Hacker News Reader

Comment by jontro | Hacker News Reader

Nice, though I wonder what is so notable with the ground track offset.

It shows that the active guidance works as expected. Active guidance is extremely difficult. The hardware is difficult because you need to design jetvanes that work under extreme pressure, and in a 3000 degree jet exhaust. The software is difficult because it needs to be extremely reliable while still compensating for a lot of factors such as wind, spin, air density, etc. and you can't test it, so there is no space for bugs.

No amateurs have ever even attempted active guidance before, and getting it right in the first go is a huge accomplishment.

Regarding testing, aren't there any possibilities of testing just parts of the software, possibly via simulation, or is everything so inter-dependent that a live physical test is the only way to go? It's pretty impressive to get something so complex right in the first try! Props!

You can of course test subsystems, how they play with each other, I/O, etc. But the main job of actually sending the rocket straight up is not testable. You could build simulations, but they wouldn't help much since there are so many parameters, unknowns, etc. that building an accurate simulation would probably be as expensive as a test.

Besides, tests are much more fun!

Just build it in Kerbal Space Program!

This is incorrect, there have been successfully flown, actively stabilised, passively unstable model rockets made by amateurs before. E.g. Gyroc in the UK.

michael.sdf-eu.org/Gyroc/

There have also been a few in the states. All passively unstable too.

It means that the guidance system kept the rocket flying straight upwards.

Reply on news.ycombinator.com