Re: Epoch

Bjoern Gimle@tt-tech.se (Gimle@tt-tech.se")
Wed, 12 Mar 1997 09:51:54 +0100

Alphonse Pouplier wrote :
>In the 2d line (21-32) of the "two-lines" TLE's the reference day
>(epoch) is given with 8 decimals, thus to .000001 of a second.

No,  86400  sec/day  is  about  10**5,  ie. 10**-3 or millisecond,
not microsecond, precision remains.

And Tony Beresford answered :
>Alphonse, the epoch is the time of an equator crossing. If you
>look at the TLe's of objects in low eccentricity orbits you
>will see that the sum of the mean anomaly and the argument
>of perigee is always very close to 360 degrees.

By definition,  this is  not a  requirement, but  NASA elsets  are
quite close.  This is handy for rough computations, but prediction
programs work just fine with any epoch (AOP+MA not close to 360).

NASA  elsets  for  geostationary,  and  some  high  orbit research
satellites, often  have an  epoch at  midnight UT.  Amateur elsets
generated  by  PELCOR  (or  FITELEM?)  can  have  any combination,
depending on the initial orbit used.

VEC2TLE,  when  used  to  convert  XYZ  ("state vector") to elset,
defaults to using the vector time  as Epoch. When used to apply  a
rocket impulse to change a TLE set, it defaults to the time of the
change.  In  both  cases,  VEC2TLE has  an  option to  re-Epoch at
Ascending Node Time (with a slight loss of accuracy,  according to 
the documentation)

In my  decay analyses,  I previously  fitted a  function to  Epoch
times, but  found that  the small  (random?) deviations  from true
node times made the extrapolation unsafe, and switched to using MM
(and ndot)

I have  now written  a program  to compute  equator crossing times
from an entire elset file. Below, three runs are tabulated: actual
NASA elsets, Alan  Pickup's and my  runs with his  SatEvo program.
The deviations from elsets'  equator crossing times are  included.
NASA Epochs appear to have a spread of about one second, close  to
zero. SatEvo elset had a  very high consistency - all  Epochs were
2.29-2.30 seconds before Ascending Node !  These two  observations
combined "prove" that these are real effects  of the sources,  not
program bugs.

OIG Epochs     -secs OIG node times +secs Alan's mails   +secs My mail

97069.17347493  .288 97069.17347159 -4.56 97069.17341873  5.18
97069.17353152
                                          97069.11293470      
97069.11300967
97069.05237822 -1.11 97069.05239118 -3.12 97069.05235497  1.41
97069.05240755
97068.99173053 -.569 97068.99173712 -2.37 97068.99170961  .832
97068.99174676
97068.99172988 -.133 97068.99173143
                                          97068.93101299      
97068.93103885
                                          97068.87027388      
97068.87029128
                                          97068.80949816      
97068.80950913
97068.74869236 -.525 97068.74869844 -.715 97068.74869016 -.185
97068.74869629
                                          97068.68785316      
97068.68785570
97068.62698098 -.760 97068.62698978                      -.001
97068.62698976
97067.28366537 -.683 97067.28367328       97068.50525876
97067.22246012 -.862 97067.22247010       97068.01728639
97067.22246007 -.592 97067.22246693       97067.52839209
97067.16124962 -.585 97067.16125640
97066.91630664 -.464 97066.91631202
97064.58359546  .260 97064.58359244
97064.27605468 -.424 97064.27605959
97064.21452423 -.580 97064.21453095
97064.21452362 -.582 97064.21453036



-- 

------------------------------------------------------------
-- bjorn@tt-tech.se                +46-8-59095783 (office)--
-- b_gimle@algonet.se              +46-8-7428086  (home)  --
-- 59.22371 N, 18.22857 E           AND member of :       --
-- http://www.algonet.se/~b_gimle   SeeSat-L / SeeSat-D   --
------------------------------------------------------------