Orbit sandbox
Turn a TLE into orbital elements and back, build a preliminary TLE for an upcoming launch, and predict passes over the network's stations or your own location. All times are UTC.
Input elements
Orbital elements
- Period
- —
- Semi-major axis
- —
- Velocity
- —
The period exceeds 225 minutes: SGP4 hands such an orbit to the deep-space branch (SDP4), while the sandbox's launch maths is written for low orbits.
The satellite's position and velocity at separation from the launch documentation. The osculating elements are turned into SGP4 mean elements by iteration, so the TLE reproduces the vector at epoch to within tens of metres.
Position at the selected time
Result
A number from 100000 does not fit the classic format: the lines use the Alpha-5 encoding.
Passes
| Station | AOS | Az., ° | Culmination | Max. elev., ° | LOS | Az., ° |
|---|
- Take the target orbit from the press kit or announcement: perigee and apogee altitude (or one altitude for a circular orbit), inclination and launch time. Pick the launch site from the list or enter the pad's coordinates.
- Choose the ascent direction: northbound (ascending arc) or southbound (descending). Sun-synchronous launches from Vandenberg go south; from Baikonur, Plesetsk and Vostochny they go north.
- "Track over the site after" is how many seconds after launch the orbital ground track crosses the pad. A rocket on its ascent moves slower than a satellite in orbit, so the point over the site on the finished orbit lags the launch moment; in Cees Bassa's experience (sattools) 300 s works well. The TLE epoch is set to that moment.
- Press "Compute TLE". RAAN and mean anomaly follow from geometry: the node comes from the pad's latitude and longitude, the inclination and the sidereal time of the crossing, then two SGP4 iterations refine it so the sub-satellite point at epoch lands exactly on the site. No trial and error as in the old faketle and satorbit guide.
- The mean motion in a TLE is Kozai's: from perigee and apogee altitudes it is computed with the J2 correction, as SGP4 itself does. Without it (as faketle did) a 500 km orbit comes out a few km too high or too low. The altitude in the "Position" line is osculating and may differ from the mean by ±10 km: that is normal.
- Check the result on the globe: "To epoch" sets the time to the crossing over the site, the ±1 and ±10 minute steps show the track. Such a TLE describes the orbit only after insertion; anything before the epoch is fiction.
- Launch delayed? Just change the launch time and compute again. The orbital plane turns with the sidereal time (about 0.9856° per day) and the other elements follow from the same geometry; no separate tool like launchtle is needed.
- If the launch documentation gives a state vector at separation, use "TLE from a state vector": it is more accurate than the geometry over the launch site. Russian documents give the vector in the Greenwich frame frozen at liftoff; pick it in the list and the vector is rotated by the sidereal time of the launch. Osculating elements cannot go into a TLE as they are: SGP4 works with mean elements, and a semi-major axis off by up to 10 km drifts the track by minutes per day.
- Compute passes over the network's active stations or your own location. The real insertion time is known to minutes and the orbit to kilometres, so look for the first passes with a ±5–10 minute margin and listen with a wide band. As soon as Space-Track or CelesTrak publish official elements, replace the preliminary TLE with them.