Second guessing their arguments

[JayUtah]

Reuse is a complicated proposition between economics and technology.  We know we can build reusable rockets.  The challenge is to build and operate them at only a marginal increase in cost.  Up until now, the cost of building and operating a reliable reusable rocket has been astronomically greater than an expendable launch system, per unit mass of payload.  Clearly SpaceX has reached the point of duplicating previous attempts to soft-land a rocket-powered vehicle on Earth.  (The SpaceX fanboys are, of course, claiming that SpaceX has "invented" this technology.)  The question on everyone's mind is whether Elon Musk has figured out the magic formula for doing it economically and reliably.

The economics comprise more than just the cost of the vehicle itself, amortized over however many launches it's good for.  The cost of each launch includes the refurbishment costs, and the more complex checkout procedures for the necessarily more complex vehicle.  Adding complexity to a system by giving it additional capabilities (especially ones not directly related to its operational purpose) almost always ends up adding more than incremental cost.

Similarly the reliability issue has to look at the systemic complexity of the newly complex design.  We have shown via many failures how bad we are (as an engineering industry) at reasoning about systemic complexity, and thus about bounding its reliabiltiy expectations.  For example, will we need tighter weather constraints to allow for the first-stage flyback?  Recovery fleet readiness?  Range safety during flyback?  SpaceX does not have a very stellar record in terms of on-time launches, all tolled.  For the types of mission they want to bid on, missing the primary launch window is considered a mission failure.

Historically how this has worked out is that you can have a fleet of relatively cheap expendable vehicles, operating them with the understanding that they will not be very reliable.  If you add money to make each unit more reliable, your costs increase far faster than your reliability.  Proposing to make a vehicle reusable makes it necessary to make it more reliable, which is a losing proposition because you not only have to make the go-up part more reliable, but also the newly-minted come-down part equally reliable.  The reliability has to be there so that you have a truly reusable rocket, not just a very fancy, very expensive "reusable" rocket that has, in practice, a high failure rate.  Lowering the failure rate is the gilt-edged proposition.  And lowering the failure-rate is not universally a matter of one-time NRE costs; it's substantially a per-flight cost as well.

So I'm legitimately interested to see whether Elon Musk has figured out the workable mix of technology and economics to make his vision come true.  If so, it really will revolutionize launch services.

[ka9q]

I too am fascinated by the SpaceX approach to (partial) reusability. I was astounded when I first learned they don't use parachutes at all. Parachutes have always been used on Mars, for example, even though the atmosphere is much too thin to give a survivable landing speed with parachutes alone; that final velocity reduction can only be done with rockets (maybe plus airbags).

This is true even on (solid) earth; that's why the Soyuz has landing rockets. But I'm still baffled as to why SpaceX doesn't use parachutes at all. Every newton-second of parachute drag should mean one less newton-second of retro-rocket impulse, and that much less extra propellant needed to produce it.

What am I missing?

[ka9q]

Since the whole purpose of reusability is economics, you really should look at it from that standpoint. And do a proper job this time, not the politically motivated whitewash job that "justified" the Shuttle.

What are the most expensive (and potentially reusable) parts of a rocket, and how might you get them back if you can't recover the whole thing?

Seems to me the most expensive parts by far of a Falcon-9 first stage (and of any liquid rocket stage) have to be the rocket engines. The rest is just aluminum plus some electronics that could probably be cost-reduced substantially through mass production.

So why not separate and recover just the engines? Cut the thrust structure free from the bottom of the stage, parachute the engines back to earth and throw the rest of the stage away.

There's still the minor problem of protecting the engines from contact with salt water and/or a hard impact on land, but those are engineering details...

Edited to add: Oh wait, I had an even better idea. Leave the engines attached to the stage, guide it back to a landing spot, and then flip it upside down before it impacts the ground. The empty tankage will provide lots of crunch space to protect the engines.

Seems to me SpaceX has already gotten this far, so all they need is to make this simple little tweak and they're there!

Or maybe they can dig a big crater and cover it with an elastic net...

I think I should go to bed now.

[Peter B]

I too am fascinated by the SpaceX approach to (partial) reusability. I was astounded when I first learned they don't use parachutes at all. Parachutes have always been used on Mars, for example, even though the atmosphere is much too thin to give a survivable landing speed with parachutes alone; that final velocity reduction can only be done with rockets (maybe plus airbags).

This is true even on (solid) earth; that's why the Soyuz has landing rockets. But I'm still baffled as to why SpaceX doesn't use parachutes at all. Every newton-second of parachute drag should mean one less newton-second of retro-rocket impulse, and that much less extra propellant needed to produce it.

What am I missing?

That it's a whole extra subsystem with all its attendant electronic and pyrotechnic requirements?

I don't know as I'm not an engineer. But it seems to me that if you add an extra system with all its bits and pieces, that's one more system you have to test and make sure it doesn't go wrong. By using just engines, all they have to worry about is the braking and steering engines doing their thing (and the landing legs extending).

[Echnaton]

It strikes me that it is a matter of weight.  I've wondered about a solutions similar to the one ka9q suggests to save and reuse the business end of the first stage while letting the tankage part go to the recycling bin.   

When compared to lifting the weight of the fuel needed for a non-parachute assisted soft landing, it naively seems preferable to take an incremental approach to reuse the most expensive parts.  Whether that be cutting the bottom loose for individual return or using the tankage as a crumple zone.   

Maybe the economics of rebuilding the engines are well enough understood that the save the whole approach is ultimately more economical.  That is, if the engine recovery is only marginally cheaper than using new, then it may only make sense to completely refurbish the first stage. 

But they don't call this rocket science for nothing.   Good luck to them, it is exciting work. 

[Al Johnston]

If they're wedded to the idea of landing on a ship, are parachutes accurate enough?

[JayUtah]

Thrust-bearing structures and propellant feeds have to be robust enough that making them also cleanly severable runs additional flight risks.  But yeah, parachutes?  I guess those aren't sexy enough for SpaceX.  More likely it comes down again to structure.  You have to hang the stage from a parachute, which means providing for that in the structural design.  Yes, you have factory hoist points anyway, but you typically can't reuse those for parachutes just because of the qualitative differences between parachuting and factory tooling.  You already have a thrust structure, so you really only have to deal with flight load paths.

Parachutes by themselves won't be accurate enough for a soft pinpoint landing.  You'll have wind drift, which you can't easily compensate for with that type of parachute.  Personally I'd just move the ship to be under the stage, but that's just me.  You need the ship because you need to soft-land.  You need to soft-land to protect your delicate hardware.  For reasons...

Yes, a parachute system is a separate subsystem with pyrotechnics and all those things that have to be checked out.  But you can often make those very self-contained -- they can have their own control logic, power supply, etc.  The big danger in adding new subsystems to the design is not that they'll just be more baggage and overhead, but that they'll interfere with the primary systems.  It's one class of accident if something happens to the parachute system and it doesn't deploy when needed.  It's a massively worse class of accident if it goes off prematurely and crashes the vehicle.

And that's why I'm still baffled at SpaceX's choice of a powered descent using the primary propulsion and primary guidance-and-control.  It means the earth-landing functions have to be built into, and share a design space with, those primary systems.  And that means complicating something that ordinarily you work really, really hard to keep as simple and foolproof as possible.  Not only that, it complicates the systems SpaceX has historically had a hard time with.

[ka9q]

The reason I'm baffled by SpaceX not using parachutes is that without them, you have to get rid of all that velocity with rocket engines, so you have to carry more propellant, and then you have to carry more propellant to lift the weight of the propellant you need to land, and so forth. Remember the form of the Tsiolkovsky rocket equation:

dV = Ve * ln(Ms/Mf)

which means the mass ratio increases exponentially with the required delta-V.

This is one reason it's so hard to soft-land on the moon. It's big enough to have substantial gravity and an escape velocity of 2.38 km/s that you have to get rid of, but not big enough to have an atmosphere for heat shields and parachutes. In some ways, it's much easier to land on Mars despite an escape velocity more than twice that of the moon, though of course it's far harder to get back off again.

Just how much delta-V is required to land a separated first stage, anyway? Could it be that they still get plenty of velocity reduction during the initial re-entry from aerodynamic drag on the body of the stage?

[ka9q]

If they're wedded to the idea of landing on a ship, are parachutes accurate enough?

Parachutes need not be your only mechanism; they're just to get rid of most of the energy so you don't have to get rid of it with rockets burning heavy propellants. You still need the rockets to get a survivable landing velocity, because the parachutes have too high a terminal velocity. And you can use those rockets for steering as well.

Apollo (and Curiosity, and Orion) do a pretty good job of steering to an accurate landing by taking advantage of aerodynamics before opening their parachutes. You could probably do the same when recovering a stage, again to reduce the propellant required for the rockets you'd still need for an accurate, low velocity landing.

[JayUtah]

The grid fins are meant for passive flight control during aerodynamic re-entry.  So I suppose they're meant to keep the stage in an attitude they've determined gives them the most flight control and/or drag.

[Luke Pemberton]

The discussion about parachutes has jolted my memory from the other day. So many thanks*.

How are parachutes different to land a vehicle on Mars, say to landing a vehicle on Earth? The reason I ask is that Ralph Rene (I think) claimed that some of probes on Mars were faked as parachutes would not work in such a rarefied atmosphere. I never really found out the answer to his rubbish.

* Does anyone else have those moments where they think of a question/problem and then 1 minute later they have forgotten it because they got sidetracked in another thought, a bit like going into a room for something and forgetting what you went in for? Or is it just me?

[ka9q]

Rene was an idiot, of course, but there was actually a grain of truth in his statement.

Because of the thin Martian atmosphere the terminal velocity of a parachute is much too high for a survivable landing. But this is precisely why every successful Martian lander has additionally used at least two other methods to lose energy.

1. You invariably start with a heat shield, which can be seen as a sort of hypervelocity parachute. The Martian atmosphere may have only 1% of the earth's atmospheric density, but entering either atmosphere at escape velocities is quickly fatal without a carefully designed heat shield. This gets rid of the vast majority of your initial kinetic energy.

2. Then you deploy a parachute, even though the terminal velocity of any practical parachute near the Martian surface is an unsurvivable few hundred meters/sec. It still dumps nearly all of the kinetic energy remaining after the heat-shield phase.

3. Just before you crash, you fire rockets to bring you to a (near) standstill. Rockets are inefficient at low velocities, but there's really no practical alternative just yet. Even though Pathfinder and the MERs (Spirit and Opportunity) had airbags, they still used rockets to arrest their descent just above the ground. The airbags did allow the use of solid-fuel landing rockets, which otherwise would have been too unpredictable to provide a reliable soft landing directly on the surface.

Maybe someday somebody will build a big net over a deep Martian crater and make landing rockets unnecessary.

[gwiz]

The grid fins are meant for passive flight control during aerodynamic re-entry.  So I suppose they're meant to keep the stage in an attitude they've determined gives them the most flight control and/or drag.

I got the impression that the fins were for active guidance, the crash was reportedly due to running out of hydraulic fluid in the fin control system.

[Allan F]

Strange - that a system that dumps the hydraulic food FLUID overboard should be better than a closed system.

[ka9q]

What was the hydraulic fluid? I know many kerosene-burning engines use their own fuel as a hydraulic fluid, e.g., for actuating valves and gimbals.