Which method STK uses to compute Umbra or Penumbra time intervals?

Question

someone knows in which way STK computes the eclipse times for a satellite?

I use a Matlab code taken from D.Vallado's library (ALGORITHM 34) but even if I use a PVT provided by STK (1 sec time step) with information about satellite and Sun position, there is a difference in seconds between the times computed by the code with respect the time listed in STKs report. I don't care about the milliseconds difference because it depends on the time sampling but I expect at least the start and end seconds to be similar.

I attach an image of the two results.

Thank you in advance

Answer 1

Although AGI freely publishes the Math Spec of their orbit determination product (ODTK), I'm not aware of a freely available MathSpec for STK.

However, the verification and validation of GMAT was done against STK. Further, Nyx is validated against GMAT. The Nyx eclipsing model is derived here: https://nyxspace.com/nyxspace/MathSpec/celestial/eclipse/ , and it achieved nearly identical results to GMAT (the errors are most likely due to ephemeris errors in Nyx, since it'll switch to SPICE file only in version 2.0). Note that the eclipse model is different than Vallado's Algorithm 34, but Vallado's Algorithm 35 is used for line-of-sight computations.