Secrets of X Prize Cup: Pixel and Texel in race for Lunar Lander Challenge
Oh good grief, even Lunar Lander Challenge hasn’t been paying attention to Lunar Lander Challenge.
I hope you all know by now to read John Carmack’s updates at Armadillo Aerospace News to find out the latest on non-goverment-funded space program hopes for taking money from NASA’s pocket in a couple of weeks.
Good luck on that second qualifying flight test, Team Armadillo!
Development, 90 second flight, X-Prize Cup
October 3, 2006 notes:
Matt has done some work on the main Armadillo web page to include our Nvidia sponsorship and X-Prize Cup mission patch design, but he also gave the picture and video gallery pages an update, so they aren’t four years out of date. If you haven’t followed all the individual updates, they are worth a look.
I’m not going to review every single flight test, but some of the things we have been changing are:
We have moved to using a crane truck instead of a big palette lift for holding the tethers for vehicle tests. This gives us more height, lifts straight up instead of tilting up, and now that Joseph has a CDL, we can load everything onto the truck and drive it to our test sites, rather than coordinating a rental delivery at the site and having us arrive separately with a trailer. It is a bit more expensive, but worth it. We have rented it all the way through the X-Prize Cup.
We fixed our shock absorber sticking problem by adding very strong springs to force the piston out even if it is a little sticky, and added round balls for rougher terrain landings. The springs actually keep the vehicle slightly off the shocks when empty.
A full load of fuel is 110 gallons (not a firm limit for blowdown, but one 55 gallon drum per tank is convenient), which we were getting very tired of moving around with our tiny little peristaltic pump, so I finally bought a 25 gpm rotating vane pump. We found that pumping ethanol dries out the pump vanes, causing it to stall on startup sometimes. The solution is to give it a shot of WD-40 in the inlet before and after use.
We moved to pressurizing all four tanks simultaneously, rather than in pairs. It is possible that we may yet trim the pressure balance a little for perfect fuel depletion, but this is generally a good simplification. I also realized that we could just use high pressure air hoses for this instead of braided stainless hoses. The new hoses are just as heavy, but they coil easier and don’t get any sharp wire bits to poke holes in your hands.
We have gone through a couple setups for the on-board vehicle video transmitters. We will have two video streams, one looking down and one looking out.
We have tested our independent thrust termination system. Russ’s company, Long Range Systems, makes various pager systems, so he adapted one of their products to be able to close our fuel isolation valve when commanded.
We shortened out lifting bars / tether attachment points so we can leave them in for future over-the-road trips, and just carry a single lever bar for tilting the vehicle up during balancing. We had originally thought it good to make the vehicle moveable by hand with four people, but since we always have a lift truck or crane for testing, it turns out to not be important.
On September 9th, we had a fairly successful flight test in Oklahoma. We used the straight unlike-impinging (45 / 0 degree) stainless steel engine with a single row of lox cooling vents. We had intended to do a 90 second flight, but we aborted the flight initially at 17 seconds to check the engine. It turned out to be ok, so we re-pressurized and flew for another 65 seconds before hitting lox depletion. We always get some propellant imbalance after we abort a flight and repressurize, so I had to trim a fair amount right after liftoff, but it generally went well. If we hadn’t vented in between and let lox boil off it would have cleared 90 seconds, but it wouldn’t have been close to 180.
The like-impinging stainless steel injector got quite hot, but did not burn. The lox cooling vents may be hurting, rather then helping. Analyzing the telemetry showed that it went in and out of mildly rough burning with a period of 15 seconds. I believe that was lox boiling in the manifold when the injector got too hot, with resulting cooler and rougher burning, which eventually let it cool down enough to get liquid lox out. This injector is on the shelf as a possible backup, but all future flights have been with “twisted element” injectors.
We did some shop hops with the first aluminum twisted injector, and the performance was better, but still not good enough for 180 seconds. We went ahead and made a stainless version to do a qualification flight for the easier challenge, while we continued work on other injectors.
90 second flight
Randall Clague from XCOR, is going to be our designated safety observer for the X-Prize Cup, so he came to witness the test, along with a couple people from FAA-AST.
This is a really big video that isn’t all that exciting:
There were two issues we were concerned about right after we did this flight: The lox pressure dropped a lot more than the fuel pressure, causing a drastic mixture ratio change, and we wound up burning some stainless from the injector at the very end, probably because with the low lox pressure not bending the unlike impinging fuel elements straight down, we got a recirculation region that we don’t get at normal pressures. There were also some attitude oscillations towards the end of the flight that I didn’t like.
We didn’t realize it until I looked at the side view video and telemetry back at the shop, but all the problem turned out to be was that the vehicle position drifted enough that it was tugging on the tether bungee cords, which caused lots of firing of the lox side roll thrusters, depleting a lot of lox ullage pressure, and also caused the oscillations. If we had been flying without tethers, we wouldn’t have seen either issue.
There was a slight drift in the integrated velocity position versus the true position over 90 seconds of about two meters, which contributed to my not realizing the exact position of the vehicle and the tethers. There were two separate problems with the on-board video reception, we have it working better now.
While the mixture ratio was pretty far off towards the end of the flight, we could still tell based on the remaining fuel that the vehicle probably wouldn’t have been able to make the 180 second flight, even if we loaded 110 gallons of fuel.
I have been making several software improvements to the flight control based on our testing. The existing auto-hover logic just aimed for a zero vertical velocity, but it could still slowly drift up or down, requiring continuous monitoring of altitude. I changed this to an explicit altitude hold when you release the hat control, which removes one of the things I needed to worry about. I also added an automatic descent from altitude based on the flight time and current altitude, so even if I’m not paying attention, it should always get itself down before it runs out of propellant (assuming nominal engine performance).
We finally made some significant strides in engine performance last week. We took the basic “twisted element” design, but broke it up into two rings of elements instead of just one, and also thickened the injector deck so the holes have a greater L/D ratio. Our first attempt at this, which bridged from the outer fuel manifold to the new inner one, had bad startup behavior because of the long trip that fuel had to make to get around the main manifold, up the bridge tube, and back down into the inner fuel manifold. We fixed this in a second revision by fabricating a (rather ugly) T-bridge directly from the valve inlet to both manifolds.
http://media.armadilloaerospace.com/2006_09_30/FOF1.jpg (FOF = Fuel Ox Fuel, the manifold layout)
Our baseline shop hover test flight time rose from 35 to 45 seconds, so we now think we have sufficient Isp for the full 180 second flight. We also did a liftoff test with a full load of propellant, which shows we have sufficient thrust, but the aluminum injector started to melt. We are cutting a stainless version now. If this doesn’t do the job, we have a couple more things to try, but time is running out…
The melted aluminum injector was actually pretty favorable, it held out for 17 seconds at full weight liftoff throttle before melting (not burning), and the melting was very even all the way around. The stainless one will probably work.
As a favor to the XPC organizers, we agreed to test a proposed launch/landing pad concrete slab by flying a vehicle over it to see how the surface held up. Local New Mexico civil engineering students did the design and fabrication of the 4’ x 4’ test slab, and one made a visit to the shop for the test. It didn’t turn out so well. We didn’t have a perfectly straight liftoff, and when I swung the vehicle back onto the test pad, the exhaust plume actually kicked the 600 pound slab of concrete around quite vigorously, causing it to break into pieces before we could really get any thermal load data on it. I wish it had been reinforced (the actual pads will be), but I wouldn’t have predicted that behavior.
http://media.armadilloaerospace.com/2006_09_30/breakSlab.wmv (this was the hardish start from the first FOF injector design)
Texel, the second Quad vehicle, is assembled and almost ready to fly. It still needs one lox tank insulated, and various calibrations. We should be hopping it any day now. I have configured everything so both vehicles and two control laptops can all be operating simultaneously, so we can actually fly them together if AST approves it. I still think Vertical Drag Racing would be a great attraction for an event. We have talked with the Rocket Racing League about it, but if anyone else is interested in doing this, let us know.
While it looks like we won’t have any actual competitors for the vertical rocket challenge and the lunar lander challenge this year, there is still a bit of drama with just our performance. If we had a couple more months it would look certain, but we are getting down to the wire now. It looks like the latest engine design has sufficient Isp and thrust to do the 180 second flight, but we haven’t demonstrated it yet. The 50 meter climb and 100 meter translation for the prizes will also be new territory for the vehicles. We can’t test those before hand, because they require a launch permit, and any failure is likely to be a loss-of-vehicle, so we might as well do the spectacular crash for the paying audience. I’m fairly confident that it will work, though.