Taming the N5800
In May of 2012, Cesaroni Technology announced a challenge to the rocketry community, to fly a minimum diameter rocket with a CTI Pro98-6GXL C-STAR N5800 engine. Minimum Diameter means that the body of the rocket is just large enough to accomodate the motor, leaving no room to reenforce the fin attachments with through-wall construction. Through-wall construction means that the fins are attached both to the airframe of the rocket, and inside; to a smaller diameter motor mount tube.
This challenge was not at all trivial. A rocket with an N5800 motor is going to break the sound barrier pretty quickly off the pad. I witnessed an attempt at our West Mesa launch site in December of 1012. It hit Mach 2.7 and went unstable as the first of three fins sheared off. These were fiberglass fins treated with tip-to-tip carbon fiber casing, using aerospace grade epoxy.
The prize for completing the flight before the close of 2012 was Two N5800 or smaller reloads. The conditions were a safe flight and recovery, and being awarded the TRA record for the flight.
In September of 2012, Mike Passaretti of the Pine Island METRA organization completed this flight at Blackrock Nevada. Here is his account.
Taming the N5800: “Don’t Debate This” a BALLS 21 Project
By Mike Passaretti, TRA 5369, METRA, MDRA 097
On September 22, 2012 the first successful minimum diameter rocket powered by a Cesaroni N5800 was flown and recovered. The flight took place on Day 2 at BALLS 21 in Black Rock Desert, Nevada. The weather was not very cooperative, but after a nearly day long delay due to cloud cover the range and skies were finally open. That distant roaring sound of long burn solid rocket motors fading into the desert sky had resumed. “Don’t Debate This” was long ready to go when the LCO got on the radio and began the countdown. The Aussie camp was on their feet and cheering when the Kloudbuster guys mashed the launch button. The motor came up to pressure in an instant as the fluorescent orange, white and black rocket exploded off the pad and angled slightly into the stiff surface winds. The N5800 screamed down at us as it hammered out every last one of its 20,000+ Newton-seconds. In an incredible hurry and on a column of fire and smoke the 15lb airframe was vaulted more than 55,000ft above the desert floor, nearly 11 miles, on its way exceeding the sound barrier three fold. The moment the motor stopped burning the rocket was already out of sight. The cheers started to get louder – up until this very point in time, no minimum diameter airframe riding on this motor had gotten this far. The on-board GoPro camera recorded every second of the flight. On the ground, all that was left was a crater beneath the launch tower and the cheers from the friends and family that helped make the flight possible.
This attempt to fly an N5800 began nearly a year earlier in a backyard workshop half way around the world. The decision to take on this motor at BALLS can be attributed to a number of things, namely confidence in my abilities, the existence of the right support crew, the right inspiration and let’s be honest - enough discretionary funds. It was early October 2011 and I had just arrived at the Couzens home in South Perth, Western Australia. Dave was fresh off a trip to the “the States” that included attending LDRS 30 with his family and a sizable Aussie contingent. By this time, BALLS 20 had concluded and the flight reports were starting to roll in. There was one flight in particular that Dave and I got to talking about. Earlier that week we watched a video of an eight inch diameter Q18,000 powered all aluminum rocket climb to over 120,000 feet, the flight was perfect and straight with little roll, it was to say the least - amazing. The flight now famous was the responsibility of Derek Deville and I’m sure a very well-staffed team. To me, aside from the awesome flight the most important thing that Derek did was thoroughly capture this tremendous accomplishment. Had he flown this rocket without a single camera, no doubt those in attendance would have congratulated him just the same. But, the expansive footage he walked away with and was willing to share is what opened the door for millions of people to experience the excitement of what can be accomplished at Black Rock. His flight in large part inspired me to trek across the country and fly the N5800 – so thank you Derek. We crossed paths at the pad just before my flight, but I never did take the opportunity to introduce myself and say thanks.
From the get-go the goals for this project never included vying for formal records or prizes. I very frankly never cared about maximizing altitude or speed and certainly not optimization of either. The overarching goal very simply was to have a successful flight and hopefully walk away with some flight video. The approach to realizing this goal was loosely modeled after the way I have grown accustomed to approaching projects in my professional career. In the United States military and NASA's engineering design life cycle, a phase of design reviews are held for technical and programmatic accountability and to authorize the release of funding to a project. In a nutshell, since its conception in October 2011 the design of the rocket and supporting systems was matured while continually identifying and addressing risks. I had absolutely zero interest in going to BALLS to fly without the utmost confidence in being able to succeed.
After several weeks of discussions and research an initial concept for the rocket was settled on. The approach for the design was to couple the payload section directly to the motor casing and to slide the fin-can over the motor. The recovery system would be “single-end, dual-deployment” where both recovery parachutes would exit the same end of the rocket. At this point, I was also heavily leaning towards an all metal fin-can. Telemetry avionics would reside inside the nose cone, with recovery avionics installed in the payload section along with at least one outward facing HD camera.
Over the next several weeks I focused the bulk of my time identifying the areas of this project I felt were going to be the most challenging and consequently the most risky. The key areas I surmised needed the most focus included the fin-can design & build along with nose cone separation & drogue deployment. The next several months were spent focusing efforts on these key areas.
Preliminary analysis indicated the rocket would be traveling in excess of Mach 3 at approximately 5kft AGL. Aero FinSim and ANSYS (FEA) were both used to analytically arrive at fin and can wall thicknesses that would survive the flight. First, FinSim was used to identify flutter and divergence velocities. My initial pass turned out to be misleading due to a number of factors namely my misinterpretation of the results. I would later come back to this analysis and find out that FinSim predicted the potential onset of flutter and divergence well below Mach 3. Further research and analysis alleviated some of my concern. I reverted back to one of the sources cited by FinSim a NACA (predecessor to NASA) technical note written by Dennis Martin of Langley Aeronautical Laboratory in 1958. This method, albeit slightly adapted indicated positive margin on flutter velocity. In the end, while I was confident that flutter should not be an issue, I did lose some sleep leading up to launch day.
Second, FinSim was used to calculate worst case forces acting on a single fin under specific flight conditions. For the given fin geometry and attachment method, the angle of attack was varied to obtain resultant loads. My approach for this analysis was to pad or add margin to the variables driving the resultant loads on the fins. Effectively, I assumed a range of flight angles likely through very unlikely and a maximum velocity that was not very likely. There are many approaches one might take but this is the one that made the most sense to me at the time.
The range for the angle of attack was 0 to 7 degrees from vertical and maximum velocity was taken as Mach 3.5 @ 4kft MSL. This was a minor error that was caught late; this altitude should have been higher. 4kft MSL reflects the approximate ground level altitude of the Playa; the rocket should be about 6-7kft above the ground when nearing maximum velocity. The slightly higher air density should equate to higher loads, however I felt it was "in the noise.” The loads that resulted from FinSim, in my opinion are "worst-case." As you can see in the graph, above 7 degrees the fin stress is no longer linear; which is a good indication that the fin material was getting close to yield. For the present design, this occurred at a relatively high angle of attack and so it shouldn’t matter. I settled at 5 degrees for the angle of attack to start the FEA. This gives a moment of 1084in-lbf.
Third, I enlisted the help of a friend at work who has a lot of experience running FEA. He put together a model and ran three separate analyses using the worst-case moment arrived at earlier. The goal was to arrive at a wall thickness for the can that would ensure survivability at the worst case load. Prior to running the FEA we both thought the loads were going to be too high for the notional values I anticipated. However, the results looked much more promising than we expected. So in a nutshell, assuming a 5 degree angle of attack, 0.189” fin-thickness, velocity of M3.5 @ 4kft MSL, single fin moment of 1084in-lbf, and a 15 degree chamfer on all outside fin edges: simulation one (left) was run with a 0.0625” wall thickness, simulation two (center) a 0.125” wall thickness, and simulation three (right) a 0.125” wall thickness and a 0.1875” chamfer fillet. In simulation one - maximum stress is approximately 66.6ksi or 167% of the material yield strength. The can, as expected, should not survive this load. In simulation two - maximum stress shifts (mostly) from the can to the fin, and is 82% of the yield strength. The fin should survive this load. In simulation three - maximum stress again is in the fin, and is 82% of the yield strength. The fin should survive this load. While this was a relatively brief analysis, I was confident that we had arrived at an adequate wall thickness for the can.
The analyses conducted above assumed that the fin to can attachment method was stronger than either the fin or can material by itself. It was desired that the fin or can would fail before the attachment method i.e. the joint. Several months of research, discussions and testing would reveal aluminum brazing as a viable option for reliably attaching the fins to the can. A strong second was aluminum welding however; I sided against this option for two main reasons. The first complexity/cost – I simply did not have within arm’s reach someone I could trust to create the welds and I was not particularly interested in doing it myself. The second reason was the reality of having to deal with the tremendous amount of heat that would be introduced into the sections and its effect on the material from a physical distortion and properties standpoint. Aluminum is a tremendous conductor of heat and post-welding this can lead to problems with material movement, change in properties, reduction in temper, etc. I concluded that welding - engineered and executed properly would likely be the best option. However for practical reasons brazing was more approachable and sufficient for the task at hand. This of course was not immediately obvious and ultimately it was favorable test results that lead to my final decision. As far as I could tell using aluminum brazing to attach fins was not a novel idea. However I could not find anyone or any instance where it was used on the fins of a rocket expected to travel in the neighborhood of Mach 3. This led to somewhat uncharted waters and so several tests were conducted to determine if it was a reliable method for fin attachment.
The first test was a static load test where a prepared section was put under increasing load. The test was not designed for specific flight condition but rather simply to see if two pieces of aluminum brazed together in a t-joint configuration could hold up under a relatively substantial load. In other words, I was playing. So two pieces of 0.25” thick 2” wide 6061-T6 aluminum plate stock were prepared and brazed in a t-joint configuration. One of the plates was modified and attached to a vertical grounded surface, while the other plate – effectively now a cantilevered beam was loaded incrementally via an attached cable supporting an Olympic barbell with weights. All total, the joint did not budge under an approximate moment of 550in-lbf.
The static load test was repeated with the introduction of heat to raise the base material/brazed joint temperature. Prior to conducting the test it was unclear if the joint would behave differently at elevated temperatures. During flight I expected the fin-can to heat from two sources. I knew that heat conducted through the motor casing from burning propellant grains and that there would be aerodynamic heating as well. After reviewing a few research papers that detailed analytical results of sounding rocket fin heating simulations, I came to the conclusion that aerodynamic heating was not likely going to be a significant contributor. However, heat conducted through the motor casing was a situation that had to be addressed. Per NFPA 1125 no part of a motor casing may exceed 200 degrees Celsius. This was the temperature I would to test to. I set-up the same t-joint test section I prepared for the earlier static load tests, introduced a moderate 180in-lbf moment and began to heat the entire section with a MAPP gas torch. The joint held and did not budge, all while raised to approx. 400 degrees Fahrenheit (204 degrees Celsius). The section was above room temperature for approx. 20 minutes, about 6 minutes of which were around or above 400 degrees Fahrenheit. The same joint while cooling down but still at approx. 350 degrees Fahrenheit sustained an increased and total moment of approx. 530in-lbf for a few seconds.
The conclusions obtained from the analytical and test work was used to mature the design and make the decision to start building the fin-can. The fins were cut out of 0.1875” thick 4” wide 6061-T6 plate stock. The 15 degree airfoil was accomplished with a belt sander. While the desired result was achieved, next time I’ll use a mill. The can was cut from a section of schedule 80 (0.375” wall) 4” aluminum pipe. The ID and OD were both machined in a lathe to achieve the desired ID and wall thickness. The fins were aligned using the “door-jam method” i.e. a piece of right angle stock as a guide. Each fin was temporarily secured to the can using a few 4-40 stainless socket head cap screws installed through the inside of the can into the root edge of each fin. The entire assembly was placed in a barbeque grill which was used to pre-heat the base material to approx. 500 degrees Fahrenheit. The goal was to get the base material as close to the brazing temperature (732 degrees Fahrenheit) as possible. Thereby reducing the amount of energy the brazing torch (MAPP gas) would have to introduce to create the necessary conditions for brazing. The entire job was a bit onerous however, after the first fillet was out of the way I began to get a good feel for brazing such a large section and the rest was simply work. All total the entire assembly was brazed in approx. three hours. Taking place on a warm July day in the northeast US, I probably lost a good 5lbs of water weight in the process.
Once brazing was completed, some West Systems epoxy + colloidal silica was prepared and introduced in typical fashion to smooth over the brazed joints, partly for aesthetics but also to decrease surface drag. Everything was painted over with white spray can primer followed up by a few coats of fluorescent orange high visibility spray paint.
Nose Cone Separation
In this hobby it is often said that the “up part is the easy part” and on most flights that is true. However, for this motor, those who have come before me will tell you the “up” part was arguably the hardest part. Either way, I wanted to have utmost confidence in the ability to get the nose cone off the rocket at apogee. Effectively the idea was that nose cone separation should give way to deployment of the drogue recovery system and thereby the drastically reduced possibility of a gut wrenching “hot” recovery. Early on my intention was to use a gas based system for separation but in the end I sided with the use of black powder. Based on the experience of others, I knew that either was a viable option for separation at 50-60kft AGL. However, this was experience that I lacked and I needed to have confidence in my approach. To gain this confidence another round of testing was designed and performed.
The ejection charge tubes were designed based on knowledge gleaned from Jim Jarvis and his experiences at high altitude. The tubes or “cannons” I employed had a relatively high internal length to ID aspect ratio – in this case this was about 15:1. The packed length of the primary ejection charge (3 grams of Goex 4F black powder) ended up being approximately 0.5” OD and 2” long. The charges were prepared and installed such that the top of the charge (end nearest the exit of the charge tube) was ignited first. In a nutshell, the idea behind the extra-long tubes and lighting of the “top” of the charge was to encourage and maximize thorough burning of the black powder. This of course would be validated through testing.
The ejection charges were prepared and installed in the tubes. They were readied along with the rest of the payload section of the rocket which included the recovery system and the nose cone shear-pinned in place. Effectively, the payload section was prepared to be tested in its “as-flown” configuration. There is an old and very well-known saying, a practice really that goes around NASA and the aerospace industry, “fly as you test, test as you fly.” Addressing this particular area of risk to the project in this manner was an opportunity to hold true to this doctrine. The payload assembly was placed inside of a 10 ft length of schedule 40 PVC pipe, capped and connected to vacuum pump. A cost effective atmospheric pressure chamber was created that was large enough to conduct separation tests. The nose cone was separated twice inside the chamber using 3 grams of black powder at an atmospheric pressure representative of 60-75kft MSL. The tests validated my approach for using black powder to create a reliable separation of the actual payload section in flight.
The center of pressure of a rocket much like the center of gravity does not remain in a fixed position throughout flight. It is relatively easy to understand why CG moves, this should be common knowledge to all high power fliers heck it’s on the level two exam. What’s not so obvious is that CP can and does in fact move as well. On your average flight the movement is not significant and can be ignored. However on flights that are expected to go fast, say faster than Mach 1, it’s a really good idea to investigate this because it might be something you need to design around. For this project I used two commonly available, widely accepted and free software programs to gain insight into this aerodynamic effect - RASAero and OpenRocket. Per the analytical results obtained from these two programs: “Don’t Debate This” had a stability margin of 3.5 calibers and never less than 1.5 calibers. That means that the center of pressure had a predicted total displacement of two whole calibers during the burn. I surmise this was the root cause of failure on more than one flight that came before me on this motor.
While the smoke was clearing and the cheers were starting to die down I was pointing the 433MHz Yagi receiver antenna into the air and listening to a steady set of tones from the on-board Big Red Bee 100mW radio beacon. At approximately, one minute into the flight the beeps ceased. It was an awful feeling but it was reality. Nothing was heard over any radio from the rocket after that point. The 900MHz Big Red Bee GPS unit was not completely functional prior to take-off and so it was of little use to us. While it occurred to me long before there was a very good chance I would never see the rocket again, I took solace in the fact that the rocket had clearly survived the ascent. Optimize that! Whether it was recovered or not, the fact remained that this minimum diameter rocket had survived the thunderous burn of the once untamable N5800.
Still smiling ear to ear we headed back to camp to help Dave bring his two stage N5800-M1450 “Starsky & Hutch” project out to the pads. It was now well past 5PM and Dave knew it was going to be tight getting Starsky & Hutch ready and off the pad. Unfortunately, time ran out and due to some last minute problems with the altimeters, Dave decided to bring the rocket down and aim for the following day. When we got back to camp we immediately heard that someone had reported a set of coordinates to the LCO marking the location of a rocket closely resembling Don’t Debate This. Simon had the coordinates in hand, so along with Blake and Drew we jumped in my truck and immediately headed towards the area … actually that’s not true. First we had to figure out how to put decimal coordinates into my vehicles navigations system and then we headed towards the destination arguing about whether we had them entered correctly. Sure enough, a few thousand feet out something appeared to be laying on the Playa surface. As we neared the excitement was brewing within the vehicle. I promptly dismissed it asking for utter silence until we were near enough to make an un-debatable identification. Once we were a few hundred feet away it became evident – Don’t Debate This had been found! A partial donut maneuver at about 50 mph encouraged additional screams, cheers and high fives. We hopped out of the vehicle and remarked at what we’d done – it was a great feeling - mission accomplished!
We got back to camp and celebrations began immediately. It was Saturday night at BALLS and the very first minimum diameter rocket to survive the N5800 was lying in the back of my truck. Everyone at camp along with many others back home and beyond in some way had a hand in this accomplishment. Those physically present that weekend are pictured. A huge thank you to everyone present and to those who could not be for all their help. This includes all of the contributing members of THE Australian Rocketry forum (ausrocketry.com/forum) who thoroughly assisted along the way.
I’m a firm believer that it’s the journey not the destination that matters. I learned a tremendous amount throughout the life of this project and it was a remarkable experience to attend BALLS for the first time – let alone with a bunch of great friends from the other side of this earth. Then again, what else could I say? How else am I supposed to justify blowing a few thousand dollars on a rocket and skipping master’s classes? Don’t Debate This!