Home / News

News Archive

First Flight of Stiga

15 December 2011 update by Ben Brockert, with photos and videos from Ben and Phil Eaton.

On Sunday morning, 4 December 2011, Armadillo Aerospace launched the tube rocket "Stiga" (aka STIG-A) to approximately 140,000 feet above sea level from Spaceport America. The rocket took just under 2 minutes to reach apogee and landed twelve and a half minutes after the launch. The landing point was 2.87 miles away from the launch pad and conveniently on a road, well within the 4.35 mile range established before the launch.

Following is the story of the launch with photos and videos. An update covering some of the new hardware that went into this rocket will follow. As always, images are links to higher resolution versions. Feel free to contact me at ben@armadilloaerospace.com if you'd like the full 18 megapixel version of any image for printing or media use.


The crew drove to New Mexico on December 1st, following the crane truck and rocket trailer which departed the day before. In a flashback to the X Prize Cup, Neil found out that the pickup and trailer have extra rules governing its transport through New Mexico, but this time thankfully the rocket made it into the state without needing to call the governor.

On Friday and Saturday we were on-site at Spaceport America setting up the launch site and doing a hardware checkout. Our flights at the spaceport are in the vertical launch area, which is about four and a half miles away from the Terminal Hangar Facility and the runway. Our chosen launch spot is one of the 10 meter circular concrete pads that the spaceport installed in the last year of the Lunar Lander Challenge, 50 meters from a small facsimile of the moon. Launch control is set up on card tables next to the temporary building from which our friends at Up Aerospace launch their rockets.

We didn't need to haul the huge flatbed trailer behind the crane truck on this trip, and the launch stand for the first tube rocket was built to mount to the rear of that trailer. We revised the design to set up on a flat pad, with the rocket sitting on a combined blast deflector/rocket stand, known as a "milkstool" in some prior rocket programs. The blast deflection was contributed by 1" thick graphite plates ducting the plume laterally.

The rocket could freely stand upright on that rocket stand, but to keep it from falling over in a gust of wind we built a cable-guyed launch tower as well. It used butterfly arms similar to what was used on Stig's launch stand, as well as lengths of aluminum extrusion (that any high power rocketeer would recognize) to act as short rails holding T buttons on the rocket. The tower also served as a ladder stand, for unhooking lifting straps as well as installing and starting cameras just before launch.

cable anchorben doing stand rigging

guyed standlaunch stand/drink holderwamore van, parachute bay

Unlike the first tube rocket, but like the last Mod, this rocket was built with a flyback parachute system from Wamore. They're used to do precision airdrops for the military, and eventually should make it possible for us to land the rocket within a few hundred meters of the pad, for quick turnaround.

The rocket passed all the post-transport tests without a problem; no loose grounds or broken tubes on this excursion. Though there are still things to be learned, we're getting better at remote operations.

We bought a new enclosed trailer to serve dual purpose as a way to haul the rocket to New Mexico and to act as a temperature-controlled mobile workshop, with tools and workbenches necessary for any last minute fixes or modifications. It also made it possible for us to keep the rocket warm overnight, to keep the onboard batteries within their normal operating temperatures. Running everything on our own generator we could theoretically set up to do a launch anywhere in the U.S.

new rocket trailertrailer interior looking to afttrailer interior looking to fore

launch area view looking westlaunch area view looking east

mike and purdue team working on payload integrationmore payload workrocket with fin can on

Launch Day

On Sunday we were on site just after sunrise. It had frozen hard overnight, and deposited a layer of frost on everything.

launch control panorama, launch stand is just right of road on left

frosty morningfrosty toophil hidden, neil, and russ do final assembly

We had to install and start the payload provided by Purdue professor Steven Collicott and his students. We hadn't given them much warning as to when we'd be flying, and they had to put their experiment together quite quickly. It looked at the interaction of fluids under varying acceleration. Unfortunately, they accidentally damaged their camera before the flight. We loaned them another camera but didn't have time to test it; after the flight we found out that the substitute camera didn't work as a replacement.

We got the rocket unloaded and installed on the launch stand. We had tested the stand with the rocket ahead of time, but it was the first time we'd bothered to unhook it from the crane. It was good to see it freestanding and ready to launch.

rocket rolloutheaded to launch sitelaunch site configured for ops

rocket uprightrocket on stand looking westrocket view to east

Running through the pre-launch checklist and doing propellant loading went nominally, aside from a slow leak from the high pressure tank through the LOX regulator, likely due to it being cold. It eventually closed, and just required a top off of the high pressure tank before the launch.

purdue teamloading heliummike and james loading lox

Around 11:15 the pad ops were complete and we pulled back to launch positions. We went through the usual process for launch, with Russ on the controls and Neil giving a countdown. Things had gone well so far, so Murphy chose this as the moment to interject: at roughly T-3 seconds, the GPS reception dropped out almost completely, and the rocket couldn't launch.

We gave it a few moment for the GPS to come back in. It was the first time it had dropped out so bad, we're still looking into what the cause may have been. Once GPS was back working well enough the countdown was restarted from 10 seconds.

At zero Russ hit the launch command on the laptop and Stiga fired up immediately and leapt off the stand! The thrust to weight of this rocket is much higher than Stig, to cut down on gravity losses, and as you can see in the videos it accelerates off the pad more like a solid rocket.

launch sequence: t-0.1 seconds, t+0.17s, t+0.44, t+0.71s, t+0.98s, t+1.25s

in flighton the up and upa zoom of t+0.71s showing shockwave off the rail and fury hitting the standthe stand and concrete, post-launch

Neil had calculated ahead of time how long it would take to reach apogee (~2 minutes), how long it would take to come down if there was no recovery deployment (~4 minutes, hitting the ground supersonic), and a nominal recovery (~15 minutes). The sky had been clear when arrived on site, but a front was slowly moving in all morning, and by launch time it was mostly cloudy with just a few holes to launch through.

Our telemetry wasn't very solid on the way up, plus the difficulty of pointing the antennas at a rocket that couldn't be seen behind clouds. We had a secondary tracking system for the nose cone that would activate when it deployed, but nothing was heard from it after two minutes had elapsed.

We stood at launch control, literally waiting for a sign from the heavens. After four minutes had elapsed with no earth shattering kaboom, we knew it was coming down with at least some of the recovery system, as otherwise it would have already hit the ground.

Shortly thereafter Mark Kusbel of Wamore received telemetry from the Air Guidance Unit (AGU) which is the return-to-launch parachute system. It had an independent GPS unit and radio link, and information coming down from it meant that the main parachute had also deployed. There was a relatively quiet boom heard at launch control about four and a half minutes after launch; later we figured out that this was a sonic boom from the few seconds the rocket was falling supersonic. People tracking the launch from north of the pad heard it much more distinctly without the noise of a generator near them.

We spotted the rocket as it came through the clouds a couple miles from launch control. It wasn't managing to fly back to the launch site, but it was certainly slowing the descent. We re-pointed our telemetry antennas and managed to talk to it before it landed. Russ commanded the propellant valves open, which dumped the remaining LOX to slow the descent rate a bit further before it hit the ground.

joseph and james look for the rocket, tommy asks neil how far along it isstiga comes through the clouds under chuteabout to land, dumping lox

Most of the crew headed out to the landing coordinates, except Tommy and James who stayed behind to get started on packing up the site. The snow that had fallen in the previous days had all melted to create a couple inches of slick mud. The rocket was conveniently sitting on a road, but it was a mud road, and all of the vehicles had a solid coating of orange mud by the end of the day.

On site we evaluated the rocket and recovery. The landing was significantly faster than intended, which pushed in the bottom of the fin can on landing. The connection just above the fuel tank had sheared most of its bolts as the rocket fell over onto a hillock. The recovery system had also managed to pull the synthetic parachute lines through the 0.090" thick aluminum tube during the deployment.

some crunch in the rearsome zipper in the frontsome screws pulled in the middle

team inspects the rocketrecovery system in good shape, onboard cam still onboardzipper detail

nose landingrocket and parachuteevaluating pickup process

The stainless steel roll vanes were still a bit warm from the flight, and one sticker on the nose had bubbled up from the heat. Other than that, there was no sign of any particularly severe flight environment; all of the stickers on the nose cone had stayed on well past mach 2.

We loaded the rocket onto the trailer and pulled it back to the launch site. Everyone managed to get unstuck and headed back as well. We spent the afternoon re-packing hardware, loading the trailer, and taking down the launch stand, and were packed up and out of the spaceport in the early afternoon.

pulling the computer sectionrocket loaded on trailersticker bubbles

Data analysis

Stiga made a maximum speed of mach 2.47 and maximum acceleration of 3.85G, hauling its 32 foot long, 15.25" diameter self to an apogee of 140,000 feet above sea level, 42.67 kilometers. Gross lift off weight was around 1600 pounds, running on liquid oxygen and denatured ethanol. The launch pad is at 4560 feet above sea level, making the climb from launch altitude about 135,400 feet.

Apogee was determined from the LEA-5S GPS unit by u-Blox on the AGU. The Novatel GPS in the rocket guidance system continued having lock issues after burnout as it did before the launch; it's highest altitude report was 128595 feet above sea level. We'll be testing it further to establish the cause of its flakiness.

As you can see in the video, the rocket wobbled at around cloud level, which was a combination of atmospheric turbulence combined with the rocket going through the transonic regime. The roll control vanes high on the rocket had good authority on the roll rate, but experienced a control inversion from 460-480m/s. That control inversion, likely from a shockwave reflecting off the cable fairing, meant that as the vanes moved to reduce the roll they actually increased it.

Given unexpectedly increased roll control in the wrong direction the rocket did a snap roll, which increased the roll rate beyond the software's preset roll rate limit. Upon exceeding that limit, the rocket went into a preprogrammed mode to maximize the burn by locking the roll vanes and engine gimbal to center. The roll rate dropped and Stiga continued on unguided, watching the GPS to ensure that it was not exceeding the preset maximum range.

The rocket's position over the pad had been held well; at 18,000 feet above launch pad the rocket was only thirty feet away from being perfectly above the pad. However, once the gimbal was locked the vehicle began a slow arc over. 38.5 seconds into the burn, when the vehicle hit the maximum range of 7km, the engine was shut down and the vehicle was left to coast to apogee. A full burn would have been closer to 50 seconds.

At apogee the nose was deployed. The air pressure at that altitude is less than 0.2% of that at sea level. While not yet in "space" as typically defined, it is well above 99% of the atmosphere, leaving a black sky. The cameras aren't sensitive enough to see stars, but they would have been out. The streak across one camera is red loctite I had used to secure the cameras to the rocket.

The nose was connected to the rocket and a balloon parachute, or ballute. The ballute was chosen as it won't deploy inverted, unlike a parachute, and is well suited to supersonic flight. Our ballute design was not quite up to the task. The vents did not stay open to keep the ballute inflated, and it was too close to the rocket which put it in extremely turbulent air. The ballute is made of very heavy fabric (5.5 ounce ripstop Nomex) and is eight feet in diameter, but dragging it through supersonic thin air it flapped around so hard that the fabric tore in multiple locations.

Because the ballute hadn't worked correctly to slow the rocket, the main parachute deployed at 250 MPH rather than the design speed of 110 MPH. The parachute deploys from the same front opening as the nose and ballute, in a free bag. The parachute has four sets of lines, and upon parachute opening the set of lines on the right front edge ripped due to high loads, which made it impossible for most of the cells on that side of the chute to inflate. With the parachute partially deflated the AGU was unable to steer it and it landed faster than intended.

The data from the flight showed other interesting behaviors, such as how much of a performance hit there is from going into a roll, and how well the helium had been stratified in the LOX and fuel tank until it was shaken up. The rocket shut down with around 150lb of propellant remaining, so there is still room for higher launches in this configuration.


  • Demonstrated rapid payload flight and return, with the scientific payload back in the hands of the experimenters within six hours of installation on the rocket (within two hours of launch).
  • Confirmed the rocket's aerodynamic design and control at supersonic velocities, up to mach 2.47.
  • Demonstrated correct sequencing on the recovery system.
  • Was the third liquid fueled rocket launched at Spaceport America, after Armadillo's other two launches in 2011.
  • Was the highest launch of a liquid fueled rocket under the current FAA Class 3 Amateur rules, and as far as I know, the highest confirmed launch of any rocket under those rules.
  • Is by far Armadillo's highest launch.

While some parts of the rocket were damaged in flight an on landing, none of them are very significant. We had two complete fin cans with boat tails built before we launched. We can fix the stripped bulkhead, and easily replace the recovery section's torn tube. The parachute lines are made to be replaceable.

crunched boattailgimbal bolts pulled outbent gimbal mount and arm

chute inspectionbroken linesballute tears

The ballute isn't worth fixing, but we have a new design in progress. We're working with some folks who are interested in ballute research, and hopefully will be able to trade expertise on supersonic decelerators for the data created by actually flying them.

We are in the process of fixing the rocket and intend to fly it again in the near future. The next new rocket will control roll differently than this one did, but there are things we can do to Stiga to improve its roll control.

The design has plenty of room for further performance improvement, by reducing weight and increasing helium load. The next rocket will be able to do space with a small payload.

Other photos

sunset north of las crucessnowcover on eastern hills
snow dusting on west hillssunset at spaceport america

random people pictures
Phil, Russ, Kelly Rogers and Joseph, Lee Hardesty, Joseph, Neil


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