In addition to LN2 cooling, an Iwatani ``Cryo Mini'' pulse tube refrigerator (PTR) is used. The main components are the cold head, the compressor, and a rotary valve in between these. The connecting cables used for characterization are shorter than the ones required for use at the telescope. Between the compressor and valve were 3 meters of flexible tube. Between the valve and cold head, a relatively inflexible copper tube was used. Experimental flexible tubes were also supplied, but their performance was found to be poorer than the copper tube, and are not further described here.
Test in external dewar
The performance was first measured with the cold head mounted in a small dewar. On the cold head, a temperature sensor and a electric heating element were mounted. The temperature of the cold head versus electrical heat load is shown in figure 5. An approximately linear relation is found, and the cooling power at liquid Nitrogen temperature is about 9 Watts. As there is also heat exchange by radiation, the actual cooling power must be somewhat higher than indicated by the figure.
A performance dependency on orientation of the cold head was found. This is shown in figure 6, where cold finger temperature is plotted as a function of orientation and heat load. It appears that efficiency is good as long as the cold finger is pointing downwards or is horizontal, but performance is degraded significantly if the cold finger is pointing upwards.
Test in NOTCam
All tests with the PTR mounted on the cold table of NOTCam were performed in the standard storage orientation, where the cold finger is pointing upwards. As seen above, this is the orientation where the cooling power is the lowest.
The purpose of the PTR is to prolong the LN2 holding time when the instrument is already cold. The efficiency was measured in two ways: By measuring the LN2 holding time and by measuring the N2 gas flow from evaporation.
Without the PTR mounted, a LN2 holding time of 44 hours was found. With the PTR running, this increased to 57 hours, a 30% improvement.
The gas flow from LN2 evaporation indicates a heat flow of 10 Watts without the PTR installed and 7 Watts with it. That is, the cooling power of the PTR when operating on NOTCam is 3 Watts. From figure 6, a cooling power of 5 Watts was expected in this orientation.
The inadequacy of the PTR for cooling without LN2 can also be seen from
figure 3, where the PTR is running for a day after the LN2
tank emptied. During the day, the cold table temperature increased by
15 degrees. As shown elsewhere, this would eventually have stabilised at
-170^
C.
In all, the performance of the PTR is disappointing, as it does not greatly increase the LN2 holding time. When mounted on the telescope, a more favorable orientation should increase the gain from operating the PTR.
Although the PTR is not intended to be able to cool the instrument from
room temperature, it has been attempted, as can be seen in figure
1. During the first two hours, the table temperature went from
+22C to +4C, with steadily decreasing rate, and much slower than what is
achieved by using LN2. The outer wall of the instrument is seen to get
warmer, as the PTR head gets quite warm, about 40^
C, and conducts heat
to the vessel.
A PTR cooling experiment of a longer duration, cooling from room temperature
for 58 hours, resulted in the following temperatures:
Cold table: -113^
C, wheel house: -105^
C, detector flange -101^
C.
During the cooling, the vacuum pump was continuously running, and a pressure
of 
mBar was reached at the end of the cooling.
As demonstrated during storage, the PTR should eventually cool the instrument
to about -170^
C.
So if plenty of time is available, the PTR can strongly reduce the amount
of LN2 required to cool the instrument completely.
