Frank Reed posted: >Incidentally, can anyone point me to comparisons of the full models versus >simple precessing ellipses? >I understand I can expect ground track errors of a >couple of tenths of a degree in simple cases, but is it ever worse? As a simple test to compare a precessing ellipse against SGP4, I created an elset for a fictitious object in a plane many degrees east of ISS so that it will make a very favorable pass over me tonight at a high elevation and low height. 1 90000U 06500A 06304.83777757 .00019838 00000-0 12002-3 0 9679 2 90000 051.6346 293.1609 0012692 158.4739 291.3758 15.77504245454722 My lat/long: 30.31N 97.86W BCRC Tonight: 2006 Nov 2 Times are UT The Quicksat prediction (modified for greater displayed accuracy): H M S Al Azi Hgt Rng R A Dec 0 47 3.3 82 317 215 217 20 33.3 35.86 The predictions from SeeSat using the original SGP4: H M S Alt Azi R A Dec Rng Hgt 0047:03.3 82.10 306.09 20 28.89 34.72 215 213 0047:03.4 82.12 306.97 20 29.28 34.81 215 213 0047:03.5 82.13 307.86 20 29.66 34.90 215 213 0047:03.6 82.14 308.75 20 30.04 35.00 215 213 0047:03.7 82.15 309.64 20 30.43 35.09 215 213 0047:03.8 82.16 310.54 20 30.82 35.19 215 213 0047:03.9 82.17 311.44 20 31.20 35.28 215 213 0047:04.0 82.17 312.33 20 31.59 35.37 215 213 0047:04.1 82.17 313.23 20 31.98 35.47 215 213 0047:04.2 82.17 314.13 20 32.37 35.56 215 213 0047:04.3 82.17 315.03 20 32.76 35.65 215 213 0047:04.4 82.17 315.93 20 33.15 35.75 215 213 0047:04.5 82.16 316.83 20 33.55 35.84 215 213 0047:04.6 82.16 317.72 20 33.94 35.93 215 213 If I take the Quicksat prediction (20 Hr, 33.3Mn, +35.86) and put into my Fitelem program, it computes residuals of 1.2 seconds and 0.09 degrees crosstrack. The best fit is thus about RA 20 Hr 33.35 Mn +35.79 at time 00:47:04.44 In general, all south-to-north predictions are off about 1 to 1.5 seconds (the satellite is later than the prediction) and all north-to-south predictions are similarly off, but the satellite is earlier than the prediction. The crosstrack error is a maximum for lower heights and gets smaller with larger heights/ranges. There are also small north/south crosstrack errors for objects in west-to-east orbits passing near the zenith. >Isn't element age a bigger issue than all these subtleties of models? This is quite possibly true for SGP4. I believe that the SGP4 model has been stable for a long time and I have glanced at the new "combined" SGP4/SDP4 code and it appears to me that there were not any changes to the SGP4 model. Of course if an object in low orbit is off 30 seconds in time, the observed position will be different because the observer has moved (due to Earth rotation) during that time. The SDP4 model is an entirely different question. I have found what I called "discrepancies" between the SGP4 model and the SDP4 model for one implementation of the SDP4 model for an orbit where the SDP4 model should be used. The fixes in the new "combined" code appear to be only for the SDP4 model. Mike McCants ------------------------------------------------------------------------- Subscribe/Unsubscribe info, Frequently Asked Questions, SeeSat-L archive: http://www.satobs.org/seesat/seesatindex.html
This archive was generated by hypermail 2b29 : Wed Nov 01 2006 - 12:02:41 EST