[evlatests] Another (different) Condensation Problem

[Rick Perley]

    Bob and I went to antenna 24 today to do antenna efficiency, 
spillover, and system temperature tests at K, Ka, and Q bands.  We began 
with Ka band at 33160 MHz, and immediately ran into a curious effect, 
not seen before:

    When we placed the cold load (liquid nitrogen in a styrofoam box) 
over the feed, the the power steadily declined over the course of about 
one minute by about 20K -- a very large amount!  We then removed the 
cold load to look at cold sky, and watched the power drop to the 
expected level, then rise over the course of about one to two minutes by 
about 20K. 

    The same effect was seen at Q-band, although the magnitude was 
considerably smaller.  It was nearly absent at K-band. 

    It was clear from these tests that the problem was associated with 
the weather window.  The outside of this membrane (GoreTex, I believe) 
was dry, although the feed was initially covered in ice and water, due 
mostly to dripping from the subreflector.  We carefully dried off the 
upper surface before these tests commenced.
    We then removed the weather window -- and sure enough, the bottom 
side was completely saturated with water, mostly likely due to 
condensation.  The air in the enclosure between the horn and the RFI 
shield which surrounds it (and which includes the weather window) is not 
vented in any significant way.  Of course, it doesn't help that the feed 
heaters on this antenna are not yet operating correctly. 

    We think what was happening is this:  The styrofoam box is a poor 
insulator (which we can confirm directly -- it was always icing over, 
even in this very dry day).  When it was placed on top of the weather 
window, the condensate on the bottom side froze over about a 1 minute 
time, greatly reducing its emissivity, and reducing the power.  When the 
box was removed, the frozen condensate melted, restoring the original 
(bad) state.  I'll show some plots of this at the next test meeting.

    With the weather window removed, we did the basic hot/cold/sky, dip, 
and Venus efficiency tests.   The results were spectacular (in most 
cases):  Tsys (cold sky) = 41K!  Tr = 15.7.  Tcal = 5.5, which is 
exactly what Bob measured in the lab.  The efficiency comes to 38%, a 
little lower than we hoped for.  (This leads to Tsys/effic  ~ 110K -- 
very close to what Emmanuel has measured). 

    During these sky tests, we dried out the weather window in the sun, 
and reinstalled it.  With a *dry* window in place, there were no further 
time-lapse, 20K changes, and all derived parameters were the same as 
with no weather window at all. 

    So -- we need to think a little about how to keep the air 
surrounding the high-frequency feeds dry, and prevent condensation.