[evlatests] Bad Correlator Behavior, last Saturday

[Rick Perley]

I had an 8-hour test run Saturday afternoon and evening, at P-band.  I 
set up the correlator to cover just 8 SPWs, spanning the mostly 
RFI-clear areas.  Each SPW was 16 MHz wide, covered with 125 kHz 
resolution.

The lower half came through just fine.  No problem noted.

Not so in the upper half (384 -- 448 MHz).  For about two hours in the 
middle of the observation, many antennas and SPWs gave zero signal.  I 
know this is not an antenna problem, because the distribution of the 
'dead' antennas was different for each scan and polarization. 
Furthermore, correlation between apparently 'dead' antennas give correct 
results.

It is extremely difficult to flag out these apparently dead channels -- 
I'll likely have to abandon this half of the data (which is a pity, as 
the upper frequencies are where the moon is strongest).

The project code is TLOW0002, in case somebody wants to look into what 
went wrong.  IAT times were 0/22:35 to 1/06:40.  The 'bad times' started 
right about at 1:00:00, and last for about 2 hours.

Rick