0 0 0
Home / News

News Archive

STIG B Program Update - Preamble

Update by Neil Milburn, February 12th 2013

It’s been some time since we did an update on our progress, February 2012 to be more precise, so here is a belated summary of our three STIG B launches. It’s quite a lengthy report so we’re going to split it up into four submissions; this preamble followed by three separate mission reports.


Following the demise of STIG A back in February, we were faced with a dilemma. The fastest way back to flight would have been to build another STIG A, 15” diameter, with just a few tweaks to nudge her over the 100-km Karman line and claim our first space shot. In retrospect this would probably have been a better course of action. Monday morning quarterbacking aside, we decided to build STIG B, a 20” variant, so that we could carry more massive and larger payloads to space. One of the determining factors in this decision was that, even if we could nudge STIG with small payloads into space under a waiver, NASA’s Flight Opportunities Program was mandating that all flights under their banner be FAA/AST licensed flights per 14CFR Part 431. If we had to license the vehicle anyway, we might as well take advantage of the unlimited total impulse … and get our first taste of the license regime.


So, a simple change in diameter with a few enhancements … NOT!


The structure of the vehicle was basically “start from scratch” although we got to keep much of our electronics intact. We opted for 20” because it was the largest airframe we could accommodate in our CNC equipment and that was plenty big enough at this stage. Not long after we made this decision we hit our first major snag. The propellant tanks were to be rated for 400-psig MAWP with proof pressures in the 500-psig range and burst pressures at 600-psig plus. This necessitated using tube wall thicknesses of 0.190” in 6061-T6 and we had to fight a months’ long battle to figure out how to heat-treat these post-welded to meet the desired specifications. But even before that we hit another problem … availability of material. Super-sized sheets are prevalent in US sales literature but non-existent in reality. We had settled on 6½-ft long tanks for both LOX and Ethanol and 20” diameter requires 63” wide material. We didn’t want to “stitch” together spools or use an intermediate spool piece so we needed material that was at least 84” long or more (we made the tanks 7-ft long so we could true them up on the lathe (our limit!) x 63” or more wide. Nada! In the end we sourced some super-sized sheets from Europe at significant cost above normal and then to compound matters the sheets arrived damaged and had to be replaced. This process caused us significant delay and our problems had only just begun.




Welding this thickness of material with the standard AC TIG weld procedures that had served us so well with the 15” diameter 0.125” thick tanks on STIG A proved to be problematic. Weld porosity was the main problem especially near tacks and it required much rework. Based on advice from a NASA welding engineer, and after a lot of practice, James developed a procedure using DC TIG and a home-made, semi-automatic welding rig that eliminated the porosity at the expense of a few inclusions which were more easily repaired. We made full use of our in-house X-Ray capability during this intense period of R&D. Solving this part of the propellant tank issue was only the first part of the story.




We had built a hydraulic rig for testing recovery system components (after the failure of a 5,000-lb man-rated piece of hardware at 5,000-lb … no safety factor!) and we used this to tensile test weld coupons from the new DC TIG process. We determined that we needed every bit of the T6 temper on the longitudinal seam welds to meet the burst pressure requirements and provide the necessary safety factor. This required that we send the finished tubes, not yet capped, for solution heat treat. There are not too many shops that have a drop-bottom oven and water quench tank of the requisite size but we did eventually find one in Wichita KS. Even this oven was pushed to its limit and the quench was somewhat uneven as the cage was dropped into the water leading to a distorted tube. This we remedied by taking the tubes to Baldwin in Dallas who re-rolled them to an acceptable tolerance. Only after this were we able to weld the end caps and inter-tank coupler on with the same DC TIG, X-Ray and inclusion repair process as on the long seam. Now we have a T4 longitudinal seam (good) and a T0 circumferential seam (bad!) The solution is to artificially age the T0 to ~T5, and the T4 to ~T6 temper in our in-house heat treat oven. This process, proven out by our hydraulic test rig produced the requisite results in just a few hours. The burst test of the sample tank proved very satisfactory at 675-psig and it failed at the circumferential weld as anticipated. The proof test of the flight tanks to 510-psig was satisfyingly non-dramatic and we have subsequently cycled these tanks dozens of time with intermediate re-proof testing after flight.


Description: E:\DCIM\Camera\IMG_20130206_074514.jpg 

      TENSILE TEST RIG                           BURST TEST SAMPLE TANK


For the other airframe sections where the weld was not critical and we didn’t require any pressure capability beyond atmospheric, we used 0.125” and 0.90” wall sections made by Argyle Industries in NJ. Their quality was more than acceptable but their automatic welding process would not pass muster for pressure sections.


The nose cone of course required a new mold for the 20” diameter … and it’s big! The shift from 15” to 20” diameter doesn’t sound like much but it’s deceptively larger. We used Air Command International, our next door neighbors, for the nose cone and also the composite fins. There was another learning exercise in making the fins. It is hard to believe that something as heavy duty as the fin molds we had machined could distort but at the elevated temperatures in the curing oven they do just that and we now have some humongous C-clamps to hold the assembly together.


Description: C:\Users\milburnmnk\Pictures\STIG B\Fin Mold.JPG 

                                        FIN MOLD                                                          NOSE CONE – FIRST ARTICLE

There were two other issues with STIG A that we wanted to address with STIG B; roll control and telemetry dropout at altitude.


STIG and STIG A had electrically actuated roll vanes for roll control and these had proved problematic on two counts. Initially we had used a single roll vane and this induced inertial coupling which led to a premature thrust termination and ultimately loss of STIG after a recovery system failure. STIG A used dual roll vanes in an attempt to eliminate this coupling problem but flight testing was about to teach us another lesson. As we approached the transonic regime we experienced a “roll control inversion.” This meant that the aerodynamic forces from the roll vanes operated in the exact opposite direction to that experienced subsonic. The result was that the vehicle went into a snap roll and we exceeded the flight rule roll limit of 360-degrees/second. Our flight control logic was to lock the roll vanes and the engine gimbals in the neutral position and continue to burn until we exceeded any other flight rule limit, which we did. With no active guidance STIG A weather-cocked into the relative wind and exceeded our expected maximum IIP. However, the vehicle still reached an apogee of 43-km and was recovered reasonably intact. For the second flight we rotated the roll vanes around the airframe to potentially eliminate any aero-effects on the fins downstream but alas this was not the problem. We experienced the same transonic regime roll control inversion and snap roll. However, this time we had allowed the engine to continue to gimbal and STIG A achieved a full propellant depletion burn which resulted in an altitude of 90+ km; our best effort yet and knocking on the door of space. Unfortunately because of a recovery system problem, this was the last flight for STIG A.


For STIG B we discussed a software change to counteract the roll control inversion but opted for a change in tactics. We knew that in the near future we would want pointing capability and there was some concern that any asymmetric thrust at thrust termination could lead to a pitch-yaw rotation that we could not counter. Therefore, we decided to use some of the additional space afforded by STIG B’s larger airframe to incorporate a cold gas thruster system using residual helium post-boost. We had a cast aluminum ring fabricated with eight nozzles … pitch+/-, yaw+/- and pairs of CW/CCW roll … and to this we mounted Parker 12-V solenoid operated valves. We statically tested this on the vehicle by suspending from the crane and it seemed to be more than adequate. In fact, so much so that we will likely introduce a modification to permit fine control by regulating the inlet pressure.


ACS MANIFOLD W. VALVES                                                INVESTCAST ACS RING


The telemetry dropout issue was something else that needed serious attention. There is something very disconcerting about launching a vehicle, have it disappear from visual range and then lose contact with it altogether. For many years we had been using Esteem 195Ed-2 (900 MHz & 2.4 GHz) radios with good luck but they had not performed well on STIG vehicles. Russ did a detailed analysis to check the link margins, uplink and downlink, and determined that “we need more power Scotty” plus some enhancements to the antennas and GSE. However, during ground testing we started to experience communications issues with the Esteems and we never did quite resolve this problem. It appeared, and got worse, we believe, after a firmware change and it led to the adoption of the Ubiquity Rocket M900 radio. Perversely, this is the same radio that is in the Esteem but it worked perfectly straight out of the box confirming our suspicions about the firmware on the Esteems.



To verify our choice of hardware we installed the entire set-up in a fin can module and flew it on a leased aircraft with the main flight computer inside. The system worked all the way out to 100-km and should be good to 150-km plus when operated vertically. The RTT (Remote Thrust Termination) system was also tested successfully out to 60-km and its required range is 30-km; the boost phase. The ground station used a dual-polarized Ubiquity yagi antenna and the RTT used a Sure helical antenna.


Description: C:\Users\milburnmnk\Pictures\STIG B\LAUNCH CONTROL VIEW OF PAD.JPG



Description: C:\Users\milburnmnk\Pictures\STIG B\Main Flight Computer.JPG



The main flight computer is starting to look “busy” now. The additional driver requirements imposed by the cold gas ACS and the need to provide trigger signals for multiple payloads required us to install additional boards. A later addition to this stack will be trace heating and/or blanket to keep the batteries and flight computer warm during cold climate operations. We are still using the Diamond boards which have proved very reliable with the only downside being long lead times for the CPU board necessitating buying in batches larger than we would like and keeping a “bin reserve.”


Having given the launch license as one of the primary drivers behind the decision to build STIG B, we should probably say a few words on that front. We have an excellent relationship with FAA/AST built over more than a decade of collaboration that resulted in multiple permits and waivers, both with extensions, for a variety of vehicles in several locations. We had hesitated applying for our first license imagining it to be a daunting task. In the final analysis it wasn’t so much daunting as it was lengthy. We already had extensive documentation for the High Power Class III Experimental Waivers for STIG and STIG A to draw on and the only element of 14CFR Part 431 “Launch & Reentry of a Reusable Launch Vehicle” that required much work was the EC analysis. We also had significant experience with TAOS “Trajectory Analysis & Optimization Software” and Missile Datcom for vehicle inertial property evaluation which served us in good stead.


The analysis was no different to that required for our Lunar Landers but, of course, our range with an errant Pixel did not cover much area whereas STIG B could definitely reach out and touch someone … in El Paso, Juarez not to mention a host of smaller population areas. Having a million plus souls in your potential impact area does bad things for your EC (Expected Casualty) value! The process is simple if mind numbingly tedious. We assume that the vehicle goes AWOL at any time during the boost phase and heads off in random directions with the gimbals locked in some unfavorable position. The vehicle then could either tumble and crash short range or execute a beautiful maximum range gravity turn sometimes in the direction of densely populated areas. The impacts were plotted onto a population density map of the area and factors such as explosive debris from residual propellant, impact angle, skip and bounce were used to determine the “body count.”


The initial runs were on the basis of “everything fails, every time!” This was unrealistic and resulted in an EC that was about twice the permitted limit. One feature of our flight safety system that has been ever present and 100% successful is the thrust termination system. We monitor several parameters to determine whether or not to terminate thrust. Two of the most important are the Ground Speed and the Max Range IIP (Instantaneous Impact Point). The latter is basically a conservative estimate of “where would I land if I flew with current vehicle vector accelerating for another full second before the thrust was terminated.” For Spaceport America the Flight Hazard Area was initially determined to be a circle of 7-km radius centered on the launch pad. Sir Richard Branson’s pretty new hangar building was 7.2-km distant. J Of course, one can’t assume that a flight safety system that has a prior 100% success record is going to remain so indefinitely and we agreed a binomial distribution approach which was more than optimistic enough to meet the required EC limit.


As a final footnote to the license process I have to add that we could not have asked for a more cooperative and helpful group than our support team at FAA/AST. Our POC was John Howell who did an absolutely stellar job of keeping things on track in Washington and I spent countless time with Steve Millard on the trajectory and safety analysis. In the field David Gerlach has been the FAA/AST observer and it was invaluable to me as the ASO (Armadillo Safety Officer) to have his advice and counsel. I said final but I should also say a word about Spaceport America and their support team. We are undoubtedly the most prolific launch company from their new facility and we’ve both learned a lot over the past couple of years. Each mission is smoother than the last and the Spaceport America team could not be any more professional or supportive. In fact, we now hold the Mission Readiness Reviews in their Las Cruces offices and we chose Las Cruces as our base of operations partly on this fact and also for the wide range of accommodation and services. We do wish that it was several hundred miles closer. That drive out on I-20 and I-10 seems to get longer each trip especially towing the 40-ft rocket and launch control trailer!


One recurring theme above is that of recovery system problems. Phil Eaton has been our lead on this working closely with Lee Hardesty of Applied Deceleration Technology, Wamore Inc of Phoenix and with input from folks at Strong Enterprises, JPL and others. Phil has written a diatribe on our recovery system learning curve which was posted to aRocket a little while ago. We’re going to post that report, which covers the entire STIG family of vehicles, in a separate post directly after this one because it is quite lengthy in its own right and attach some photographs and video.






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