Re: generating new TLEs and finding lost satellites...

From: Björn Gimle (
Date: Wed Oct 23 2002 - 15:59:33 EDT

  • Next message: George Amos: "Re: Freeware sky chart program"

    For these tasks I use SkyMap (Matson), FIRST/FIRSTORB, VEC2TLE, ELCOR/IDE and/or
    Since you do not know the range (unless you own a high-power radar/laser), you
    can't really 'convert' to XYZ. You may guess (a constant height or range?) and
    use your Siderial time + geocentic latitude to obtain your XYZ, and the RA/Dec +
    range to add the relative XYZ, then get approximate XYZ-dots (speed) required by
    VEC2TLE. If the range/speed is too high, you get a hyperbolic orbit that it
    can't solve, too low (or too differing heights) and the orbit will intersect the
    Earth surface.
    FIRST/FIRSTORB produce the circular orbit from two observations.
    ELCOR needs only one or more RA/Dec  AND  an approximate TLE, so you may need to
    perform one of the steps above. If the prediction errors are too large, even in
    an intermediate "solution", ELCOR will diverge, so you may have to first iterate
    your prediction program to get the time accurate enough.
    IDE may be better, but does not estimate drag terms (?), so it is only suitable
    for very stable orbits, or single-pass observations.
    FINDSAT is tedious, because you have to manually select each element and its
    change until all residuals are near zero.
    INT2 is useful when the elset is good, but the orbit is mainly perturbed by Sun
    and Moon gravity, rather than drag.
    When you have a reasonable TLE, the procedure depends on the reason for error.
    If the error is likely caused by drag, and you expect the satellite to be found
    on the first attempt, you can simply load one elset multiple times in SkyMap
    F7/F8 table with the 'A' or 'C' command, and alter the drag (F1 for SGP, F3 for
    SGP4). The effect of a 20% change is given on the display.
    If you think you may have to run predictions more than once, it is better to
    prepare an elset file with changes to multiple copies of the same elset. Edit
    the drag term if it is so large it will have had a considerable effect on the
    orbital height, otherwise (if a small drag or period error has accumulated over
    a long time) simply add increments to the Mean Anomaly. For a complete orbit
    scan, it could be sufficient to have nine elsets, with 40 degrees difference in
    MA. The procedure was discussed on SeeSat-L in April as 'Predictions with many
    TLEs for a single object' and mid-August as 'Generating search elements' and on
    DSat in 'Re: SkyMap'
    You can also handle the case of large maneouvres, if you can guess the size of
    the change. Pick a reasonable date and use VEC2TLE with zero Impulsive DeltaV to
    find the TLE and the Mean Anomaly. The MA, Epoch and orbital period determine
    when it will pass perigee. Guess an alongtrack DeltaV, apply at perigee time,
    check the resulting MM, iterate until a reasonable MM is obtained. Then use
    previous methods to generate search elsets or predictions.
    Unsubscribe from SeeSat-L by sending a message with 'unsubscribe'
    in the SUBJECT to

    This archive was generated by hypermail 2b29 : Wed Oct 23 2002 - 16:30:06 EDT