Comment by thehappypm
2 days ago
Does anyone have a good sense for the actual purpose of reusing rockets?
Call me skeptical of the actual cost savings. You now need systems in place to catch them, the rockets need more components and fuel and such to control descent, the rocket then need to be refurbished back to launch-ready after going through hellacious stress. It seems like it’d be cheaper, lighter, and simpler on the whole to just make a new one. But I would love to learn more.
Rocket engines are very complex machinery that's difficult and expensive to make. Reusing the rocket allows you to reuse the engine.
SpaceX has also managed to show that reused hardware can be more reliable than brand new hardware. You run the hardware through a number of tests before launch, but there is no better test than an actual launch. Satellites still cost a lot more than rockets, so reliability is a big deal
I do wonder if there's a way to reuse engines without reusing whole rockets, that could wind up cheaper. Pop the engine out of the rocket with some small fuel tanks, spin round and do a retro burn, then ... somehow land. I have not worked out all the details.
ULA calls this "SMART reuse" but has not yet successfully achieved it with their Vulcan rocket.
https://planetary.s3.amazonaws.com/web/assets/pictures/_2400...
In some ways this is simpler, but in others it is more complex since you need a way to catch the engines before they dunk in the salt water. A rocket booster that is already set up with avionics, engines, and fuel can guide and land itself without needing a precise helicopter catch etc.
I'm not sure how leaving the rest of the booster would make a propulsive landing easier. It seems like more complexity than just leaving some fuel in the tank, for less benefit. But ULA's Vulcan (which incidentally uses the same BE-4 engine as New Glenn) plans on using a helicopter to catch it's engines as they parachute down [1]
[1] https://www.planetary.org/space-images/smart
This concept is generally known as "Stage-and-a-half", and it was used on the Atlas rockets
As Diane Rehm once lamented on her radio show about two decades ago, as she was interviewing some cutting-edge engineer about his work, “why does it have to be so haaaard?!?”
A NASA planetary probe or a NRO spysat cost more than an order of magnitude more than launch but your run of the mill comm sat is only a bit more expensive than a SpaceX launch.
For the booster I think it’s fairly well established that it’s cheaper to reuse it. It’s not that hard on a booster to come down again, and it doesn’t seem like it’s that much work for SpaceX to refurbish their boosters.
For the upper stage it still remains to be seen. Having the upper stage return from orbit is a lot harder, as we’re seeing with the Starship. You need a lot of thermal protection. Maybe flaps depending on the design.
Jeff Bezos mentioned in a recent interview that for the upper stage they’re working on a reusable and a cost reduced version of the upper stage in parallell. They can’t determine yet which will end up being cheaper. That tells you something about how hard it is to do second stage reuse.
Rocket Lab has shown with the Neutron rocket what’s possible if you’re really innovative with optimising for cost on the second stage. Their second stage will probably be cheaper than the satellites they send to orbit
This was the argument a decade ago when SpaceX was the only one seriously talking about reuse. Given SpaceX is now responsible for the vast majoriy of mass to orbit it is hard to argue with the results, and now everyone serious is investing in reuse.
this is survivorship bias. SpaceX made a choice, and it succeeded; but SpaceX also made a million other design choices. Reusability could well be sub-optimal.
Your argument seems logical but the conclusion is wild, and I’m pretty sure extremely wrong. What other design choices, SpaceX made have been more significant for capturing 70% of all rocket launches in the world?
Reusing seems like the obvious and biggest design decision that led to their success
1 reply →
Note that SpaceX's whole schtick at the beginning was to specifically go for a very cheap single use rocket. All of their design decisions were based around making the Flacon 1 and later the Falcon 9 easy to mass produce. This was the prevailing philosophy after the Space Shuttle's reusability failed to bring down launch costs. SpaceX tried that and they failed to achieve the cost savings they wanted. That's when they pivoted to reusability, which did work, propelling them from well branded newcomer to dominant player in the launch market.
For SpaceX to keep up their current launch cadence without reusability, they would need to build a new Falcon 9 booster every 2-3 days instead of every two months. It would also require an 8x increase in their build rate of Merlin engines.
It's true that in principle, you need a stronger rocket (the whole rocket, not just the motor) and a bunch of extra infrastructure to be able to reuse.
However, without being able to recover a rocket, it's actually quite difficult to figure out just how much corners you can cut, while remaining reliable. Since blowing up revenue payload is an awful way to optimize this, I think this means that disposal rockets will be inefficient in a different way - there will be excess safety margin in the wrong areas.
Reliable re-use also changes the operating model of the company. Since each rocket in stock represents many customers over time, you don't need to be nearly as stressed about exactly matching your manufacturing pipeline to predicted demand. This likely also enables generally faster turn around time (as in from cheque signed to launch).
Finally, as it turns out, it's not unreasonable to expect a rocket to be reused like 20+ times. I think you're point would be reasonable if it turned out that reusing a rocket more than ~3-5 times was difficult. But like... it's REALLY hard to do disposal anything better than something that can be reused 20+ times.
By that logic there should be launchers of similar capabilities cheaper than F9... which there aren't.
Reason 2: Enables increased launch cadence.
There is an older (2015) discussion on this at https://forum.nasaspaceflight.com/index.php?topic=37390.0
Rocket engines are very expensive, and the whole rocket is fairly expensive in traditional construction methods. The cost of the rocket engines dominates the others.