Arrangement A in Figure 1 has been discussed in References [12] through [14].
This configuration permits on to feed the condensed liquid to the bottom of the helium
tank while feeding the boil-off helium to the space just above the condenser. If warm
helium gas is fed into the drop in cooler tube at the top of the cooler, this gas can be pre-
cooled using the pulse tube and regenerator tube before it is pre-cooled using the cooler
first stage. After further pre-cooling by the cooler first stage, the pulse tube and
regenerator tube between the first and second stages can be used for more gas pre-
cooling. Arrangement A is well suited for the pre-cooling of the helium tank and the
liquefaction of helium, even without extended heat exchange surfaces on the first stage or
the cooler tubes [20], [21]. Using arrangement A, we were able to cool down the
experiment and liquefy helium into the tank using the cooler alone. Ideally one wants to
use the configuration shown in arrangement A to cool all of the MICE magnets. The
spectrometer magnet vendor felt that arrangement A was difficult to assemble on the
magnet. As a result, arrangement B was tested for use in the first spectrometer solenoid.
Arrangement B as shown will not liquefy helium. If gas is brought into the cooler
sleeve at the room temperature end as shown in arrangement A, helium liquefaction may
be possible. This was not tested. The disadvantage is that the return gas from the tank is
brought to the bottom of the condenser. When tested in the drop-in cooler experiment,
arrangement B appeared to re-condense about as well as arrangement A. In both cases
re-condensation was achieved even when 1.3 W was put into the helium in the tank. The
sleeve around the cooler the vent pipes and the instrumentation wires going into the
experiment accounted for 0.3 to 0.4 W of extra heat leak that is seen by the cooler second
stage. The tubes and cooler sleeve added an additional 2 to 4 watts to the first stage of
the cooler. On the second stage, vent tubes, the sleeve, and the instrumentation wires
account for 20 to 25 percent of the cooler second-stage capacity at 4.2 K. The tubes and
cooler sleeve reduced the effective capacity of the cooler first-stage by 5 to 10 percent.
As a result of the second cooling test using arrangement B, the first spectrometer magnet
was connected to its coolers using arrangement B.
Magnet 1 was tested in 2008. We found that the cooler second stage was not
connected to the magnet. The tube that connected to the bottom of the tank was plugged
with nitrogen ice, which broke the connection between the condenser and the magnet
cold mass [22]. Between tests, magnet 1 was allowed to warm up to about 65 K with the
coolers running, as a result, the tubes that connected the gas space to the condenser
became clogged with nitrogen ice. (There was lots of nitrogen in the tank.) As a result,
the region around the condenser was at temperatures in the range from 3.2 to 3.5 K,
which indicates that there was virtually no cooling being transferred from the cooler
second-stages to the cold mass. During the magnet cool-down using liquid cryogens, the
tank was allowed to collect liquid nitrogen. The tank was not properly pumped and
purged. As a result, the connection between the coolers and the cold mass was clogged
with nitrogen ice. Ideally, liquid nitrogen should never have been allowed to collect in
the tank. The nitrogen flow should have been stopped when the resistance of the
superconducting coil went down a factor of six. To prevent the connection from the
cooler to the cold mass from being blocked by nitrogen ice, arrangement C was adopted.
The drop in coolers seemed to work about as well using arrangement C as with
arrangement B. We were able to re-condense with up to 1.25 W being added to the liquid
helium tank. As a result, a modified form of arrangement C was used for spectrometer
magnet 2. In magnet 2, the liquid dripped into a manifold, then into the helium tank.

-4-