[teqc] GLONASS navigation error

Lou Estey lou at unavco.org
Wed Jun 18 12:21:58 MDT 2008


Hi Uwe,

> According to everything I know, and the fact that the other MDB GLONASS
> messages in your .m00 file are being decoded by teqc correctly, teqc is
> "doing the right thing".  What is odd about the several GLONASS messages
> that are in error is that the seconds are (according to the info in the
> MDB messages) 59.999 seconds (rounded to "60.0" in the RINEX -- which is
> certainly not the way the Leica Spider software treated the seconds.

Well, teqc _isn't_ "doing the right thing" when there are these types of non-integer
seconds in nav time tags.  Going back to the teqc vs. Spider translations:

teqc:
23 08  6 16  0 15 60.0-1.270640641451D-04 3.637978807092D-12 0.000000000000D+00
     1.060660107422D+04 1.154708862305D-01-9.313225746155D-10 0.000000000000D+00
    -3.738262207031D+03 3.132190704346D+00 0.000000000000D+00 3.000000000000D+00
     2.290192138672D+04 4.613704681396D-01-1.862645149231D-09 0.000000000000D+00

Leica Spider:
23 08 06 16 00 15  0.0-1.270640641451D-04 3.637978807092D-12 0.000000000000D+00
     1.060660107422D+04 1.154708862305D-01-9.313225746155D-10 0.000000000000D+00
    -3.738262207031D+03 3.132190704346D+00 0.000000000000D+00 3.000000000000D+00
     2.290192138672D+04 4.613704681396D-01-1.862645149231D-09 0.000000000000D+00

The correct time for these messages (using the full resolution in the MDB records)
is actually 2008 June 16 00:14:59.999, which the RINEX format can't represent,
so the Spider software rounds this to 00:15:00.0.

This is the first time I can recall seeing either a GPS or a GLONASS nav message
that had any non-integer second for the epoch time of the message -- which is
throwing the decoding I have in teqc which assumes integer seconds for a loop.

More disturbing, even though the Leica Spider translation is doing the best
job possible here, the representation of the time in Spider's RINEX is still off
by 1 millisecond due to the format limitations of RINEX, assuming that the epoch
time in the MDB records is accurate.  Does anyone know if this situation is
something new that we're going to start seeing from GLONASS? (GPS, too?)

I'm also wondering if nav epochs which can't be accurately represented in the
RINEX format (such as these) shouldn't be discarded and not force-fit into RINEX.

--lou


More information about the teqc mailing list