Terrapin Rocket Team Project Karkinos Team 127 Project Technical Report to the 2023 Spaceport America Cup Hailu Daniel∗, Andrew Bean†, Ezra Bregin‡, Dasam Gill§, Nathan Roy¶, Khalid M Jaffar‖, Garret Alessandrini ∗∗, Michael Mallamaci††, Adin Goldberg‡‡, Howard Zheng§§, Saim Rizvi¶¶ University of Maryland - Department of Aerospace Engineering, College Park, MD, 20742 This document presents the University of Maryland’s 10,000 ft SRAD Motor Category rocket, Karkinos. It is the third time that the team will be attending the Cup in person since 2018, and the current team has built on the lessons learned at the 2022 Cup. The design process for Karkinos is centered around manufacturability and reliability coupled with a rigorous testing process. This report also documents the design of our Air Brake system that actively trims the rocket’s altitude during flight. The CubeSat payload for this rocket will test the release of a vehicle from the rocket during drogue descent. ∗Team Leader, Terrapin Rocket Team. †Chief Engineer, Terrapin Rocket Team ‡Air Brake Team Leader, Terrapin Rocket Team §Aero-Structures Team Co-Leader, Terrapin Rocket Team ¶Payload Team Leader, Terrapin Rocket Team ‖Air Brake Team Member, Terrapin Rocket Team ∗∗Air Brake Team Member, Terrapin Rocket Team ††Payload Team Member, Terrapin Rocket Team ‡‡Aero-Structures Team Member, Terrapin Rocket Team §§Aero-Structures Team Member, Terrapin Rocket Team ¶¶Solid Propulsion Team Member, Terrapin Rocket Team Contents I Introduction 6 I.A Academic Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 I.B Stakeholders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 I.C Team Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 II System Architecture Overview 7 II.A Propulsion Subsytems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 II.A.1 Material Selection and Pressure Vessel Design . . . . . . . . . . . . . . . . . . . . . . . . . 11 II.A.2 Propellant Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 II.A.3 Full Scale Motor Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 II.A.4 Motor Manufacturing and Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 II.A.5 Motor Test Stand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 II.A.6 Full Scale Motor Tests and Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 II.A.7 Mixing Sessions and Motors Made . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 II.A.8 Igniters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 II.B Aero-Structures Subsystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 II.B.1 Body Tubes and Couplers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 II.B.2 Fins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 II.B.3 Bulkheads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 II.C Recovery Subsystems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 II.C.1 Recovery Devices and Shear Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 II.C.2 Flame Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 II.C.3 Recovery Harness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 II.C.4 Recovery Attachment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 II.C.5 Parachute Deployment System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 II.D Avionics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 II.D.1 Commercial Deployment Computers and Battery Selection . . . . . . . . . . . . . . . . . . . 38 II.D.2 Commercial Tracking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 II.D.3 SRAD Avionics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 II.D.4 Sled and Mounting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 II.D.5 Avionics Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 II.E Air Brake . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 II.E.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 II.E.2 Deployment Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 II.E.3 Air Brake Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 II.E.4 Avionics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 II.E.5 State-Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 II.E.6 Controller Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 II.E.7 Parameter Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 II.E.8 In-Flight Tests and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 II.E.9 Wind Tunnel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 II.F Payload . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62 II.F.1 Release Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 II.F.2 Vehicle Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 II.F.3 Payload Bay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 II.F.4 Payload Sizing and Mass . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 II.G Full-Scale Test Flights . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69 Page 2 III Mission Concept of Operations Overview 73 IV Conclusions and Lessons Learned 75 V Appendix A - System Weights, Measures, and Performance Data 76 V.A Rocket Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 V.B Predicted Flight Data and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76 V.C Recovery Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 V.D SRAD Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 VI Appendix B - Project Test Reports 78 VI.A Recovery System Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 VI.B SRAD Propulsion Systems Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 VI.C SRAD Pressure Vessel Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 VI.D SRAD GPS Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110 VI.E Payload Recovery System Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 VI.F Additional Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 VIIAppendix C - Hazard Analysis 137 VIIIAppendix D - Risk Assessment 139 IX Appendix E - Assembly, Preflight, and Launch Checklists and Procedures 141 X Appendix F - Engineering Drawings 155 Page 3 Nomenclature ABS Acrylonitrile Butadiene Styrene CAD Computer-Aided Design CFD Computational Fluid Dynamics 𝐶𝐺 Center of Gravity CNC Computer Numerical Control COTS Commercial of the Shelf CP Center of Pressure CSV Comma-separated Values D Drag FEA Finite Element Analysis GNSS Global Navigation Satellite System GPS Global Positioning System GUI Graphical User Interface HTPB Hydroxyl-terminated polybutadiene I2C Inter-Integrated Circuit ID Inner Diameter IMU Inertial Measurement Unit L/D Length to Diameter MDI Methane Diisocyanate MDRA Maryland Delaware Rocketry Association MEMS Micro-electro-mechanical Sensors MPC Model Predictive Control OD Outer Diameter OLED Organic Light-Emitting Diode PCB Printed Circuit Board PETG Polyethylene Terephthalate Glycol PID Proportional–Integral–Derivative PLA Polylactic Acid PSRAM Pseudostatic (Random-Access) Memory SAC Spaceport America Cup SF Safety Factor SPI Serial Peripheral Interface TWR Thrust-Weight Ratio USB Universal Serial Bus M Mach SRAD Student Researched and Developed T Thrust UMD University of Maryland 𝑔 Gravitational Acceleration Page 4 SF Safety Factor 𝐶𝑝 Center of Pressure 𝐶𝐺 Center of Gravity 𝑃𝑚𝑎𝑥 Max Operating Pressure 𝜎 Stress 𝜏𝑠 Shear Stress 𝐾𝑛 Burning Surface Area/Nozzle Throat Area 𝑃𝑐 Chamber Pressure 𝐶∗ Characteristic Velocity 𝜌 Density 𝑇𝑠 Sampling Time 𝑔𝑖 Turn Rate Along Axis 𝑖 𝛼 Tilt Off Vertical Axis ℎ Height 𝑣 Velocity 𝛿 Flap Deployment Angle 𝑣∞ Far Field Velocity 𝜏 Time Constant 𝐶𝑑 𝑓 Coefficient of Drag Of A Flap 𝐶𝑑𝑅 Coefficient of Drag Of The Rocket Body 𝐴 𝑓 Reference Area of a Flap 𝐴𝑅 Reference Area of the Rocket Body 𝑄 Process Covariance Matrix F State Transition Matrix P(ℎ, 𝑣, 𝛼, 𝛿) Apogee Predict Function 𝑅2 Coefficient of Determination Page 5 I. Introduction A. Academic Program The Terrapin Rocket Team is a student organization at the University of Maryland, College Park. The team was established with the goal of providing students hands-on opportunities to learn about rocketry, and to gain experience designing, manufacturing, and testing engineering projects. The team is made up of over 50 undergraduate and 5 graduate members across many disciplines such as Aerospace Engineering, Mechanical Engineering, Computer Science, and Physics. While this team is comprised of several different academic backgrounds, the club is sponsored by the A. James Clark School of Engineering - Department of Aerospace Engineering under the advisement of Dr. Christopher Cadou. B. Stakeholders There are two main types of stakeholders for our project - academic and professional. The academic stakeholders are mostly engaged with the development of the team, the members, and the team’s reputation. These individuals, with their more complete understanding of rocketry, assist the team through hours of mentorship and explanation of complex topics. These individuals are the ones who provided us the critical feedback during our design reviews and assisted in developing new skills to better design our project. One of the most valuable of these mentors was our Tripoli advisor and President of the Maryland Delaware Rocketry Association (MDRA), Dennis Kingsley, who has been working with the team for the last three years and has been monumental in our success. Scott Szympruch, the former MDRA prefect, has acted as our Solid Propulsion mentor for the Solids Subteam. Additionally, our academic advisor, Dr. Christopher Cadou, has been instrumental in our team’s success. Our professional stakeholders have been those that have helped to ensure our continued success. Their support and engagement are in the form of donations and access to engineering software. These include the A. James Clark School of Engineering, The University of Maryland Student Government Association, Maryland Space Grant Consortium, The Maryland Robotics Center, Glenn L. Martin Wind Tunnel, Boeing, Northrop Grumman, Kratos Space and Missile Defense Systems, Inc., and Blue Origin. C. Team Organization TerpRockets is a rocket organization that promotes rocketry and engineering principles. One of the ways we accomplish these goals is by participating in the SAC. The executive board of TerpRockets oversees all of TerpRockets’ activities including the hands-on involvement of the SAC team. This team was partitioned into the natural subteams: Aerostructures, Air Brakes, Solids, Recovery, Avionics, and Payload. Aerostructures was tasked with the design, testing, and fabrication of the main rocket body as well as all subsystems not covered by recovery or payload. Air Brakes designed, tested, and manufactured the Air Brake Module. Our Solids Subteam designed and fabricated the solid rocket motor for Karkinos. The Recovery Subteam took responsibility for designing our recovery system and commercial avionics. Avionics developed an SRAD flight computer that behaves as a data logger and broadcasts telemetry. The Payload team was tasked with the design, testing, and fabrication of the payload for Karkinos. During the 2022-2023 season, we approached SAC 2023 with the goal of flying early and often. This meant following previously proven rocketry techniques and modifying them accordingly for Karkinos. To delegate the workload and ensure rapid testing and flying, we developed a separation between competition-critical systems and non-critical systems for the subteams. A SAC team focused on building and prepping Karkinos for early flights had the ability to work on the competition rocket without relying on the finalized sub-systems. The full breakdown of the team can be shown in Fig. 1. Page 6 Fig. 1 Team organization II. System Architecture Overview Fig. 2 Karkinos layout Karkinos is composed of 6 separate subsystems: Propulsion, Aerostructures, Air Brake, Avionics, Recov- ery/Tracking, and Payload. The propulsion subsystem is an SRAD solid rocket motor made by the team. It is a 98 mm motor with approximately 37” of propellant length and 11,500 N-s of impulse designed to take Karkinos to 10,000ft. The Aerostructures subsystem is responsible for the air frame and structural components of the rocket including the fin can, drogue tube, main tube, electronics bays, and nosecone. All external air frames except for the nosecone were made by students from either carbon fiber or fiberglass. Next, the Air Brake system sits directly above the fin can and is designed to trim the rocket’s flight to a precise altitude. A variety of sensors onboard allow it to control when to deploy flaps into the air stream to induce drag and slow down the rocket. The SRAD avionics system is within the primary electronics bay and is a custom flight computer collecting data and sending it back to the ground. While it is not deploying any charges, it can transmit the status and location of the rocket to a ground station. Also in the avionics bay are the commercial recovery electronics and GPS trackers. This consists of two Altus Metrum EasyMinis, a Featherweight GPS, and AltusMetrum TeleGPS. Karkinos is a dual break - dual deploy rocket with the drogue parachute directly aft of the electronics bay and the main parachute forward of it. The drogue is deployed at apogee and the main is deployed at 1,000 ft during descent. Directly below the nosecone is a secondary electronics bay with a single Page 7 EasyMini and an Eggtimer Wifi Switch. This deploys the nosecone at a set altitude to allow the payload to deploy. The payload is on the top nosecone electronics bay bulkhead and is split up into two sections, a vehicle, and a vehicle release mechanism. The release mechanism is made of an aluminum frame and holds the vehicle during ascent. Once the nosecone is ejected and the release mechanism detects light, the vehicle is pushed out and recovers separately. The vehicle will attempt to control its descent using a flat parachute and moving the shroud lines. The vehicle performance parameters for Karkinos can be found in Table 2. The motor being used for Karkinos is an SRAD 98 mm N2500. Details Table 2 Vehicle Parameters Predicted Apogee (ft) 10,969 Total Impulse (N-s) 11,548 Peak Thrust (N) 3,620 Takeoff Mass (lb) 82 Takeoff TWR 9.6 Velocity off Rail (ft/s) 92 Max. Velocity (ft/s) 890 Max. Acceleration (g) 9.4 Stability Margin (Calibers) 2.75 about the performance of the motor are listed in Fig. 3 and Table. 3 and further discussed in the propulsion subsection. Fig. 3 TRT N2500 motor characteristics Flight Simulations Karkinos was primarily simulated in OpenRocket. This program gives us an easy way to design our rocket before detailed design in CAD. It also lets us get fairly accurate weight estimates before building anything, which is important for designing our motor. The base design of Karkinos was based on our competition rocket from 2022, Terpulence II with a few major changes. The primary goals we had were to decrease the weight substantially, Page 8 Table 3 TRT N2500 Motor Characteristics Loaded Weight (g) 12,609 Propellant Weight (g) 6,667 Burnout Weight (g) 5,942 Total Impulse (Ns) 11,548 Average Thrust (N) 2,531 Burn Time (s) 4.5 decrease length where possible, and increase the stability margin slightly. The primary weight savings for Karkinos were gained by switching to student-made tubing and removing unnecessary weight. Our custom tubes are around 50%-60% the weight of commercial filament wound fiberglass tubing from Wildman. Especially since we do not have a very demanding flight profile, reducing the load capacity with lighter tubing was worthwhile. This was especially important in the Air Brake module which has a 22” coupler and 10” airframe which added lots of mass to the aft end of the rocket last year. Switching both to custom tubing decreased the weight significantly. We were also able to remove some unnecessary parts that were increasing weight previously. Last year, an extension was added to the fin can so that a longer motor could be accommodated in the middle of the year. This meant another 24” of commercial airframe and 12” of coupler needed to be added since remaking the entire fincan would be excessive. Since we knew what size casing we would need from the start this year, we were able to remove that extension and instead use a single carbon fiber tube that is the proper length. The drogue airframe was also reduced from 32” to 24” while the main tube was increased from 33” to 35” to give more space for the parachute. Weight was also saved throughout the rocket by switching from steel hardware to aluminum hardware. Especially for the multiple threaded rods in both Ebays, switching to aluminum saved an unnecessary weight for little loss in strength. The model was then finalized with the mass of parachutes, harnesses, payload, and estimated epoxy mass. Fig. 4 OpenRocket simulation for Spaceport America conditions Since Karkinos has a fairly high L/D for a rocket, the standard rules of stability are not consistent. While the rocket may have a stability margin above 1, since it is fairly long this may not be adequate. A new rule of thumb for stability that has become more popular in recent years has been to calculate stability in terms of percent of body length. The formula for this value is, 𝑆𝑀 = |𝐶𝑝 − 𝐶𝑔 | 𝐿 , (1) where 𝐶𝑝 is the center of pressure, 𝐶𝑔 is the center of mass, and 𝐿 is the total length of the rocket. For a stable rocket, this safety margin should be between 8%-18%. This means that a long rocket will have a traditional stability Page 9 margin much larger than one and a very short rocket can have a margin below one. The fins were designed around these restrictions and targeted stability on the far end of this margin. Since our air brakes will pull the center of pressure forward when deployed, bringing the center of pressure as far back as possible is important to ensure stability throughout the flight. As seen in 5, Karkinos lifts off with a minimum stability margin of 2.7 calibers, equivalent to 10%. As propellant leaves the rocket, the center of gravity moves forward, and the stability peaks at 5.4 calibers, which is equivalent to 20%. Fig. 5 OpenRocket simulation of stability over time for Karkinos In general, our past simulations in OpenRocket have overpredicted our actual flights. This is for a variety of reasons such as extra protrusions like cameras or bolts outside the rocket as well as not flying perfectly straight. Because of this, the simulation was made to be fairly conservative. Launch conditions were set for Spaceport America with an altitude of 5,000 ft and a rail angle of 2 degrees. Since the rail buttons are spaced 3 ft apart, a rail length of 14ft was used since the buttons can no longer guide the rocket once the first one has left the rail. Since Karkinos contains an air brake module and cannot add any extra thrust, the rocket must overshoot the 10,000 ft target. Theoretically, our air brake can decrease a flight’s altitude by up to 1,000 ft. However, it is much easier and less demanding to remove less altitude, so we simulated the rocket to target between 10,200 ft to 11,000ft. Once the rocket was built, the center of gravity and total mass were updated in OpenRocket, and following our third test flight, the surface finish was changed to better match actual flight conditions. Simulation results can be found in 4 below. Table 4 Open Rocket Simulation Results CG Location (in from tip) 97 CP Location (in from tip) 122 Stability (Caliber) 4.1 Stability (%) 16.3 Takeoff TWR 9.6 Velocity off Rail (ft/s) 92 Max. Velocity (ft/s) 890 Altitude at Max Velocity (ft) 2,000 Max. Altitude (ft) 10,969 Page 10 A. Propulsion Subsytems The motor for Karkinos is a 98 mm SRAD N motor developed and tested by students in the Terrapin Rocket Team. It uses a custom purple propellant called Terple Nebula and is designed to launch Karkinos to 10,000ft at the Spaceport America Cup. Fig.6 shows a cross-section of the motor. It is made up of an aluminum casing from Fisher Research, an aluminum forward closure, a graphite nozzle, and XX phenolic liner. The nozzle and closure are retained on either end with steel snap rings. The motor casing is then slid into the motor mount tube and thrust is transferred from the lip of the case to the thrust plate. Fig. 6 Motor Grains Assembly Cross Section 1. Material Selection and Pressure Vessel Design Since this is the team’s first time making large experimental solids, we wanted to minimize the risk to hardware and our testing timeline. The team completed four full-scale test flights this year; a failure of the motor in testing or flight would lead to a delay of potentially months as new hardware would need to be ordered and the motor redesigned and retested. Because of these requirements, the decision was made to stick to common experimental solid practices Page 11 Table 5 Motor Case Dimensions Inner Diameter (in) 3.53 Outer Diameter (in) 3.88 Wall Thickness (in) 0.18 Length (in) 42 Distance from Edge to Snap Ring Groove (in) 0.375 and components. Our motor casing is an aluminum 98-13,000 snap ring casing from Fisher Research. These casings are very common within the hobby and are designed to hold the pressure of standard experimental motors. We also purchased a full graphite nozzle and aluminum forward closure from Fisher shown in Fig.7, also common choices. Steel snap rings retain the nozzle and forward closure on either end and a steel washer is used to distribute the force on the nozzle. On both the nozzle and forward closure, silicone o-rings are used for their good temperature resistance. Fig. 7 Nozzles and Forward Closure In order to protect the casing from the heat of the motor, we used XX Phenolic Liners from Franklin Fibre. XX liners provide much more protection compared to spiral wound phenolic liners from Always Ready Rocketry, and while our propellant does not run very hot, extra protection is helpful. The propellant is packed into standard cardboard casting tubes and then the grains are glued into the liner before firing, as shown in Fig.8. Since we are using a commercial casing, we are confident in its ability to hold reasonable pressures, but its failure modes were calculated to determine our burst margins. Table 5 shows the most relevant measurements of our cylindrical casing design. In order to calculate relevant safety factors, we noted various strength metrics of our chosen material, Aluminum 6061-T6, and recorded them in Table 6. The decreased strength of the material at increased temperatures was also recorded as an addition to the safety factor calculation. The rightmost column in the table takes into account the temperature in the casing at high pressures when the motor is burning (approximated as 392 ◦F). An additional useful metric for snap ring casing designs is the max allowable thrust for the snap ring itself. This was provided by the manufacturer of the snap ring and is given to be 54,700 pounds. We can also define Max Expected Operating Pressure(MEOP) to be 𝑃𝑚𝑎𝑥 = 655𝑝𝑠𝑖 (2) According to the OpenMotor simulation for the static-tested N motor, the peak chamber pressure was expected to be Page 12 Fig. 8 Glued Test Motors Table 6 Motor case Strengths 75◦𝐹 392◦𝐹 Ultimate Tensile Strength (ksi) 45 19 Yield Tensile Strength (ksi) 40 14.9 Shear Strength (ksi) 30 11.2 about 720 psi. During the completed static fire in January, the pressure transducer data showed an actual peak of about 655 psi, so we used the practical result. For a thin-walled pressure vessel, the hoop stress can be calculated as 𝜎ℎ𝑜𝑜𝑝 = 𝑃𝑚𝑎𝑥 𝐼𝐷 2𝑡 (3) where ID is the inner diameter of the pressure vessel and t is the wall thickness. This results in a hoop stress of 6,422.64 psi, and to calculate the hoop stress safety factor, we simply divide the yield strength of the material by the hoop stress, or 𝑆𝐹ℎ𝑜𝑜𝑝 = 𝜎𝑦𝑖𝑒𝑙𝑑 𝜎ℎ𝑜𝑜𝑝 (4) The above equation results in a hoop safety factor of 6.23 for a room temperature yield strength, and 2.32 for a yield strength at 392 degrees F. Similarly, the axial stress in a thin-walled pressure can be calculated as 𝜎𝑎𝑥𝑖𝑎𝑙 = 𝑃𝑚𝑎𝑥 𝐼𝐷 4𝑡 (5) Page 13 Table 7 Motor Case Safety Factors 75◦𝐹 392◦𝐹 Hoop Stress S.F. 6.23 2.32 Axial Stress S.F. 12.45 4.64 Snap Ring Groove Distance S.F. 4.87 1.81 This results in an axial stress of 3211.32 psi, and to calculate the axial stress safety factor, we simply divide the yield strength by the axial stress. 𝑆𝐹𝑎𝑥𝑖𝑎𝑙 = 𝜎𝑦𝑖𝑒𝑙𝑑 𝜎𝑎𝑥𝑖𝑎𝑙 (6) The above equation yields in an axial safety factor of 12.45 at room temperature, or 4.64 at 392 degrees F. For snap ring pressure vessel designs, another useful safety metric is the distance from the snap ring groove to the edge. The minimum distance can be calculated as 𝐸𝑚𝑖𝑛 = 𝑃𝑚𝑎𝑥 𝐼𝐷 𝜏𝑠 (7) where 𝜏𝑠 is the shear strength of the casing material. Using our values, our minimum distance is 0.08 inches. Our actual snap ring groove distance is 0.375 inches. To calculate the snap ring groove distance safety factor, we simply take the fraction of these numbers as shown. 𝑆𝐹𝑔𝑟𝑜𝑜𝑣𝑒𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 = 𝐸 𝐸𝑚𝑖𝑛 (8) where E is our snap ring groove distance. This safety factor turns out to be 4.87 at room temperature, or 1.81 at 392 degrees F. The final safety factor calculated was related to the max thrust on the snap ring. The thrust on our snap ring was calculated by taking the MEOP and multiplying it by the cross-sectional area of the casing, or 𝐹𝑠𝑛𝑎𝑝 = 𝑃𝑚𝑎𝑥𝜋( 𝐼𝐷 2 )2 (9) which results in 6,301.84 lbs. In order to calculate the thrust safety factor, we divide the allowable snap ring force by 6,301.84 as shown. 𝑆𝐹𝑡ℎ𝑟𝑢𝑠𝑡 = 𝐹𝑚𝑎𝑥 𝐹𝑠𝑛𝑎𝑝 (10) which yields a safety factor of 8.68. Even at relatively high operating temperatures up to 392 degrees F, our stress safety factors stay above 1.81. Each safety factor is about triple its 392-degrees F counterpart assuming room temperatures. The closest point of failure is the snap ring groove shearing axially outwards, but even that is designed to withstand over four times the MEOP. Propellant Choice and Formulation The propellant in the motor powering Karkinos is called Terple Nebula. This is a fairly common experimental propellant that was modified with oxamide in order to slow the burn rate. The chemical proportions and purposes for Terple Nebula can be found in 8. Since this is the first year that the team is mixing motors, we wanted to have enough time to learn how to properly test motors. By using a preexisting formula, we can have a propellant characterized much faster instead of constantly tweaking the formula. In order to properly characterize this propellant, thermo-chemical values were calculated using a program called ProPep 3. With this program, we could input our chemicals and nominal percentages and receive values such as density, c*, and chamber temperature as shown in 9. When calculating our burn rate coefficients, these values are important to find the burn rates at different pressure for each motor. Page 14 Table 8 Terple Nebula Formula Ingredient % of Mixture Purpose R45m, HTPB 13 Fuel/Binder Dioctyl Adipate 2.6 Plasticizer Castor Oil 0.1 Grain Hardener Tepanol 0.1 Bonding Agent Magnesium, 325 mesh 2 Metal Fuel Strontium Nitrate, 200𝜇 15 Secondary Oxidizer Oxamide 2 Burn Rate Suppressant Copper Oxide 1 Burn Rate Catalyst Ammonium Perchlorate, 200𝜇 62 Primary Oxidizer MDI 2.2 Curative Table 9 Outputs from ProPep 3 Density (𝑙𝑏/𝑖𝑛3) 0.0632 𝑐∗ ( 𝑓 𝑡/𝑠) 4,442 Chamber Temperature (K) 2,368 2. Propellant Characterization Before manufacturing of the full-scale N motors could begin, a set of characterization motors needed to be made to characterize the propellant and determine its burn rate coefficient and burn rate exponent. The planned configuration for the full-scale motor utilized a 98 mm six-grain casing, which has a relatively high length-to-diameter ratio (L/D). High L/D values make it easier for erosive burning to occur, which can cause dangerously high mass fluxes and therefore pressure spikes. It was important to find out if this propellant was prone to these erosive spikes during the characterization phase before the full-scale tests. Therefore, the characterization motor casings had higher L/D ratios as well, roughly in the range of 10.5-13.5. The set of characterization motors consisted of three 38 mm motors and three 54 mm motors and was tested in configurations that covered a range of possible Kn values that the full-scale motor would experience during its flight. From the OpenMotor simulation, the full-scale motor would reach a peak Kn value of 312. By choosing various case sizes, nozzle throat diameters, and core diameters, a wide range of Kn values and peak pressures can be achieved. Between all six motors, the Kn value varied between 198 and 308, and the peak pressures varied between 369 psi and 723 psi. Page 15 Fig. 9 Pc vs. Kn curve The six characterization motors were tested at the MDRA Sod Farm launch site on September 8th, 2022. For these smaller-scale tests, the team utilized a test stand, pressure transducer, load cell, and DAQ to record pressure and force readings for each of the burns. All test firings were nominal, except for two slight erosive spikes which were not a cause for concern. In order to determine our Kn-Pc and Burn rate relations, the values from the static tests needed to be plotted to find best-fit lines. For the Kn-Pc curve, the known largest Kn from OpenMotor and the largest non-erosive pressure were plotted. The best-fit curve could then be found from these points. To find our burn rate coefficients, a, and burn rate exponent, n, for our burn rate Eq. 11, values from the static tests as well as ProPep outputs were used. 𝑅𝑏 = 𝑎𝑥𝑛 (11) Density and characteristic velocity, 𝑐∗, were found using ProPep. Using the peak Kn and maximum non-erosive pressure, burn rate values were found for each motor. These values were then plotted against pressure in Fig. 10 and a best-fit line was found to model Eq. 11. 𝑅𝑏 = 𝑃𝑐ℎ𝑎𝑚𝑏𝑒𝑟 𝐾𝑛𝑐 ∗𝜌 (12) Fig. 10 Rb vs. Pc curve This resulted in a burn rate coefficient 𝑎 of 0.013 and a burn rate exponent 𝑛 of 0.461. One of the major issues we experienced with characterization was an overestimate of our pressure compared to the static fires. When the Page 16 Table 10 Erosivity Conditions [1] Non-Erosive: Chamber Pressure (psi) Maximum Mass Flux (lb/in2 ∗ 𝑠) 400-600 1.0 800 1.75 1400 2.0 Erosive: 400 2.0 600 2.5 800 3.0 configurations for the motors were originally designed, the nominal propellant length of the case was used. Once the motors were cast and assembled, all of the required grains could not fit into the case. To fix this, the forward grains were shortened to fit, however, their lengths were never measured. This meant that the motors were running at a slightly lower Kn than we were predicting, which meant a lower pressure. While this aspect of the characterization was not accurate, we found that the thrust values from our simulations were fairly consistent with the values from the load cell. 3. Full Scale Motor Design The full-scale motor needs to fulfill multiple requirements for a stable flight and also to satisfy design requirements set by the DTEG, primarily an apogee of 10,000 ft and a rail exit velocity of 100 ft/s. Since Karkinos contains an air brake system, the rocket needs to overshoot its target so that the system can adjust and hit 10,000ft precisely. The team’s rocket in 2022, Terpulence II, flew on a Loki N3800 which is a 98 mm snap ring motor with 12,500 Ns of impulse. Since Karkinos is a similar, but lighter rocket due to a much lighter Air Brake system and SRAD tubing, we could target a slightly lower total impulse in a similar casing to reach the desired apogee. Due to these constraints, we chose a Fisher Research 98-13,000 casing that can hold 6 standard 98 mm grains. This casing is slightly longer than the 98-12,500 casing that was used for the Loki N3800 and gave us the option of using spacers if necessary to decrease the amount of propellant if we were overshooting too much. The second consideration that we had to design for was our rail exit velocity of 100 ft/s. Based on previous experience with rockets of this size, we would need a 10:1 thrust-to-weight ratio. For an 80 lb rocket, this meant we needed 800 lbs of thrust at liftoff, or around 3,500 Newtons. We were also expecting a small erosive spike which would help the rocket get off the pad, so there was some margin in our initial thrust estimates. The next design requirements were not important for the rocket overall but for the motor itself. Since this is the team’s first 98 mm motor, we wanted a fairly conservative design to minimize the destruction of hardware and the rocket. Since we cannot predict erosive burning within OpenMotor, we could follow a few rules of thumb for erosivity. The first is mass flux-based erosive burning, which can become a major problem in motors with higher L/D. Longer motors have much higher mass flows through the aft grains which causes local increases in burn rate. Charles Rogers provides some design criteria for different chamber pressures based on peak core mass flux in his article Erosive Burning Design Criteria for High Power and Experimental/Amateur Solid Rocket Motors shown in Table 10. For a target Kn of 300 and a chamber pressure of 700-800 psi, we aimed to keep the mass flux below the recommended limit of 1.75 lb/sec·in2. Since our propellant is fast burning, in a long motor like this, our mass flow rate is going to be fairly high at the aft end of the motor. Since mass flux is mass flow divided by area, by increasing the core diameter of the aft grains where the mass flow is the highest, we could decrease our mass flux. Widening the cores at the aft end also allowed us to increase our initial Kn for a higher initial thrust to get off the pad. However, this also has a few major disadvantages. First, the wider cores decrease our volume loading significantly. We are losing valuable space inside the case for propellant that may be necessary later on. Increasing the core diameter at the aft end also means that since those grains burn out before all of the others, the liner at the aft end will encounter much more heat. This could Page 17 Table 11 Simulated Motor Parameters Motor Designation N2733 Initial Kn 291 Total Impulse (Ns) 11808 Volume Loading (%) 80.37 Motor Class 15% N Propellant Mass (lb) 14.4 Burn Time (s) 4.3 Propellant Length (in) 36 Peak Pressure (psi) 704 Peak Mass Flux (lb/(in2 · 𝑠)) 1.58 Average Pressure (psi) 538 Port Throat Ratio 2.79 Peak Kn 312 Peak Thrust (N) 3700 lead to the remaining casting tubes becoming dislodged and clogging the nozzle or even a liner burn through. However, since our propellant is not very hot and we are not burning for too long, these risks were not especially important to us. The second design constraint in regard to erosivity is the port throat ratio. This is a comparison of the core area of the grains and the throat area of the nozzle. If the port-throat ratio is too small flow can choke within the grains which will locally increase the burn rate. Generally, this value should be above 2:1 in order to prevent major erosive burning. Since we had already increased the core diameter of the aft grains, that also increased our port throat ratio above this limit. After many iterations in OpenMotor, the final configuration decided for the motor is: • 3 x 6” long 1.25” cores • 2 x 6” long 1.5” cores • 2 x 3” long 1.75 cores Results from OpenMotor using this configuration can be found in table 11. 4. Motor Manufacturing and Assembly To make predictable, reliable motors, a set procedure, and chemicals from the same batch are important for batch-to-batch consistency. Our mixing procedure was set before mixing the characterization motors and was kept the same for the following two mix sessions. Before mixing, propellant batch quantities were determined based on the total propellant mass desired along with a precautionary extra quantity. These totals could then be input into an Excel spreadsheet that automatically calculates the required quantities of each chemical. During the mixing process, the actual weight was also recorded on this sheet so that any deviations during firing could potentially be traced back to issues during the mixing process. During the 2022-2023 year, the team completed three mixing sessions. The first is for characterization, the second is for two full-scale N motors, and the third is for three full-scale flight motors. The entire procedure, from mixing to casting, lasts around six to seven hours. The procedure starts with weighing out measuring bowls, recording their mass values, and taring their masses on a scale. Since ingredient quantities for multiple motors were larger than what a single bowl could hold, multiple 20 qt mixers were used to contain fractional batches of propellant. Wearing nitrile gloves, members poured all of the liquid components except the curative as specified by the mix sheet into each bowl one at a time. Next, the bowls were placed into a 20 qt stand mixer for 20-30 minutes. Special care was taken to scrape any built-up Tepanol on the bottom of the bowl that was not getting picked up by the mixer. Tepanol is a chemical that is used to chemically bind the ammonium perchlorate particles to the HTPB. Tepanol has a shelf life, and if not stored in cold temperature conditions, becomes less effective over time and much more viscous. The Tepanol supply of hobbyists making experimental motors is often a decades-old surplus that has become extremely viscous. This meant that while all the other liquids had already been fully incorporated, the team often had to scrape off the Tepanol from the paddle or bowl to get it mixed. While the liquid components were mixed, the solid powdered components were measured out in clean bowls in a similar fashion to the liquids. Each solid is measured in its own separate bowl, and dry solids are never mixed together. Another ingredient of special note was the Strontium Nitrate, which can absorb moisture and clump together. Before adding to the mixture, the Strontium Nitrate needed to be broken down with a mortar and pestle and pushed through a sieve in order to make sure there were no clumps remaining. Before adding the solids to the liquids Page 18 Fig. 11 Mixing Propellant mixture, the mixing bowl was placed under vacuum for ten minutes. This helps get the majority of the air out of the mixture and also gets rid of any water that may have been absorbed. Removing the air leads to a denser propellant and fewer voids, which leads to a more consistent product with a predictable burn rate. After each of the solids is added, the mixture is vacuumed again, then, while wearing solid particulate respirators, the ammonium perchlorate is added. This portion spends the longest under the mixer, so the extra time is used to cut and prepare the casting tubes and mandrels. Once the ammonium perchlorate is mixed in, the batch is put under vacuum one last time and then the curative is added (with organic vapor and solid particulate respirator masks on) and mixed in. Finally, the packing process can begin. While still wearing gloves, the propellant is shaped into elongated clumps that fit evenly around the central mandrel and within the casting tubes. Then, wooden dowels are used to evenly distribute the propellant around the mandrel after each additional amount is added into the casting tube. Once each casting tube is filled, the propellant is left to cure for at least four hours. Fig. 12 Liquid Components Under Vacuum Fig. 13 Filled Casting Tubes Curing At this point, the grains needed to be cut to size as per the grain schedule in Fig. 6. A hand miter saw was used with a backing block to get straight cuts. Once the grains were prepared, they were then test-fitted with the forward closure and nozzle into the liner and slid into the casing to check for snap ring clearance. Lastly, grains are glued into the liner with silicone caulk. This prevents castings tubes or grains from coming loose and clogging the nozzle. For the final motor assembly, o-rings and the outside of the liner were thoroughly greased, snap ring grooves are checked for Page 19 Fig. 14 Cut Motor Grains Fig. 15 Greasing the Liner sharp o-ring damaging burrs, and the assembly is secured in place into the casing by fully seating the snap rings into the casing’s grooves. 5. Motor Test Stand To test the solid motor, an SRAD test stand was chosen based on the size of the motor as well as the scale of its thrust output. It also needed to be compatible with the data acquisition setup of our choice, which will be discussed further in this report. This means having a valid location to place the load cell and pressure transducer for each static test. After thorough research, a suitable test stand design was found that was relatively simple to assemble with the materials we had. It also met the criteria of being strong enough to withstand the thrust of the motor to be tested. An assembly of the stand was developed in Solidworks to generate a straightforward bill of materials, which confirmed we had every component necessary to build the stand. The stand was assembled with mechanical fasteners such as spring nuts and bolts. Figure 16 is the assembled test stand with the motor inserted. The orientation of the stand was chosen to be upright as opposed to horizontal because the downward force from the motor would be absorbed by the ground. This way, the stand simply needed to be stabilized from tipping over, which was done with ratchet straps that were securely staked to the ground. Stakes were also placed along the base of the stand to further secure its orientation and position. This stand allowed simple placement of the load cell and pressure transducers, shown in Fig. 17. The NPT tee fitting screwed directly onto the forward closure of the casing, and had a flat surface that could be used to push onto the button load cell placed underneath the casing. Two pressure transducers were attached (one on each side of the tee fitting) to confirm the reliability of the pressure data. An important criterion for choosing a data acquisition device was its ability to connect our computers to the sensors via Ethernet. During the static fire, in order to stay a safe distance away from the motor and collect data, a long-range Ethernet cable was used to provide a direct connection from our computers to the data acquisition device. Ethernet was more reliable than a wireless connection setup and much simpler to put together. For this reason, a Labjack T4 was Page 20 chosen, as it provided reliable Ethernet capabilities. Fig. 16 Static Fire Setup Fig. 17 Data Acquisition Setup A test procedure was developed for the pressure transducer and the load cell to calibrate its readings and confirm they were accurate prior to the static fire. For the pressure transducer, a hydraulic pump was attached to an analog pressure gauge and the pressure transducer. The voltage readings from the pressure transducer were then compared to the pressure readings from the gauge, which allowed us to generate a scaling equation that would convert voltage readings to accurate pressure readings in pounds per square inch. For the load cell calibration, the load cell was attached to an amplifier that connected to the Labjack T4. Next, we measured and recorded the weights of various aluminum blocks. These blocks were then placed on the load cell via the setup in Fig.18, and the voltage readings on the load cell corresponded to the amount of weight we placed on it. This allowed us to generate a scaling equation that could convert load cell voltage to weights in pounds for the static test. Fig. 18 Load Cell Calibration Setup A problem that was encountered with the load cell was small spikes and choppiness in the data after a static fire. This was caused by low resolution in the data acquisition input port. We changed the port from an analog input to a flexible input/output port on the same Labjack. It is important to note that this port needs to be configured to its analog form (as opposed to digital), or the data will only show in the form of a 1 if a force is applied, or a 0 if no force is applied. Page 21 However, this port has a voltage range of 0-2.5 V, so if the load cell output exceeds 2.5 V (which it did according to the static fire data), the Labjack would max out and fail to record data above 2.5 V. Another solution to the choppiness was to increase the sample rate. This would of course increase the file size, but it was worth it to improve the resolution of the data. We decided it was an acceptable trade-off to have minimal spikes in the data to ensure data was collected for the entire thrust range necessary. Fig. 19 First Static Fire Fig. 20 Second Static Fire 6. Full Scale Motor Tests and Results Fig. 21 Pressurve Curve from Static Fire 1 For static testing of the full-scale motor, two motors were manufactured. These would both be ground tested, however, they would have different nozzles. The first motor would have a larger nozzle for a lower Kn while the second motor would have a smaller throat for the targeted Kn. This was done in order to verify that there would be no major differences from moving up to 98 mm from 54 mm. By testing at a lower Kn, the motor would run at a more conservative pressure in case there were any unexpected erosive spikes. For the testing, the motors would each have two pressure transducers being read by entirely separate systems for reliability. These motors were manufactured in November 2022 and were originally planned to be tested in December 2022. Unfortunately, after setting up the test stand and electronics, when assembling the motor the o-rings could not be found and the test firings would have to be delayed until January. During the January testing, setup and assembly of the first motor went nominally. The motor fired without issue, however during review it was found that the data acquisition system did not record force readings. Because of this, the load cell was switched to a different and more proven data acquisition system. This test was also nominal and both pressure and force readings were recorded. During the review of the recorded data, the results were fairly close to our expected values. Similar to the characterization motors, the pressure was slightly lower than expected, but the force was similar. There were minor erosive spikes in both motors, but neither was much larger than the normal burning max pressure. The motors also came Page 22 up to pressure quickly, even though there were no grain spacer o-rings or concave faces. The force readings for the second motor matched the simulations, with a slightly longer taper toward the end of the burn. Fig. 22 Thrust Curve Comparison of Second Static Fire Fig. 23 Pressure Curve Comparison of Second Static Fire 7. Mixing Sessions and Motors Made Over the 2022-2023 year, there were a total of three mixing sessions in preparation for the Spaceport America Cup. During the first, six characterization motors were made. Two N motors were made during the second session for ground testing and 3 N motors were made during the third session, two for test flights and one for the competition. During the first full-scale mixing session, everything went nominally, however, we ran into an issue near the end of our casting session. We had oversized the casting tubes so that we could cut the grains to the desired lengths but did not include enough extra in the batch size to accommodate this. We did include 5% for waste, but this was not enough to make up for it. Because of this, we had filled some casting tubes fully while others had 1” remaining. This led to the lengths of the grains changing slightly to accommodate the cast propellant. While some grains ended up shorter and others longer, the simulated thrust curve was not changed too much and the change in Kn was minimal. Since we did not want to retest our motors, we recorded the grain lengths and used those measurements for the flight motors. During the second large mixing session, the excess was increased to 10% and casting tubes were more carefully cut so that there wouldn’t be too much extra. Table 12 Mixed Motors Measured Values 24 Propellant Mass (g) Propellant Length (in) Test Motor 1 6568 36.375 Test Motor 2 6662 36.75 Flight Motor 1 6654 36.125 Flight Motor 2 6690 36.25 8. Igniters For igniting the motors, we have been using QuickBurst ProCast BKnO3. The ProCast is inserted into a straw and then an MJG Firewire igniter is pushed inside. It is then left for the acetone to evaporate from the mixture before being used. Generally, 1.5” to 2” of the mixture is inserted into the straw for the larger N motors. We have used these on many motors from H to N and have had great reliability with ignition. For inserting the igniter into the motor, the wire is Page 23 taped to a thin wooden dowel. If the igniter leads are not long enough, they are extended with enough solid core wire to come out of the nozzle. The dowel is then taped to the launch rail so that it cannot fall out and stays at the top of the motor. At the Spaceport America Cup, 2 of these igniters will be used in parallel to ignite the motor. Fig. 24 Example BKnO3 Igniter B. Aero-Structures Subsystems The purpose of the Aerostructures subteam is to design and manufacture Karkinos to be lightweight, cost-efficient, and manufacturable. Karkinos is nominally six inches in diameter and stands approximately 12 feet tall. The nose cone and electronics bay coupler are made of wound fiberglass. The recovery section and upper fin can are made of student-rolled fiberglass. The fin can is a student-made hand-rolled carbon fiber airframe and the fins and centering rings are made of G10 fiberglass. The main parachute is housed in the recovery section, whereas the drogue is housed in the upper fin can above the Air Brake. Fiberglass was chosen as the primary structural material as it is lightweight and very strong, which makes it very popular in high-power rocketry. Fiberglass is also transparent to radio frequencies, which is important for receiving telemetry from flight computers. Carbon fiber was chosen for the fin can as it is very strong and stiff while being lightweight. Minimizing weight at the aft end of the rocket is important for stability, and its strength is useful for keeping the rocket together when it touches down. 1. Body Tubes and Couplers Airframe Manufacturing All airframes on Karkinos except the nosecone are student-made fiberglass or carbon fiber tubes. While we have used commercial filament wound tubing in the past, we have found that it is often fairly heavy, and custom-made tubes can be around 50% to 60% the weight of a commercial tube. The process of making a tube starts with the correct mandrel. It is important for the airframes to have the correct inner diameter so that the interface with the couplers is just right. If it is too loose, the airframe will not be rigid and the rocket will bend during flight. If it is too Page 24 tight, then assembling the rocket for launch will be difficult and there may be too much interference for the sections to separate in flight. With these constraints in mind, we chose to use a 65" long section of a 6" coupler from Wildman Rocketry. With a piece of Mylar, a clear plastic sheet, over the coupler we can provide adequate tolerances and provide a surface to protect the mandrel from epoxy. The Mylar is then cleaned with acetone and thoroughly coated in epoxy, which for these tubes is Aeropoxy 2032/3665. Fiberglass or Carbon cloth is wrapped around the tube tightly and epoxy is pressed into it using a brush and plastic squeegee. Since carbon is stiffer than fiberglass, we need fewer layers for a tube with similar strength, in this case, we use four wraps of 3k 2x2 twill carbon and six wraps of 6 oz plain weave fiberglass. Fig. 25 Layup of carbon fiber airframe Once the cloth has been fully wrapped around the tube, a layer of peel ply is pressed into the cloth. This helps fill in the weave and makes it much easier to give the tube a smoother finish. We also attempted to finish a tube with shrink tape for an easier glossy finish but found it difficult to get a consistent wrap and finish. In previous years we had many issues removing the tube from the mandrel. We had tried waxing the mandrel in order to give it a slippery surface, but this was not effective. Instead, this year we coated the mandrel in a dry graphite lubricant. This lets the Mylar slide much easier along the coupler and makes it possible for one person to take it off. Tubes were then cut to size using a hacksaw and the edges deburred with sandpaper. Couplers Karkinos contains six separate coupler connections throughout the rocket. With so many concentric interfaces, it is important to make sure that they are all sufficiently stiff to minimize bending in flight. For a well-fitting coupler, a coupler-airframe overlap of one caliber is generally sufficient. This is the case for all coupler connections except for the one between the nosecone and nosecone Ebay. Since the nosecone is not experiencing any major bending loads, this connection is only 0.5 calibers of overlap. The majority of the couplers in Karkinos are commercial filament wound fiberglass from Wildman Rocketry. However, since we are trying to minimize mass towards the aft of the rocket, the coupler in the Air Brake module was custom-made and is discussed later in this report. When originally made, this coupler was fairly loose in our airframes which led to major bending during fit checks. This was remedied by adding a thin layer of fiberglass on top of the coupler to help fill in the gap and stiffen the connection. Another non-conventional coupler in Karkinos is the forward end of the nosecone coupler. In order to have enough room for the payload to deploy cleanly, a window was cut into the coupler. While this is removing lots of material from the coupler, we have not seen any major decrease in strength or stiffness over three flights in this configuration. Nosecone The nosecone for Karkinos is a commercial 5-1 Von Karman filament wound fiberglass nosecone from Wildman Rocketry. There are not many options for commercial fiberglass nosecones, so this nosecone from Wildman was chosen so that it would be compatible with our couplers. An eyebolt is threaded into the aluminum tip for connection Page 25 Fig. 26 Carbon tube after removal from mandrel to the nosecone harness that is connected to the top of the payload. Holes are drilled into the sides for the light sensor on the payload as well as for ambient pressure. 2. Fins Design The fins used on Karkinos are similar in proportions to the fins on the team’s rocket from the previous year, Terpulence II. Based on the changed dimensions of Karkinos from Terpulence II, the fin shape was altered in Openrocket until the desired stability criteria were met. Stability at lower speeds was less than satisfactory so the span was increased from 4.5” to 6”. The larger fin size also meant that the static margin of stability was greater, which is important for keeping the rocket stable while the Air Brake flaps are deployed. The final design consists of four fins with a swept-delta shape and a semi-span of approximately the same width as the airframe. Additionally, the fins include tabs that extend all the way from the leading edge to the trailing edge for maximum through-the-wall attachment length and ease of manufacturing. Page 26 Table 13 Fin Parameters Root chord (in) 15 Tip chord (in) 4 Sweep (in) 10 Semispan (in) 6 Thickness (in) 3/16 Table 14 FinSim Simulation Options Altitude 7,000ft ASL Elastic Axis 0.5 CN-Alpha NACA TN 4197 Fin Materials G10 fiberglass, AS H3501 Carbon Fiber Flutter Analysis and Construction The team utilized the aeroelastic fin analysis program AeroFinsim to estimate the flutter speed of the fins on Karkinos. Settings used for analysis can be found in Table 14. Fig. 27 Aerofinsim Analysis with G10 fiberglass Fig. 28 Aerofinsim Analysis with H3501 cf Using 3/16" G10 fiberglass as the material was problematic initially because the predicted flutter velocity shown in Fig. 34 was very close to the vehicle’s max projected velocity of 1000 ft/s. Since the fins should be designed such that neither the divergence nor flutter velocity was reached, the lower of the two was the primary design factor for the fins. Therefore, the composites team performed a carbon fiber vacuum-bag layup over the fins, adding two layers of 3k 2x2 twill carbon fiber to each side of the fin, the first at 45 degrees and the second at 0 degrees. The fin thickness increased from 0.1875" (3/16") to 0.23". Since carbon fiber was now the outer material, the flutter velocity could be computed using this material instead of the G10. Utilizing the carbon fiber composite option that gave the lowest flutter speed, CFRTP AS-4 PEEK, the new fin flutter velocity now reads 2,098 ft/s and the divergence velocity is 3,073 ft/s as shown in Fig. 28. More weight has been added to the rocket’s design since the first flutter estimation, so the max velocity Page 27 of the rocket has decreased and is currently projected to be 996 ft/s from the OpenRocket simulation, far below the predicted flutter velocity. Fig. 29 Cutting the rough shape of the fins on a bandsaw Fig. 30 Carbon, peel ply, release film, and breather prepared The fins for the rocket were initially cut out of 3/16" thick G10 plate on a bandsaw. These were cut with a fairly large margin so that the true fin shape could be more carefully achieved. The fins were then stacked together using alternating layers of painter’s tape and super glue. This allowed us to temporarily glue the fins together for sanding. The entire stack could then be sanded down to the desired dimensions using a belt sander. Next, the fins were laid up with carbon fiber in the vacuum-bagging process detailed above. The epoxy was cured to a leathery state and excess carbon was cut using a sharp knife. After the epoxy had fully cured, the fins were finally beveled using a 15-degree chamfer bit on a router table. Fig. 31 G10 cores after sanding Page 28 Fig. 32 Fins under vacuum Fig. 33 Fins after removal from bag and trimming Fig. 34 Fins after bevelling on router Fig. 35 Karkinos fin dimensions (in) Fin Can Manufacturing The fin can for Karkinos must fulfill multiple requirements overall. It must be able to hold the fins rigidly, retain and transfer the thrust of the motor, provide strong recovery attachment, and be built light enough that the rocket’s stability margin is not negatively impacted. The airframe is custom-wrapped carbon fiber as discussed in the previous section. The motor tube is a commercial filament wound fiberglass tube that is 23” long and there are three fiberglass centering rings along the motor tube. Two ¼”-20 threaded rods run through the centering rings to help distribute recovery loads. At the top of these rods, there is an eye nut for recovery attachment and the aft ends extend past the end of the airframe to keep the thrust plate secured. A 98 mm Aeropack screw-on retainer is secured to the thrust plate and the thrust plate is retained and held on with the two threaded rods. Assembly of the fin can begin with the airframe. Fin slots were cut out using a rotary tool with a cutoff wheel. A paper template was taped to the airframe and then the correct markings were drawn on to make sure the cuts would be straight and at the proper angle. These were initially cut undersized and then were test fit with the fins. The slots were widened with a file until the fins could be pushed through the airframe. This was done to minimize large gaps where epoxy could flow through when making external fillets. Next, the motor mount assembly was made. This assembly is made up of two centering rings forward of the fin and one aft. First, all three centering rings were stacked together and holes were drilled for the threaded rods. This makes sure that each of the holes will be straight and the threaded rods will not have any major twisting. The centering rings were then slid on the tube with the threaded rods with nuts keeping the spacing correct. The forward-centering rings and nuts were then epoxied and set to cure. The aft centering ring remained on the motor mount, however, it was not epoxied on as it would need to be removed. With the motor mount assembly complete, the airframe was prepared for bonding. The inside was sanded with 220 grit, Page 29 and then 400 grit sandpaper where the centering rings would interface and then cleaned with acetone. This is also the process that was followed for all bonded surfaces throughout this rocket. Epoxy was then added to the inside of the tube at the forward end only and the motor mount assembly was pushed in. The rocket was then stood vertically, which allowed the epoxy to settle around the perimeter of the centering rings and form a fillet when cured. Once fully cured, the aft centering ring was removed and any threads were taped over to prevent any epoxy from getting on them. Fig. 36 Motor mount assembly with forward two centering rings epoxied on Fig. 37 Template taped to centering rings prior to drilling Fig. 38 Inner diameter of the airframe is sanded for motor mount bonding Page 30 Next, each of the fins was tacked into place using Aeropoxy ES6209. In order to make sure that they would each be straight and at the correct angle, two fin guides were cut out of foam board and slid over the airframe. The root edge of each fin was generously coated in epoxy and then pushed onto a motor tube through one of the slots. It is firmly pressed in and wiggled around and then removed. This process is repeated two to three times in order to make sure there is enough epoxy to hold it in. During the final one, epoxy is added to the forward end of the fin tab in order to seal the fin and the centering ring. Once that process has been completed, it is pushed in for the last time and then aligned with the two fin guides. This process is then repeated for each of the four fins giving enough time for the epoxy to cure. After each of the fins is tacked on, internal fillets are made with Aeropoxy 2032/3665. The epoxy is thickened with colloidal silica and then poured into a syringe. The syringe is used to inject the epoxy from the aft end along the root of the fin tab. Two internal fillets are done at once and the entire fin can be rocked back and forth to evenly distribute the epoxy. This was then repeated for each of the fins taking care to make sure the epoxy was not leaking where it should not have been. External fillets were then done using Aeropoxy ES6209, however, this was thickened with colloidal silica and dyed black. The aft centering was then secured with nuts on the threaded rods and epoxied into place. Finally, the thrust plate was added and secured with two nuts and thread lock. Once completed, the entire fin can be wet sanded up to 1000 grit and then clear coated in order to give it a glossy carbon look. Fig. 39 A small fillet is formed when the fin is tacked Fig. 40 All fins tacked on with fin figs Airframe Finishing In order to give the airframe a smooth finish for both aerodynamic and aesthetic reasons, a few methods were used to complete the airframes. For the carbon fin can, both the airframe and fins were sanded smooth with sandpaper. A thin coat of laminating epoxy was then applied to the surface to fill in any major surface imperfections. This was then sanded down again to remove bumps. A clear coat was then used to give the surface a smooth glossy finish. For the fiberglass airframes, they were first given a coat of high-build spray primer. This helps fill any small divots and also makes imperfections visible. Spot putty was then used to fill the major gashed and then the entire airframe was sanded smooth. This process was repeated until there were minimal imperfections visible. The airframes and nosecones were then given a coat of gloss white spray paint and finished with two coats of gloss clear coat. 3. Bulkheads Electronics Bay Bulkheads The electronics bay (Ebay) bulkhead shown in Fig. 42 is a circular plate manufactured from aluminum 6061-T6. The overall dimensions of the bulkhead are as follows: the outer diameter is 5.998” and the total thickness is 1/4”. The bulkhead is separated into two layers with a 1/8” thick and 5.998” diameter outer lip and a 1/8” thick and 5.82” diameter ribbed pattern to strengthen the bulkhead while minimizing weight. Three sets of holes were cut from the bulkheads: two 0.386" clearance holes for a 3/8" u-bolt located at the center, two 0.323" clearance holes for two 5/16" threaded rods located 2.50" from the center, and two 0.164" holes for wiring of ejection charges. Two Ebay bulkheads enclose the electronics bay and are fixed together with the two threaded rods. A waterjet Page 31 Fig. 41 Internal fillets was selected as the fastest and cheapest means of manufacturing. The outer lip and inner ribbing were manufactured separately and joined together using the u-bolts and threaded rods; no adhesive was used. Fig. 42 shows a CAD of the bulkhead and Fig. 43 shows a FEA of the bulkhead. Fig. 42 CAD of Ebay bulkhead Fig. 43 FEA of Ebay bulkhead Finite Element Analysis (FEA) was conducted on the Ebay bulkhead. The bulkheads experience the largest loads at two points during the flight, apogee deployment, and main deployment. The force at apogee is primarily from the jerking of the recovery harness as it separates. For Karkinos, this is minimized by having a very long harness and z-folding the lines. During the main deployment, as the parachute fully inflates, it also induces a jerk on the rocket. Page 32 The main lines are also z-folded and the way the parachute is packed leads to a more gradual opening to minimize this effect. Based on accelerometer data collected from last year’s test flights, the peak loading condition was discovered to be approximately the weight of the rocket from main parachute deployment. Given that the structural design of our rocket underwent few changes this year, it can be reasonably assumed that the maximum force applied is the weight of the rocket which is about 83 lbf. The analysis shown in Fig. 44 progressively increased the element density around high-stress locations. The results show a maximum stress value of approximately 6.7 ksi. Aluminum 6061-T6 has a yield strength 40 ksi and an ultimate strength of 45 ksi. Thus, the bulkheads have a safety factor of 5.97 and will have no issue handling the peak loading condition. During our test flights this year, the bulkheads performed as expected with no deformation visible after the flight. Fig. 44 Convergent analysis through refinement of mesh around concentrated stress loads Thrust Plate At the very aft end of Karkinos is the thrust plate and motor retainer. Standard high power rockets transfer their motor thrust through the motor tube which then goes through the fins, then centering rings, and finally into the airframe. With a thrust plate, this force is not only put through the motor tube but also directly into the airframe. The thrust plate also provides a convenient way to attach the Aeropack 98mm screw on retainer. The 12 holes can be difficult to drill accurately by hand, so these are cut with a CNC machine and then tapped by hand. Since the thrust plate is secured using two ¼” nuts, it can be easily removed and switched between rockets as well. We have flown with this thrust plate multiple times, even with a motor that has almost 5000N of peak thrust and has not seen any deformation or issues. Fig. 45 Thrust plate attachment Fig. 46 Interface between Retainer, thrustplate, and motor mount tube C. Recovery Subsystems Recovery is one of the most important parts of the flight, and often the hardest to get right. For this reason, our recovery system is made to be as simple and as reliable as possible. We are using techniques that are proven not only by the high power rocketry community but also from many test flights by our own team. We conducted four successful flights with this setup in 2022, two successful L3 certifications, and four successful flights this year with this combination of parachutes, recovery harnesses, and ejection charges with a very low failure rate. This experience and simplicity Page 33 allow the recovery system to be integrated quickly and easily and allow the focus to go to other more complicated parts of the rocket. 1. Recovery Devices and Shear Pins Parachutes for Karkinos were chosen based on manufacturer-rated descent rates, packing volume, and previous experience by the team. After burnout, Karkinos weighs approximately 67 lbs. The team has used Recon Recovery and Skyangle Cert3 chutes in the past and is very familiar with folding and packing them into 6 in. airframes. For this burnout weight, the Cert3XXL from Skyangle was the best fit for a safe descent. This parachute is sized for a rocket weighing between 60 and 130 lbs for a safe descent of 17-25 ft/s [2]. With our rocket on the low end of this scale, Karkinos lands at a speed of 20 ft/s, however this will be slightly faster at the higher altitudes at Spaceport America. This was the same parachute that was used for our competition rocket in 2022, and we knew from experience that it could fit in the 6” diameter by 18” space of our main recovery tube. This parachute is also readily available and more economical compared to others in its size class, which is especially important if it needs to be replaced. For our drogue parachute, the team selected a 24” parachute from Recon Recovery. We have used this parachute as a drogue on multiple large 6” rockets including our competition rocket from last year as well as two L3s. Previously we have found rockets to descend around 90 ft/s with this parachute, and that is consistent with this year’s flights. It is important that the rocket descends at a high speed under drogue to minimize drift. This makes recovery much easier, especially on the east coast where trees can be an issue. During flight, Karkinos has three separate separation/ deployment events. At apogee, the rocket separates in half and deploys the drogue parachute. At 1,700 ft, the nosecone of the rocket is ejected and the deployable payload is released five seconds later. Finally, at 1,000 ft the main parachute is deployed. Each of these sections are held together with shear pins to prevent premature deployment both during descent and ascent. Between the drogue section and primary Ebay there are four 4-40 nylon shear pins. Not only could the drag on the fin can cause premature separation at burnout, but our Air Brake system is also producing up to 150 lbs of drag in a worst-case scenario, so it is necessary to retain these sections together. Each 4-40 shear pin provides around 50 lbs of shear force, which means 200 lbs of force total are required to separate that section. The nosecone is held together with two 4-40 shear pins. This section does not have much force trying to separate it in flight, so it is held together with the minimum amount of shear pins. Having only one shear pin could cause the coupler to jam in the airframe, so two are used to keep the forces symmetric. While smaller shear pins could have probably been used here, it is simpler to keep everything the same across the rocket to minimize the number of unique parts necessary and streamline assembly. The nosecone Ebay is held to the main airframe tube using six 4-40 shear pins, for a required total shear force of around 300 lbs. Because of the heavy 9 lb payload in this section, these shear pins keep the inertia form separating the Ebay if the drogue line snaps taught. 2. Flame Protection Since the black powder is being used for deployment events, the parachutes, harnesses, and electronics need to be protected from the high heat of the charges. All of our harnesses are made from kevlar, which is naturally flame resistant. It can withstand many ejection charges without weakening like a nylon harness. Since our parachutes are made primarily of nylon, they require extra protection to prevent melting or burning. This is done by wrapping the parachute in a nomex blanket. The drogue parachute uses a 24” x 24” blanket while the main uses a 30” x 30” blanket. Both parachutes are wrapped like a burrito, which protect the parachute and shroud lines entirely and allows it to slide easily into the airframe. Once deployed, the blanket easily unravels and lets the parachute inflate. Both electronics bays are protected primarily by the bulkheads on either end, but there are pass-through holes for charges as well. These holes are partially sealed by the e-match wires, and the remaining space is covered with a piece of electrical tape. This prevents the hot gas from passing through the ebay, potentially damaging the electronics and venting the pressure leading to no deployment. In the nosecone, since there isn’t a coupler interface with the top bulkhead, there needed to be an alternative way to seal the payload from the ejection charge. The top bulkhead has a layer of adhesive foam around the outer edge. When the nosecone is slid over the coupler, the foam is compressed sealing the top of the nose cone from the rest of the compartment. The charge that is wired in this section is similarly Page 34 sealed with electrical tape. Fig. 47 Main parachute wrapped in nomex with lines Z-folded 3. Recovery Harness Karkinos has four sets of recovery harnesses throughout the rocket. Starting from the aft end, two 7 ft sections of ” tubular kevlar rated for 3,600 lbs [3] are connected to the forward centering ring of the fin can form a y harness. These cords are routed through the Air Brake via two carbon tubes and then figure eight knots are tied at each end. A quick link connects both of these to the primary drogue harness. The drogue harness is made up of two 35 ft 7/16” 5300 lb tubular kevlar sections from OneBadHawk in series [4]. There are multiple reasons our drogue shock cord is so long, primarily because of the very heavy nosecone. When the drogue charges fire at apogee, we want to avoid the nosecone coming out too fast and snapping the shock cord taught. This shock can cause the nosecone Ebay shear pins to break, deploying the main at 10,000ft. A longer shock cord lets drag slow down both sections of the rocket, greatly reducing the forces of the shock. A long shock cord also helps keep the sections of the rocket further away during descent which minimizes the chance of sections hitting each other and causing damage. The final reason for a long shock cord is an issue specific to East Coast flying. While MDRA has a fairly large recovery area, there are multiple large patches of trees surrounding the launch site. Especially if there are high winds, which is common for the eastern shore of Maryland, our rocket can end up drifting out of the recovery area and into trees. This occurred for 3 out of our four test flights this year and two of those times our long shock cord let us recover the rocket from the trees. Unfortunately, even a long shock cord can not always help and the recovery of our third test flight required hiring tree climbers to cut the harness and recover the rocket. These kevlar harnesses have three loops, and the drogue parachute is connected to the middle loop on the forward harness. The main harness is made up of one 7/16” 35 ft. harness from OneBadHawk. This is connected from the forward end of the Ebay to the aft of the nosecone Ebay bulkhead. The main parachute is connected to the end of the cord instead of the middle of the cord here. We have found on test flights that letting the main parachute be the top of all the sections falling prevents it from tangling during descent. The final section of the shock cord is between the top of the payload and the nosecone tip. The nosecone needs to separate so that the payload can be deployed, but everything needs to stay together so that another section does not need to be tracked. These two sections are connected by 15 ft of ¼” tubular kevlar rated for 3,000 lb. This section does not encounter any major recovery forces and the goal is to get the nosecone away from the rocket so that the payload can deploy cleanly without tangling. When packing the drogue and main into the rocket, both harnesses are z-folded together. This is accomplished by making multiple bundles throughout the shock cord and using painter’s tape to hold them together as seen in 47. This makes packing it into the airframe much cleaner and also helps provide some shock absorption by ripping the tape off. Page 35 Fig. 48 Karkinos descending under its main parachute 4. Recovery Attachment Recovery mounting for Karkinos starts in the fin can. There are two ¼-20 steel threaded rods running through the three centering rings and out the aft end of the rocket. These threaded rods are tightened with nuts to each centering ring and then epoxied in place. On the forwardmost centering ring, a ¼-20 steel eye nut rated for 3750 lbs was then screwed on and epoxied in place on top of both threaded rods [5]. While these eyenuts are rated for much higher loads than other components, they were the only ones with a small enough form factor that could fit between the motor mount and airframe without interference. Last year’s competition rocket used much larger ” eye nuts instead which made it very difficult to connect and disconnect the Y harness. This is necessary to do if the kevlar becomes damaged and needs to be replaced. The Y harness then connects to the eye nuts using ¼” steel quick links rated for 1,200 lbs [6]. These quick links are connected to the y harness with a Figure-8 knot. We have used this knot many times over multiple rockets Page 36 and have found it to be very resilient. While we have not found it to have a tendency to untie itself during flight, it is wrapped a few times in electrical tape to keep it all together. After passing the Y harness through the Air Brake module, Figure-8 knots are tied on that end and a 5/16” steel quick link rated for 1,700 lbs is used to connect the Y harness to the drogue harness [7]. Since the drogue harness is made up of two 35 ft. lengths, both sections are connected with another 5/16” quick link. A ¼” quick link is used to connect the drogue parachute and nomex blanket at the middle loop of the forward drogue harness, and then the end is connected to the aft end of the electronics bay with another 5/16” quick link. Both electronics bays in this rocket have ” U bolts rated for 1,075 lbs on their bulkheads for recovery attachment [8]. The main follows the same pattern of quick links and U bolts as the drogue harness with the exception of the nosecone ebay and main switching which loop they are connected to. At the top of the payload frame, there is an adapter plate. A ¼” eyebolt is secured to the middle here and a ¼” quick link connects it to the payload shock cord. On the other end, another ¼” eyebolt is secured to the nosecone tip and kept in place with high strength threadlock. Fig. 49 Y harness mount in fin can 5. Parachute Deployment System All deployment events on Karkinos are done using 4F black powder charges. The team has lots of experience with black powder and it is by far the easiest and simplest method to deploy parachutes. Charge sizing for Karkinos was primarily based on charge sizes from our 2022 competition rocket, which had similar ejection volumes. They were then adjusted based on our four test flights. The charge sizes for this rocket are the following: • Main Primary: 6g Swiss 4F Black Powder • Main Backup: 7g Swiss 4F Black Powder • Drogue Primary: 7g Swiss 4F Black Powder • Drogue Backup: 8g Swiss 4F Black Powder • Nose cone: 2g Swiss 4F Black Powder The charges are made using nitrile glove fingers and electrical tape. The black powder is first measured and placed into a gloved finger and then an ematch with leads shunted is inserted. The glove is then twisted around the ematch wire and secured with a piece of electrical tape. More electrical tape is then tightly wound around the glove radially and axially until it is hard and cannot be squeezed. We have flown to 10,000 ft with these charges multiple times and have not had any issues with parachute deployment. Since the rocket is not at a very high altitude, the black powder can fully combust, producing gases that pressurize the air frame sufficiently to overcome the shear pins. Page 37 Fig. 50 Charges prepared for flight D. Avionics 1. Commercial Deployment Computers and Battery Selection Karkinos uses three Altus Metrum EasyMinis in total, two for the main and drogue parachute and one to deploy the nose cone. These were primarily chosen based on the team’s familiarity with them. They have been the default choice for many of our certification and test flights for two years and have been very reliable for us. They are very simple and low profile so integrating them into an electronics bay is easy and leaves lots of room for other components such as GPS or custom avionics. However, one of the major disadvantages of the EasyMini is that it only contains a barometer for data collection during flight. This does not provide nearly enough information that we would like for post-flight analysis such as tilt angle, acceleration measurements, and roll rate which can be very helpful to not only take an in-depth look at our flight but also tweak our various other flight computers on the rocket. In previous years, we solved this by using an Altus Metrum TeleMega. In theory, this flight computer has all the sensors and features that we are looking for, however, we have found its reliability to be very questionable. In the 2021-2022 year, we flew two TeleMegas in our rockets. At some point, the first one became unusable as the accelerometer stopped working and would not calibrate. The second TeleMega flew in our competition rocket three times flawlessly, and then at SAC 2022, its barometer completely failed and deployed the drogue parachute a few seconds after motor burnout. In addition to our own issues with the TeleMega, we have also heard from multiple other teams and individuals about their own issues with the flight computer and as a result have stopped using them. Instead, we decided it would be safer to use two simpler, more reliable flight computers and develop a separate avionics system instead which will be discussed later. All of the EasyMinis on Karkinos use 400 mAh batteries specifically made for Altus Metrum flight computers. EasyMinis are very simple flight computers and do not draw lots of current, which makes small batteries perfect for them. While we have not tested the full battery life of an EasyMini on this battery, we have found it to be more than enough for launch operations. On our second flight, Page 38 Fig. 51 Wiring diagram for COTS flight computers our EasyMinis were still beeping 3-4 hours after launch until we were able to pull the rocket out of a tree to turn them off. We have also flown the same batteries without charging them between flights as the voltage on the battery had not decreased much from nominal. Our Featherweight GPS trackers also run on 400 mAh batteries. This is the size that comes in stock from Featherweight, and according to their website can last up to 16 hours on one charge. The TeleGPS uses an 800 mAh battery made for AltusMetrum devices. Since the power draw on the TeleGPS is likely higher than the EasyMini due to radio transmissions, it was decided to use the next largest battery size for this device. 2. Commercial Tracking GPS tracking on Karkinos is completed by a Featherweight GPS and AltusMetrum TeleGPS. The deployable payload also has its own Featherweight GPS so that it can be found once it lands. The tracking solutions for Karkinos have changed drastically over the year. Originally, we had planned to use TeleGPSs on our rocket and payload for our primary tracking and a ComSpec radio beacon as a backup. Unfortunately, both of these products became very difficult to obtain. AltusMetrum has had major supply chain issues with many of its products leading to redesigns. While other units with GPS such as the Telemega and Telemetrum came back in stock, there has been no news about new TeleGPSs. Luckily we still have one, but with no way to replace it, we needed other options. We had planned to use ComSpec radio beacons because of their simplicity and robustness. There is no need for a GPS lock or complicated telemetry, it just needs a battery and then it works. The person running ComSpec retired earlier this year and stopped producing trackers and receivers. While there are other radio beacons on the market, such as the LL Electronics XLF series, they are much more expensive and you can instead buy a Featherweight GPS for $10 more. We also attempted a flight with a BigRedBee 70cm GPS unit, however, its large size and difficult user interface meant it was abandoned in favor of the Featherweight GPS. Our multiple test flights with the Featherweight GPS have worked very well and we have received good tracking data throughout the entire flight. 3. SRAD Avionics The goal of the SRAD avionics board is to be able to perform the telemetry and data logging for the flight that a COTS flight computer and GPS can normally perform and is added to the rocket as a redundant flight system for data logging and telemetry. A system overview is pictured below in Fig. 53. Page 39 Fig. 52 Flight summary from our third test flight Fig. 53 System overview of the different parts of the SRAD Avionics Sensor Array A variety of sensors and other components are used to collect data for the rocket’s flight. An Adafruit BMP280 barometer breakout board is used as an altimeter, which gives the altitude by measuring air pressure and converting it into height. The Adafruit BNO055 breakout board is then used as a 9-DOF sensor that is used as an accelerometer and gyroscope and transmits data over the 𝐼2𝐶 protocol. These breakout boards were chosen due to their relative reliability in measuring accurate data, as well as for their ease of integrating into the system, due to the standards and documentation Adafruit provides. In order to receive GPS positional data, a Sparkfun NEO-M9N GPS breakout board is implemented into the flight computer. This breakout board is centered around the u-blox M9 engine GNSS receiver chip. We chose this GPS because with it being based on a NEO-M9N chip, it can be configured to be in Airborne mode, preparing it for the high-velocity situation of a rocket launch, as well as for giving high-accuracy position details. This board has a 1 Hz refresh rate for checking if it has received a full NMEA sentence from its satellites, which is then simplified to UBX data to avoid transmitting NMEA noise over 𝐼2𝐶 to the microcontroller. This packet is then parsed and extracted for data on fixed quality, satellites in use, latitude, longitude, velocity, and time, which is then recorded or transmitted by Page 40 the PSRAM and the radio module. A Teensy 4.1 is used as the main microcontroller for the computer to handle all this sensor input, estimate the rocket’s state, and then feed the data to the radio module for transmission. The Teensy was chosen for the large number of pins it provides that can be used to handle the sensor data, as well as for its processing capabilities as a micro-controller, helping to avoid backlogging and delays. PCB Overview To combine all the sensors with the microcontroller we decided to create a custom PCB, which was done for the benefits it has. A custom PCB simplifies electrical assembly and allows for better organization of breakout boards, as well as reduces the space taken up by physical connections. In addition, a PCB is more reliable than wired connections and minimizes assembly error. An electrical schematic of the final PCB is shown below in Fig. 54. Fig. 54 Schematic of routes and pads for the PCB Data Storage In order to store data collected by the sensors, an 8 MB PSRAM chip is used for short-term storage during the flight, with the data then being offloaded to a Micro SD card for persistent storage. Since writing data directly to the SD card is relatively slow and may be blocking to the sensors, it’s first transmitted to the PSRAM chip where the data is stored in a buffer until the board detects that the rocket has landed or it has been 20 minutes from launch, whichever comes first, at which point it will transmit the data to the SD card through the SDIO protocol where it is stored in a CSV file for retrieval. The rocket’s state is recorded and written to the PSRAM every 100 milliseconds, so as to be often enough to have a good data set, yet not so often that it is storing redundant data or is blocking the execution of the sensors. In order to prevent meaningless pre-launch data from filling up the PSRAM, a rolling buffer is used to hold 10 seconds worth of data, which is then dumped to the PSRAM upon launch. Battery The Tenergy 4.8 V battery pack was selected as the power source for the flight computer due to voltage, runtime, and ease of use. Each of the main components of the flight computer that this battery would power are listed in Table 15 below. This battery pack runs at 4.8 V, which meets the operational requirements for all our components, and has 2 Ah of charge, which will provide around 15 hours of runtime on standby and around 9.5 hours of runtime when in use, which meets the needs of the flight computer. Telemetry For telemetry transmission from the rocket, we selected the TTGO Meshtastic T-Beam V1.1, primarily because it uses a SX1278 LoRa module and has an integrated battery holder. The SX1278 allows the module to transmit in the Ham radio frequency range using the LoRa (Long Range) protocol, which is capable of longer range than other methods Page 41 Table 15 SRAD Avionics Components Current Draw Parts Idle Operating Current (mA) Max Operating Current (mA) Idle Power Consumption (mW) Max Power Consumption (mW) Operating Voltage (V) Teensy 4.1 100 100 500 500 5 BNO055 0.4 1.23 1.44 4.43 3.6 Sparkfun NEO-M9N GPS 31 100 116.6 360 3.6 BMP280 Barometric Pressure Sensor 0.325 7.20 7.2 7.2 3.6 TOTAL 131.9 209.0 while using low power. The radio module itself is running custom firmware that takes in data from the flight computer over a standard Serial connection using a custom packet-based protocol for minimal overhead. It then formats this data for APRS transmission and sends it to the SX1278 over SPI using separate third-party libraries. The module is set to transmit on 433.775 MHz with 20 dBm transmit power, a spreading factor of ten, a 125 kHz signal bandwidth, and a coding rate of 4/5, all of which were chosen to optimize transmission range. The custom firmware is also capable of quickly changing frequencies or other settings in case of a conflict with other teams. The radio module is powered using a 18650 3700 mAh Lithium-ion battery. Based on an average power draw of -55 mA measured by the AXP192 power management chip on the radio module, the radio should have over 67 hours of battery life. Data from the radio module onboard the rocket is first received by a separate TTGO Meshtastic T-Beam V1.1 on the ground. We chose this board for the ground station receiver because, since it is the same as the transmitter, we were already familiar with it. This board is connected to a 70 cm Yagi antenna to improve signal range and reliability over the included antenna for longer distances. Once it receives a transmission, the module then deconstructs the APRS message using the same third-party library used for encoding, and the data is sent to a connected computer over USB Serial using the same custom packet-based protocol used by the onboard radio module and the flight computer. A custom ground station application programmed in Node.js and running on the computer is connected to the receiver’s Serial port and saves the received data to a CSV file on the computer. It then displays the data in a GUI that includes live altitude, speed, heading, stage, and position data, along with a live graph of altitude and speed, and a map with a marker that represents the rocket’s current position. If there is an issue with the GUI, the program also has a text-based output for the data, and the radio displays transmissions on the attached OLED. Test Flight Results The entire SRAD avionics system has been test flown twice, with the second test flight being a successful one where the system was proved to work. During this test flight, radio transmissions stayed in contact throughout the duration of the flight, effectively range testing the radio up to 9,800 ft, with the distance between the radio transmitter and ground station at its maximum at apogee (9,570 ft in altitude and 2090 ft along the ground). The onboard sensor array also yielded useful data for characterizing the rocket’s flight. As seen by the plot of altitude and vertical acceleration versus time in Fig. 55, key moments can be identified as to represent stages in the rocket’s flight, which is then leveraged by the onboard sensor-detection code. The initial peak in acceleration represents the transition from powered ascent to coasting, the zero acceleration occurs at apogee, and the spike in acceleration and change in velocity indicates main deployment. 4. Sled and Mounting All electronics are mounted to a 3D-printed PETG sled. This sled has guides for the two threaded rods to keep it stationary within the electronics bay. The flight computers are mounted with brass standoffs to the sled and batteries have printed enclosures to keep them from moving. The screw switches are mounted to the sides of the sled on shelves and also have their own standoffs. The switches are placed so that they are close to the coupler wall and they are easy to turn on once the rocket is on the pad. PETG was chosen for the sled as it is more temperature resistant than PLA and we have seen PETG work well for rockets in high-temperature environments. In order to fit both the commercial and Page 42 Fig. 55 Plot of SRAD board recorded altitude and acceleration during a Karkinos test flight custom avionics within the electronics bay, the sled had to be made in two parts in order to fit on our 3d printer. Since the sled is not a structural component of the rocket, its strength is not critical. One of the major difficulties with this sled is spacing the various radios properly. During the first iteration of the sled, the SRAD avionics antenna and BigRedBee antenna were basically on top of each other on opposite sides of the sled. Since they are both on the 70 cm band, major interference meant we did not get good telemetry during the flight. This was fixed by moving the devices around and they are not on completely opposite sides of the sled pointing in opposite directions, which has worked well during our testing. 5. Avionics Testing While avionics were primarily tested via flight tests, the EasyMinis were also tested on the ground to verify functionality after potential damage. During the third flight of Karkinos, the rocket landed in a tree and could not be recovered by the team from the ground. A tree climber needed to be hired to recover it which took about a week from the launch. During this time there were some rain showers, so we wanted to confirm that the flight computers still worked properly. Both EasyMinis were turned on and placed in a vacuum chamber. The vacuum was then turned on, held for five seconds, and released. We then confirmed that the "flights" on the altimeter recorded a pressure change and "fired" the charges. E. Air Brake Karkinos comes equipped with an Air Brake module to increase its performance to reach a desired apogee. The module is a drag-modulating closed-feedback system to vary the drag of Karkinos to allow it to reject the many uncertainties that cannot be accounted for in simulation such as motor thrust uncertainties, tilt angle off of the pad, and wind conditions. This project was started during the 2021-2022 academic year and has continued up until the 2023 SAC. Much of the development was discussed in the 2022 Terrapin Rocket Team Technical Report, "Team 9 Project Technical Presentation to the 2022 Spaceport America Cup". The only items discussed in this report are the current year’s development and analysis as well as items deemed critical to the understanding of the system. 1. Overview The Air Brake is designed as a separate module in order to minimize the complexity while integrating it with the rest of the rocket. Taking inspiration from traditional Ebays the Air Brake’s casing is a 10" long 6" airframe tube and a 22" in length 6" coupler. This design was chosen over the lighter design of building the Air Brake into the frame of the rocket for two main reasons. First, due to its complexity, there was a significant chance that part or all of the Air Brake Page 43 Fig. 56 Air brake module Fig. 57 External CAD of the Air Brake module. Fig. 58 Internal CAD of the Air Brake module. would not be certified as flight ready. In this scenario, it was important to have a contingency plan to allow Karkinos to still compete at the 2023 SAC without Air Brakes. Since it is designed to be modular a boilerplate with the appropriate fairings wou