[evlatests] Tests, July 21

[Barry Clark]

We had only antenna 16.  Antenna 14 is turned off for want of baseband
converters.

However, antenna 16 has four basebands.  So I set out to try the four 
IF software, first at L band.  At first, there were fringes only on
IF C.  Jim and Rob traced the missing IF A to a dead fan, which caused
the IF A deformatter to cook and die.  (They managed to find a fan and
another deformatter later in the day.  For reasons unknown, the 
replacement deformatter doesn't do slotid right.)  There were obvious
software problems.  The first one that occurred to me was that I had
forgotten to turn on phase switching for IFs BD.  So I turned off 
phase switching at the correlator (and in IFs AC).  When I did that
the fringe amplitude on IF C increased by a factor of two.  Therefore,
we probably have some sort of timing problem at the antenna; I didn't
take time to investigate, except for verifying that the situation did
not change if I resynced the antenna.  I just spent the rest of the day 
with phase switching turned off.

With phase switching turned off, we had fringes on IFs B and C!

There were still obvious software problems.  I switched to X band to
work on them.  Software was painful, because at this time the link
to the AOC was very slow.  One cycle of compiling and bringing the new
code into execution took more than fifteen minutes.

Finally got IFs BD working at X band.  Sort of.  Phases are still garbage.
Not clear if this is something to do with BD or if it is the weird phase
phenomenon that affects this antenna at L band in AC.  Finally got tired
of looking at code and let it be.  I can check out the BD code in the 
simulator before next test time.

Started solving for delays at X band.  They were extremely wrong, many 
tens of nanosecs.  Might be due to the floundering around Mark and I did
week before last, when we were plagued by unstable amplitudes.  Since
I changed the antenna delay, in order to get reasonable values in the 
IF file, the delays at other bands are now probably thoroughly messed up.
It is quite likely that I was peaking on a delay sidelobe in IF D; the
amplitude was low.  Might be a messed up passband, though.

Starting at about 23:35 UT, I took what should be pretty definitive passbands
for IFs A and C, for purposes of defining the phase response we want to 
put into the FIR.  Ran out of time before I could get to B & D.

Also, never got around to going back to L band to see if IF D would fringe 
now.