From brideout at haystack.mit.edu  Tue Nov 23 11:48:44 2004
From: brideout at haystack.mit.edu (William Rideout)
Date: Tue Nov 23 11:48:48 2004
Subject: [gps-developers] Modifications to GPS Processing toolkit
Message-ID: <41A369EC.9060906@haystack.mit.edu>

Anthea,

I've made a number of small changes to the GPS processing toolkit.  If you want 
the latest version, its called gps_install.tar on my home directory on hyperion.

Changes:

1. New option added to process_gps_kduh2.py and createMonthTec.py:

	--limit=<secs>

Decimates data so frequency is never less than secs.  This reduces the size of 
the data set.

2.  If gps_retrieve.py can't find the satellite bias file for the given day, it 
will step back in time until it does find a copy, up to a set number of days 
(now 100).  This should help us process data less than a week old.

3. Occasional isolated sites were coming through with bad receiver bias values. 
  This problem was due to the last-ditch use of the 0 tec method, but without 
the rigorous error checking of the old version.  This problem has been fixed in 
two ways:
	- The minimum scallop technique used to search around the 0 tec value for the 
real receiever bias.  It would occasionally fail if the 0 tec value was wildly 
wrong.  Now if the 0 tec value starting point search fails, it trys a wider 
search around a receiver bias of 0.  This reduces the number of sites that must 
ultimately fall back on the 0 tec method.

	- A list of sites that ultimately needed to use the last resort 0 tec method to 
get their receiver bias is kept.  Then only these sites are subject to the 
rigorous error checking of the old version.

Bill


-- 
Bill Rideout
MIT Haystack Observatory
Email: brideout@haystack.mit.edu
Phone: 781 981-5624