Chaminda B Senavirathne.
SULECO (Pvt) Ltd,
House No.6, ROSETON PARK,
No.275/1, Divulpitiya,
Pepiliyana,
Sri Lanka.
Mob: +94 77 2440789
Tele/Fax: +94 11 2828100
E mail: info@corsnet.lk
Web Site:
sulecoltd@sltnet.lk,www.corsnet.lk,
Dear Colleagues,We received raw Topcon data files for April 3rd that were translated incorrectly into RINEX files labeled March 27 (DOY 087). Therefore the good files DOY 087 were overwritten on April 4th with data from April 3rd. Those bad RINEX files were just replaced
with the correct files for DOY 087. The IGS stations impacted were the following:CHIL, CLAR, CMP9, HOLP, ROCK, SNI1, SPK1, TORP, UCLP, WHC1We apologize for the inconvenience. A detailed description from Lou Estey is provided below.Best Regards,David Maggert[2735] teqc +quiet +mds rg070403a
week: 1890
[2736] teqc -week 1890 +quiet +mds rg070403a
2016-03-27 00:00:00 2016-03-27 23:59:30 2441027 rg070403a
... which is one week earlier than when the data is valid. This can be traced
to the code changes I made on March 25th (see
http://www.unavco.org/software/data-processing/teqc/log/log.html) to correct the
GPS week reporting from SBF with navigation messages from the previous week.
The logic changes made on March 25th were absolutely necessary and are correct.
However, in correcting all that for every format, including Topcon TPS, this
apparently brought to light an even more severe problem that these Topcon receivers
have (that just happened to work with the older incorrect logic in teqc).
E.g. file rg070403a is data for 3 Apr 2016, which is GPS week 1891 (this week).
And this file does contain may GPS navigation messages for week 1891. However,
it has a bunch of the previous week 1890. This in itself is not a problem
and is what I was fixing details on March 25th. The problem is that these
Topcons, on Sunday, are reporting navigation messages also collected a week
prior to that. In particular, there are a bunch collected in the wee minutes
of GPS week 1889 (since ToW is just under 604800 == seconds in a week), but
these are messages for 00:00 on 27 Mar (GPS week 1890). The first old
message below does not cause trouble. because it is clearly for end of week
1890, but any of the following ones would trip teqc up and cause it to report
the starting week as 1890 because they are for the start of week 1890:
866 -> 1890: ToC= 604784 ToE= 604784 ToW= 604770
866 -> 1890: ToC= 0 ToE= 0 ToW= 601110
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
866 -> 1890: ToC= 0 ToE= 0 ToW= 601140
866 -> 1890: ToC= 0 ToE= 0 ToW= 599880
866 -> 1890: ToC= 0 ToE= 0 ToW= 601110
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
866 -> 1890: ToC= 0 ToE= 0 ToW= 601140
866 -> 1890: ToC= 0 ToE= 0 ToW= 599880
866 -> 1890: ToC= 0 ToE= 0 ToW= 601110
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
866 -> 1890: ToC= 0 ToE= 0 ToW= 601140
866 -> 1890: ToC= 0 ToE= 0 ToW= 599880
866 -> 1890: ToC= 0 ToE= 0 ToW= 601110
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
866 -> 1890: ToC= 0 ToE= 0 ToW= 601140
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
866 -> 1890: ToC= 0 ToE= 0 ToW= 604410
Now to figure out a stable way to code around this ...
_______________________________________________
IGSSTATION mailing list
IGSSTATION@igscb.jpl.nasa.gov
https://igscb.jpl.nasa.gov/mailman/listinfo/igsstation
