[teqc] qc bug & improvements related to GLONASS

Lou Estey lou at unavco.org
Fri Apr 3 12:57:20 MDT 2009


If you are interested in GLONASS, keep reading; otherwise
you can probably skip this email.

----------

A group looking at GLONASS data with teqc (during their vetting
of the 23 Mar 2009 version) have discovered a long-standing qc
bug that's been around ... probably for 10 years.  The problem
occurs when there is data (e.g. from a RINEX obs file) for
a GLONASS SV, but there is no corresponding frequency channel
number for that SV.  As you know, the frequency channel number
for each SV slot # is needed to convert GLONASS phase data from
units of cycles into distance.  Without a corresponding frequency
channel number, supplied e.g. in the GLONASS nav RINEX file or
in a file with slot # vs. fc# read in with the '-glonass_n'
option, then a frequency channel number of 0 was accidentally
being applied.  This leads to very erroneous detection of slips
and multipath rms for any GLONASS SV in this situation (unless
the frequency channel number for that SV actually happened to
be 0).

The current workaround, for any version of teqc that you probably
have, is to make a small file of the slot #s vs. frequency channel
numbers, e.g. for the current GLONASS constellation see
http://www.glonass-ianc.rsa.ru/pls/htmldb/f?p=202:20:806691743196686987::NO

[826] file s2fcn
s2fcn:          ascii text
[827] cat s2fcn
  2  1
  3  5
  4  6
  6  1
  7  5
  8  6
  9 -2
10 -7
11  0
13 -2
14 -7
15  0
17  4
18 -3
19  3
20  2
21  4
22 -3
23  3
24  2

and then do your qc with any GLONASS data including this filename
with the '-glonass_n' option:

[828] teqc +qc -glonass_n s2fcn ...

(Remember: the above values are true right now, but aren't necessarily
the correct mapping if you are looking at much older GLONASS data.)

---------

Changes made to teqc since 23 Mar 2009 related to GLONASS:

The next version of teqc will correctly detect a missing GLONASS
fcn and report this during the qc in various ways:

a) to stderr, you will get a notice of the form:

! Notice ! unknown GLONASS frequency channel number file for slot # = 18

b) in the ASCII time plot, any GLONASS SV with an unknown fcn will
be marked with an 'N', e.g:

R10|              coooooooooo                  oooooMoIoooooooo             |R10
R18|              NNNNNNNNNNNNN                       NNNNNN                |R18
R 3|               NNNNNNNNNNNNNNN                        NNNNNNNNNNNN      |R 3
R19|                NNNNNNNNNNNNNNN                  NNNNNNNNNNN            |R19
R 4|                   oooooooooooooooo                      cooooooooooo   |R 4
R11|                   NNNNNNNNN                    NNNNNNNNNNNNNNNNN       |R11
R20|                     NNNNNNNNNNNNNN                  NNNNNNNNNNN        |R20

where here the fcns for R10 and R4 are known and for R18, R3, R19, R11, and
R20 they are not known.  With this marking, you can at least see in time
where the GLONASS SVs with an unknown fcn have data occurring.

c) You'll also have this line:

     GLONASS SVs w/o fcn :  3  11  18  19  20

d) and if you are using the '+ssv' option, then:

SSV R 2     0     0     0   0  0.38  0.44      0
SSV R 3     0     0     0   0   n/a   n/a      0
SSV R 4     0     0     0   0  0.39  0.35      0
SSV R 6     0     0     0   0  0.42  0.36      0
SSV R 7     0     0     0   0  0.42  0.36      0
SSV R 8     0     0     0   0  0.45  0.34      0
SSV R 9     0     0     0   0  0.46  0.75      0
SSV R10     0     0     0   0  0.39  0.51      0
SSV R11     0     0     0   0   n/a   n/a      0
SSV R13     0     0     0   0  0.36  0.35      0
SSV R14     0     0     0   0  0.39  0.41      0
SSV R15     0     0     0   0  0.43  0.37      0
SSV R17     0     0     0   0  0.44  0.35      0
SSV R18     0     0     0   0   n/a   n/a      0
SSV R19     0     0     0   0   n/a   n/a      0
SSV R20     0     0     0   0   n/a   n/a      0
SSV R21     0     0     0   0  0.44  0.37      0
SSV R22     0     0     0   0  0.39  0.36      0
SSV R23     0     0     0   0  0.39  0.36      0
SSV R24     0     0     0   0  0.49  0.53      0

the mp1 and mp2 rms will be marked as 'n/a' -- which doesn't
necessarily mean there is missing fcn for each GLONASS SV with
it, but it's a pretty high likelihood.

The other qc item that I noticed while fixing all this was that
the new '+' indicator on the '-dn' and '+dn' lines were not correct
when there was any GLONASS, and are especially messed up if any
GLONASS has no fcn.  Using the above dataset with the 23 Mar 2009
version of teqc:

[992] teqc.20090323 -G1 +qcq *.09o 2> /dev/null | grep "^[-+]dn"
-dn|++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++|-dn
+dn|++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++|+dn

(very wrong! and obliterates any meaningful information), but is ok if
GLONASS is eliminated:

[993] teqc.20090323 -G1 -R +qcq *.09o 2> /dev/null | grep "^[-+]dn"
-dn|   + ++  ++ +++       + ++    +++++ ++  +   ++  + ++     +++ + +  + +++ |-dn
+dn|              1 22  11 1                        3 2 11                  |+dn

Now I get the same result leaving GLONASS in with the fixed teqc:

[994] teqc -G1 +qcq *.09o 2> /dev/null | grep "^[-+]dn"
-dn|   + ++  ++ +++       + ++    +++++ ++  +   ++  + ++     +++ + +  + +++ |-dn
+dn|              1 22  11 1                        3 2 11                  |+dn

(very correct -- but, remember that these metrics are only for those SVs for
which SV orbits are being determined, and for the time being that is only
GPS SVs.  G1 was eliminated in these runs because the receiver was set not
to track the new G1, probably because G1 is still in the commissioning phase
and is set as unhealthy.)

Next related item: in the next teqc release, the option '-glonass_n'
no longer exists but is replaced with the more explicit name of
'-glonass_s2fcn'.  Same functionality, different option name.

cheers,
--lou

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Louis H. Estey, Ph.D.              office:  [+001] 303-381-7456
UNAVCO, 6350 Nautilus Drive           FAX:  [+001] 303-381-7451
Boulder, CO  80301-5554            e-mail:  lou  unavco.org
    WWW:  http://www.unavco.org   http://jules.unavco.org

"If the universe is the answer, what is the question?"
                                                -- Leon Lederman
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~


More information about the teqc mailing list