Hacker News Reader

Comment by wffurr | Hacker News Reader

So what changed to make re-usable rockets reality? Is there a specific advance in materials science or computing that made this possible not just for SpaceX but other companies too? Or is it just an array of advances in a variety of fields that came together?

Reading about the DC-X, it seems like this was possible in the 1990s but NASA had no interest in funding it for whatever mysterious reason: https://en.wikipedia.org/wiki/McDonnell_Douglas_DC-X

View on HN

Propulsive landing has been possible since the 70s. But cheap and easily available sensors (especially but not exclusively navigation--gyroscopes, accelerators, altimeters, GPS), cheap and light high-performance compute hardware, and existing packaged software and other advances in stuff like convex optimization and fluid dynamics make a real difference in practicality of doing so, both on the rocket and during design.

There's also significant improvements in metallurgy that make it easier to do things like long-life turbopumps, carbon-fiber composites help with mass fraction, CNC machining and additive manufacturing that make complicated parts affordable, and other such things that also help. But really it's pocket-sized sense-and-compute that really kicked off the low-cost retro-propulsion thing.

Also, SpaceX being the first to demonstrate it was feasible. There were naysayers even after they landed their first booster in December of 2015. ("refurbishment will cost too much!")

Now that SpaceX regularly flying a booster 5 times or more and have at least 10 boosters in the fleet, the only people arguing that reusable space hardware is not worth doing are people with a vested interest in old-space-hardware designs.

Thanks - sounds like it was a lot of things but mainly low cost compute and sensors. Not to mention Elon's deep pockets and appetite for risk.

Yes; someone willing and able to try was also a missing ingredient. None of the incumbents seemed interested in trying anything particularly different (and for the most part, still aren't) so who knows how long we'd have waited for a new stab at reusable launch without SpaceX.

Nit: propulsive landing was not only feasible but actually done in the late 1960s. On a human-rated Single Stage To Orbit vehicle, no less.

Okay, fine, so it wasn't to Earth orbit, but still...!

That argument really doesn't work. Most of those censors are needed in orbital rocket too.

Shuttle software was far more complex then Falcon 9 and it runs on 70s computers.

GPS is not really needed, its just there as a backup. Launching on a ship might be tricky, but on land you can target a pretty big landing area.

> advances in stuff like convex optimization and fluid dynamics make a real difference in practicality of doing so, both on the rocket and during design.

True, but if you do more iterations on first stages that you drop into the ocean anyway, you can build design and iterate.

Shuttle didn't fail because of they could figure out reentry conditions.

> There's also significant improvements in metallurgy that make it easier to do things like long-life turbopumps, carbon-fiber composites help with mass fraction,

SpaceX uses very little carbon fiber.

There is truth to the improvements in metallurgy but its not metal issue that caused the RS-25 to not work well in practice is not really the metal.

> CNC machining and additive manufacturing that make complicated parts affordable

While true, that makes the benefit of making it reusable even more worth it.

It seems to me that SpaceX simple had the right design and the right setup to test and iterate. There was nothing fundamentally stopping NASA from doing this instead of the Shuttle. Nobody else even tried to make a reusable rocket.

I would argue this could have been done in the 70s and it would have worked nearly as well.

The Space Shuttle was reusable, and it was designed in the 70s. Even it's side boosters were partially reusable.

It wasn't usefully reusable, refurb costs were too high, but that was as much due to design mistakes as it was to technology limitations in the 70s.

Better CFD, better knowledge of engine design, limited scope of requirements/capabilities, better material science.

NASA is a government agency. It isn't in governments' interest to lower the cost to orbit since the space-capable large governments enjoy a monopoly on space when it is expensive. Once NASA opened up commercial resupply contracting, a door opened a crack for people and organizations without a similar monopoly interest and with ideas for cost reduction.

It was possible. Like others have commented already, it's much more affordable to do that today, given sensors, computing and advances on our understanding of the whole field. But given large enough pockets, I'm pretty sure by the time we had the capability to build Space Shuttles, we could do reusable rockets instead.

There's one point I would like to add. This is something that's very difficult to do for well established companies. They tend to have their workhorse rockets - which were NOT designed with reusability in mind. Modifying these rockets is not feasible in many cases. Even when it is, it may not be cost effective. So the usual solution is to design a new "reusable rocket program".

Now, the problem with such a program is: the success criteria is reusability. If the new rocket cannot be put into service quickly enough and demonstrates reusability successfully, it will be scrapped. If it delivers payloads just fine but can't land (or can land but refurbishment is costly), it will be a failure. After all, on one hand you have a working system, which is generating revenue. On the other, you have a problematic R&D program that's draining resources and engineering cycles for the promise of potential savings. They end up getting scrapped on the first resource crunch.

SpaceX was developing their rocket. Their success criteria was that it would deliver payloads into orbit. Given that they started from scratch and had no existing workhorse, they also added reusability as a goal and designed the rocket to allow for that. If it achieved reusability, great! If not, it's just another single-use rocket. They also got "early" adopters, companies that were willing to launch their payloads on SpaceX, on new rocket designs, that only had a handful of flights. So they were, by definition, a little less risk-averse.

On every flight, SpaceX got closer and closer to the reusability goal. But that didn't matter to most customers (they would pay for new boosters anyway), because that part of the mission happened after their payload was already on its way by the second stage, so who cares what happens to the first. All the while cementing their reputation.

In essence, SpaceX got companies to finance the R&D for the reusable boosters, because they only had one rocket. If they tried to start with a single-use rocket, and then created a "Falcon 9-reuse" version, they would have faced the same difficulties. Namely, who would fly on the untested new design?

Rocketlabs might be able to pull it off still, because - while it's a new booster - it's also one that's intended to increase their capabilities (much like Spaceship), it's not just a "reusable rocket program". If it can't be reused, then it's a more expensive rocket, but one that still adds value.

Shuttle was fully reusable. The boosters came back down on parachutes.

Main tank

Main tank is lost and side booster and shuttle required extensive refurbishing. It sorta counts, but against the point of reusability

Starship moved to tiles too, instead of transpirational cooling, it will likely require refurbishment.

If shuttle had continued on tile inspection could have likely moved to machine vision based approaches, maybe they will be able to use something like that to keep refurb costs down on spacex's shuttle/starship.

They are created with better materials and also do not require special one offs for every single tile, unlike the shuttle.

Replacement is supposed to be "break it and attach a new one" without needing special glues, curing times which was the reason why shuttle heat shield refurbishment took so long at high cost.

That sounds like an improvement, but not enough for passenger air travel replacement between the cost of coach and business class (New York to shanghai), promised for 7 years from now. They need many orders of magnitude of reliability improvements over shuttle for that, though it would be slightly suborbital reentry.

The Dear Moon mission in a couple years will have full reentry at extreme speeds, assuming they weren't just ripping off that Japanese billionaire guy. I think they still planned transpirational cooling at that point?

Dragon was supposed to have at least one mission to Mars every transfer period from 2020 onwards and that seems to have been scrapped.

I dont understand where this comment is coming from. What sources are you using the make the cost claims you are making and/or the reliability requirements and/or heatshield requirements?

> Shotwell estimated the ticket cost would be somewhere between economy and business class on a plane — so, likely in the thousands of dollars for transoceanic travel. “But you do it in an hour.”

https://www.vox.com/2018/4/11/17227036/flight-spacex-gwynne-...

I based reliability requirements on passenger jet travel and the risk people would be willing to take for non space-tourism transport.

Also, I don't see how the tiles can be fully uniform except on the cylindrical part. The geometry of the nose part I don't think would allow it mathematically. Shuttle tiles often only differed in thickness based on needed heat withstanding, and could be generated by cnc processes automatically. Spacex may want something similar to optimize weight, especially since they are planning computer controlled install as well for most of them.

The orbiter was reusable, the boosters were refurbishable, and the external tank basically crashed into the ocean.

Always thought it was a shame they didn’t push the main tank into orbit. Just think of an ISS made from a few of them rather than what we have now or the Ares from Red Mars.

There were some orbital concepts that basically inflated a living space inside of an orange tank and stuck a few solar panels and radiators on the side. Shuttle trips were just too expensive to make it practical and with the ISS it would have been redundant.

Alternative designs could have worked, but the Shuttle used the OMS after main engine cutoff, so keeping the tank would have eaten a ton of delta-vee

The shuttle required refurbishments on the thermal protection system. Every single one had to be inspected.

The engines also required refurbishment.

The boosters had to be fished out of the ocean. Those same boosters that used solid fuel and that could not (and cannot) be shutdown in case of issues. And the Shuttle happened to have two of them, with a decent moment arm on each, so a solid booster failure (even a partial one, with less thrust than expected) meant a mission failure. They had to be highly scrutinized. Given that the Shuttle was also a jobs program, they were built in segments, to allow for transport across long distances. So the o-ring seals were also problematic.

All in all, the Shuttle was very brittle. If they could have been mass-produced, it would have probably been more economical to just throw them away on every flight and rely on economies of scale. At least, that way, one would only have to account for manufacturing issues, not every conceivable stress that could happen during a mission.

Reply on news.ycombinator.com