camass@smith.edu>
Xiong J (Jincheng) <xiongj@ucarb.com>
Robert (Dave) Scott <scott@iastate.edu>
Daniel W. Bearden <bearden@manta.chbr.noaa.gov>
John Oatis <oatisj@musc.edu>
Karen Ann Smith <karenann@unm.edu>
Alan Kook <meadhbh@io.com>
Marc Stchedroff <Marc.Stchedroff@inorg.lu.se>
Rick Wintermute <rdw@interaccess.com>
Valerie Robinson <robinson@chembio.uoguelph.ca>
Elwood Brooks <Elwood.Brooks@uc.edu>
Roger A. Kautz <rkautz@lynx.dac.neu.edu>
Mark Stevenson of Varian
Jerry Hirschinger <hirsch@purdue.edu>
Andy Myles of Varian
Patrick Hays <STRLLIB@aol.com>
Don Barrett <barrettd@cofc.edu>