[evlatests] Software time February 1

Barry Clark bclark at nrao.edu
Fri Feb 2 11:04:56 EST 2007


Before software time was started, system parameters were transfered from
the Modcomp system files to the EVLA parameters database.  This worked
well enough that, after a little fiddling, under Executor control, the
antennas pointed well enough for 6cm.  When the correlator was given to
the Executor, after a little fiddling, delays were adequate for operation 
at 12 MHz bandwidth, though some decorrelation was seen at 25 MHz.

We had another incidence of a CMP hangup a couple of hours into the 
software time.  Rebooted and restarted, and it behaved itself for the
rest of the time.

For reasons unknown, the F display was not working.  Telcal claimed 
it was sending data out, and there was some indication that it was being
received by F, but nothing showed up on its screen.  Will diagnose at
leisure.

So when we switched the datapath to Idcaf, excluding the Modcomps, we
were pretty well flying blind.

Idcaf seems to work reasonably with a modest number of antennas, but 
seems to get in trouble with more.  After a few fumbles on my part
getting a calcode set, Ken managed to get his delay solver running.
The numbers it provided were not believable, but it is as yet unclear
whether this is due to the Idcaf problems mentioned above or something
in the delay solver itself.  (On a fresh Idcaf invoked with 24 antennas,
a lot of delays came out 'nan'.  With 10 antennas they were too small, 
and different from those suggested for 13 antennas.)

The one attempt to connect Idcaf to the datamgr resulted in seg fault
on the datamgr side - Idcaf said something it didn't like.

This is the first time I've had an O-stream listener turned on for a
long interval.  Some curious things showed up on it.  Some device 
occasionally emits a record with a single character in the location field,
and there are intermittent records with values of 'nan'.  These are 
harmless but annoying.  I shall diagnose at leisure.

All in all, not as productive a day as I had hoped, but still progress
was made.

Barry & Ken



More information about the evlatests mailing list