Armadillo Aerospace - News Archive

News Archive

April Update: Tube Rocket Completion

March 31, 2011 notes, photos, and videos by Ben Brockert:

The Tube Rocket 'Stig' is Complete
The tube rocket has come together. It presented some new challenges. On a diagrammatic level, it's a supermod with cylindrical tanks, but in reality it wasn't that straightforward. Parts started arriving in January, so it's taken us most of three months to put it together and get it ready to fly. The planned first flight is this Saturday, I'm posting this update from a hotel in Truth or Consequences, NM, the closest town to Spaceport America.

All of our most recent vehicles have been 36" diameter Mods and quads of 36" or 48" tanks, which have a stout form factor and a lot of space to put stuff. The tube rocket has all of the same hardware, plus the full helium pressurant section that the blowdown mods didn't have, but it all has to fit as much as possible into a 15.25" cylinder. It didn't entirely work; the most convenient position for the valves puts them largely into the airstream, but for the most part the entire mechanism was crammed into a relatively tiny volume. Joseph and James did a good job putting together the 3-d jigsaw puzzle.

We switched to a dome loaded regulator for pressurizing the tanks, rather than the motor driven valve that was used on Scorpius, the Lunar Lander Challenge mod. The setup has more parts but less 'smarts', the computer doesn't have to do active regulation. It also fails safer. It probably came out a bit heavier, but has room for easy improvement.

This rocket is also drastically taller than any previous rocket, about 24 feet long with the payload section taken off. It precludes any vertical work on it in the shop. There is something useful in being able to walk around a rocket in flight position while working on it.

We decided to go with a pressurized section for the computer box, it resides in the fin can between the engine and the LOX tank. The last mod had issues of overheating the computer with it sealed in a metal nose cone on top of the fuel tank, and this one had higher power disspation in the computer box because of more powerful radio electronics.

So for this vehicle we added fans to keep it circulating on the ground and in microgravity. With the top of the computer section being the bottom of the LOX tank, the temperature in the computer box drops after propellant loading. It steady-states at around -20F; around 0F for a normal sequence of launching right after propellant loading. The rocket's fins also act as good cooling fins, so even without the lox on top the sealed computer section never gets more than warm.

The sealed computer box also meant we had to have an airtight wire passthrough for the first time. The first bulkhead connector Russ found was $1000 per set, with us needing one or two sets for the rocket. We did a couple experiments with just sealing wires through a metal wall, and ended up going with a ~$150 off-the-shelf connector. In the grand scheme $1000 isn't a huge percentage of the rocket, but they do add up.

Once it was all together we ran a good series of tests. The test program went something like this:

Actuator checks, making sure the computer, electronics, valves, &c. are all working correctly. This also included range tests, where we had the base station and the vehicle transmitters at increasing distance, to see how far it would work. We have redundant communications over 900MHz and 2.4GHz; with relatively small ground station antennas we exceeded 20km with the 2.4GHz and made it all the way to 100km with the 900MHz.
Chilldown test, loading fuel and liquid nitrogen into the vehicle and running it into the engine. The LOX tank gets about a quarter inch shorter when it's cold, so the propellant lines, cables, and pressurant lines on the outside of the vehicle have to be able to survive that.
Hot fire hold-down testing. Loading the rocket with propellants and firing the engine with the vehicle unable to go anywhere. Here we developed a loading sequence for the helium, liquid oxygen, and fuel that had to be customized to this rocket. It's a "common bulkhead" rocket, meaning that the top of the LOX tank is the bottom of the fuel tank. The challenge is to have the LOX tank completely full of -300 deg F LOX while keeping the fuel from getting cold and viscous. We managed to find a sequence that accomplished this without needing insulation on the bulkhead.
Hover testing. There was a short series of tests here, where we varied how sensitive the computer is to the movement of the rocket. We want it to fly generally straight up, but allow it to drift a bit; trying to keep the rocket's attitude or trajectory too tightly constrained leads to a lot of engine movement, which wastes impulse. The rocket has fins and is passively stable by a good margin, so it's going to want to fly straight.

The hover tests included the trademark Armadillo test of hooking a cable to the rocket and having someone give it a good pull, to see how it responds to a disturbing force. Here Mike yanks its cord, as it were.
Remaining task: fly the rocket to a high altitude.

The rocket and seemingly half of the shop is loaded up and in New Mexico right now, for a first launch attempt on Saturday. If something doesn't check out after the trip out there, or winds are too high, we can push the date into the next week. I believe it will be the first liquid fueled rocket launched at Spaceport America. We're actually setting up the launch stand on one of the pads they installed for the Lunar Lander Challenge.

The stand
The launch stand concecept started with Russ modeling a cool launch stand that looked like one of the Russian launchers, with vertical support arms hinged with the load support arms, so that as it took off the vertical supports would fall away. I did a parallelogram version of it that had an interesting movement, but the thing we ended up building was quicker to get together and works fine for now.

The Mk. 1 concept for the launch stand was just a rail, very similar to what a lot of folks use. However, there was worry about what it would do after the first rail attachment came off but before the second one. Since it is relying on the engine movement to stay stable at low velocity, being pinned at one spot could do weird things to the control.

We decided to have the launch stand move away from the rocket after it was released, which led to the Mk. 3 design that got built. It's essentially the rail/button concept inside-out. The rocket has short channels in it, which support the vehicle laterally with a butterfly arm mechanism.

The weight of the rocket is held up with an arm at the bottom bulkhead, which swivels up to get out of the way. The butterfly arms just hold the rocket from falling over and torsional loads. That said, the assembly of 1000+ pounds of rocket hanging off a ~100 lb steel structure in 30MPH wind sways a bit when you're at the top of the ladder.

The rocket is put on the stand with the crane truck, which puts the propellant connections right at a convenient height for the ground crew. Propellants are loaded, crew evacuates, the engine starts, it looks ok, it throttles up, and when it moves up about three inches it suddenly finds itself completely free and a foot and a half from the body to the nearest object.

The stand is a bit of trouble as the rocket is worked on horizontally in the shop and it has to be rotated to vertical with manpower and the crane truck before putting it on the stand. So the next version will have a hydraulic lift mechanism, like classic mobile rocket launchers.

Parachute
The other new system for this rocket is a full parachute recovery, with a drogue from near apogee to near the ground, followed by a large main parachute. Phil went through a lot of tests on this, repeatedly blowing prototype noses off the recovery section, and working out how the drogue is attached and how the main deploys.

We dropped a ~1/2 scale tube rocket from an airplane five times, establishing the behavior in flight under drogue and main, with variation in CG and CP. They didn't all go well, and I think Tommy was getting tired of rebuilding the broken parts, but it was quite instructive.

The rocket lands on the engine, but it will be cool by then, and it's quite a stout engine. A gimbal actuator may end up broken, a fin bent, or a propellant line dinged, but if we launch the thing to near vacuum and get it back with just a few dents it will qualify as a success.

The nose is sealed on with o-rings and classic HPR nylon shear screws, and the internal pressure is bled down as it ascends into higher air at a specific leak rate. Once over the top it's blown open with ullage gas from the fuel tank. The nose has space for payload. It would need to survive the ejection shock, but it will be a good spot for a released payload or one that needs to be exposed to vacuum. I'm flying a personal amateur radio experiment in that space on the first flight, we'll see if it works.

There's a second payload section below the parachute. For the first flight, in there we're flying a student payload organized by Professor Steven Collicott of Purdue.

Future tube rocket plans
For this flight we're hoping to exceed 100,000ft. The models suggest we should, but there's always room for Murphy. If we get a successful flight we'll be bringing it back and modifying it toward reaching space, i.e. 100km. There is apparently a Metric line somewhere below the Kármán line, where altitudes go from feet to kilometers.

Modifications will include reducing the weight in some places where it came out heavier than necessary, and putting a larger and more efficient engine on. The next full iteration of the tube rocket will include some research we did on improving the tanks another step, allowing us to run at higher pressure without increasing the weight.

For other ideas with what we could do with tube rockets, here's John:

Yes, I very much want to cluster some of these!

This vehicle should be able to handily get past 100km, even though we are going to be conservative on the first flight attempt and just be happy to exceed 100,000'. We have lots of fairly straightforward axis for improving the performance -- fair over the exposed bits to reduce drag, increase the nozzle expansion ratio, move to one of our newer injector designs, and, of course, take weight out of it.

Once we show that the vehicle can clearly fly to space, clustering will improve the payload fraction -- two tubes = a single computer box and no roll vane (differential gimbal like titan), four tubes = fixed engines with better plumbing flow and less actuators, etc.

The first order of business is launching and recovering it successfully. Doing sub scale drop tests from an airplane has been helpful, and may have already saved the real vehicle with what we learned.

One advantage of the current single setup is that it falls under the federal regulations as an "amateur rocket", not requiring a permit or license to launch. The clustered rockets will be over that limit, which adds new levels of paperwork, analysis, and proof that things will work as expected. Doing paperwork is definitely not the hardest part of making a reliable space rocket, but it adds delays to the process.

Some updates from John
I know a lot of folks miss hearing from John in these updates so I'll finish the post with some other quotes from John on ARocket recently. He's been very busy with the new RAGE game at Id Software. He's missing this tube rocket launch as well as Space Access next weekend, for the first time in a decade.

For reasons that still aren't exactly clear to me, this vehicle took longer to build than any of our previous ones. With all the full time salaries, that also makes it the most expensive vehicle we have made. I have said that watching one of our serial-produced mods crash from an altitude flight is "like watching a BMW fall out of the sky". Losing this vehicle will be more like planting a Ferrari. A turbo Ferrari. Realistically, it is almost inevitable. If we get a successful first flight, everything will work out fine, but I imagine the mood in the shop will be pretty grim while building up a new version of this vehicle if all we got out of the previous one was a couple hover tests and a crash.

Despite being extremely upset with how long it took us to get to flight-ready, I find myself now wishing we had another week to run additional tests, but we have already postponed launch once at Spaceport America, and we really should take our shot now.

However, testing is not without drawbacks. Two days later:

Yesterday I did wind up running a final set of tests on the rocket to put my mind at ease about the sequencing of the recovery system by the flight computer. The test showed that it was doing exactly what we wanted, but in the process of modifying things to fake the computer into thinking it was falling from 35km while it was in the shop, I wound up breaking into the debugger while one of the igniter solenoids was powered on, and burned it up, resulting in the need to make a last minute part swap on an otherwise ready-to-ship rocket.

I hate using NOS solenoids because they aren"t continuous duty, but the initial configuration of the vehicle had the igniter in an orientation where that was the only solenoid that would fit. We wound up changing the igniter position later for unrelated reasons, so we could have used a GEM or predyne solenoid if we had thought about it, but we didn't.

The rocket is on the road to Spaceport America today, everything looks good. Because we came in overweight and are running a fairly low tank pressure, we will be eating a lot of gravity losses on the vehicle, but that also means we won"t have a really crushing max-Q. Sims say we should comfortably clear 100,000", but I wouldn't be stunned if it underperformed.

Finally, advice for anyone who thinks this is all amazing and wants to start their own rocket project:

The mention of weight prompts one of my usual bits of advice for newcomers:

Essentially every single time you make a decision with "flight weight" as a deciding factor, you are making a mistake. Pushing for flying something instead of just doing test stand work is important, but you are many, many generations away from building some mass-ratio 8 vehicle where you are shaving every gram of weight. There is over an order of magnitude difference between what commercial orbital vehicles need to care about with respect to weight and what a new comer to experimental liquid rocketry should be caring about. A liquid fueled rocket can fly just fine with a mass ratio of 1.25.

Go for the solution that looks easiest, with almost no regard for performance. It will turn out not to be easy, but if you lowballed your expectations enough, you might actually succeed in making it fly, instead of winding up with a failed and abandoned project.

Ok, one more video: