Deployable mechanism for CubeSat antennas

Deployable Mechanism for CubeSat antennas Degree Thesis submitted to the Faculty of the Escola T`ecnica d’Enginyeria de Telecomunicaci´o de Barcelona Universitat Polit`ecnica de Catalunya by Alejandro Garc´ıa Morilla In partial fulfillment of the requirements for the degree in TELECOMMUNICATIONS SYSTEMS ENGINEERING Advisor: Prof. Adriano Jos´e Camps Carmona Barcelona, June 2021 Acknowledgements First of all I would like to express my special thanks of gratitude to my advisor, Professor Adriano Camps, who gave me the opportunity to do this project and also to open the doors of the exciting world of CubeSats and Space. Also I would like to thank the whole NanoSat Lab team who helped me whenever I needed, specially to Lara Fernandez who gave me a lot of advises and Albert Morea, who started his thesis simultaneously, and with whom I shared really great moments. Finally I would like to dedicate some words to my sister, my grandmother and specially my parents, who always believed in me and supported me. You are my example and real inspiration. i Contents List of Figures iii List of Tables iv 1 Introduction 1 1.1 GanttDiagram ................................. 3 2 State of the art of the technology: 4 2.1 3Cat-4 ...................................... 4 2.2 RainCube .................................... 4 2.3 FalconSat-7 ................................... 5 3 Previous model 6 4 Model development 9 4.1 Firstiteration.................................. 9 4.1.1 Telescopic rod support . . . . . . . . . . . . . . . . . . . . . . . . . 9 4.1.2 Basesupport .............................. 10 4.1.3 Modelassembly............................. 11 4.2 Seconditeration................................. 12 4.3 Thirditeration ................................. 14 5 Final model 16 5.1 Liftsystem ................................... 16 5.2 Telescopic deployment redesign . . . . . . . . . . . . . . . . . . . . . . . . 18 5.3 Toothedbeltreels................................ 20 5.4 Lift system locking mechanism . . . . . . . . . . . . . . . . . . . . . . . . . 20 5.5 Electronics.................................... 21 5.5.1 Dyneema lines burning . . . . . . . . . . . . . . . . . . . . . . . . . 21 5.5.2 DCmotorscontrol ........................... 22 5.6 Software..................................... 24 5.6.1 Initialization and dyneema lines cutting . . . . . . . . . . . . . . . 24 5.6.2 Telescopic deployment . . . . . . . . . . . . . . . . . . . . . . . . . 24 5.7 Modelassembly................................. 26 6 Tests and results 27 7 Budget 29 8 Conclusions 30 9 Future Work 32 References 33 Appendices 35 ii List of Figures 1 CubeSatStandard ............................... 1 2 FSSCat dual-frequency L-band patch array . . . . . . . . . . . . . . . . . . 2 3 Fresnelzones .................................. 3 4 Fresnel Zone Plate model prove . . . . . . . . . . . . . . . . . . . . . . . . 3 53Cat-4NADSdeployment ........................... 4 6 RainCube KaRPDA deployment . . . . . . . . . . . . . . . . . . . . . . . . 5 7 Deployed Peregrine payload . . . . . . . . . . . . . . . . . . . . . . . . . . 5 8 Firstmodel ................................... 6 9 Configurations of the structure . . . . . . . . . . . . . . . . . . . . . . . . . 6 10 Previous model behaviour . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 11 Three main pieces from the previous model . . . . . . . . . . . . . . . . . . 8 12 Carbon fiber telescopic rod . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 13 Carbon fiber rod support . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 14 Carbon fiber rod support break . . . . . . . . . . . . . . . . . . . . . . . . 10 15 First iteration model assembled . . . . . . . . . . . . . . . . . . . . . . . . 11 16 Previousandnewmotor ............................ 12 17 Seconditerationmodel............................. 13 18 Attachment of the motor to the base support . . . . . . . . . . . . . . . . . 14 19 Final model of third iteration . . . . . . . . . . . . . . . . . . . . . . . . . 15 20 Isispace CubeSats launcher . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 21 Onespringmodel................................ 17 22 3UCubeSatframemodel............................ 18 23 Basebreaksystem ............................... 18 24 Final design of base supports . . . . . . . . . . . . . . . . . . . . . . . . . 19 25 Deployed angle calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 26 Toothedbeltreel ................................ 20 27 Kanthal power supply system . . . . . . . . . . . . . . . . . . . . . . . . . 21 28 Lift system locking PCB . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 29 Switch for toothed belts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 30 DC motors control development . . . . . . . . . . . . . . . . . . . . . . . . 23 31 DCmotorscontrolPCB ............................ 24 32 Dyneemalinescut ............................... 24 33 Telescopicdeployment ............................. 25 34 Interruptions management . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 35 Finaldesignassembly.............................. 26 36 Lateralview................................... 26 37 Topview..................................... 26 38 6U CubeSat dimensions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 39 Liftsystemtest................................. 27 40 Telescopic deployment test . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 41 TunaCan .................................... 31 42 Project’sGanttdiagram ............................ 35 iii Listings 1 Initialization and dyneema lines cutting . . . . . . . . . . . . . . . . . . . . 36 2 Driver pins initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 3 Telescopicdeployment ............................. 36 4 Stop telescopic deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 List of Tables 1 Radius of the Fresnel zones . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2 Budget for deployment system prototype . . . . . . . . . . . . . . . . . . . 29 3 Costoflicenses ................................. 29 4 Personnelsalary................................. 29 iv Abbreviations AOCS Attitude orbit control system CAD Computer-aided design EO Earth Observation ESA European Space Agency GNSS-R Global Navigation Satellite Systems Reflectometry KaRPDA Ka-band Radar Parabolic Deployable Antenna NADS Nadir Antenna and Deployment Subsystem NASA National Aeronautics and Space Administration PLA Polylactic acid RTOS Real time operating system SMAP Soil Moisture Active Passive SMOS Soil Moisture and Ocean Salinity satellite UPC Universitat Polit`ecnica de Catalunya v Abstract Nowadays, nanosatellites are gaining momentum in the space sector thanks to their develop and launch reduced costs, attracting the attention of many universities and companies. The NanoSat Lab, located in the Universitat Polit`ecnica de Catalunya, is focused on the design and development of nanosatellite missions as well as developing and integrating subsystems and payloads for Earth Observation and communications. This type of spacecrafts have quite limited dimensions and shape, forcing companies to invest in new technology focused in miniaturization. These challenges include the integration of antennas for communications and Earth Observations applications. This thesis is part of a NanoSat Lab project in which the main goal is to develop an antenna that enhances the features of the current solutions. In particular this work is in charge of the development of a modular deploying mechanism that could be used in the future for different types of stowable antennas. vi Revision history and approval record Revision Date Purpose 0 24/04/2021 Document creation 1 16/06/2021 Document revision DOCUMENT DISTRIBUTION LIST Name e-mail Alejandro Garc´ıa Morilla [email protected] Adriano Jos´e Camps Carmona [email protected]c.edu Written by: Reviewed and approved by: Date 13/06/2021 Date 16/06/2021 Name Alejandro Garc´ıa Morilla Name Adriano Jos´e Camps Carmona Position Project Author Position Project Advisor vii 1 Introduction During the last years, Earth Observation (EO) satellite missions have gained interest. Their periodic orbits and large coverage make them very useful to obtain continuous data which is crucial for natural disasters prediction and prevention. Nowadays, most of these payloads are based on three techniques. Some use hyperspectral cameras that obtain images from the Earth such as the Sentinel-2 [1]. Other satellites use Microwave Radiometry or Global Navigation Satellite System - Reflectometry (GNSS-R) which have multiple applications for ocean monitoring, ice thickness, flooding, soil moisture and detection of water ponds as is the case of Soil Moisture and Ocean Salinity satellite (SMOS) [2]. Other spacecrafts use radars such as the Sentinel-1 Synthetic Aperture Radar [3], or the Soil Moisture Active Passive (SMAP) [4] that carries also a microwave radiometer. Although, these satellites have considerable dimensions, the current trend in this industry is the miniaturization of the payloads and the whole spacecraft as much as possible. The appearance of the CubeSat [5] form factor standardized the satellites envelope and weight, and has played an important role in the miniaturization process. These satellites are described by Units (U), so a 1U satellite is 10 cm x 10 cm x 10 cm, with a weight of approximately 1.3 kg. Depending on the number of units that are used, there are different types of CubeSats. There are some other structures like the 16U, and up to 27U, but they are not as common as the ones presented in the Figure 1. With CubeSats, opportunities of faster and more cost-effective development and launch have appeared. Figure 1: CubeSat Standard [6]. One of the main requirements of these payloads is to have a high spatial resolution, since the resolution is too low, the area covered is large, and the small features cannot be identified. In the case of radiometer payloads, this spatial resolution is directly related to the directivity of the antenna in the following way: the higher the directivity, the smaller the beamwidth, and the better the resolution. The antenna directivity is proportional to the effective area, and this one is proportional to the antenna dimensions. For example for a reflector antenna, the directivity increases with the square of the diameter of the reflector, or for a Yagi antenna it increases with the number of passive elements. Thus, there is a trade off between the directivity of the antenna (i.e. having a large antenna) and space available inside a CubeSat, making it difficult to allocate them. Most of the current solutions for these Earth observation CubeSats are based on arrays of patch antennas that are normally located in the outer faces of the satellite, but the number of elements of the array is limited, and therefore the payload performance is also limited. An example mission is FSSCat [7]. It is a mission composed of two 6U CubeSats flying together for the monitoring of polar ice, and the measurement of soil moisture. The first 1 (a) Base. (b) Telescopic rod support. (c) Base support. Figure 11: Three main pieces from the previous model. Despite the good performance that this model achieved, there were some limitations that must be solved. First of all, the final structure did not fit in a 10 cm x 10 cm space since one of the telescopic rod supports sticks out from the satellite frame. The second and most important limitation are the deployed dimensions of the rods. The used car antennas had a deployed length of 800 mm, but that’s not enough since the Fresnel Zone Plate model needs at least a mast length of 1750 mm. Last, but not least, there is the consumption limitation. This model has a consumption during the deployment around 12 W which is a good achievement, but taking into account that the power budget of a 6U CubeSat is around 7 W, is still high. These features and some others that have been found during the new model development are the ones that must be improved. The idea of the telescopic rods seemed to have a lot of potential because of the few space that they occupy when they are stowed and the dimensions that they can reach when they are deployed. For that reason three carbon fiber telescopic rods were purchased. They have a outer diameter of 25 mm, a folded length of 170 mm and a unfolded length of 1900 mm. Figure 12: Carbon fiber telescopic rod. 8 4 Model development During the model design, three different iterations have been exploded before arriving to the final model. The results of the first and second iterations were not the ones expected, but the third one gave interesting results, and some of its features were used in the final model design. All this development process, and the characteristics of each iteration are going to be explained in more detail in this section. 4.1 First iteration The first iteration was mainly based on adapting the previously developed model to the new carbon fiber rods. Before it, all the previous Sketchup designs were converted to SolidWorks. The first one is a good software to kick off with 3D design, but it is a bit limited, so starting the new designs with SolidWorks, which is a much more professional software, will be beneficial for the future. From that point it was decided to only modify the pieces that were strictly necessary to be able to use the carbon fiber rods instead of the car antennas. The first one to be modified were the telescopic rod supports which hold the car antennas. These telescopic rods supports were also fixed by their axis of rotation to the base supports, so these ones must also be modified. 4.1.1 Telescopic rod support The new telescopic rods supports have the same behaviour than the ones used for the car antennas, but they have been completely redesigned taking into account the characteristics of the new telescopic rods which have a considerable larger dimensions. The first difference is how these supports hold the carbon rods. Instead of having a completely closed hole as before, the hole is vertically cut into two halves and has a diameter exactly the same as the carbon rods (25 mm) so that they are perfectly adjusted to the hole but they can be also easily extracted. These cuts also helped with the occupied space, since the available space between the base support and the face of the CubeSat is a little more than 25 mm in the worst case so there is only space for the carbon rod itself. The axis of rotation of this piece has been completely modified in order to implement a break to stop the rotation movement when it arrives to its final position. This new system is explained with a bit more detail later. The last modification is the way how the toothed belt enters to the telescopic rods. As it can be seen in Figure 13b, in the lowest part of the rod support there is oval hole through which enters this toothed belt. The oval shape is to avoid that the wire bends too much when the rod support rotates. 9 (a) Lateral view. (b) Bottom view. (c) Front view. Figure 13: Carbon fiber rod support. 4.1.2 Base support The base supports are known as the part that holds the DC motors, and also the telescopic rod supports. The only modified part was the one where the rod supports are attached in order to adapt it to the new design, while the part of the DC motor remained the same. The most significant modification came with the introduction of a break. For a satellite antenna it’s only important the deployment part since it is not going to be stowed again. Taking advantage of it, a break for the angular movement was designed. Every rod support has a flange on its axis of rotation. Once each carbon rod support has rotated to its final position, a little piece attached to the base support is pushed with a spring and it blocks the carbon fiber rod support avoiding it to return to its initial position. In Figure 14 it could be understood better. This system was designed taking into account the modularity that the system should have to be adapted to the deployed antenna. Only modifying the position of this flange, it’s possible to modify the final angle that the three masts will have. Moreover, each branch could have different deployed angles depending on the flange of its rotational axis. With Figure 14, this behaviour can be better understood. (a) Stowed break position (b) Deployed break position Figure 14: Carbon fiber rod support break. 10 4.1.3 Model assembly In the previous model, at the top of the three base supports, there was a cover that made sure that the DC motor did not move. It also gave rigidity to the structure since it was screwed to the three base supports. This cover has been modified to be able to attach at the top the antenna feeder. Basically, it was given a 30 mm of height to ensure that the antenna connector has enough space. In one lateral of the piece there is a hole through which the coaxial cable passes. In one lateral of the base support, it can also be seen the gear and the pulley that push the toothed belt inside the rod support. Figure 15: First iteration model assembled. The final design accomplished part of the main goals since it has the deployed dimensions needed and it occupies less than 10 cm x 10 cm when folded. The previous car antennas were deployed with a toothed belt of 1 meter of length. The new telescopic rods have an unfolded length of 2,9 m, so this toothed belts could not be used. A longer nylon wire with a similar rigidity was started to be used instead of the toothed belt. Although this new wire was completely flat, the intention was to pull it with the same gear used for the previous one. To solve the diameter differences, the gear was slightly redesigned making its diameter 0,5 mm larger. At that moment, a deployment test was performed, but the results were not the appropriate ones, and some problems were encountered. First, only the deployment of one of the telescopic rods was attempted. The first problem encountered was that the nylon wire was not performing well since some times and specially at the final part of the deployment it skidded, and it did not push the telescopic rod properly. Even so, there was a second test trying to deploy all the telescopic rods and a more important problem appeared. The DC motor did not have enough strength to deploy the three telescopic rods at the same time, and the current consumption was excessive for a CubeSat. Besides the motor transfers its rotational movement to the gears with an endless screw and there was a lot of friction that turns into an important loss of power. At that moment it was decided to change radically the design and explode another solutions. 11 4.2 Second iteration For the second iteration a radical change of the design was executed. Three new DC motors were purchased with some interesting characteristics. As the other one, they have an operating voltage of 12 V but in this case the power consumption is 10 times lower since they have a reducer that makes the torque much larger and reduces the rotational velocity to 18 rpm. This low rotational velocity also benefits in our application where the deployment must be smooth and controlled and there is no need to be quick. (a) Old motor. (b) New motor front view. (c) New motor lateral view. Figure 16: Previous and new motor. In order to continue using the nylon wire and do not have problems with the friction of the wire and the gear, it was intended to push it from the end instead of pull it. In this iteration only one of the new motors is used. In this case the motor axis is attached to three reels where the nylon cables should be coiled and with the rotation of the motor axis, the reels also rotate. The covers of the three reels are fixed to the base so they do not rotate. The wire goes out of the reels by a little aperture that is in the reel’s cover. With this idea it is possible to eliminate all the gears of the system. The motor and the reels are located under the base and at the top of it there are located the carbon fiber rods supports and the base supports. These ones and their attachment to the base are practically the same than in the previous iteration. When the wires leave the reels, they cross the base each one through a different hole and after that they enter into the carbon rods by the bottom part of the telescopic rods supports. The linear actuator is in the same position as in the previous iteration and it has the same behaviour. 12 Figure 17: Second iteration model After having this design, it was decided to print the base, one reel and one cover, and it was tested how strong the reel could push the wire to verify if it was able to deploy one of the carbon fiber antennas. The result was not the expected one because the strength that the motor was doing was too high and, therefore, the current consumption was higher than expected for having only one reel. On top of that, the movements that the nylon wire had to do between the reel and the carbon fiber rod support were a future cause of problems. At that point it was decided to look for another solution using the same motors. 13 4.3 Third iteration After the bad results obtained in the previous iteration, it was decided to change completely the design taking profit of the new motors. The new model is going to use three DC motors, one for each telescopic rod, and it will return to the toothed belt system which avoids the skidding problem. The idea is to attach a motor directly to every base support with the requirement that the axis of rotation of the motor is located under the telescopic rod support. The gear that pulls the toothed belt should be directly attached to the motor axis avoiding the friction between gears that there was in first iteration. The toothed belt is compressed between the gear and a pulley making sure that it does not move when the motor is stopped. Using three motors has some important benefits. First of all, the deployment of the branches will be sequential instead of deploying all at the same time. The idea is to start deploying one branch for a few seconds, after that it will stop and the second branch will deploy during the same period of time. When the third one has completed its period, the cycle will start again until the deployment of the three branches is completed. In this way, each motor will have to push only one branch so the power consumption peak will be considerably reduced, which is an important requirement taking into account the CubeSat power budget. That is not the only benefit of using three motors. For example, there are some antennas, like the reflectors, that may need an angle different of 90ºrespect to the feeder, and with this solution every mast could have different lengths. The conclusion is that three motors increase a lot the modularity of the system. To implement the idea, the most challenging part was the space limitation of 10 cm x 10 cm. The motors have a width of 25 mm and the base supports 14 mm. These dimensions made not feasible the attachment of the motors directly to one face of the base support. The solution to the problem is the following one. 12 mm of width of the base support were extruded to gain space for the motor and this one was positioned with the cylindrical part downward in order to not interfere the carbon rod support movement. This part could be better understood with a CAD representation. Figure 18: Attachment of the motor to the base support. 14 As it can be seen in Figure 18, the rod support break has been relocated. In previous iterations it was located under the axis of rotation of the rod support. Now it has been placed in a lateral to minimize the distance between the gear and the rod support aiming to avoid that the toothed belt collapses before entering to the rod support. Finally a little support for the feeder was designed. Basically it is a little peace of 8 mm of height with a big squared hole in the middle for the antenna connector. It has three holes that each of them coincide with one base support to screw this piece and fix it. Now there is no need for a lateral hole for the coaxial cable since there is no the DC motor under the feeder as in iteration 1 (4.1) and the antenna cable can go directly from the feeder to the base. Figure 19: Final model of third iteration. After having the full assembly and verify that all the pieces fit properly, a 3D model of the base and one motor support were printed to test the deployment of one branch. The results were satisfactory since the telescopic rod was deployed perfectly. The current consumption of the motor was not more than 200 mA in the worst case, which is a power consumption of 2.4 W taking into account that the motor works at 12 V. After that, the parts needed for the other two masts were printed to verify that they can fit properly in the available space. At that point it was the moment to start working on the wires placement. It was expected to have three belts of 1,9 meters each one, so they must be rolled up somewhere to ensure a good deployment. The only place where they could be is under the base, and when the structure is inside the CubeSat, there is no much space there because the motors occupy big part of this space. At that moment we came up with a new idea that lead to the final model. 15 5 Final model After analysing the previous iteration and the result of the telescopic deployment test, it was concluded that having three DC motors have lots of benefits, and seemed to be possible to fit them in the available space, so this was a good line to continue working on. During the deployment, before start deploying the telescopic rods, the system should be lifted up some centimeters. Up to that moment the responsible of this was the linear actuator. The linear actuator was attached to the CubeSat frame and also to the base, and with a stepper motor and a lead screw, it pulled up all the deploying mechanism. This performed well, but there were some drawbacks that have not been considered up to that moment. This system occupied a lot of space because it could not be put nothing above neither below it. Moreover, this system was responsible of an upwards movement that will be done once in the life of the satellite so it is not worth having such a complex system. At that moment it was decided to start analysing if there could be found any other possible solution to substitute the linear actuator. Looking for ideas with other CubeSat missions, it was found the FalconSat-7 [16]. As explained in the State of the art [2.3], the purpose and dimensions of this deploying system was not the same that the ones we are aiming for, but the deployment idea has lots of similarities with this thesis concept. After analysing the system in detail it was noticed that the external part of the payload goes out form the satellite with some impulsion of a spring or similar system and without using any motor. While looking for a similar solution, the idea of the CubeSat launcher came to us. These ones are standardized boxes where CubeSats are located during the launch. When it is time to deploy them, the door opens and a spring pushes them out of the box. This mechanism made us think that a similar system could be implemented, but adding a break which ensures that when the base arrives to its final position, it blocked, and does not go back. Figure 20: Isispace CubeSats launcher [19]. 5.1 Lift system The first idea was to include one big spring below the base, similarly to how CubeSat launchers are implemented. This spring should be compressed and fixed with some mech16 anism during the launch storing an elastic potential energy and when it is time to deploy, the mechanism should unblock it and the potential energy must become kinetic energy which should be enough to lift up the base and all the deployment system. In Figure 21 there could be seen that there is one tab at each corner of the base. Inside of them there is a little spring similar to the ones used in pens. The idea was that, when the system is stowed these tabs are compressed, and during the deployment, when the spring pushes and the base arrives to its final position, these tabs should go out and fit in a hole of the CubeSat frame. With this it is intended to avoid that the base moves back to its initial position or starts going up and down with the spring inertia. Figure 21: One spring model. This idea seemed to have a lot of potential, since it leaves a lot of free space on the top of the base and it also leaves free space in the spring inner part. It also reduces the complexity of using a motor, and last, but not least it could help to reduce a lot the weight of the final structure which is also an important requirement in CubeSat standard. In this design, the spring has a diameter of 95 mm and a length of 200 mm but with a wire width of 1 mm. This type of springs are not easy to be found, and they need to be manufactured. Finally we arrived to the final solution which is based in the same idea. The final implementation, instead of one spring, uses four but with different dimensions than in the previous case. Now they are located in the corners of the satellite structure, which until now were quite wasted. These 300 mm long and 12 mm diameter springs have inside a inner 8 mm diameter aluminum circular guide that is part of the satellite frame. Since these guides go from the low part to the top of the satellite structure, they give the frame a lot of rigidity, giving the possibility of considerably simplifying the side walls of the structure and thus freeing up space for the entire deployment mechanism. These guides are also used for the base guiding, which has four linear bearings through which the guides pass, and they make sure that the base doesn’t move in the lateral directions. The prototype frame of the CubeSat is made of four lateral profiles with L shape and a length of 300 mm. They are fixed in the lower part by a 3D printed base and in the top part by four 3D printed flat pieces that are screwed in the lateral. Next to the lateral profiles there are the already mentioned circular guides that also have a length of 300 mm and are also attached to the top and low part of the satellite structure. 17 (a) PCB top layer. (b) PCB bottom layer. (c) Manufactured PCB. Figure 31: DC motors control PCB. 5.6 Software The STM32 microcontrollers are programmed with C. To do so there is a development software called STM32CubeIDE that simplifies some of the tasks, specially the configuration of the GPIO pins that are going to be used. All the code is located in the main file and there are three differiented parts. The first one is the GPIOs configuration and is generated automatically by the software. Then, it comes the part for the unlocking of the lift system and finally the telescopic rods deployment. 5.6.1 Initialization and dyneema lines cutting The first part of the code is where all the GPIO pins are configured as inputs or outputs and other characteristics depending on their functions. There are three inputs interruptions with rising edge for the switches of the toothed belts, and eight GPIO outputs, three for each motor driver and two of them for each of the MOSFETS. After that, the microcontroller starts with the dyneema lines cut. First of all it sets to 0 the GPIO connected to the first MOSFET to activate it and supply the first Kanthal. After four seconds it sets it another time to 1 to stop the current flow. Then the same is done with the other MOSFET to lift up the structure. After that, the thread task that controls the motors is initialized. It has been tested the amount of time needed to cut the dyneema lines and with four seconds its enough to ensure that they will be cut. Figure 32: Dyneema lines cut. 5.6.2 Telescopic deployment The second part of the code is related to the second part of the deployment, which is the telescopic unfold of the carbon fiber rods. This part of the code is based in freeRTOS, a real-time operating system (RTOS) for microcontrollers and small microprocessors. 24 Basically there are two parts, the part that controls the motor drivers and therefore the DC motors and the part that manages the interruptions of the toothed belt switches to stop the deployment of each branch when necessary. This second task has higher priority than the first one so when an interruption occurs, the microcontroller stops executing the default task and it makes the changes related to that interruption. After that it continues with the default task execution. This software part, firstly initializes the GPIO outputs that are connected to the drivers, setting the STBY pin to 1 to initialize the drivers and the two input control signals for each motor to 0 to ensure that the motors stay still. After the initialization, the code starts with the telescopic deployment. First it checks if the toothed belt of the first motor has arrived to its end with the verification of the variable that later will be explained in a bit more detail. If the deployment has not arrived to its end, it sets to 1 the first input of corresponding driver and to 0 the second input of the same motor to start its forward movement. After 3 s it returns the two inputs to 0 to stop the motor. After that, the same procedure is applied for motors 2 and 3, respectively. Then, it checks if the deployment of the 3 branches has finished. If not, it starts the loop another time with the deployment of the first branch, otherwise, it finishes the deployment finishing the default task. Figure 33: Telescopic deployment. Finally there is the interruptions part. When an interruption with higher priority than the task that is on course occurs, the system stops that task to attend the interruption. In this case there are three equal interruptions. Each switch system is connected to a GPIO input of the NUCLEO board that will generate the interrupt in case that there is a rising edge. When this happens, the code checks which switch has generated the interruption and it immediately stops the motor setting to zero the two inputs of the driver and sets to true the variable that indicates that the branch should not be deployed any more. This variable called finish Mx is the one that is continuously checked in the default task before deploying each branch. Figure 34: Interruptions management. 25 5.7 Model assembly Here is presented the final model assembly with all the parts. Actually, this assembly was created at the moment when the final model was started and it has been updated during all the designing process. This helps to identify any incompatibility between pieces and to find out new ideas that could perform well. Figures 35a and 35b show two views of the model when stowed, while Figure 35c shows the model after the lift up movement and without having the telescopic rods deployed. (a) Stowed lateral view. (b) Stowed low view. (c) Deployed lateral view. Figure 35: Final design assembly. As it could be seen in Figure 35a, the carbon fiber rods stick out from the CubeSat frame and therefore from the standard length of 300 mm that a 3U CubeSat must have as it was told in the introduction. But that’s not true at all, because in the launcher there is left a distance between 1U CubeSats to ensure that they do not touch each other. This small distances can be used by bigger satellites so a 3U CubeSat has a usable length of 350 mm approximately. In this model, the stowed distance from the end of the carbon fiber rods to the lower part of the frame is 330 mm, so there is still 20 mm of margin to be used. Figure 36: Lateral view. Figure 37: Top view. Figure 38: 6U CubeSat dimensions [21]. 26 6 Tests and results During the model development, several tests have been done. Some of them to test the compatibility of the different 3D printed pieces and others directly to test the deployment. Finally, the final model pieces were all printed and assembled and it was time to test the performance of the whole system. Note that the toothed belts that are used are the ones of the car antennas which have a length of one meter. During all the model development some other toothed belts of two meters were found, but the cheapest ones that fit better with this model dimensions had a price of more than 60 €each. At that point it was decided to test the deployments with the 1 m toothed belt since there is no difference in the behaviour with a two meters one and when the model is verified with the 1 m toothed belt it will be the moment to purchase the ones of 2 m. In the belt reels there is enough space to roll up a 2 m toothed belt, so there will not be much difference between them apart form the deployed dimensions and cloth dimensions. First the lifting up mechanism was tested. Basically all the parts were assembled and the dyneema lines were tied to the base and lower part of the frame. They were left in tension for one day to ensure that they do not break. After that it was time to test the burning system with the NUCLEO Board and ensure that the base stays fixed at the top part of the frame with the neodymium magnets once the dyneema lines are cut. Notice that the rotational springs that must ensure that the rods supports rotate to their final position are not in the structure. They were purchased a month ago but they have still not arrived so this movement was done with our help. (a) Completely stowed. (b) Ready for the telescopic deployment. Figure 39: Lift system test. Note that in the previous figures there is not the cloth. That is because the cloth that was being used was not the final one and it occupied a lot of space when the system was stowed. It was decided to first test the lift mechanism without the cloth and then, attach the triangular cloth to each of the ends of the telescopic rods and start with the telescopic deployment test. This one ended up with a satisfactory result since it reached the desired 27 dimensions and with a current consumption of the DC motors of 350 mA as the worst case, which is a reasonable value for a CubeSat. (a) Ready for telescopic deployment. (b) Completely deployed. Figure 40: Telescopic deployment test. 28 7 Budget This section contains the approximate cost of the project including the pricing of the components prototyped, used and the cost of the engineering process. Item Quantity Price/unit (e) Total (e) 3D printer PLA 1 20,00 20,00 Carbon fiber rods 9 17,25 155,25 Aluminium L profile 1 14,00 14,00 Aluminium 8mm guides 4 3,16 12,64 Dyneema reel 1 14,99 14,99 Nodymium magnets 12 0.22 2,64 DC motors 3 3,76 11,28 Motor drivers 2 4,45 8,90 NUCLEO Board 1 14,97 14,97 Dyneema reel 1 14,99 14,99 Linear bearings 4 0,92 3,68 Total 273,34 Table 2: Budget for deployment system prototype Program License Type Price (e) Solid Works University 6.600,00 Total 6.600,00 Table 3: Cost of licenses Personnel Hours Price/hour (e) Net salary IRPF+SS Gross salary Engineer 700 8,00 5.600,00 6%+30% 7.616,00 Lab Technician 40 8,00 320,00 6%+30% 435,20 Total 8.051,20 Table 4: Personnel salary 29 8 Conclusions It would be important to remark how difficult is to allocate any innovative system in the dimensions of a CubeSat. When thinking in the electronic part, these components have achieved a miniaturization level that make easier this work, but when its time to allocate new mechanical components, this work is much more difficult since they are not normally standardized for these applications and their dimensions are considerably larger. Almost all parts of this model have been made using the PLA 3D printer. This type of printing is very useful as it has considerable precision, a speed to make complex pieces of a few hours and with a final material that has considerable stiffness and hardness. At the beginning of the project, it was intended to take as much advantage as possible from the previous model, but comparing this previous model with the final one of this thesis it could be noted that all the parts have been redesigned from scratch. On the top of that, during the development process, several iterations were performed. The ideas developed in the some of them, like the second one, were completely discarded, but during others, like the first or the third one, important parts were developed and this parts were later used in the final model. In most of this parts there could be found innovative functionalities. 1. The new telescopic rod supports have several features like the new rotation axis or the oval hole through which enters the toothed belt to avoid it bends too much. 2. The break system for the telescopic rod support that ensures that when each rod has rotated, it cannot return to its initial position. 3. Use of three motors to decrease the power consumption an to increase the modularity of the system. 4. Redesign of the base supports to adapt them to the new motors. 5. New lifting system with the four springs and the circular guides as part of the frame of the satellite. 6. Implementation of the dyneema lines to lock the deploying mechanism during the launch, similar to the one used in the NADS. 7. Introduction of the neodymium magnets to fix the base when the first part of the deployment is completed. 8. The toothed belt reels have implemented a tube that guides the belt properly and enables to allocate them parallel to the base taking advantage of the free space. The modularity of the system that has been taken into account gives the opportunity to use it not only for antenna deployments but for other applications. For example, nowadays magnetometers are improving substantially their precision up to nanoteslas. When these magnetometers are used in satellite applications to measure the earth magnetic field, they must be located some meters away from the satellite since all the electronics can cause interferences in the magnetometer measures. Using only one telescopic rod it could be possible to achieve this goal. 30 As it could be seen in Section 6, in the final model there have been left 20 mm of margin to be able to fold the cloth, but in case that that’s not enough, some satellites dispensers gave the opportunity to use an extra volume located at the top of the satellite structure. This extra volume called Tuna Can fits with the central part of the launcher spring. Figure 41: Tuna Can [21]. During the project development, there has been an approach by some external companies and startups to colaborate and present a proposal to a contest at European level. It is intended to launch a single stage rocket that carries a 6 CubeSat as payload. This satellite could carry this deployment mechanism as a technological demonstration. In the event that the project goes forward, the external financing could help to develop it more comfortably. 31 9 Future Work The first think that must be done in the future is to test the performance of the rotational springs when they arrive. The model has already taken into account how this spring should be attached to the base support and how it should push the carbon rods. The second and one of the most important things is to purchase a toothed belt with a length of 2 meters that could fit with this model. There is a trade-off with the rigidity that it should have, since it should be able to push the carbon rods without collapse but at the same time it should be folded in the reels. For the prototype model, cheap DC motors have been purchased, but they cannot be used for space applications. A good thing to start working in immediately could be to look for DC motors for space applications with similar dimensions and characteristics to those that have been used. This is a delicate job since this type of motors are considerably expensive. Once the deployment system is completely finished, it would be the moment to start with the testing. Lots of deployments must be done to identify any weakness that could be improved. After that it would be a good idea to start with the vibrations tests. This tests simulate the vibrations of the rocket during the launch and they are useful to identify any damage or weakness that the structure could have. Another type of tests are the thermal vacuum chamber tests. This type of the chambers simulate the space vacuum conditions and can perform temperature cycling to simulate the different temperatures to which the satellite is submitted during the orbit. Up to now, all the prototyping has been made with PLA, but it cannot be used in space despite the good performance it has during the development process. Once the deployment is completely verified it would be a good idea to start with the mechanization and manufacturing of some parts in an appropriate material like aluminium. It would be also a good idea to start looking at metal 3D printing. This technique is gaining a momentum in the metallic manufacturing sector since they can produce really complex pieces that cannot be produced with mechanization. In this model, there are some pieces that cannot be mechanized, like the toothed belt reels because of the guiding tube. 32 References [1] ESA. Sentinel-2. https://sentinels.copernicus.eu/web/sentinel/missions/ sentinel-2, 2015. [2] ESA. Soil moisture and ocean salinity satellite (smos). http://www.esa.int/ Applications/Observing_the_Earth/SMOS, 2009. [3] ESA. Sentinel-1. https://sentinels.copernicus.eu/web/sentinel/missions/ sentinel-1, 2015. [4] JPL. Soil moisture active passive (smap). https://smap.jpl.nasa.gov, 2015. [5] CubeSat. Cubesat standard. https://www.cubesat.org. [6] NASA. What are smallsats and cubesats? Available in: https://www.nasa.gov/ content/what-are-smallsats-and-cubesats/. [7] NanoSat Lab. Fsscat - towards federated eo systems. https://nanosatlab.upc. edu/en/missions-and-projects/fsscat, 2020. [8] Lara Fern´andez. Dual-frequency L-band patch array for GNSS-R andMicrowave Radiometers onboard a 6 Unit CubeSat. Master’s thesis, Universitat Polit´ecnica de Catalunya, 2019. [9] Berkley Lab. Fresne zone plate behaviour. Available in: http://zoneplate.lbl. gov/theory. [10] Adri´an M´arquez. Deployable Fresnel Zone Plate antenna for CubeSats - Electrical design. Master’s thesis, Universitat Polit´ecnica de Catalunya, 2021. [11] COMET. Deployable reflector antenna for cubesat missions. https://comet-ingenieria.es/actualidad/ comet-ingenieria-is-looking-for-cubesat-and-small-satellites-operators/, 2020. [12] NanoSat Lab. 3cat-4. https://nanosatlab.upc.edu/en/ missions-and-projects/3cat-4, 2022. [13] NanoSatLab. 3cat-4 render. Available in: https://nanosatlab.upc.edu/en/ missions-and-projects/3cat-4. [14] JPL. Radar in a cubesat (raincube). https://www.jpl.nasa.gov/cubesat/ missions/raincube.php, 2019. [15] NASA. Karpda deployment. Available in: https://directory.eoportal.org/web/ eoportal/satellite-missions/r/raincube. [16] USAFA (United States Air Force Academy). Falconsat-7. https://directory. eoportal.org/web/eoportal/satellite-missions/f/falconsat-7, 2019. [17] USAFA (United States Air Force Academy). Deployed peregrine payload. Available in: https://directory.eoportal.org/web/eoportal/satellite-missions/ f/falconsat-7. 33