[evlatests] EVLA Antenna Feed Moisture Condensation Conflaguration

[Bob Broilo]

Chuck says:
> moisture in the very bottom of the feeds. No moisture was observed at or
> near the top of the feeds, which rules out feed leaks as the culprit.
> The problem is due to condensation.

I don't think you can rule it out.  Gravity, time, and
evaporation/recondensation could move water around in the feed.  I'll
bet if you checked them on a cold night the drops would be located at
the top, on the cold metal.  I don't think you can rule out leaks from
purely the location of the condensation.

My concern is that if the feeds are leaking water, the dry air system
could easily be overloaded and become useless.  This happened at the
VLBA and we had to ensure that the feeds were tight.  The pressure
that can be applied to the feed is limited by the strength of the feed
window.  1" WC (approx 1/27 PSI) on the L-Band window is 290 pounds of
force.  This severely limits the flow rate of dry air that can be
supplied through the purge port.  Each unit of dry air can only absorb
so much moisture at a given temperature before saturation.

A positive pressure is not a impermiable barrier against water
ingress.  The partial pressure of water in the air wants to pump
moisture into the dry space (just like those mystery houses, it is not
really going uphill, it just looks like it :-).  So there has to be
enough flow to keep the air exchanged and dry.

It sounds like the EVLA feeds have been verified water tight by your
inspection, so hopefully that would help the dry air system to work
correctly.

The dry air system incurs maintenance costs as well.  There is a
compressor, a membrane or reciprocating dessicant system, filters,
regulators, etc. that all require maintenance.  The system must be
monitored for proper operation.

The VLA is a very dry site compared to the some of the VLBA antennas.
The dry VLBA sites (including PT and LA) rarely had moisture problems
as the dessicants were PMed twice a year.  The dry air system was
primarily for NL and HN, where the purge procedure was hopelessly
inadequate for the amount of moisture in the air, SC and BR, where
water was ingressing the FRM motors, and then every other site because
all the VLBA antennas are the same :-).

The EVLA vertex room is a slightly moister environment during parts of
the year that the VLA was.  The cooling coils are at 13C with the
glycol instead of 6C that the R12 operated at.  This means that the
dewpoint in the EVLA vertex room is limited to 55F instead of 42F for
the VLA.  This may not sound like much, but when the diurnal cycle
brings the night temperature below 55 degrees then BAM, you have
condensation.  If the dessicant is saturated, then that moisture is in
the feed, even after the ambient air becomes dry again.

So if the feeds are truly sealed tightly against rainwater, and the
dessicants really are becoming saturated relatively quickly, then we
should install a dry air system on the EVLA.  Or, if the maintenance
on the dry air system is less than changing dessicant packs regularly.

I believe that the dry air system will require LESS maintenance than
dessicants.  A diaphram/membrane setup will probably go three years
without trouble, and is fairly easy to rebuild.  The feeds MUST be
sealed well and we may need to plumb the feeds serially which means
drilling a second purge port.  Another benefit is that if water does
ingress into the feed, a dry air system will slowly remove it without
human help.  The existing dessicant packs also do this in a less
active manner, and cease to work when saturated.

I'm just making sure that the prerequisites are met and the
performance, installation and maintenance issues of the dry air system
are understood.  Also I want to instigate discussion so I get free
engineering support from ya'll, at least those of you who survived the
above droning.

>Conflaguration

What a great word!

Bob.