Home / News

News Archive

Engine work, Methane work, Selling vehicles, Lynx

Engine Development

April 2, 2008 notes:


Engine Work


Our top priority remains getting an extremely reliable engine that we can use for all of our projects. We have been tuning startup / shutdown procedures and trying different injector patterns to avoid heating zones on the stainless steel chamber and face, and we think we are pretty close now.


We did a better rocket crane truck run with an improved startup sequence and a longer burn time:




Normally during a test, either of a tethered vehicle or on the test stand, I manually shut it down as soon as I see the engine “chug” when the fuel line gulps its first helium. We had always been worried about the effect on our graphite engines, and there were similar worries about red hot stainless steel engines, but we hadn’t seen any real problems so far. We decided to do a test to see what would happen if we just let the engine continue firing as it ran completely out of fuel. We warned everyone that this test would almost certainly wreck something, and might very well result in a significant explosion if the engine completely flamed out, then dumped a pint of mixed propellants into a red-hot chamber. The results were destructive, but not explosive:




For vehicle flights, we want the engine to keep firing, even if it is eating itself alive. An extra couple seconds with stainless steel hybrid assist might get the vehicle back on the ground. The manual shutoff has seemed to work out ok for us on the test stand, but an automated shutoff would be safer and more consistent. We investigated several different options, and wound up testing an optical liquid sensor from McMaster-Carr, part number 5094K31. This worked like a charm. We did see some bubbles in the line on an initial start once, so I now require a ten millisecond no-liquid signal for shutdown. We don’t have high hopes for it working in cryo fluids, but we will probably give it a try.


We wound up with two rings of F-O-O-F split triplets on the injector to reduce the recirculation zones heating the injector face. This engine worked well for both many restarts:




And for sustained long burns:




The injector face is still clearly getting hot, and shows a little pitting, but it hadn’t changed any after the long full throttle run, which is a harsher test than any of our normal hovering flights. This engine works well, but the mixture ratio is quite a bit rich, giving unequal propellant depletion and lower Isp. Our next engine test will be with modified orifice sizes to get higher Isp and an equal depletion. If that doesn’t melt anything, we should be basically done.


Methane Work


We have started our cooperative work with NASA involving methane propulsion. Getting some LNG to work with turned out to be a much bigger hassle than expected. In the short term, we are getting the LNG brought up from Houston by the dewar, which is certainly non-optimal. We are getting a decent sized ISO tank installed soon, which will allow us to top off our own dewars.


The day after the dewar showed up, we tested an igniter with it. We decided to use a thermal vaporizer for the methane, just like we do for the lox, to give us gas-gas combustion in the igniter. Our vaporizers are old peroxide catalyst pack chambers with all stainless screens in them (no silver), which take a cryo liquid up to room temperature gas for ten or fifteen seconds of flow. We painted the methane one red all over so we would never mistakenly use it for lox again. We normally use a 0.088” diameter gas orifice for gox to go with a 0.032” liquid alcohol orifice and out a ¼” or 5/16” diameter nozzle exit. We used the same 0.088” sized orifice for the gaseous methane, which gives a mixture ratio well rich of stoichemetric, but we probably can’t run it for tens of seconds in a row without burning it out.


The one surprise we had was that with the gox inlet closer to the igniter nozzle than the methane and the spark plug opposite the methane, it wouldn’t light. Swapping the lox and methane inlets made it work fine. We made a series of tests over a range of feed pressures:




We did this set of tests by loading the methane with the low tech “two vented ball valves with an AN fitting in between” arrangement that is common for a lot of small cryo operations, and we hated it. As soon as we got the parts, we converted over to the Parker micro-bulk cryo quick connects that Tex Air Cryo has always used for our lox equipment. They are pretty damn expensive, but it is so much nicer. They don’t have a specific keying for LNG, so we used the nitrogen keying, because it is farther away from the lox bayonet keys than the argon keys.


We don’t have a second valve set up for cryo conditioning the methane line for engine testing, so we tried to elevate the lines as much as possible to let the warmer cryo fluids rise, but we still had a noticeable gas bubble in the first couple seconds of firing.





Selling vehicles


We have had a number of inquiries from government agencies and Google Lunar X-Prize teams about possibly buying or commissioning one of our vehicles for research or prototyping purposes. The way to look at it is as a “rocket trainer”, rather than a vehicle that can perform any kind of real lunar or suborbital mission. We don’t pretend that the vehicles could actually land on the moon, but if you want to hack on a real, flying system, there is a lot of value to be had.


The price is $500k. The experience of the Lunar Lander Challenge shows quite clearly that you aren’t likely to do it yourself for less, even if you spend a couple years at it. Several intelligent and competent people thought otherwise, and have been proven incorrect.


You can have either a module or a quad, at your choice. The quad has more hover duration, but it is more of a hassle to operate. A module could be fulfilled right now, a quad would take about three months to build, since we are still planning on using Pixel for LLC this year and other tasks. The engine will be one of our new film cooled stainless chambers, and we will warrant it for ten flights. If it blows up or burns through in that time frame, we will replace it. We will not replace the vehicle if it crashes, but historically our engine problems have been visible at startup, and you should have an opportunity to abort the flight. Ground support equipment is included, except for the lox dewar(s), which would be specific to your local lox vendor. We will test the vehicle ourselves, then train your crew to operate it. You get copies of our experimental permit applications and information about the insurance policies we use for permitted flights. Details on modifications to the flight control software are negotiable.


Going to different tanks, engines, or propellant are possible, but the cost will go up.




There was a lot of talk recently on the web and at Space Access ’08 about XCOR’s Lynx announcement. I am an investor in XCOR, but Armadillo and XCOR are also now pretty much direct competitors, so read whatever combination of bias you want into the following statements.


XCOR and AA firmly agree on several points – build the smallest possible vehicle that can serve a market, use only inexpensive liquid propellants, and design the vehicle to cycle several times in a day instead of using a big vehicle that flies more people less often. We split on the classic HTHL / VTVL argument, but the marginal operating costs of our vehicles will likely be quite similar.


I might as well run through and give the John Carmack opinion on all the usual suspects in suborbital space here:


Scaled / Virgin is the safest bet for success. Outside of the X-15, Space Ship One is the only example of a reusable, 100km class manned vehicle. Everyone else, us included, requires a lot more extrapolation for an investor to believe in, and the problem isn’t nearly as trivial as some people like to make it out to be with the “There are no technical challenges, just give us the money!” lines. It is not true that any old team could have won the X-Prize if Paul Allen had given them $20 million.


Air launch and feathered reentry do have some real safety advantages. However, they are spending a frightful amount of money, their schedule has slipped a lot, and the configuration is inherently much lower in operability than the others. The fatal explosion at Scaled has also forever ended the “nitrous hybrids are inherently safe” argument. I am confident that they will fly customers, and if they don’t have any competitors, they may eventually turn a profit. If a more highly operable vehicle is competing, SS2 probably won’t “earn out”. Virgin has no exclusive ties to Scaled, and they will be happy to fly their customer backlog on any vehicle that they can.


Blue Origin is still shrouded in secrecy, which lets a lot of people project their fondest hopes onto them with little evidence to the contrary, but the federal experimental permits give us some insight into what they are doing. Last year, they conducted one short experimental-permitted launch that they showed online, and one other one for which we have no public data. One could speculate that they might have done a spectacular 100km flight on their second shot, but that would be, well, silly. They have the funding to go head to head with SS2 if they choose, but they have not stated that as their plan and there is evidence that they don’t intend to. It looks like they want to play in the orbital space, and perhaps don’t want to get bogged down competing in suborbital space. I think that is a mistake. If Jeff Bezos is reading this, drop me a line sometime…


Rocketplane was kicking around, raising money long before we showed up on the scene. They have burned through tens of millions of dollars of funding, and nothing has ever left the ground. Would another ten or twenty million dollars do it? I doubt it very much.


SpaceDev is a real company with real revenue, they have hardware in space, and they did significant work on the propulsion system for Space Ship One, but their only in-house vehicle work was a hybrid lunar lander that flew up a set of guide wires last year, and didn’t look all that good in the process. I also think the hybrid powered VTHL DreamChaser is about the worst design for commercial suborbital flights. They aren’t going to build it on their own dime, and it looks like it would be a very expensive development project that would end up with quite high operating costs if it was ever completed.


EADS Astrium. Oh, please.


XCOR has flown two rocket powered airplanes and gone through many engine development cycles. The Lynx is designed around engines that are essentially in-hand right now. XCOR has stated that they need to raise nine million dollars to build the mark 1 version, and an additional twelve million for the mark 2 that can actually reach 100km. I think they can probably do it.


Armadillo Aerospace. Anyone can go back through this website and watch us crash vehicles and blow up engines over and over, while never getting more than a few hundred feet off the ground (and that one was, uh, unintentional). It isn’t hard to discount what we do and gaze admiringly at a slick press release speaking to the dreams of the future. Some people doubt the value of experience, or think that the experience will just materialize when the investment lands, but lets go over a few things:


We have built and flown so many different vehicles now that I don’t even count, but lessons are learned from each one. We have experienced and solved issues eerily similar to the things that downed the first two Space-X Falcon launches, among many others. We have more reusable rocket powered vehicle flights than everyone else here put together, although XCOR is gaining ground with rocketplane flights. Scaled made the X-Prize flights under a launch license, and Blue Origin has made two flights under experimental permit, but we have made many more free flights under AST experimental permit than both of them combined. As far as a rocket engine is concerned, a three minute hover is as hard as a 100km suborbital flight, and we have had weekends where we did more vehicle burn time than all the flights of Space Ship One combined.


Total cost to date, about $3.5 million. I think it takes two million more of dedicated funding to put a man to 100km on our Six-Pack, and another million dollars for a flight test program of a hundred launches before entering commercial service. There are lots of challenges, and we will probably crash a few more vehicles before we get there, but I continue to think we are a good bet, and the virtues of modular rocket systems will probably turn out larger than most people expect.


Some people will never be comfortable in a VTVL, but the fishbowl cabin will be an experience like no other. I think the market would be quite happy if Scaled, XCOR, and Armadillo all delivered vehicles. I still think there is a chance we might be first, depending on how Scaled chooses to play things.





© 2001-2011 Armadillo Aerospace, LLC. All rights reserved.