Craig D Mackay, Institute of Astronomy, University of Cambridge, England: 22/05/01.
In case there is a disaster inside the dewar it is important to be able to warming it up, strip it down, effect the repair, put it back under vacuum, and cooled again as quickly as possible. The new dewar is much easier to manage because the vacuum quality is generally very good. If it dewar starts cold then it is important to be able to warm it as quickly as possible. Under no circumstances whatever must the dewar be opened to air with the temperature inside any part of the dewar significantly below ambient temperature. In fact it is only necessary to make sure that all internal parts are above the local dew point. Most observatories have a weather system which indicates the relative humidity as well as the local dew point. On a recent INT run the ambient temperature was 22 C whereas the dew point was only about 1 C. The last part to warm up is the radiation shield and that must be the part checked for temperature most carefully. If air is introduced that has any moisture in it at all then that moisture will condensed on colder spots and can be very difficult to remove. It could cause damage to the infra-red detector chips and it will certainly cause staining on the surface of the chips and on the filters. In addition if there is any residual moisture inside the dewar the vacuum pump will take a geological time in order to remove it.
The fastest and most efficient way of warming up the dewar seems to be to use a supply of dry air under reasonable pressure directed into the very center of liquid nitrogen cylinder through the liquid nitrogen fill hole. At the LPO there is a good supply of dry air at a decent pressure. Small diameter tubing may be used and inserted directly into the dewar. Tests of this method show that very fast warm of times may be achieved. Running at a pressure of 4 bar it was possible to go from fully cold to a point where the dewar could be opened in a time of 5 hours.
It is possible to speed the slowest part of the warming process which is waiting for the heat shield and filter wheel assembly to reach an acceptable temperature. This can be done as follows. Attach to the closed vacuum valve a length of flexible stainless steel tubing. It is essential that the air inside this to must be completely dry. What it is possible to do is to flush it out with the boil-off gas from the liquid nitrogen cylinder used to fill CIRSI. Once you're sure all the damp air has been flushed out, seal the end of the pipe with either a metal plate and O-ring or with one of the plastic cover pieces used top protect these vacuum systems. Then gently open the vacuum valve until the pressures equalizes and then close it again. This process may be repeated three or four times to bring the pressure inside the dewar up significantly. This allows convective heating of the heat shield and filter assembly and rapidly speeds it up. It is also possible to introduce not simply boil-off gas from the liquid nitrogen cylinder but also to include some liquid nitrogen itself. While this is boiling off there is positive pressure inside the hosepipe. By placing a cover over one end and releasing the vacuum valve it is possible to bring the dewar up to virtually atmospheric pressure without introducing any moisture.
Vacuum pumping can be done fairly quickly, but it really should be done for a few hours and include firing of the getter. Cooling can be done to reach 100C in about 4 hours, but the dark current will still come down considerably for a further 4-6 hours. In practice it means that if the system has to be taken off the telescope, as soon as possible start the warm-up procedure, and it should be possible with help to strip it down, effect repairs, pump and start cooling so it is back on the telescope for that night (remember that pumping and getter firing may be done on the telescope) and that while the sytem is still cooling do alignment, focus and standards measurement, without loosing very much of that night.
Prepared by cdm@ast.cam.ac.uk Please contact with comments, etc. Last updated 22 May 2001.