There’s a much easier way to handle the first 100 meters: a moving pipe to keep the propellant topped off as the rocket rises. Trivial compared to a machine that accelerates the whole rocket, and only a little more complicated than the current “chopstick” system.
And… it’s surprisingly helpful. Starship accelerates at about 1.5 gees initially. So it takes 3.65 s to rise 100 m.
The Raptors burn about 700 kg/s each, or 23.1 tons/s total. So 84 tons of prop is burnt during that first 100 m.
At MECO (stage separation time), the “dry” mass is about 100 t payload, 1300 t wet upper stage, 200 t dry mass of lower stage, and perhaps 50 t of reserve propellant for landing. That’s 1650 t total. But let’s say there’s an extra 84 t of prop available.
Plugging into a rocket equation calculator, we get 514 m/s. That’s not trivial. The total needed to make orbit is about 9 km/s. And the rocket equation makes that last bit of performance even more important.
Were you to be able to top off the 84T of propellant over the first 100m of flight, the acceleration would be reduced by that much throughout the rest of the first stage flight profile. Or at least it would be during the full throttle portions. During the partial throttle portions you could offset the extra weight with more throttle. And hence more fuel consumption.
Almost certainly you’d gain overall, but there’s a lot of pain alongside that gain. 5 steps forward and 4.5 steps back.
True. I was assuming that would be negligible, but now I’m not so sure. It’s small compared to the total mass (5000 t at liftoff), and fairly small compared to the final mass of 1650 t, but still… hard to say how the different factors contribute.
I’d say the way to calc it is to determine the vehicle velocity at the 3.65s / 100m altitude. Which would be something north of 30m/s and WAG less than 60m/s.
Then plug that into dear old Tsiolkovsky as the v0 with all other parameters the same as a fully fueled launch from the pad.
55 m/s by the previous numbers. But you get more benefit than that, because, going faster in the first place, you’re less subject to gravity losses.
Also, I screwed up in a different way. The thrust is equivalent to 1.5 gees. But it’s going near-vertical then, so it’s only accelerating at 0.5 gees. So it actually spends 6.3 s going the first 100 m, which sounds like a larger benefit at first, but since it’s not accelerating as quickly you only get about 32 m/s of improvement.
But like I said, I don’t think that’s right either, except as a lower bound. Ugh, rocket science is hard .
There’s also the factor that if you’re pumping propellant into the rocket as it’s rising the pumps doing so are doing the work of lifting that 84 t of propellant and presumably getting it to match the rocket’s velocity at that point.
At this point the rocket sled going up the side of a mountain is sounding like the less complicated option.
Just put a “water tower” at approximately the final altitude (a little above, since it needs to overcome the ullage pressure). You need to move the hose along a vertical track, but SpaceX is already doing that with the chopsticks. With a counterweight, it should be easy to slide up and down, though you do need to slow it down when it reaches the topmost point.
That’s the whole thing, BTW. It’s not that they just haven’t added all the extra tubes and wiring and such yet–all that stuff has been either deleted or made internal. I anticipate that the flat sandwich-looking piece contains a lot of complication that’s hidden from view.
Apparently, they’ve added a lot of external regenerative cooling. Many/most modern liquid-fuel engines use regenerative cooling for the nozzle–that is, they run the propellant through internal channels to keep it cool. But this engine needs to survive reentry without extra heat shielding. So some of the external components have cooling channels as well. I’m not sure if this is unique, but it’s at least unusual.
Comparison pic vs. previous Raptor models in case you aren’t convinced yet:
Very interesting evolution. I have always thought, looking at rocket motors in the NASA visitor center for example, how ridiculously complicated they look for what they need to do.
Basically take in fuel and oxidizer, burn it & shove the exhaust out the back end?
Even the Raptor 3 has a rat’s nest of pipes… it’s just hidden. They’re making heavy use of 3D printing and its ability to make blocks of material that contain all kinds of hidden pathways. So this probably couldn’t have been done a decade ago.
It’s still heavily simplified, though. The Raptor 1 was to some extent a development engine and contained a lot of extra sensors so they could learn how the engine behaved. Once they learned what they needed, they were able to reduce that a great deal on Raptor 2. Raptor 3 undoubtedly removes even more.
The basic design itself is surely also simplified. In what ways exactly, I couldn’t say. But even on a rocket engine, there are a lot of “just in case” elements that prove unnecessary once you’ve built the thing. Check valves, purge lines, that kind of thing. You have to make a concerted effort to take them out.
Raptor 3 does make one big sacrifice: it’s harder to repair. Far fewer bolted flanges, which means to take it apart you have to cut it and weld it back together. Musk says this works fine. And it’s surely the right way to go when the engine is reliable enough. Might be some growing pains, though.
All true, but when you’re building the engines at the volume SpaceX plans, much easier just to pull the broken engine and fix it at leisure, while a new engine takes its place. Cadence is king.
Yep. They’re pretty much at that level already, even with R2. A couple times now they’ve swapped in a new engine right on the pad. Still worth repairing, but as you say, it can be done at their leisure. They can have a rework shop sized for average repair load, not peak.
They hope to eventually achieve >300 tons thrust (a 200:1 thrust/weight ratio). It’s already roughly matching the Merlin 1D at ~180:1, but with a much more advanced engine cycle and much higher thrust per unit area.
Also, they’ve apparently already produced 569 Raptor 1+2 engines (they restarted the numbering with R3, but not R2).
I’ve said it before and I’ll say it again, for all his many personal faults (and they are legion), Elon cracked the code on space flight in a way that may never be repeated. And frankly, if you aren’t working towards high-reliability, high-cadence, highly-reusable, and (relatively) cheap launch vehicles, then you’re a buggy-whip maker in the age of automobiles - a dinosaur doomed to extinction. We need options to SpaceX, but we need them to be as good (or nearly so) as SpaceX at all 4 requirements.
It helps beyond measure that Musk wants these rockets and engines for his dream project of colonizing Mars. To do that they have to work and they have to do so economically. SpaceX is largely its own customer with the Starlink project. The contrast between this and a contractor who is simply building hardware to spec as ordered cannot be overstated.
.
.