Hacker News Reader

Comment by yellowapple | Hacker News Reader

I got it on the first try, but only because I've got 600+ hours of Kerbal Space Program under my belt (where docking is much harder when you consider that docking in KSP entails actually getting to the target and doing it with imbalanced RCS thrusters that I placed on the craft willy-nilly; SpaceX's online docking sim was a breeze in comparison).

Trick was to get rotation aligned first, then worry about translation. And yeah, might as well ignore visual; just went by the numbers.

View on HN

I had a terrible first attempt. Then discovered the same procedure, sort the rotation to 0, then minimise the translation and finally approach the dock. Head down And concentrate on the numbers. I didn’t have the KSP experience to fall back on. I wish they had added a timer as I am now trying for my fastest time.

Yeah, the live stream mentioned the ability to switch between coarse and fine controls, and that ability would be invaluable for a speed run.

You can do that by clicking in the middle of the controls.

Well shit. That would've been useful to know. Time for some speed-running!

If you do rotation then translation, you're either going to waste fuel or time. Just aim for the target and get within 20 meters before doing any of that, and the amount of translation you'll have to do is minimal. It's like traveling on the hypotenuse instead of the other two sides of a right triangle.

Rotating before translating should have little to no impact on fuel consumption. You still have to do the same amount of translation no matter how the craft is oriented. That means the same amount of energy, and thus the same amount of fuel. It took a little bit more time, but at the timeframes typical of orbital docking it's negligible.

In my first-try-success attempt, I was doing all translations at once, not something silly like lining up the numbers one at a time (that won't work well anyway, due to the differences in orbit between yourself and the docking target). That is: moving toward the ISS while simultaneously moving up/down and left/right to line up, and thus moving along the hypotenuse.

And by the way, this seems to be the exact approach Crew Dragon Endeavour took, judging by what I saw in the livestream early Sunday morning. They were already lined up rotationally before starting the final docking approach.

Yes, you're right. You can of course travel on the hypotenuse even if you're not oriented along it. I was probably stuck in some fixed front wheel thinking. I would think it's a lot easier to get the the proportion of lateral speed exactly right if you're aimed along the hypotenuse though.

True. Still, if you wanted to get really exact about it, you could surely work it out mathematically (i.e. by making sure your velocity components are proportional to your distance components).

So for example, with the simulator loaded up in a tab right now, after lining up the rotation, I'm seeing:

    X | 200.0 m
    Y | 12.0 m
    Z | 30.0 m

So if I then apply an X thrust of 20 m/s, Y thrust of 1.2 m/s, and Z thrust of 3.0 m/s, then I should travel along the hypotenuse and reach the docking port in 10 seconds. Likewise, a thrust of X=2,Y=0.12,Z=0.3 would get me there in 100 seconds (and X=200,Y=12,Z=30 would get me there in an insane single second). Of course, some corrections will be necessary along the way to account for small differences between my orbit and the target orbit (which is more relevant the longer it takes for you to get to the target), and the velocity change ain't instantaneous so I'll need to adjust for that, too (making 10s or 1s runs kinda implausible with the current interface, unless perhaps I do some kind of browser-level scripting to click the buttons and read the numbers).

Crucially, this makes it feasible for a computer to do automatically (though automated docking tends to stop shy of actually docking from what I understand, instead using CanadArm to grab the craft and put it on the dock).

You're right that aiming along the hypotenuse makes this a little bit easier to eyeball, but it also makes it a bit harder in other respects since now you'll have to adjust both rotation and translation along the way to stay along that slightly-curving hypotenuse.

Thta's an interesting point, and something I thought a lot about while I was doing this.

Basically, this simulator just cares about the physics of it, but not resources (Fuel, oxygen, etc).

I did it exactly that way, first rotation then translation. The 2nd fail it took me so long that in the 3rd (and successful) run I just hit the metal after getting the rotation correctly.

Side note, it's much easier to get the rotation right first and then moving the ship.

It was fun, thought.

Reply on news.ycombinator.com