'Doom' creator turned rocket pioneer
Best known for creating two of the world's most ground-breaking video games, John Carmack is quietly breaking ground in another new field: commercial rocketry.
Nearly six years ago, Carmack founded Texas-based Armadillo Aerospace to build next-generation vehicles for transporting people and payloads into suborbit. Still under development, his experimental rockets were put to the test at last month's X Prize Cup, as the lone competitors in a NASA-funded contest to build and fly a lunar vehicle.
Armadillo's rockets, called Pixel and Texel, didn't win the $1 million-plus prize money (their legs buckled upon landing). But they did fly and, in doing so, offered hope that a hyped commercial space tourism industry could get off the ground.
Carmack is a rocketry hobbyist, thanks to a "geek childhood with model rockets and science fiction." He only really got serious about the prospect of building space vehicles after funding two participants in a small contest hosted by the Space Frontier Foundation. He still works full time at Doom- and Quake-seller id Software, which defined the genre of first-person shooter games and remain two of the all-time best-selling video games. He devotes the rest of his time to Armadillo, along with about eight part-time engineers and rocketry specialists. Armadillo is funded solely by Carmack.
CNET News.com caught up with Carmack after the X Prize Cup 2006.
Q: You caught some flack for Armadillo's performance at the X Prize Cup. Was it merited?
Some people have derisively said that I'm trying to make aerospace development like software with the implication that that's like an awful way to do it. I would say they're exactly right, I am trying to make this like software because software has progressed so many orders of magnitude better than aerospace has it's not even funny. If we can make aerospace anything more like software--where you can just try things out, hack it out, make it work, improve it as necessary--that will be a tremendous advance. And that is very much what we're trying to do.
How do you feel about Pixel's performance at the X Prize Cup this year?
It's worth noting, there were two absolutely critical things we had to test before the team but couldn't: ground liftoffs and landing; and horizontal translations. When we got there we were pleased that the horizontal translation worked. We were not so pleased with the legs breaking off. We really needed another 30 or 60 days to test, and if we had a little bit better weather, we could have caught the (problems).
The fact that we were able to get the vehicle up--lift it off right on the count down--is a reasonably significant accomplishment. The propellant loading, liftoff, elevation to 100 feet and translation of 50 feet all went flawlessly. It's the landing we had the problem with.
Will Pixel be revived for next year's Lunar Lander Challenge, or will Armadillo engineer a new vehicle?
Yeah, since we got back from Cup we've re-hydrotested everything, and it was fine. It's pretty likely that both these vehicles will be in flight shape for next year. We want to use these vehicles for our vertical drag racing event this year, so we might wreck one before then.
What are the vertical drag races?
We've been pitching that as an adjunct event to the Rocket Racing League event (a NASCAR-inspired rocket racing event expected to launch in late 2007). It's taking two vehicles, similar to the ones we're talking about now, and set them out at an air show. You'd have a vertical racetrack with a Christmas tree that counts down. The two would launch and burn at full acceleration for a quarter mile or so, or about 9 to 10 seconds, and then coast for another mile. Bright lights on the vehicle would change when they hit the finish line.
We've been negotiating with RRL since early this year about forming a commercial contest. We've retired all the risk, shown vehicles can fly and come down. It's a pretty good bet we're going to do it this next year.
Can you describe your next-generation rocket?
Carmack: What we've got right now are the current quad vehicles, Pixel and Texel. They have four spherical propellant tanks and one engine in the center.
What we're going to be doing is reconfiguring the same basic systems, the same basic tank and engine, but making modules out of them. So you have two tanks and one engine as a discrete unit, and then you'll be able to bolt them together in any configuration you want, where you could start with four modules, which would look sort of like two quads stacked on top of each other with more engines on the bottom. That would be enough to take one person to 100 kilometers.
We'll mass-produce these modules and apply them in lots of different configurations. That's our focus in the coming year.
What business are you aiming Armadillo at?
Carmack: We're looking at some ways to commercially exploit the current generation of vehicles, like vertical drag racing and space diving. The space diving record was set in the '50s by an Air Force colonel from 105,000 feet, and he needed to be in a full-pressure space suit. All the records have been set by riding in a helium balloon, which would take a few hours, and it's extremely weather dependent. But if you were in a rocket, it would only take you two minutes to get there.
One of the great aspects of our development process is that we've been doing this so cheaply that a few stunts can make the company profitable and pay for the development.
But we think the first really significant business opportunity is with the suborbital space tourism market, taking people up to 100 kilometers on a rocket. Virgin Galactic has really proved that that market exists by taking in over $20 million of hard-cash deposits. But it is worth noting that they do not have any kind of an exclusive arrangement with Burt Rutan's development company and that if somebody else comes up with a vehicle (with) worthwhile capability, they'll be more than happy to work with other companies. So it's not out of the question that we might wind up flying some of the Virgin passengers at some point.
Why would Armadillo's vehicle be superior to Virgin's or others'?
Carmack: Well, SpaceShipOne and SpaceShipTwo are built from Burt Rutan's perspective as an aircraft designer, and he's probably the best person in the entire world to build a vehicle like that. But fundamentally, I'm from the camp that says that spaceships really shouldn't look very much like airplanes (because) they're going to be much more expensive to operate, have lower relative performance and probably cost more to fabricate.
We have some idea of their expenses per flight, based on their engine technology and their operational cost. And they could certainly turn a pretty good profit at $200,000 per passenger, but it's not likely that they could turn a good profit if the price pressure pushes it down to $50,000 or lower.
A good rocket, in my opinion, looks like a flying fuel tank. If you just fly it straight up and straight down, the operational cost is very, very low. It's essentially just the ground crew time and the consumables that you wind up burning on the way up, and there's no technical reason why a vehicle can't wind up being profitable with the ticket price of $10,000 per passenger.
Now that's not going to happen for a few generations (because) you'll have lots of experimental issues. But 10 years from now, if the market doesn't completely evaporate, you'd wind up with $10,000 ticket prices. And it would be for a vehicle that looks more like a vertical takeoff and landing, flying fuel tank, which is better operationally than a carrier-aircraft-based system.
Do you think that there is a big enough market for daredevil rocket diving? It seems like some new event for the X Games.
Carmack: There are probably a dozen people who are interested in one-upmanship: somebody wants to go to a 110,000-feet drop; the next person wants to go to 120,000 feet. It may be something that can be done more like a high-priced stunt, rather than a passenger thing, with a whole different level of informed consent because it would be on a very experimental vehicle and essentially a no-net stunt.
Would the rocket be like Pixel with an open-air chair?
Carmack: We were just going over this on Saturday and after a little bit of joking, the first thought was that (the seat) would be a saddle. That would be a very Texas and Armadillo. But then we decided that a racing motorcycle seat is probably the best thing on the current generation vehicle, mounted on a frame above the current computer.
We will probably wind up doing this for next year's X Prize Cup. Before that, we'll fly a dummy on the vehicle as our payload ballast.
How do you imagine it will work?
We'll probably have an "AeroShell" over (the rocket) just to reasonably streamline it with the person underneath. The rocket would go ahead and burn to--I don't have real simulation numbers on this--but the engine throttles down a little bit at maybe 70,000 feet and coasts up over your target altitude, so you wind up coming back down with the rocket and letting it pick up a little bit of air resistance. The air is very thin up there, but you get up to a few hundred miles per hour coming down. The person riding it would probably just toss off the AeroShell and let it flutter to the ground and then, while sitting on the rocket, open the parachute and get pulled away from (the rocket) rather than actually leaping off of it.
How does your work in software and games translate into what you're doing at Armadillo, if at all?
Carmack: The cornerstone from the very beginning at Armadillo was that we're going to attempt to take advantage of modern control systems. There are a few specific things that are much better today than they were in the '50s or '60s, when rocketry was originally starting. And foremost among those is computers. The computer in your cell phone is more powerful than everything that was in the Apollo program going to the moon. And you want to leverage this.
For the longest time, most of the people that were pursuing rocketry were more physics and plumbing guys--they want to make rocket engines, and they don't think about control systems. They are content to make something go really fast and maybe get it back, if the parachute happens to work. But from the very beginning, we want to leverage computers and use as much software for automating flight control, automating the landing, so that we can do double duty for all the systems, where you have one propulsion system that flies it up and takes it back down.
A lot of people in some of the rocket companies are mistrustful of computers because a lot of them come from aviation backgrounds, where there's still this kind of pilot clique. I firmly believe that while there are challenges and issues, there's just no way you will reach long-term goals of flying to orbit and the moon without embracing computers. So you might as well embrace them early on and leverage them as much as you can.
Conversely, is your work at Armadillo seeping into any games you're designing for id Software?
Carmack: We keep talking about trying to collect some media shots of real rocket engine thrusts for a current game, but it's unlikely that there will be any real direct crossover there.
Do you plan to be one of the first space tourists with Armadillo or any of the other competitors?
Carmack: If you talk to a lot of people in the new space companies, a lot of them are real true believers, and they've always wanted to go to space. That's not really a motivator for me. I would want to do it, but it's more along the lines of like driving one of my Ferraris 200 miles an hour. It's this huge rush, but it's not so much a lifelong dream or anything.
I probably wouldn't be one of the very first ones, but I would want to be in like the first 1,000 to reach space. I think that there is a cache to that that would be worth $200,000 to me. But I certainly plan on doing it on our own vehicle.
What do you think the threshold will be for space tourism to become a wider sport for the public?
Carmack: I think that you'll get at least 500 people that will pay the $200,000. And then I think the price will start to steadily go down when you get two vendors out there. They'll start undercutting each other, but the early generation of ships won't go down much below $100,000. (When the industry) builds a more cost-effective vehicle, it will start coming down more, and eventually, maybe 10 years from now, it will be a $10,000 ride located someplace like a quick ride from Vegas, where people can just go and do their mad-money thing, dropping $10,000 on a ride.
Will Armadillo be one of the first generation of vehicles out there?
Carmack: Yes, I honestly think that we are the closest, next to (Burt) Rutan (president of Scaled Composites and winner of the $10 million Ansari X Prize to fly into suborbit twice in a week). Obviously, everyone is behind him because he has demonstrated the capability, so I think it's almost a forgone conclusion that they will succeed on this. The distressing fact about most of these other companies is that they've never flown anything.
But you're in the same boat.
Carmack: But, we went into this knowing, with our eyes open, that we don't know everything about this. So instead, we incrementally tested, developed and learned all of these things. We've made 13 different flying vehicles in the last six years, and we've learned something different from each of them. We've explored just about every control authority system possible--altitude-control jets, differential throttling--and we've actually flown vehicles with this, not just studied them.
One of the things I'm most proud of with the X Prize Cup (2006), is that those two vehicles went from concept to flight in six months, and it only cost us about a quarter of a million dollars. Each of those vehicles, while they only went up that 50 meters over a 100 meters there, they have more delta-v than SpaceShipOne. They're very potent vehicles. Our next generation of vehicles is going to leverage this exact same set of systems--just kind of double everything up and modularize it, and that will be a commercial 100-kilometer vehicle, and it won't have cost us that much.
How much have you invested in Armadillo since founding it six years ago?Carmack: I have spent about $3 million.
Is it worth every penny to you?
Carmack: Yeah. After I found out that this was definitely something holding my interest, I stopped buying Ferraris. I wound up selling all of my turbocharged Ferraris, and this satisfies all of the mechanical urges.
A lot of the companies try to put on this big corporate face about how they're going to go out and conquer the galaxy. We don't try to make people take us too seriously because we just go about our own business and people's opinion of us doesn't really matter all that much.
But now with this latest generation of vehicles, a lot of people in positions of importance--in the Department of Defense or the Air Force Research Laboratories or NASA--are really looking at us and being very impressed by what we're doing. We're still probably a couple million dollars of investment away before we start to wind up flying your average $200,000 ticket holder, but we're doing this at a level that's sustainable for me. The fact that we've been doing this on essentially a volunteer, part-time basis with the funding coming from essentially my salary at id Software, it puts us in a position where we can do the right thing technically.
What do you think of being a sole competitor at this year's Lunar Landing Challenge? Is that indicative of how difficult it is to take on developing next generation lunar vehicles?
Carmack: Well the price values are too low to encourage any of the conventional aerospace companies to wind up doing any of this. The million-dollar price for level two (of the competition) is enough to attract people like Acuity that have some technical credentials, and it could probably mount a reasonable effort had it given sufficient time and money to put toward it.
To some degree, we intentionally plan on scaring off most of the competitors because obviously we want to win.
What else is in it for you besides the prize money?
Carmack: The major motivator for me really is the engineering challenge.
Right now is an opportunity for key people to make a real difference, too. The fact that much of rocketry hasn't been thought about clearly since the '50s or '60s; things have just progressed down this one expensive, large government-funded path, and people have ignored all the other possibilities. Someone can go in now and make a real difference.
It's far from a sure thing. The general rule is that 90 percent or more of every single start-up is going to fail, and there is no difference here. It's far from the foregone conclusion that the new space age is here.
Do you think the space tourism business is at the beginning of an Internet bubble or something similar?
Carmack: I don't think it's going to be. I mean I'd be thrilled if it does turn into a little bubble and people start throwing stupid money around like they did the dot-com era, but I'm certainly not relying on that or expecting it.
It is still going to be a niche business for quite some time and it's not something that's going to impact every man in the street. The more likely scenario would be that some government agency decides that this development path is really significant and the DOD or NASA makes a significant investment there.
I think it will make its way one step at a time. We can go from air shows or vertical drag racing to space diving to suborbital space tourists to orbital space tourists, to voyages to the moon. We're finally at the point where we have capabilities that allow us to do things that people are willing to pay for. So I think the real bootstrapping starts here, and hopefully I've sunk most of the money that I'm going to into this and we can start slowly crawling up to being a successful business venture.
