Problem with SDP4?
Bruno Tilgner (Bruno_Tilgner@compuserve.com)
Tue, 17 Nov 1998 16:41:53 -0500
On 13 Nov Ron Dantowitz wrote:
>I have a question for all: has anyone noticed any problems with SDP4 aft=
er
>September 21 1998 at 12:00 UTC? Although STSPlus uses SGP4 for
>satellites like Intelsat 1-1 r or Comstar 4 r, when I use SDP4
>there is a discontinuity at the time mentioned above...
I don't think there is any particular date for which SGP4 or SDP4
would produce a discontinuity. Both algorithms don't use dates
explicitely. The time variable T is measured in minutes since the
epoch and is derived via Julian dates.
However, T occurs up to the 5th power, and if elsets are several
weeks old this leads to very big numbers which may result in compiler
errors. I encounter occasionally crashes because of eccentricities
greater than 1 which are invariably due to excessively large values
of the time argument.
About a year or so ago, an error in the code (not the algorithm) of
SDP4 was discovered. The difference between two angles (right
ascensions of ascending nodes) was calculated incorrectly if one of
the angles went beyond 360 degrees. This occurred in the Lyddane
section of the deep-space routine of SDP4, both in the FORTRAN and
PASCAL implementations. It only took a few extra lines (suggested by
Rob Matson) to put the code right.
If you haven't followed the discussion on SEESAT-L at the time and
made the necessary corrections it may well be that the discontinuity
you noticed stems actually from this faulty code and not from the
date. The fact that you don't notice the discontinuity with SGP4
is an indication that this may be the case.
Bruno Tilgner
Paris, France