Keep in mind that every millilitre of fuel burned has to be carried from the start.
If you want to start burning a few seconds early on landing by policy, you need to lift that fuel, and the fuel to lift that fuel, from the surface to space and then decelerate all of that. Just for those seconds. If you’re doing this from Mars, you also have to accelerate that mass around the Solar System and out of the Martian gravity well.
Not a bad thought. Just wanted to illustrate how energetically leveraged these systems are.
That's true for missions beyond LEO, but for routine LEO missions there's no expectation to do refuels for every flight. Also vehicles returning from missions beyond LEO will not be able to refuel before re-entry and landing.
That's a good question actually, given that they had another engine as well. Or maybe light up all three and then shut down the one they decide not to use.
But these are all in the 'why don't they' category, SpaceX has some pretty clever cookies, experts in their field, the chances that anything we can think of has not been thought of by them are nil.
I'm reminded of the story of two economists walking down the street. One sees a dollar bill on the ground, and the other decides "that can't be right; if there was a dollar bill on the ground someone would've picked it up already".
Not saying that a bunch of random folks on HN are seeing dollar bills in this analogy, but I've never quite been satisfied with the "if we're thinking of it then they probably already tried it" explanation :)
"Question: Why only light 2 engines for landing? Any engine failure means loss of vehicle, so you have two single points of failure. Why not light all 3, do the flip, then pick the best two and turn off the other?"
To be fair they selected it quite some time ago now. There's a lot of very promising research now though, so hopefully they will have better options available down the line.
> To be fair they selected it quite some time ago now.
They've had better fuel options from the very beginning. I guess they didn't have space cycles to additionally solve this issue; hope they'll come back to it.
The landing propellant comes from very small header tanks (to reduce slosh), and those tank sizes constrain how long they can run a landing burn. Future iterations may end up with a bigger header: these are still early prototypes.
Or does it not matter because a few more seconds won't increase the chances they'll be able to relight?