NASA X-33 space plane

I'd like you to think about: if you had an additive manufacturing idea and you were doing a startup, what would your strategy be? What industry would you go after? What material would you go after? Why did we not talk about aluminum? Because aluminum has an oxide and it's hard enough to weld. Try to take little particles of it and join them together is going to be extremely difficult. Titanium doesn't have a surface active oxide that's hard to get rid of, so titanium is nice. Stainless steel is not bad from that point of view. Why haven't we talked about copper alloys? You could do copper alloys or brasses, but what's the volume there? The volume is in steel. Or if you're going to aerospace, you're looking at aluminum or titanium. Or you might get out of metals and into composites. So you might talk about plastics additive manufacturing, which probably has a number of applications in the aerospace industry to make excellent composite structures. Or maybe you've got an even better idea, if you can figure out how to repair those composite structures, because that's one of the big Achilles heels for composites in aerospace. We can make them, but we sure don't know how to repair them very well, and they do get dinged up every now and then. Look at the Space Shuttle.

A $50M structure with manufacturing defects; NASA reduced the safety factor to 1.05 to keep the program alive; the tank passed at room temperature but failed on liquid-hydrogen thermal cycling. Used to illustrate that safety factors are not magic — they exist to absorb the manufacturing imperfections of real fabrication.

The X-33 spaceplane started out with a 2.0 safety factor. They had manufacturing problems — this was a fifty-million-dollar structure — and they knew they had some big defects in it. NASA sharpened their pencils and said, "Oh, we still got 5% safety factor above." They tested it, and it actually passed until they started to warm it up, and the thermal stresses from going from liquid hydrogen temperatures to room temperature — it failed because it didn't have enough safety factor. If it had been manufactured properly it may not have failed.

Used to illustrate how aerospace weight-savings drives composite adoption — and how the $12,000/lb fabrication cost of the composite liquid hydrogen tank ($50M for a 4,000-lb tank) killed the program. Tank size: "two-story house." Set up rapid-prototyping context from the first Gulf War.

In aerospace structures, an example of weight savings being critical to the whole structure is the X-33 spaceplane, which was supposed to replace the Space Shuttle. It was a $1.3 billion NASA project. The X-33 was a half-size vehicle, and it was going to fly all the way from Edwards Air Force Base to Dugway Proving Grounds in Utah. It was going to get up above a hundred miles, which makes it space, but it wasn't going to go around the Earth. It was just a demonstration vehicle. Before SpaceX came along, this was going to be a reusable vehicle. The Space Shuttle, as you remember, was not completely reusable — the main tanks were disposed of every time, and they tried to recover the solid rocket boosters. The actual thing that got up into space was a lot smaller. This whole thing would have gone into space.

Tom uses the X-33 as the central morality tale of post-Gulf-War rapid prototyping. NASA awarded Lockheed Martin Skunk Works a $1.3B contract on a ~33-month schedule for a single-stage-to-orbit replacement for the Space Shuttle. The team spent six months designing a 3D-woven carbon composite ring-stiffened liquid hydrogen tank, discovered the design would blow the budget 4–5×, then improvised a Nextel-honeycomb / carbon-tape sandwich design in one month at $50M per tank, ~2,000 lb each. Cryogenic fill test at Huntsville appeared successful until video showed frost popping off — the carbon composite was permeable to hydrogen. Project cancelled; the tanks still sit at Palmdale. The X-33 piece passed around the class ($12,000/lb) anchors the cost-per-pound argument.

[Tom hands a piece of HY-80 around.] This is a piece of HY-80, made by our friends down there in Groton. You'll feel the weight and density. As-fabricated, that's probably two or three bucks a pound for the cost of what you're going to float or sink, as the case may be. [Tom hands a piece of X-33 composite around.] That is a piece of the X-33 space plane that cost $12,000 a pound for that composite. So you've got $2 a pound and you've got $12,000 a pound. But it's light, isn't it? It's not as strong and it's more brittle. It wouldn't do well in an explosion. But you can get things into space and still have some payload left over. You can afford it at $12,000 a pound. You could make an Indianapolis racer out of that material, but you can't afford to make a Ford Taurus out of it and sell it to anybody.

Tom's extended case study of the failed $1.3B Lockheed Martin Skunk Works program. Used to teach: (1) rapid prototyping as a buzzword that can override engineering judgment; (2) how a 2.0 safety factor degraded to 1.05 through autoclave delamination; (3) how NASA misread the 3M epoxy data sheet (10% remaining strength after 10 days, not 80–90%); (4) the unverified hydrogen permeability number off by three orders of magnitude that became the final Achilles heel.

The lecture's flagship case study. Three failure modes braided: (1) the 3D woven graphite-fiber composite cost overrun that ate the budget before tanks were built; (2) the 10-day-shelf-life adhesive bullet point that hid a 100%→20% strength curve; (3) the post-test leak from CTE mismatch between graphite-epoxy, Nomex, and adhesive joints under cryogenic-to-room-temperature cycling. Outcome: $50M tanks scrapped, $1.3B program cancelled.

$1.3 billion half-size demonstrator. Two hydrogen tanks ($50M each composite) and one aluminum oxygen tank ($15M), totaling about $12,000/lb as fabricated. Intended successor to space shuttle, targeting $1,000/lb-to-orbit. Tank physical sample passed in class.

And this is the X-33 space plane. When you're going to twenty thousand dollars a pound for material into space — it always cost about ten thousand dollars a pound to get a pound of useful material into space. In the early 1970s, NASA had a goal to drop that to a thousand dollars a pound, and so they designed the space shuttle. The goal of the space shuttle was to drop the cost to about a thousand dollars a pound of useful material into orbit. So whenever you read about these things — any Course 16 folks in here? — they say, "Oh, we're going to colonize the moon, and we can send all the rich people to the moon." The only people who can afford ten thousand dollars a pound — it's going to cost $1.8 million to send an average person, just to get them there without their clothes or anything else. So there's a little problem with this "we're going to colonize Mars." It's because you haven't looked at the cost of these things.