Headlines > News > Armadillo Aerospace: Development work

Armadillo Aerospace: Development work

Published by Sigurd De Keyser on Thu Jan 4, 2007 5:57 pm
More share options

We found that one of Pixel’s gimbal linear actuators was broken in the last crash. The anti-rotation block was broken into several pieces, giving a large amount of lash and allowing the actuator end to freely twist. We have temporarily replaced it with a spare Electrak, but I have been worried for a while that even our hot-rodded Electrak actuators are really too wimpy for 3000+ lbf motors. I have ordered two Ultramotion Bug linear actuators as replacements, which should be several times stronger. At almost $1000 each, they are also several times as expensive, even counting our replacement motors.

The “saddle” is done for Pixel. We will be doing the upcoming tethered flight test with it in place, but empty. After that, we will probably strap a dummy on for future flights. I doubt we will do any untethered, manned flights with this vehicle class, but I am pretty curious how much affect leaning induced CG movement has on the control system, so there is a decent chance someone will get on with it attached to both hanging crane tethers and ground rocket anchors. Carrying a dummy as our “payload” for the Lunar Lander Challenge next year will also be fun.


The first test hemispheres of 5383 alloy are in, and should be welded up and hydrotested within a week. We negotiated a firm spec with AMS Industries for this order (24 hemispheres), so the tolerances should be closer than on our previous orders.

We have found a supplier for 5059 (Alustar) alloy in the gauges we use, and it isn’t all that expensive, but the minimum order is enough for 18 spheres, which is a bit much for us at the moment. We still need to do some trade studies to examine the performance tradeoffs in different operating pressure / weight tradeoffs before we place a large material order. A nice power point on Alustar: http://aluminumnow.com/Downloads/AluminumNowAlustar.ppt

The regen engine development has been a bit slower than we hoped, but we think it is almost wrapped up.

The welded injector face shown in last month’s update didn’t work out for two separate reasons: the angled spot faces for the injector elements went deep enough to get below the weld depth and make a leak path, and the fact that the flange and lox manifold weren’t completely parallel after assembly and welding made the chamber pocket uneven enough with the flange that the o-ring couldn’t seal there.

I figured out how to modify the design to avoid any buried welds and still leave the injector face as a single monolithic part. The new design does require welding things together in a specific order, and some of the welds are hidden by other pieces after assembly (but they are still only propellant to air welds, not propellant to propellant). The first try at this did have a weld problem that couldn’t be fixed, so I added pockets around all the mating points on the next try to make the welds easier to do. Designing for consistent manufacturability is a big deal for us now, because we are going to be flying vehicles with at least four engines this year, and we are looking at possible orbital vehicles with 64 or more modules.


I would really like to follow up on the coaxial injector design, which would have a lot of manufacturing benefits, but it could be an unknown time sink, and we want to get back to doing flights ASAP.

The big worry for the regen engines is the startup sequence, since there is a lot of post-valve volume in the cooling jacket that needs to fill before anything comes out of the injector elements. It also turns out that we don’t have clearance under the Quads to mount the fuel valve directly to the lower fuel manifold, so we had to add an extension pipe, which adds more post-valve volume to fill before the engine will start.


The upgrade to this engine uses a “tear drop” corner to the bottom fuel manifold, similar to the lox manifold, to avoid the fuel pipe dog leg in that picture.

Since we have had our testing at both the shop and my local land shut down, we have to drive a significant distance to do an engine test. We don’t have any permanent facilities for static tests now, so we set up for a horizontal firing off the trailer hitch at half operating pressure so we don’t jump the wheel chocks.


As expected, ignition didn’t work with the existing settings, since main flow fuel didn’t reach the chamber in the 1000 msec allotted from valve motion start to igniter shutoff. Operating at only 200 psi was certainly not helping. The non-start left a big puddle of alcohol in the engine, which the helium purge couldn’t clear out at that angle, so we had to stick a rag in and mop it out by hand. I extended the igniter running time to 1500 msec, but it still didn’t light. We tried again with a longer lox pre-chill, and got a mild hard start. The graphite chamber was cracked, and a couple welds were broken on the injector and jacket bottom.

It turns out that the longer lox pre-chill had allowed enough of a lox puddle to accumulate in the horizontal engine that some of it ran back into the film cooling holes, and combustion was able to start inside the injector, probably with some left-over alcohol, when the igniter started. We never had this problem on our shop test stand, because it was angled down at least ten degrees.

I was tempted to just remove the film cooling holes on the new injector, since we may not need film cooling at all now. The graphite is slightly porous, so some fuel will seep through to the chamber from the cooling jacket during operation (which is also why we can’t put the fuel valve post-cooling-jacket), giving some degree of transpiration cooling, which probably provides enough of a rich zone to avoid worries about oxidizing streaks on the graphite. I am trying not to fall into the trap of chasing performance that we really don’t need, so we are keeping the “known good” design for the time being.

We decided that we are just going to try and fly the vehicle with the new engine, leaving it a vertical firing and avoiding the problem. If we can get the crane, we are going to try for a flight this weekend. Our launch permit amendment for doing untethered flights at OKSP is still being processed. The initial application is just for flights like the lunar lander challenge, but we want to do 4000’ flights there as well.

I had concerns that letting the igniter run for too long could melt the aluminum housing, but we did some tests with runs over five seconds without a problem (combustion efficiency is really really bad, but who cares?), so I have extended the igniter run time to three seconds, which should be plenty of time. I’m not thrilled with the extra delay, especially since when we run at higher pressures the engine will start earlier, leaving it cooking the ground and legs for a couple seconds before liftoff. We might adjust the time based on tank pressure.

I’m sure everyone that follows the Armadillo updates is already aware, but Blue Origin finally posted some public information about their first test flight:

They have experimental permit number 1, we have number 2. We did the first permitted flights at XPC shortly before their first test flight. It will be interesting to see how the relative test flights go in the coming year. We operate at a higher tempo, but their experimental permit already allows them to go up to higher altitudes.

I have zero inside information about Blue Origin, so my comments are strictly from the peanut gallery here. It’s HUUUUGE! I honestly think they are making a mistake doing a development vehicle that big, because it is going to cause much more anguish when it eventually crashes. While bigger doesn’t always mean more expensive, over broad ranges there is a strong correlation, and at least for suborbital tourism, I still don’t think bigger is actually better. I am in the demographic of potential space tourism customers, and I would rather have “my rocket flight” in a smaller vehicle than be stuffed in “the space tourism bus” with a half dozen other wealthy strangers. When we saw the weight listed in the papers filed with the FAA, I thought that the only reason to build a suborbital vehicle that large would be if you intended to also boost upper stages for orbital work, but it doesn’t look like the shown design would be appropriate for that. Maybe it is a subscale version of an SSTO, or a nearly-SSTO upper stage intended to be boosted by an even larger straight-up-straight-down VTVL (my preferred RLV path to orbit).

I still think peroxide has some good advantages, and I assume that an operation of their scale didn’t have the same difficulty dealing with FMC that we did (at least I would expect they went that way instead of running their own concentrator). Flying a monoprop vehicle and later transitioning to biprop is eminently sensible. Fixed (?) landing gear is sensible. I’m still not sold on pumps, which they are advertising for engineers to work on for future vehicles. It looks like they are spending a lot of money, which will be hard to recoup if the market is competitive. Of course, while being profitable is nice, Jeff Bezos doesn’t exactly have to worry about it if he doesn’t want to, and he can proceed as “slow and steady” as he feels like.

I would love to pick over all the technical decisions some time (hint hint!). Also on that note, for the Blue Origin folks that read these updates: if you have finally stepped out of your cone-of-silence, you should join the Personal Spaceflight Federation, so we can all have a unified front in dealing with the regulators, insurance companies, and so on.

No comments
Start the ball rolling by posting a comment on this article!
Leave a reply
You must be logged in to post a comment.
© 2014 The International Space Fellowship, developed by Gabitasoft Interactive. All Rights Reserved.  Privacy Policy | Terms of Use