March 9th, 1995
Automatic ice-maker servicing
Dear Dr. Shapiro,
We have a venerable Oxford 1.89T/31cm horizontal bore magnet [model 80/300, project #B26694]. It has been brought down only twice, once in '88 when it needed to be pumped (TAMU Newsletter 367-15) and in '90 when it swallowed a peristaltic pump (ibid. 388-47). So, we take its operation very much for granted and are surprised when it decides to call itself to our attention.
In retrospect, the sound associated with filling the magnet with liquid nitrogen had acquired a "whistling" quality. After this subtle change had been ignored a few months, the whistling became less subtle. Liquid squirting out of the fill port upon disconnecting the inlet hose after a refill prodded us to act.
Everything had the classic signs of a plug in the outlet of the nitrogen chamber. [The big knurled "nut" holding the liquid nitrogen level sensor adjustment circuits was difficult to turn after five years, but we have a very strong visiting scientist (Adolf Feinauer) in the lab.] We found water ice in the exhaust port blocking about 75% of the opening a short way down. Jim Carolan did not think we could chip the ice and; blow it out with nitrogen from the other port because things would be too cold but he thought the ice could be knocked into the nitrogen without any harm. Unfortunately, the wire for the nitrogen level sensor was imbedded in the ice which complicated the ice removal.
We connected a 1/4" copper tube to house vacuum (generated by a mechanical pump in the far recess of our building) through Tygon tubing, and warmed the end with a heat gun. Then, the warm end was inserted it into the plug and the Tygon was unpinched to pull some vacuum. When it was retracted, there was a solid plug of ice in the end of the tube. Two repetitions poked a hole through the opening, knocking down the remaining ice onto the connector for the liquid nitrogen sensor. We retrieved most of that ice on the next try. The last bits came flying out when liquid nitrogen was refilled.
So, how did the ice form? Our setup is the simplest possible; flexible tubing venting to atmosphere on both ports. The port that had the problem is the exhaust port during filling and its hose freezes and breaks once in a while. We did not always cut the replacement hoses long enough for their ends to droop below horizontal, thinking that water from the ice that forms on the hose always froze before it got close to the magnet. Evidently we were wrong. Another possibility is that water leaks through the stack from the iceball which forms even in dry New Mexico. A positive pressure setup would prevent this problem but we like the simplicity of venting to atmospheric pressure. If the problem recurs in five years, we will reconsider this view and write another note to this publication.