Cryogen expiry alarms

[roger smith x294]

ccd-world, 

(In response to Paul Jorden)

At CTIO we have had similar experiences and I was going to write an
almost identical reply to that of Carter/Koorts/Evans at SAAO.

We use a neck tube which goes to the center of the flask so that hold
time is more or less independent of orientation.  With about 18-20
hour hold time, we fill twice a day.

There are occasional human errors or vacuum problems, so we too find
an alarm useful.  When I have the opportunity, I glue a sensor to the
neck tube (on the vacuum side) to sense the presence of gas flow from
the boiloff.  Placing the sensor near the warm end provides very good
detection of loss of gas flow, within a few minutes of liquid nitrogen
expiry, and with no risk of false alarms as dewar orientation, or
thermal load on the cryogen flask change.  I found that (in our
dewars) the N2 flask temperature was too slow moving and varied
strongly with orientation.  If one set the threshold high enough to
avoid false alarms in all orientations, then the response came late
enough that the detector temperature was soon to rise.  Using the vent
gas temperature proved better in our system.

The next level of protection we have dreamed up is to log the
temperatures to a file so we can plot the thermal history of any/all
dewars in real time from anywhere on the network.  This will tell us
time since last fill and hold time, as well as current status, and the
gaps will show when the system /software was down.

I would like to stress the importance of the warning from SAAO about
not refilling too soon after if the Nitrogen flask and/or moleculear
sieve have released water vapour into the vacuum since this will
condense on the detector until it catches up thermally.  Though this
is very rare, this is our most serious human error: it is usually only
discovered after cooling and taking flats and requires another thermal
cycle to fix.  It can cost a full night of observing.

Roger Smith, CTIO