Date: Mon, 4 Mar 2013 13:45:57 -0500 (EST)
From: Ken Young 
To: cqi@cfa.harvard.edu, jzhao@cfa.harvard.edu, mgurwell@cfa.harvard.edu, aargon@cfa.harvard.edu, tcooper@cfa.harvard.edu
Subject: Test SWARM Correlator data files
Parts/Attachments:
   1 Shown     31 lines  Text
   2   OK     105 KB     Image, ""
   3   OK      19 KB     Image, ""
----------------------------------------

Dear SMA Data Stakeholder,

   I believe that since last Friday night (March 2 UT) we have been
producing data files in the new format which have valid data in the new
SWARM correlator chunks, s49 and s50.   The test version of dataCatcher
will continue to produce these files in parallel with the current format
files, which will eventually give us a large set of test files to process.

   Chunks s49 and s50 will contain copies of the data in chunk s01,
interpolated onto the innermost 14,000 channels of the 16,284 channels in
chunks s49 and s50.   The other channels in those chunks are set to a
fixed, nonzero value.   Chunks s49 and s50 will not exist if the
correlator is configured with 0 chips/chunk on chunk s01.   I've attached
two plots to this missive, the first showing the amplitude on chunk s01
on the first baseline of scan 25 in the track 130304_04:14:38.   The
seconf plot shows the s50 chunk for that scan and baseline.   Please note
that the channel numbers on the second plot are offset by 100, because of
the way I extracted the visibilities from the sch_read file.

   I've dumped the sp_read, in_read, bl_read and spot-checked the sch_read
files, and they appear to have valid data.   But I would not be at all
surprised if one of you finds additional problems with these files that my
checking missed.

   We should all write our software to be able to accomodate data sets
without s49 or s50, and with any combination of those two new chunks added
to the existing chunks.   There's no telling exactly how the new
correlator will come online, and we may run for some period with, for
example, s49 present but not s50.

Taco