Tankers. A new price category?

[93143]

Sea Dragon was fully reusable, despite dropping into the ocean.  Nothing fancy, just built like a battleship...  and equipped with an inflatable drag enhancer on the tail...  Independent studies indicated extremely low cost to orbit.

Paper rockets have a habit of having extremely low cost to orbit, ESPECIALLY with "independent" studies which are usually made by consultants who don't actually have to fly the thing.

Do you actually know anything about Sea Dragon, or are you just shooting from the hip?

I'm not in this for an argument; I just don't like getting hit with derisive generalizations from people who usually turn out to be posturing without data.

[93143]

Well, at _that_ extreme the assumption is obviously wrong. If a flight was close enough to free there are a lot of things _I_ would want to launch so there would indeed be a new industry exploding.

http://forum.nasaspaceflight.com/index.php?topic=21867.msg608276#msg608276

It's just a feeling but I believe the real potential might lie in mass production as Ed writes - after all we've seen it when LVs were ICBMs and if you _know_ you have a case for massive mass-production you might also be able to get some economies of scale by investing in more scalable production facilities, do more automation and all these things that make planes and card affordable.

Planes are about as expensive as rockets.

[mmeijeri]

Planes are about as expensive as rockets.

But you don't throw them away after a single use.

[Jason1701]

Planes are about as expensive as rockets.

But you don't throw them away after a single use.

Exactly, that's why we need a RLV.

[pippin]

Do you actually know anything about Sea Dragon, or are you just shooting from the hip?

Yes. And it's full of assumptions that are nice on paper but have never been even remotely tested. A lot of concepts look very good at that stage.
The biggest issue with SeaDragon is that it relies on a very low margin and tries to compensate this with massive scaling but this can go down the drain very fast if your margin doesn't come by.

[pippin]

Planes are about as expensive as rockets.

Depends on the plane.

But seriously: what's your point? An expensive plane can be profitable and an expensive rocket can be profitable, too. The point is that are a few dozen times more of them being built every year even though you don't throw them away after each use.

[93143]

That's the point.  You don't throw them away after each use.

You can get some economies of scale through mass production, but it's not going to bridge the cost gap between air travel and space travel by itself.

It's not necessarily just about reuse either - I read an interesting essay on launch costs recently in which the author decided that the lack of intact abort modes at all flight stages was a major cost driver.

[RanulfC]

Why plan to build "unreliable" launchers?  I'm not sure how making a launch vehicle less reliable would save money.  I'm not even sure how to save money by making a launch vehicle less reliable! 

If you design from the outset for mass manufacturing, regular mass manufacturing QA methods, low personnel overhead costs you could, in theory, have a rocket that starts out being 80% reliable and improves as you improve your manufacturing process.

That, by the way, is how a lot of new consumer tech enters the market. The yields on the first generation products are often crap.

As an FYI/Example; First Gen Plasma TVs had around a 50% manufacturing "failure" rate or higher and was "acceptable" at that time. Today's rate is less than 10% IIRC the last figure correctly and that's considered "quite-high" in the industry.

WRT rockets, there is a historical precedent : V2 started out very unreliable, but they did improve it later on.

Relevent article:
1993: "A Rocket A Day, Keeps High Costs Away"
http://www.fourmilab.ch/documents/rocketaday.html

Randy

[RanulfC]

Rather than developing a BDR, just fly Atlas 15 times per year instead of 5, year after year for decades.  The per-launch cost would drop by one-third or one-half, probably, and 98 percent of the payloads would get where they're supposed to go.
 - Ed Kyle

Careful Ed, "generalizations" have a habit of coming back on you

Increased manufacturing rates do NOT always maintain reliability rates and tripling the amount of Atlas CCBs built WILL introduce, (even encourage if you will) lower reliability overall. Increased workload, less time available for QA, increased pacing of assembly, check-out, and operations tempo will all contribute and you'll of course actually have to increase personnel for both manufacturing and operations which introduces training errors, management errors and other "personnel" problems that have been "reduced" under the current methods of operation.

Enough to cause the Atlas to suddenly be "unreliable"? Probably not, but it WILL lower your "98%" success assumption.

Randy

[pippin]

That's the point.  You don't throw them away after each use.

You have no proof for that, it's just your personal assumption. Do you have data?

So far, re-usability did not help getting costs down and again - rocket makers are not dumb.

Rockets are not the only example of high-value assets that get thrown away after a single use. There are other cases (e.g. in construction, manufacturing etc.) where perfectly fine equipment gets thrown away because it would be more expensive to refurbish it and to guarantee the quality you need with a second use or where you simply build something new instead of using something that exists just because it's cheaper.

A very different but hands-on example is that it's generally cheaper to tear down an old house and build a new one than to refurbish an old one and that's ESPECIALLY true for cost-optimized buildings like storages, factories etc.

Processes usually make for a bigger part of system costs than most people believe at first glance, _especially_ if you have high quality requirements due to low margins.
It can be much, much cheaper to build a new engine in mass production than to verify that used one will work again.

[charliem]

Going back to the root of my original post.

SpaceX does, and I suppose all the other mayor players also do, test multiple times each of the main systems of its launchers.

For instance every engine is fired first for a short period, then longer, then integrated, and again, etc.

That level of testing, piece by piece, once and again, has to be expensive.

What would be the consequences of relaxing it a bit, in costs and reliability? Anyone with an estimation?

In the long run RLVs may look a better promise than making cheap expendables, but those are not mutually exclusive strategies (and I think, just my opinion, that full reusability is going to take more time than what some people believe).

[edkyle99]

Rather than developing a BDR, just fly Atlas 15 times per year instead of 5, year after year for decades.  The per-launch cost would drop by one-third or one-half, probably, and 98 percent of the payloads would get where they're supposed to go.
 - Ed Kyle

Careful Ed, "generalizations" have a habit of coming back on you

Increased manufacturing rates do NOT always maintain reliability rates and tripling the amount of Atlas CCBs built WILL introduce, (even encourage if you will) lower reliability overall.

Russia/USSR's R-7 reliability actually increased when its launch rate grew.  Reliability was best during the 1980s when R-7 flew nearly 56 times per year.  The same is true of the retired Atlas Centaur, which became more reliable during the 1990s when it flew more frequently than it ever had before.

 - Ed Kyle

[edkyle99]

Planes are about as expensive as rockets.

But you don't throw them away after a single use.

Because they don't reenter the atmosphere at Mach 23 during each flight.

 - Ed Kyle

[93143]

That's the point.  You don't throw them away after each use.

You have no proof for that, it's just your personal assumption. Do you have data?

Of course.  You should know this too.  Aircraft are reused multiple times.

QED.

Read what I type, not what you think I mean.

...

The reason aircraft are cheaper to use than rockets is that you don't have to build them new each time.  This is not the same thing as saying that the reason rockets are more expensive to use than aircraft is that you do have to build them new each time.  I didn't say that.

[Proponent]

"Minimum cost design", if you are talking about the classic "Big Dumb Booster" idea, has never been employed.  There are several reasons.  One reason is because the payloads cost more than the rocket - sometimes multiple times more than the rocket.  The Dumb rocket might be able to tolerate an above-average failure rate, but the payloads cannot.  Even if the payloads are propellant tankers, they still have great value in the context of a complex, costly mission buildup sequence.

To reiterate what others have said above BDR and MCD aren't the same.  Minimum-cost design's originator, Arthur Schnitt, did not suggest trading reliability for cost.  His point was rather that minimizing the cost of an expendable launch vehicle is not the same as minimizing its mass.  Hence, particularly for lower stages, he suggested the use of larger margins and heavier materials to minimize cost without affecting reliability.

EDIT:  "CDR" -> "BDR"

[savuporo]

"Minimum cost design", if you are talking about the classic "Big Dumb Booster" idea, has never been employed.

I'm not talking about BDR. Big, by its very definition can never be minimum cost, because of the associated infrastructure, and inherently lower flight rate for the same fixed size of the payload market.

Perhaps a better candidate for MCD rocket design would be a simple, low performance, small booster, that can still make orbit, not necessarily dumb though ( not dumb in the sense of integrating a ton of electronics and instrumentation sensors that have actually very little materials cost, so you can get good feedback loop about failure modes )

And to your second point, yes current payloads are very expensive comparatively, but this entire thread is based on the premise of emerging market of propellant tankers, or other bulk payloads.

[pippin]

The reason aircraft are cheaper to use than rockets is that you don't have to build them new each time.  This is not the same thing as saying that the reason rockets are more expensive to use than aircraft is that you do have to build them new each time.  I didn't say that.

OK, but what's the point then, if you do _not_ want to imply rockets should be reused, too?

[wolfpack]

So far, re-usability did not help getting costs down and again - rocket makers are not dumb.

I think we need to define re-useability. I agree completely that just because something is re-used does not in any way shape or form imply that it costs less. Shuttle was a perfect example. SRB's dropped in the ocean, had to be retrieved, disassembled, shipped to Utah, reloaded, reassembled, etc. The orbiter itself required that the engines be removed and reinstalled, new tires installed, often times TPS replaced among many other things. Yes it was mostly re-used. But it certainly cost a ton of money!

Compare that to SpaceX's *plans* (yes, I said plans) to flyback at least a first stage booster. I think that *could* reduce costs. No ocean recovery, no elaborate TPS. If they can simplify it down to basically flushing out the plumbing and reloading the propellants (heck, throw in a wash-n-wax too) then I think it will be cheaper. Time will tell.

[pippin]

Well, I didn't mean to imply it's impossible to bring down cost through re-use, I just believe it's quite some way left until somebody gets there.

But your example is a good one. Let's look at what that means, essentially what they have to do for the Grashopper:

You need:
- additional structure and fuel to sustain re-entry (even for the first stage, currently they are breaking up) and to RTLS
- You probably DO need some TPS
- You'll have much,much stricter range requirements when flying back the empty stage which can add significantly to cost and availability.
- You'll probably want to still disassemble the engines to re-certify them, after that you probably want to test-fire them again. I don't believe that margins in the current design are big enough that you can assume to be able to re-use them without further checks. If you have more operational experience with the engines _this_might be a good point where you may be able to relax requirements for a tanker. If you don't max out payload and you have a decent level of engine-out capability you might be able to live with losing an engine each flight on average or something like that if the payload is cheap enough.

[93143]

The reason aircraft are cheaper to use than rockets is that you don't have to build them new each time.  This is not the same thing as saying that the reason rockets are more expensive to use than aircraft is that you do have to build them new each time.  I didn't say that.

OK, but what's the point then, if you do _not_ want to imply rockets should be reused, too?

The point is that since airplanes are as expensive as they are despite mass production, mass production of rockets is not necessarily going to help a lot.

Now, it is reasonable to point out that rockets are less complex than airplanes.  But the manufacturing costs are still significant even at very high flight rates; assuming an 85% learning curve, you'd need to build several thousand rockets, maybe a couple tens of thousands, to reduce manufacturing cost to 10% of the cost of a one-off.  And an Atlas V is not a one-off; it's been launched 28 times according to Wikipedia, meaning that reducing its manufacturing cost to 10% of its present value would involve building about half a million rockets.

Best case is that the whole incremental cost of launching a rocket is manufacturing costs, which can then be moderately reduced - remember, rockets are already mass-produced, just not at a particularly high rate; they aren't custom-built prototypes.  If a large part of the incremental cost is due to other factors, even that moderate improvement is diminished.

Obviously reuse isn't the whole story; look at Shuttle.  But I do happen to think that reusability is going to be part of any truly cheap launch system.  Personally I'm a fan of Skylon, which is being designed for low operational overhead, low maintenance requirements and fast turnaround, as well as airplane-like failure characteristics (lose an engine and you have to abort, but "abort" means turn around and land, not explode).

You'll probably want to still disassemble the engines to re-certify them, after that you probably want to test-fire them again.

Eugh, no.  If I'm not mistaken, the SSME Block III (well along but never deployed; probably the starting point for RS-25E) was supposed to be robust enough and well enough instrumented that you could just leave it in the orbiter for about ten flights (and yes, it would have been cheaper to make than the Block II).  If SpaceX can't even do that, I'll be rather disappointed.