Re: Argo Navis

From: Gavin Eadie (gavin@umich.edu)
Date: Fri Jun 07 2013 - 15:18:43 UTC

  • Next message: wkitty42@gmail.com: "Re: Fw: Argo Navis"

    George's is a good way to locate the problem, but it's not debugging, which involves fixing the root cause (to my early morning, slightly over-zealous thinking)!  So, when you the find the problem, please tell the authors of the software to improve their act -- there's far too much crappy software in the world.
    
    Possible places to look, though I'm probably over-thinking the issue with (2) -- it's likely to have a simpler cause like (1):
    
    (1) "Line 0":  Mike's "line 0" (informal object name) has trailing blanks removed, Celestrak's do not.
    
    Mike: "OBJECT"
    Celestrak: "OBJECT                  "
    
    You could try padding out Mike's name to 20 characters with extra spaces.
    
    (2) Comparing TLE's:
    
    Mike's:
    
    1 03334U 68063A   13096.45276387 0.00000000  00000-0  00000-0 0    00
    2 03334  15.0142 339.7199 0931023  92.4347 278.2486  0.99238613    04
    
    and at Celestrak's:
    
    1 00694U 63047A   13141.85021544  .00002006  00000-0  26072-3 0  2225
    2 00694  30.3581  47.3372 0604916 119.9763 320.6634 13.97910174474066
    
    you'll notice there's a lot of zero values in Mike's.  Line 1 has zeros (or blanks) from column 34 (First Time Derivative of the Mean Motion) onwards; line 2 is likewise from col 64 (Revolution number at epoch [Revs]).  There's nothing wrong with that -- that data is either not computable, not knowable or not used for predictions -- and zero/blanks are nominally OK, but this would not be the first software I've seen (or written, for that matter) which tripped over zero values.
    
    It's difficult to use George's technique to try changing characters because of the checksum at the end of each line, but you can try (I personally think this not a likely cause, but you eliminate everything else ..).  I would change the second last character of each line from "0" to "1" (element set number and rev count are not significant for predictions, so you're not impacting output), and add 1 to the checksum character to keep it legal.
    
    1 03334U 68063A   13096.45276387 0.00000000  00000-0  00000-0 0    11
    2 03334  15.0142 339.7199 0931023  92.4347 278.2486  0.99238613    15
    
    More than you ever need to know about TLEs: http://celestrak.com/columns/v04n03/
    ___________
    
    On Jun 7, 2013, at 9:35 AM, George Roberts <gr@gr5.org> wrote:
    
    > Standard debugging technique:
    > 1) get it to not work (you acheived this)
    > 2) get it to work (you achieved this) e.g. did you try just a single TLE? Or 
    > use a TLE that works.
    > If you can get to this point (you already did) you are guaranteed success.
    > 3) Start with either of these samples and move it towards the other, 
    > shrinking down to one TLE if possible and then down to one bad character if 
    > possible trying to alternate getting it to work and then not work.
    
    _______________________________________________
    Seesat-l mailing list
    http://mailman.satobs.org/mailman/listinfo/seesat-l
    



    This archive was generated by hypermail 2b29 : Fri Jun 07 2013 - 15:19:47 UTC