The NASA Spaceflight article Extended Russian EVA complete – conducts satellite deployments says:

The first order of the day was for Ryazanskiy to head into the manual deployment of five nanosatellites from a ladder outside the airlock, following set up tasks including the hosting of a Go-Pro camera. (emphasis added)

The satellites, each of which has a mass of about 11 pounds, have a variety of purposes.

One of the satellites, with casings made using 3-D printing technology, will test the effect of the low-Earth-orbit environment on the composition of 3-D printed materials. Another satellite contains recorded greetings to the people of Earth in 11 languages.

It sounds like they were just tossed by hand into space. What are the orbital mechanical consideration behind "hand-launching" things from the ISS? What keeps it from coming back and hitting you in the back of the head 93 minutes later? (or from the side 46.5 minutes later for that matter)

Seriously, are there rules and constraints applied to launching satellites by hand from the ISS so that there is very little risk of it intercepting the ISS at some point later in time? Certain directions and/or speeds only?

More background on the nanosatellites:

A third satellite commemorates the 60th anniversary of the Sputnik 1 launch and the 160th anniversary of the birth of Russian scientist Konstantin Tsiolkovsky.

  • 1
    $\begingroup$ If you toss a satellite by hand from ISS, you modify its orbit slightly. But on a different orbit it could not hit the ISS. The effect of drag in the low orbit will be different for the large ISS and the tiny nanosatellite, therefore a hit after 93 minutes later is unlikely. The orbit height decay of the ISS is about 2 km in half a month, that is 8 m per cycle. $\endgroup$
    – Uwe
    Commented Aug 20, 2017 at 16:32
  • $\begingroup$ @Uwe sounds good but remember, when hand-launching you control both the magnitude and direction of the $\Delta v$, and an 8m loss of altitude corresponds to only a few millimeters/second. Also the non-uniformity of Earth's gravity introduces another source of uncertainty. I'm not sure "unlikely" is good enough when it comes to the safety of the ISS, so I'm guessing that there are " rules and constraints" so that there will be "very little risk". If you are convinced that there are no rules, you could leave that as an answer and see what happens. $\endgroup$
    – uhoh
    Commented Aug 20, 2017 at 23:23
  • 2
    $\begingroup$ Atmospheric and other effects aside: Even if the satellite encounters the ISS again at some point their relative velocities won't be much different from the initial toss. $\endgroup$ Commented Aug 21, 2017 at 9:53
  • $\begingroup$ @RomanReiner and all the photovoltaic panels and ammonia-filled thermal radiator panels, and high-gain antennas (some shown in youtu.be/PJzjs4EI22k) are all designed and rated to accept impacts from astronaut-thrown and returning 5kg cubesats? That's not considering that the velocity difference could change different after a few orbits - the Earth is not a perfect monopole. If you think it's safe to throw things off the ISS any way you feel like because nothing could go wrong and so there are no rules, consider posting that as an answer! $\endgroup$
    – uhoh
    Commented Aug 21, 2017 at 10:18
  • $\begingroup$ @uhoh If the orbit of the ISS and the nanosatellite differ only by 1 m in height by different decay, their periods differ only by 1.2 millisecond for a circular orbit in 400 km height. But at a speed of 7.67 km/s, 1.2 millisecond means a distance of 9.2 m. As you wrote the difference in speed is only 0.56 mm/s, but the circumference of the orbit is 42543 km. If the difference in orbit height increases by one 1m for each orbit, the tangential distance increases by another 9.2 m. I think it is very unlikely that the decay of the orbits of the ISS and the nanosatellite would be equal. $\endgroup$
    – Uwe
    Commented Aug 21, 2017 at 21:38

1 Answer 1


If you launch nadir and retrograde, you will put the object into a lower energy orbit such that, barring ISS deboosts (they do happen, but they are rare), it will never again intersect ISS's orbit. Launching prograde, zenith, or out of plane will set you up for a potential recontact scenario.

These objects are also large enough that they can be independently tracked once they've gained enough separation from ISS, so any unlikely potential recontact scenario can be averted.

  • 1
    $\begingroup$ Is this what's actually done in practice? $\endgroup$
    – Chris
    Commented Aug 21, 2017 at 22:53
  • 1
    $\begingroup$ Arguments based on intuition assume the Earth's gravity field is a perfect monopole; $\mathbf{r}/|r|^3$ but this isn't the case, so hand waving arguments on the level of meters and milliseconds are not necessarily valid. Also I've asked "...are there rules and constraints applied to launching satellites by hand from the ISS...?" rather than "Do you personally feel that the chances of something bad happening are small?" Is it true that you can just do whatever you want, or are there rules and constraints? $\endgroup$
    – uhoh
    Commented Aug 21, 2017 at 23:42
  • 4
    $\begingroup$ There is a pre-prescribed acceptable range of jettison trajectories that is already analyzed and acceptable to the program. There's a cone that faces down and aft, approximately 45 degrees in width, where as long as something is launched in that direction with a good velocity (at least 1 m/s or so), recontact is all but eliminated as a possibility. I promise you that these aren't hand-wavey arguments but are informed by actual analysis, which I cannot post here. $\endgroup$
    – Tristan
    Commented Aug 22, 2017 at 13:06
  • 3
    $\begingroup$ @Tristan Would the answer be improved by having the information in that comment edited into it? $\endgroup$ Commented Aug 22, 2017 at 19:37
  • 1
    $\begingroup$ @Tristan the wording in your comment is better than in the answer; can you consider porting some of it back up there in some form? 'there is a cone' and 'at least 1 m/s'. I think that's certainly good enough, and is easy enough to be checked with some simulations that in this particular case would not need a further link to make it verifiable. Thanks! $\endgroup$
    – uhoh
    Commented Sep 2, 2017 at 11:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.