[deepblueocean]

I find this confusing. It's awesome, that's for sure. But it seems like lifting all the fuel needed to do a soft landing is inherently a tradeoff for less payload capacity. What is the benefit, then? Is it just so much cheaper to be able to re-use the first two stages?

[InclinedPlane]

Almost the entirety of the cost of an orbital launch is in manufacturing and operational costs. Cost of fuel is basically just noise. If you can reduce operational turnaround time and operational complexity, even at a cost of payload, then you'll save so much money it'll be worth it in the long term. SpaceX is aiming for a re-assemble, gas up and go workflow. If they can pull it off it might reduce their per flight launch costs by a factor of 10 in the short term and perhaps as much as 100 if they get really good at it (although that's likely several generations of hardware down the line, at best). With that sort of thing on offer a reduction in payload is easily justifiable.

Imagine if somebody replaced your car with a version that had twice the carrying capacity, but would only run for one tank of gas. It wouldn't be a worthwhile trade would it?

[mikeash]

To put it in perspective, the fuel used to launch a Space Shuttle cost in the neighborhood of $1 million, while the cost of the entire launch was somewhere around $500 million to $1 billion.

[wiredfool]

Another perspective is that the Falcon9 carries about as much kerosene as a 747.

[747facts]

I decided to fact check you. 57,285 U.S. gallons in a fully loaded 747. Jet-A is 6.84 lb/US gal. That's about 196 tons. Looks like stage 1 of the Falcon 9 is estimated to use 239 tons of RP-1 and stage 2 49 tons.

http://www.boeing.com/boeing/commercial/747family/pf/pf_400_...

http://www.spacelaunchreport.com/falcon9.html

[simonh]

Cool, so they're well within an order of magnitude of each other.

[mikeash]

That's a good one, although a 747 gets its oxygen for free.

[schiffern]

LOX is literally cheaper than dirt (8¢/lb, or 67¢/gallon).

[DavidSJ]

Yes, but oxygen's pretty easy to come by it turns out.

[InclinedPlane]

For the Falcon 9, it costs about $200k for propellant per launch (Kerosene is cheaper than LH2 and the F9 puts about 1/10th as much mass in orbit) while the price of the launch is about $50 million.

[slacka]

Could and real or kerbal rocket scientists here explain why SpaceX isn't doing something like the curiosity rover landing? Sure there's 3 extra steps, deploy parachute, cut parachute, and evasive maneuver, but the fuel saving would massive.

[JshWright]

I'm not sure the savings would be as massive as you think (and would be offset by other costs).

Parachutes (and associated equipment) are heavy, so you'll burn fuel lifting that extra weight.

Parachutes are complicated, and would be an extra system to develop, test, and validate.

Parachutes are annoying to repack/replace (increasing turnaround time).

Parachutes put odd stresses on large objects when they deploy (increasing the amount of inspection you would have to do after each flight).

All that hassle to reduce the terminal velocity by a couple hundred miles an hour. That's not that big a win for a pretty high cost.

[Florin_Andrei]

> All that hassle to reduce the terminal velocity by a couple hundred miles an hour.

That's the TLDR. The delta-vee from the parachute is not worth the trouble.

[mikeash]

Or put it another way, parachutes are generally less effective than using a system you already have for other purposes. A parachute probably beats a rocket engine if your only task is landing, but when you already have the rocket engine, you're better off using it for landing than building a completely separate landing system.

In short, same basic reason why we use wings and wheels to land airplanes rather than dropping them from a parachute when they reach their destination.

[schiffern]

Not "massive". The empty stage is very light, so a little propellant goes a long way. It's less than 10% of the fuel for the entire descent phase (and the optimal upper stage length goes up to compensate somewhat).

[smackfu]

Well, one reason is that parachutes at this scale aren't that simple. For instance, check out the ones on the Shuttle SRBs: http://en.wikipedia.org/wiki/Space_Shuttle_Solid_Rocket_Boos...

[samstave]

Also, the first stage put up the second and third stage units.

What if you have a payload container that was the size and weight limits of the second and third stage, which was put up into LEO by the first stage. The payload would then be picked up by vehicles already in orbit, and the first stage unit returns to earth.

Assuming that the payloads are simply building materials, to be assembled by bots in orbit....

[rbanffy]

The first sage alone won't take your payload to LEO - you'd still need a second stage. However, orbital tugs with electric engines could take your payload from LEO to other orbits with less reaction mass than a chemical rocket could.

You'd still have to refuel their reaction mass, unless we are talking solar or magnetic sails.

[6ren]

It's also extra weight to lift.

[EthanHeilman]

Fuel is very inexpensive (~2% the cost of a rocket launch), rockets are very expensive. If you can trade off fuel for not having to build a new rocket you drop the cost to orbit by orders of magnitude.

Additionally you can build more expensive rockets that are more efficient since rocket creation becomes a capital rather than a reoccurring cost.

[petenixey]

But why not just use a parachute and airbags? They're going to be a vast amount lighter than the fuel you'd need for a rocket landing.

[InclinedPlane]

Parachutes are a pain in the ass to replace. But more than that taking a dunk in the ocean isn't so good for rockets, and spending the better part of a day with a search and recovery team and all the relevant equipment necessary out looking for the rocket and hauling it back home isn't cheap, it also adds a lot of delay to the whole workflow. Compare that to a propulsive landing on the launch pad next to the assembly facilities. You don't have to have a whole special team of folks on hand. The first stage is back in your hands in a matter of hours, where it can be transported back to the assembly building using a simple crane or other specialized equipment. And it can be inserted into the processing workflow much faster than if it had been out in the ocean, and with less overhead of having to clean it, refurbish it, inspect it, etc.

[anonymfus]

But why everybody compares landing solely on parachutes in ocean with landing solely on engines on land? For example, Soyuz use multiple parachutes for slowing down and engines for guiding and final soft landing on land.

[InclinedPlane]

The Soyuz does not "soft land" anywhere. It basically crash lands at a speed which doesn't cause any injury. For a first stage rocket the advantages of coming down on land via just parachutes are pretty much non-existent to negative.

First off, if the goal is to save fuel from having to do a propulsive return to the launch site then that's not going to happen (except for the 2nd stage and capsule). The US has a lot of sparsely inhabited land but it doesn't have the same huge swathes of uncared for steppe that Russia/Kazakhstan have where they can just dump spent rocket stages everywhere with nary a care. There are range safety issues there that can't easily be avoided. Second, a giant rocket stage coming down on just parachutes is going to be damaged more on land than at sea. If you're trying to avoid the weight of landing gear you're just going to end up with the rocket engines crunching into the ground, which isn't going to be good at any speed. OK, so you can't save RTLS fuel, and you can't avoid having landing gear, at that point the only difference is a tiny little dribble of fuel to bring the stage in for a controlled powered landing. So you might as well just do that and be done with it.

[JshWright]

The Soyouz lands 'somewhere' in a fairly large target area. The recovery overhead is still non-trivial.

Something as big and 'light' as the F9 first stage will be slowed down tremendously by the atmosphere. No need to add all the extra complexity and weight of parachutes.

[EthanHeilman]

You can't direct where the rocket ends up, highly likely that the rocket will get damaged or much less likely that the rocket will damage something else (This is why NASA dumps it's rockets over the ocean). Then you have find a way to get the rocket home, which involves time and money. Why not just burn fuel (rocket fuel is only a few times more expensive that burning water) and bring the rocket home?

If you need more payload build a bigger rocket.

You could also add wings to the rocket and fly it home which has it's own set of trade-offs and benefits(see space shuttle).

[TeMPOraL]

> highly likely that the rocket will get damaged

And because you don't know whether it is damaged or not (sometimes the damage may not be obvious), it's likely that the rocket will have to get a long post-flight inspection to check if it is suitable for another flight. It's something you can almost completely avoid if you land the rocket gently.

I read somewhere, that the costs of recovering SRBs of Space Shuttle from the ocean and then inspecting and fixing them were many times greater than building another pair of boosters.

[InclinedPlane]

The SRB thing isn't true. The costs ended up being pretty much the same for new vs. refurbed. The thing is, with solid boosters the "rocket engine" tends to just be giant aluminum cylinders, almost all the complexity of the job is in casting the fuel and putting the segments together, which is completely orthogonal to the reusability aspects.

Also, part of the allure of SRBs is that they are cheap to manufacture, comparatively (this is a false savings, due to increased operational complexity, but it's still very tempting), so even if a significant amount of money could be saved per SRB through reuse it wouldn't have affected the cost a launch much.

[bfe]

The SRBs don't just impose higher operational complexity, but also far higher acoustic load that requires much higher structural strength and therefore weight on everything else.

[danielweber]

Maybe you are thinking of the costs of "re-using" the Space Shuttle. $250 Million to refurb it each time. But we do it for cost savings!

[InclinedPlane]

Fun fact, the cost of building a Shuttle orbiter from scratch was about $1.7 billion.

The cost of launching a Shuttle including the amortized development costs ended up being $1.5 billion per launch.

[rst]

Parachutes were the first thing they tried, on the first stages of the first Falcon 9 launches. It didn't work; the stages broke up on re-entry before they slowed down enough to deploy the parachutes. The response was to try active control to keep them intact --- and I guess they figure they might as well keep it all the way down.

[deletes]

Parachute and/or airbags are not necessarily lighter than fuel on a rocket as big as falcon. Parachutes don't scale linearly with the weight of the cargo. Plus they introduce an additional complexity and a point of failure.

[simonh]

On large heavy objects they also impose significant stress on it's structure.

[artursapek]

I don't think a parachute would give them the amount of control they want over where it lands. Maybe the idea is to eventually have many of these things docking in the same area, like an airport. If you have a fleet of multi-million dollar rockets, you want to be in as much control of them as possible.

[cryptoz]

You cannot land a large payload on Mars with a parachute due to the very thin atmosphere.

[EthanHeilman]

When I first read this I thought you were missing the point, then I remembered Musk's ambition to live on mars and the planned landing sequence for the Red Dragon. Good point!

[danielweber]

Parachutes are really hard. Ask Carmack at Armadillo, but make sure he's in a good mood first. It's not because he's an idiot or anything.

Plus, parachutes create a new failure mode: parachute deploying when it shouldn't.

[schiffern]

…or not deploying when it should.

[VLM]

Think like systems administrator. They've gotta have a rocket engine and a nav system or its not much of a 1st stage. However adding parachutes, airbags, etc, is new systems. Which means much lower reliability and higher expense.

[HodCarrier]

Elon wants rapid (think aircraft), turn around times. That's what's driving the decisions.

[jordan0day]

I imagine there's a lot less fuel required to land than is required to take off. That is, on launch, the rocket is probably (eventually) reaching speeds of several thousand miles per hour.

Upon landing, the atmosphere does most of the slowing-down for you, so you only need enough fuel to reduce the speed from terminal velocity (not sure what this would be for a rocket, probably several hundred miles an hour, at least) to 0. So, I'm sure it's not an insubstantial amount of fuel, but maybe less than you'd think?

[JshWright]

Not to mention that the giant flying fuel tank (er... 'rocket') is _substantially_ lighter now that you've burned most of the fuel off.

[InclinedPlane]

And separated from the 2nd stage and the payload, of course.

[Laremere]

Two other points to consider:

Atmosphere is slowing you down as you go up as well, so lift off requires fighting that as well. Therefore the energy you've expended is not entirely stored in potential energy. (Fun fact: If it weren't for the atmosphere, rockets would actually take off almost horizontally. They go up at first to get out of the thickest air before going sideways.)

The craft is much lighter, because it no longer has the payload, and also has used almost all of its fuel.

[Florin_Andrei]

> I imagine there's a lot less fuel required to land than is required to take off.

Right.

At take off, the delta vee is the escape velocity PLUS all the losses due to friction, which are tremendous.

At landing, the delta vee is only equal to terminal velocity. EDIT: Okay, plus some flying time on top of the landing point.

[Someone]

Another way to state this: at launch, air resistance is your enemy. When landing, it's your friend.

[cbhl]

"With an advanced rocket you can do maybe two to three percent of your lift-off mass to orbit, typically. And then reusability subtracts two to three percent. So then you've got nothing toward or negative and that's also not helpful. So, the trick is to try to shift that from two to three percent in the expendable configuration, to make the rocket mass efficiency, engines efficiency, and so forth so much better so that it moves to around three and half to four percent in the expendable configuration and then try to get clever about the reusablility elements and try to drop that to around the the one and a half to two percent level so that you have a net payload of about two percent."

http://youtu.be/vDwzmJpI4io?t=27m

(Watch from 27m for about 50 seconds.)

[stuff4ben]

I thought the same thing. One of the Youtube commentswas pretty insightful in that it mentioned that when you do a parachute landing into the ocean, it's landing in salt water. Therefore the craft has to be taken apart inspected, cleaned, and put back together. So I'm assuming they've done the cost benefits and figured it's cheaper this way.

I wonder if a combination parachute/thruster landing would be more feasible though? Sort of like the Mars Science Lab without the sky-crane.

[InclinedPlane]

Perhaps on Mars, unlikely on Earth. With our dense atmosphere a vehicle like a mostly empty Falcon 9 first stage is going to have a fairly low terminal velocity, in the low hundreds of km/s range. Slowing down from that speed to a controlled hover/landing is pretty easy. The cost/benefit on Mars might be different though, since the atmosphere is thinner.

[jlgreco]

> in the low hundreds of km/s range

I think you might mean m/s? low hundreds of km/s is very supersonic. ;)

[InclinedPlane]

What am I, a physicist? Jeez, boffing a few orders of magnitude like that is pretty bad, and it's too late to edit.

I meant kph, which was HOPEFULLY obvious from context.

I'll just leave this here: http://www.smbc-comics.com/?id=2679

[ryanbrush]

Not to mention the operational costs of a small navy to recover the rocket and transport it to a launch site.

[mikeash]

I believe the "small navy" was a political and technological requirement and not something you'd need today. A single ship ought to do now.

Political, because it was the Space Race, and a lot of it was about shows of force. And hey, why not? You have a bunch of carrier battle groups just waiting around for a hot war, and you need to have them out training anyway, so why not use them to pick up spacecraft from time to time?

Technological, because guidance wasn't necessarily very accurate. It's interesting to look at the miss distances here:

http://en.wikipedia.org/wiki/Splashdown_(spacecraft_landing)...

Some of these landed hundreds of miles from their target. However, by the time Apollo came around, they were all very close. You definitely want a big recovery fleet to cover a lot of area when you can't be sure it'll land on target, but that's not so much of an issue these days.

[emp_zealoth]

I remember reading about the issue in a book called "This New Ocean" apparently by the time Apollo was on the go NASA had to ask the military to position the ships off to the side of the splashdown zone - the guidance improved so dramatically they were afraid of hitting the carrier directly

[Lambdanaut]

Basically, yes. Fuel costs make up a marginal part of the rocket's cost. One reason the Space Shuttle was such a massive failure was because it was supposed to be "re-usable", yet it really wasn't. It threw away most of its volume every time it launched, yet it still had to carry more fuel into space for the soft landing. It had downsides from both sides of the equation.

[bdonlan]

According to http://www.popularmechanics.com/science/space/rockets/elon-m... :

> "The payload penalty for full and fast reusability versus an expendable version is roughly 40 percent," Musk says. "[But] propellant cost is less than 0.4 percent of the total flight cost. Even taking into account the payload reduction for reusability, the improvement is therefore theoretically over a hundred times."

[cdash]

There is an estimated 40% reduction in payload to become reusable but since the fuel is only less than 1% of the cost in launching the rocket it is pretty acceptable.

[Florin_Andrei]

As Elon Musk put it, it's the difference between filling up your car at the end of a trip, and then going on another trip immediately, versus having to build a whole new car almost from scratch at the end of each trip.

[mtgx]

Fuel cost is only 2-3% of the mission's cost, according to Elon Musk.