A telemetry communication system for UCubeSat

The goal of U CubeSat club at the University of Utah is to launch the University of Utah's first CubeSat into space, and thus inspire future engineering students at the U to develop and launch increasingly capable CubeSats. My contributions to this project was to determine the requirements for maintaining a wireless communication system for the UCubeSat. Achieving this objective required multiple link budget analysis for links in the S band and UHF frequency bands. The RF simulation program CST was used to analyze the behavior of different antennas in the S band and UHF frequency bands. Through the results of these analyses and research, it was determined that the frequency of 436.4 MHz would be the most optimal frequency for the link between the UCubeSat and ground station. A pair of dipoles on the UCubeSat and a Yagi Uda on the ground station will also be able to provide sufficient gain to receive a signal with a bit rate of 20kB/s. These results will be a useful guide to determining the required Commercial Off the Shelf (COTS) hardware for a 1U (10 cubic centimeter) CubeSat.

ABSTRACT The goal of U CubeSat club at the University of Utah is to launch the University of Utah’s first CubeSat into space, and thus inspire future engineering students at the U to develop and launch increasingly capable CubeSats. My contributions to this project was to determine the requirements for maintaining a wireless communication system for the UCubeSat. Achieving this objective required multiple link budget analysis for links in the S band and UHF frequency bands. The RF simulation program CST was used to analyze the behavior of different antennas in the S band and UHF frequency bands. Through the results of these analyses and research, it was determined that the frequency of 436.4 MHz would be the most optimal frequency for the link between the UCubeSat and ground station. A pair of dipoles on the UCubeSat and a Yagi Uda on the ground station will also be able to provide sufficient gain to receive a signal with a bit rate of 20kB/s. These results will be a useful guide to determining the required Commercial Off the Shelf (COTS) hardware for a 1U (10 cubic centimeter) CubeSat. ii TABLE OF CONTENTS ABSTRACT ii INTRODUCTION 1 METHODS 8 RESULTS 11 DISCUSSION 24 REFERENCES 25 APPENDIX A 27 APPENDIX B 30 iii INTRODUCTION Satellite technology, once only a subject of research, has now become vital to modern day life [1]. Many facets of everyday life, such as GPS guidance and weather forecasting have been greatly improved through implementation of satellite technology [2]. Satellites now contribute to many fields of study, such as resource economics, where satellites have allowed for precise measurement of deforestation rates [1][2]. Satellites have made great contributions to mankind’s scientific and technological understanding through obtaining data on celestial bodies and testing new technologies like ion propulsion [1][2]. All satellites, no matter their application, are highly complex and multifaceted devices, requiring many different systems to fulfill their intended objectives [1][2]. One such system is the wireless communication system a satellite uses to relay information back down not earth. The goal of this project is to design a link budget and simulate antennas for a communication system for a satellite to be designed by University of Utah engineering students. The reductions in size and cost of current space-capable hardware are now allowing for small scale satellite design to be within the budgets and skill level of electrical engineering students [3]. This is a considerable development for engineering students, as it opens a plethora of new scientific exploration opportunities that a generation ago was only available to full time engineers working in aerospace companies like NASA [1]. Satellite design and creation also necessitates collaboration between people specialized in many technical fields, such as wireless communications, power distribution, heat dissipation, and structural design [1][2]. Thus, a satellite design project can also provide engineering students with an opportunity to obtain experience in collaborating with people of other disciplines, which can prove to be valuable experience in their future technical careers [1]. One of the most current types of small scale satellites currently being designed is called the CubeSat. CubeSats, as their name suggests are a type of small scale satellites shaped into a cube. The sizes of these satellites are classified into units called U’s, with a 1U satellite being 10 cubic centimeters in volume and weighing from 1 kg to 1.33 kg [3]. Larger CubeSats are usually described in terms of U’s, with 2U, 3U, and 6U CubeSats having 2,3, and 6 times the volume and maximum possible weight of 1U CubeSats [3]. The value of satellite design for engineering students is becoming increasingly apparent, and it is now becoming trendy for universities with accredited electrical engineering programs to design their own satellite using Commercial Off the Shelf (COTS) components [3]. This trend is visible in Fig. 1, which describes the number of small scale satellite launches per year since 1998, including CubeSats [4]. This graph shows a near exponential increase in the number of CubeSats launched since 1998, with many of the launched CubeSats having been designed by university students [3][4]. 1 Fig 1. This graph shows the number of number of CubeSats designed and launched per year since 1999 [4]. This can be seen in Fig. 2, which shows that universities are the second most common type of organization designing and launching CubeSats. Fig 2. This chart divides the total number of CubeSats produced into categories based on the organization that made them, see source [4]. 2 The goal of the UCubeSat club at the University of Utah is to follow this trend by launching the University of Utah’s first CubeSat into space. This CubeSat is going to be 1U in size, and the mission of this satellite will be to broadcast at least one photo of the earth from space to a ground station located at the University of Utah itself. It is hoped this accomplishment will inspire future engineering students to build upon the experiences of 2021-2022 U CubeSat team and develop and launch their own increasingly complex and capable CubeSats. The ability for the CubeSat to communicate with the ground station on the earth is paramount for the successful completion of the mission [3]. Though a CubeSat can be detected and tracked using sophisticated hardware, the small size of the CubeSat makes this method difficult [5]. Detection of radio transmissions from the CubeSat is amongst the most proven means of tracking a CubeSat after launch [3]. Thus, a wireless link is needed to provide a means for the CubeSat to successfully located and tracked as well as fulfilling the goal of transmitting a photograph from space [5]. Determining the optimum value of the communication link parameters, such as antenna gain and required broadcast power, is a fundamental step to ensuring that a wireless link between the CubeSat and earth can be successfully established. One technique that CubeSats often use to communicate is to have two separate wireless links at different frequencies. One link is typically used to receive information from the ground station, and another link to broadcast payload data back down the earth. Though the there are multiple 3u and larger CubeSats that use a two-link system [7], there are also examples of 1U CubeSats similar to the projected UCubeSat that utilize two wireless links [8]. This indicates that using a transceiver with two different frequencies is possible even with the size and weight constraints of 1U CubeSat, and will likely be the approach that the UCubeSat will also employ for its wireless communication needs. Though both links are necessary for the UCubeSat to function, the main focus of this work is going to be on the payload link. This is because the payload link usually has more stringent design requirements due to the much more limited amount of power the CubeSat can provide to its antenna, as well as increased path loss due to higher operating frequency [6]. Thus, if the more difficult payload link can be successfully implemented, then it should also be possible to implement the less stringent link to send data to the CubeSat [6]. As research was invested into designing basic wireless links, it was observed that basic wireless links tend to have a similar procedure The power is produced from a transmitter, which is then sent through an antenna with a certain amount of gain. Gain is a parameter of all antennas and is a measurement of the degree of which the antenna concentrates electromagnetic energy in a specific direction. The higher the gain, the narrower the beam the produced electromagnetic energy travels in. Narrowing this beam increases the concentration of the electromagnetic energy within it, increasing the ability 3 of the electromagnetic energy to overcome losses. This electromagnetic energy radiating from the transmitting antenna experiences path loss attenuation predictable by the Friis transmission equation. This loss represents a minimum loss from transmitter to receiver, with new losses being introduced by physical obstacles in the path of the electromagnetic energy. The electromagnetic energy is then picked up by a receiver antenna, which then converts the electromagnetic energy back into a signal that can be processed by electronic hardware. Fig. 2 is a graphical representation of this process. Red represents the losses the signal encounters, which are subtracted from the power at the transmitter. Green represents the gain of the antennas, which is added to the power at the transmitter. Blue is used to represent the difference between received power and the minimum signal power the receiver needs to detect a signal [9]. The 10 dBm Fig. above the transmitting antenna refers to the power at the transmitting antenna, whilst the 14 dBm above the receiving antenna refers to the power received at this antenna after the signal has encountered path loss. Fig 3. A visual example of propagation of a signal from point to point [9]. With this information in mind, the next step was to begin determination for an exact value for the antenna gains and required transmission power on both the CubeSat and the ground station. It was decided that the parameter that should be determined first was the operating frequency. This is because all the parameters for wireless links, such as the path loss, the gains of the transmitting and receiving antennas, and attenuation due to obstacles are all affected by the choice of operating frequency of the system [9][10]. The frequencies that are used by CubeSats are often located in commercially allocated bands of frequencies, such as VHF, UHF, and S Band. [6][9][10]. The usage of 4 a frequency within a particular band depending on parameters like desired data rate, available licenses, available hardware and CubeSat size. The frequency band first considered for the CubeSat communication link was the S Band. CubeSat’s that use this band tend to have operating frequencies within the range of 2.0 Ghz to 2.3 Ghz. This decision was made because a higher operating frequency allows for a channel to have a wider bandwidth [10]. In the case of the S band, the large bandwidth makes data rates as high as 3.4 Mb/s possible [10], which could potentially allow for a video feed connection between the CubeSat and the ground station. As seen in Fig. 4, the usage of the 2.0 Ghz to 2.3 Ghz range in the S band is that it is also lower than the frequency where atmospheric attenuation sharply rises. Fig 4. This Fig. shows the atmospheric attenuation given an operating frequency. Signals below 3 Ghz generally have neglible atmospheric attenuation. Taken from source [11]. In the case of a signal travelling through the atmosphere, the primary obstacles of the electromagnetic signals are weather related phenomenon such as water droplets and other forms of precipitation. As the wavelength of the transmitting signal approaches the size of precipitation in the atmosphere, the precipitation begins to absorb more of the electromagnetic energy in the signal. However, all frequencies below 3 Ghz tend to have negligible atmospheric attenuation [10][11]. Getting through the atmosphere with higher frequencies will require significant amounts of extra power at the CubeSat antennas, which 5 can be strenuous to engineer considering that the CubeSat can only operate on internal batteries after being deployed into orbit. However, the S band does have a significant disadvantage that limits practicality in a 1U CubeSat application. The main disadvantage is that in the case of the S band, the higher operating frequency necessitates a much larger amount of power to be dissipated within the transceiver [12]. With some exceptions, it seems that S band communication is largely applied in CubeSats larger than 1U, as larger cubesats have afford the increased energy cost due to additional volume for battery storage and heat dissipation [12]. One solution to the issues behind using S band was shifting the CubeSats operating frequency from being in the S band to one in the UHF band, which being much lower in frequency uses considerably less power even when transmitting. Research revealed that UHF is a common frequency band for 1U CubeSats [6][7][10]. Usage of frequencies in the UHF band also has the advantage that antennas operating in this frequency band tend to be omnidirectional, or send electromagnetic energy in all directions. This trait would allow the UCubeSat to communicate information verifying its status no matter its orientation in space after release into low earth orbit. While the CubeSat will need to orient itself to aim its camera towards earth, the ability to transmit to the ground station despite orientation will ensure that at least contact and verification of status can be achieved if the UCubeSat’s orientation adjustment systems fail. However, using a frequency in the UHF band will reduce the available bandwidth per channel also reduce the data rate. Though the data rate will be considerably lower when compared to a frequency in the S band [6][10] this limitation would not be fatal to the mission. Five minutes is a conservative estimate for the amount of time the CubeSat can communicate with the ground station per orbit [6]. With a 9.6Kb per second data rate, which can be expected given a UHF link [6][10], approximately 2 M bits of data can expect to be transmitted within the 5 minute time interval per orbit. A 640 × 480 pixels VGA image, assuming an encoding of around 1.5 to 2 bits per pixel, requires around 614K bits to be transmitted [6]. This Fig. indicates that there is a strong likelihood that a UHF operating CubeSat can transmit at least one image per orbit. This rate is much too slow for a video feed, but still sufficient for the original goal of transmitting a photograph of the earth from space. Though it was known that most CubeSats of the 1U size rely primarily on the UHF and VHF band for their wireless communications [6][10], it was not known how to choose an exact frequency within the UHF band. This problem was solved by consulting an online open-source database known as Satnogs, which contains design information CubeSats and ground stations, as well as information on the downlink frequencies of hundreds of different CubeSats [13]. The Satnogs website organized these downlink frequencies into a histogram. In the UHF band, the tallest column corresponded to a downlink frequency of 6 436.4 MHZ, indicating this frequency is the most common downlink frequency for CubeSats using the UHF band. Since at least seven 1U CubeSats have used this UHF frequency for their downlinks, such as the Tevel 1 [8][13], it was decided to proceed with making a link budget with 436.4 Mhz as the operating frequency. 7 METHODS Before deciding on a type of antenna, the first step was to account for the known losses and gains in a document known as a link budget. The link budget is a list of all gains and losses a signal will encounter as it travels from the transmitter to receiver, or in this case, from CubeSat to Ground Station [10]. The link budget takes the transmitted power and compiles all the gains and losses to estimate the amount of power transmitted to the receiving antenna [10]. Since the operating frequency is in the UHF region, losses due to precipitation particles is neglectable, less than a hundredth of a decibel [10][11]. Thus, in the case of UHF, the main loss is the path loss defined by the Friis transmission equation [6][10][11]. The main gains that can be expected come from the antennas on the CubeSat and the ground station [6]. 𝐺𝐺𝑡𝑡 𝐺𝐺𝑟𝑟 𝜆𝜆2 (1) 𝑃𝑃𝑟𝑟 = 𝑃𝑃𝑡𝑡 (4𝜋𝜋𝜋𝜋)2 (1) describes the Friis transmission equation [10][14], where 𝑃𝑃𝑟𝑟 is the received power in Watts [W], 𝑃𝑃𝑡𝑡 the transmitted power in Watts [W], 𝐺𝐺𝑡𝑡 the linear magnitude of the gain of the transmitting antenna, 𝐺𝐺𝑟𝑟 the linear magnitude of the receiving antenna gain, 𝜆𝜆 being the wavelength of operating frequency in meters per second [m/s], and 𝑅𝑅 being the distance between transmitter and receiver in meters [m] 𝜆𝜆 (2) 4𝜋𝜋𝜋𝜋 (2) is the result of taking log base 10 on both sides of (1) [14]. Leaving the 4*pi*R term unchanged is common practice and often defined as space loss [14]. The fact that Equation 2 is a sum of separate terms indicates that the received power in dBm can be computed by measuring and summing the path loss and the gains of the antennas separately. This idea is the basis for the procedure of making a link budget, computing each gain and loss separately and summing them together. The sum of the gain terms 𝐺𝐺𝑡𝑡 (𝑑𝑑𝑑𝑑) + 𝐺𝐺𝑟𝑟 (𝑑𝑑𝑑𝑑) can also be seen as the net total gain needed to ensure the received power remains above a particular threshold. 𝑃𝑃𝑟𝑟 (𝑑𝑑𝑑𝑑𝑑𝑑) = 𝑃𝑃𝑡𝑡 (𝑑𝑑𝑑𝑑𝑑𝑑) + 𝐺𝐺𝑡𝑡 (𝑑𝑑𝑑𝑑) + 𝐺𝐺𝑟𝑟 (𝑑𝑑𝑑𝑑) + 20 ∗ log The largest distance that can be expected between the ground station and the CubeSat is 2600km [6]. By using this distance for R, the link between the ground station and the CubeSat can be ensured to function over all expected distances [6]. By assuming a speed of light as the propagation speed, the wavelength of 463.4 MHz is about 0.69 meters, which is the value for lambda. The above equation generated a propagation loss of about 153.5 dB, which agreed the results of the IEEE essay [6]. 8 Before determination the gain of the antennas can be accomplished, it is necessary to determine the minimum amount of received power required. The minimum required received power is a function of bit rate, bandwidth, modulation scheme, noise Fig., and desired bit error rate [15]. The Fig. of -116dBm was calculated using the thermal noise Fig. at room temperature at using the UHF frequency. A detailed derivation of this Fig. is described in this INtersil application note, and is further described in the context of the U CubeSat project in appendix B [15]. With the noise floor defined, it was now possible to determine the needed gain that the antennas must provide. In general, the maximum amount of power that a CubeSat can transmit is 2 watts or 33dBm at its antenna [10]. Further increasing the power risks discharging the CubeSats internal batteries or preventing other systems from functioning [6][10]. If the path loss is subtracted from 33dBm, then the received power is -120 dBm. This is about 4 dBm below the noise floor; however, it is generally recommended to have at least 8 dBm of margin above the noise floor to maximize the chances that the signal reception will remain stable, as well as ensure that any unexpected losses will not deter reception [15]. Fig. 5 shows the UCubeSat link budget made to sum all the expected gains and losses to estimate received power. According to this link budget, this raises the needed gain of the antennas to have a total of at least 12 dB. The figures used for calculating the noise floor are also present. Fig 5. UCubeSat Link budget for project implemented in Microsoft Excel. 9 Since the U CubeSat must transmit electromagnetic energy in all directions, the gain of the antenna on the UCubeSat needs to be 0. An antenna with gain of 0 is an omnidirectional antenna [14]. This requires the ground station antenna to have atleast 12 dB of gain to ensure the sum of the gains is atleast 12 dB. Most CubeSats in the UHF range use a pair of dipoles to produce an antenna that radiates close to evenly in all directions [6][8]. Other sources also point out that the Yagi-Uda antenna usually has a gain of 12 dB to 15 dB [6]. To verify whether these antennas can be used, the next phase was to assemble the antennas in a program called CST. This program is capable of providing extremely accurate simulations of how an antenna can be expected to behave given its shape and material type, and can be used to determine exact radiation patterns and gain. 10 RESULTS The first antenna that was constructed in CST for the UCubeSat was a basic quarter wavelength dipole. Each arm being an eighth of a wavelength, or about 8.6 cm, in length. Since each arm of the dipole antenna are typically folded alongside the edges of a CubeSat when not in use [16], the maximum length of each arm is restricted to less than 10 cm or the length of one edge on a typical 1U CubeSat. A quarter wavelength dipole was specifically chosen at the 436.4 MHz operating frequency, the length of each arm of a quarter wavelength dipole antenna is about 8.6 cm. Quarter wavelength dipole antennas still radiate effectively despite their relatively small size [14]. Fig. 6 shows a quarter wavelength dipole antenna in CST. Each arm is made from aluminum and separated by a gap that is very small when compared to the wavelength of the operating signal. The diameter of each arm is constant, and their thickness is approximately .005 of a wavelength to ensure that the current distribution is stable. Fig 6. capture from CST workspace. Fig. 7 shows the results of the dipole antenna simulation in CST. The radiation pattern revealed that the electromagnetic energy is relatively spread out, as confirmed by the large amount of red around seen in Fig. 8. This confirms that a dipole antenna can receive and transmit well in directions perpendicular or near perpendicular to the axis alongside the length of the dipole antenna. The blue near the top seen in Fig. 8 confirms that the dipole is unable to send radiation directly below or above itself, as the blue represents a loss of -4 dB to –8 dB, a large portion of the anticipated link margin in the link budget [14]. The presence of a region where the antenna can not radiate makes a single dipole a poor choice for an antenna on the UCubeSat. 11 Fig 7. CST capture of the radiation pattern of a quarter wavelength dipole antenna. The dipole antenna itself is visible in the center of the pattern. Fig. 8 contains the same results as Fig. 7, but compressed into a 2D image. The large dip in signal magnitude at angle 0 and angle 180 further confirms that signal transmission along the vertical axis of the dipole is comparatively weak. Fig 8. Capture of 2D radiation plot generated in CST. 12 Though it is a small chance, it is possible that the UCubeSat can be oriented in space such that the axis where radiation is weak is pointed towards the earth. Such an event would result in an intermittent connection between the UCubeSat and the ground station. One means of solving this issue is that adding a second dipole oriented 90 degrees with the other dipole. Fig. 9 shows the radiation pattern of a dipole antenna oriented 90 degrees with respect to the other dipole. The radiation plot of this second dipole shows strong radiation along the 0 and 180 degree axis, where the previous dipole was unable to radiate. This finding indicated that if the two dipoles could be run at the same time, then the directions each dipole radiates well in would overlap the axis where the other dipole cannot radiate, thus producing a more omnidirectional antenna. Fig 9. Capture of 2D radiation plot generated in CST. Fig. 10 shows the implementation of the dual dipole approach, with two dipoles of identical dimensions being oriented 90 degrees from one another with a small gap in the center. 13 Fig 10. Image of original dipole simulation with extra dipole added. As illustrated in Fig. 11, the results of dual dipole test are similar to that of the single dipole. The reason behind this was not initially clear. Fig 11. CST capture of radiation pattern produced by two orthogonal dipoles. With further research, it was determined that the simplest solution would be to adjust the phase of the input signal feeds for each dipole antenna to be 90 degrees apart [17]. 14 As shown in Fig. 12, adjusting the phase of the inputs was an effective solution. The radiation pattern now indicates that the ability to receive a signal is close to consistent in all directions, which is highly desirable for the UCubeSat antenna. The -.9 dB loss shown in the green is acceptable as it is well within the loss margin anticipated in the link budget. Fig 12. CST capture of radiation pattern produced by two orthogonal dipoles with altered phase to feeds. However, one detail neglected in this simulation is that the dipole antennas used on CubeSats’ do not follow the assumption of small spacing between arms or alignment of the arms on the same axis. In most UHF CubeSat antennas like in Fig. 13, this is to allow for easy stowing within the chassis of a CubeSat [16]. The change in spacing and offset should not affect the ability for the antenna to radiate in all directions [18]. However, simulation was required to verify of such factors would not significantly affect the radiation pattern. 15 Fig 13. Image of actual UHF antenna deployed on a CubeSat [16]. Fig. 14 shows the dual dipole simulation adjusted to account for detail like spacing and offset. The armatures of both dipoles were offset by 10cm and spaced apart by 10 cm. Fig 14. Capture of adjusted dual dipole CST simulation. 16 As seen in Fig. 15, simulation revealed that the radiation pattern of the adjusted antenna still radiates in all directions, similar to the original dual dipole simulation seen in Fig. 12. Fig 15. Capture of Radiation pattern produced by adjusted dipole. Fig. 16 shows a 10 cubic centimeter aluminum block that was added to the simulation. This was added to approximate the effect of the UCubeSat itself on the radiation pattern of the dual dipole antenna. Fig. 16. CST capture of dual dipole simulation further modified with 10 cm^3 aluminum block. 17 Fig. 17 illustrates that the addition of the aluminum block hardly affects the radiation pattern. This is a strong indication that the dual dipole antenna system can be adapted to the UCubeSat, just as it has been done for other 1U CubeSats [8][16]. Fig 17. Radiation pattern of Dual dipole simulation with 10 cm^3 aluminum block. With an optimum antenna arrangement for the UCubeSat determined, the next step was to focus on determining the ground station antenna. The Yagi Uda type antenna was selected as a promising candidate for this role. This is because the Yagi Uda type antenna is relatively light weight when compared to alternatives like the dish antenna. The Yagi Uda antennas also has considerably lower wind loading when compared to alternatives like the dish antenna. The Yagi Uda antenna also has a gain of typically 12 to 15 dB, which meets the required gain to facilitate communication between the UCubeSat and ground station according to the prior link budget analysis. As seen in Fig. 18, the Yagi Uda antenna consists of a typical dipole antenna in parallel with multiple passive conductors known as elements [14]. These elements help guide the electromagnetic radiation from the dipole in a single direction [14]. The dipole antenna itself, marked in Fig. 18 with a length of L, is commonly known as the driven element [14]. The element with length 𝐿𝐿𝑅𝑅 is known as the reflector, and it is typically longer than the driven element [14]. The element with length 𝐿𝐿𝐷𝐷 is known as a director element, the directors tend to be shorter than the driven element, and in many Yagi Uda antennas there is more than one director to facilitate higher gains [14]. Length 𝑆𝑆𝐷𝐷 refers to the distance between each individual director element and the distance of the first director element from the driven element [14]. 𝑆𝑆𝑅𝑅 refers to the distance between the excited and reflector elements [14]. These spacings also affect the overall gain of the antenna, as well 18 as the total length of the antenna which is an important parameter for physical construction of the antenna [14]. Fig 18. Typical Yagi Uda element arrangement [14]. The interaction between each of the elements in the Yagi Uda antenna is complex, as each element affects the other elements to some degree. Effective design of Yagi Uda antennas is often based on simulation and experimentally determined values. Many studies on the Yagi Uda antenna have been performed been since its invention, and such analyses have yielded tables that accurately relate the gain of the antennas to the number of elements, the length of each element, and the spacing between elements [14]. Fig. 19 depicts a chart published in the paper Design Data for Short and Medium Length Yagi-Uda Arrays, by H.E Green [19]. This chart has been cited by numerous sources and is a standard that subsequent Yagi Uda antenna designs been compared too [14]. This chart describes a feasible gain for a Yagi Uda antenna with a certain number of elements. This chary indicates that a Yagi Uda antenna with eight elements can get very close to the 12 dB requirement established through the link budget analysis. 19 Fig 19. Yagi Uda Antenna Gain chart [18]. Fig. 20 depicts simulation data from H.E Green. This information goes into more depth than the chart in Fig. 19, and includes further measurements for the eight element Yagi Uda antenna. It can be seen from this data that the gain of an 8 element array can be increased above 12 dB given a .25 wavelength spacing. With such spacing, an eight element Yagi Uda antenna can have a gain as much as 13.1 dB, which more than meets the requirements established by the link budget. These values can be seen in the row of Fig. 20 corresponding to 8 elements and a spacing of .25 wavelengths. This data assumes that the distances 𝑆𝑆𝑅𝑅 and 𝑆𝑆𝐷𝐷 are equal. Fig 20. Yagi Uda Antenna simulation data [19]. 20 At the intended operating the frequency of 436.4 MHz, the wavelength is short enough that the overall length of an 8 element Yagi Uda antenna with a gain of 13.1 is roughly 5 feet in length, which is advantageous for construction out of PVC pipe. A 5-foot hollow PVC pipe of ¾ inch diameter can easily hold its own weight and that of the antenna elements. The low dielectric constant of PVC is also advantageous to ensure that the boom holding the elements will have minimal effect on the antenna radiation pattern [19]. Fig. 21 depicts a Yagi Uda antenna constructed in CST. Though the exact measurements provided in Fig. 21 were followed, the gain of the antenna was considerably lower than indicated in Fig. 21. Fig 21. CST capture of a Yagi Uda antenna with a boom made of ¾ inch PVC pipe. Note that this is made from copper. However, H.E Green’s paper also mentioned that Yagi Uda antennas designed with the provided table could have less gain than expected due to approximations made in the model used to derive the design tables, such as neglecting dipole end effects [9]. The paper suggested increasing the spacing and adjusting element lengths for a frequency 1 to 2 percent higher than desired [19]. By adding 10 additional millimeters onto the spacing in between elements and adjusting the lengths of the elements to correspond to a frequency of 447.31 Mhz, which is 2.5 percent above 436.4 Mhz, the simulated performance of the antenna was greatly improved. The simulation now predicted that the particular Yagi Uda antenna now had a gain of 13.1 dB at 436.4 Mhz, which exceeds the 12 dB requirement. This increased gain is indicated in the radiation pattern shown in Fig. 22. 21 Fig 22. Radiation pattern of Yagi Uda antenna after modifications. According to the scale, the darkest shade of red corresponds to a maximum gain of 13.2 in the x direction. These promising results indicate that the Yagi Uda has the required gain to meet the requirements defined in the link budget. This fact makes a strong argument for the use of this Yagi Uda antenna design in the ground station. Fig. 23 depicts the results shown in Fig. 22 compressed into a 2D image. Fig 23. Radiation pattern of Yagi Uda antenna compressed into a 2D image. The blue lines indicate the -3db beam angle. 22 Fig. 23 also confirms that the gain is 13.2 dB, but it also shows that in consequence the -3 dB beam width is a relatively narrow 36 degrees. The narrow beam width can be compensated though by designing the ground station to be able to point the antenna in a specific direction, which is a common ability amongst CubeSat ground station designs [13]. Since the Yagi Uda antenna design was shown to be promising in CST, a prototype Yagi Uda antenna was made using the measurements tested in CST was made and can be seen in Fig. 24. The white boom was made from ¾ inch diameter hollow PVC pipe, with the antenna elements made from copper dowels. A small hole was also cut in the boom to allow wires to be attached to the driven element. Fig 24: The prototype Yagi Uda antenna. A future goal of the UCubeSat club is to test this prototype Yagi Uda using an RF signal generator and a receiving monopole antenna attached to a spectrum analyzer. If the results of this future test are similar to that provided in CST, then it is possible for this prototype to be used for the future UCubeSat ground station. One drawback to using the dual dipole arrangement for the UCubeSat antenna is that the polarization of the signal transmitting from this antenna will vary from circular to linear depending on the orientation of the UCubeSat. This change in polarization can affect the ability for the designed Yagi Uda antenna on the ground station to receive the signal transmitted from the UCubeSat. In the case of the signal from the UCubeSat being linearly polarized, which is expected to be the most common case, it is possible for reception to be reduced or entirely cut off if the signal transmitted has no electric field component parallel to the conducting elements of the Yagi Uda antenna [14]. The plan of overcoming this is to add a mechanism to allow the Yagi Uda antenna to be rotated around the axis of its 23 supporting boom, thus aligning the antenna elements of the with the electric field component of the incoming signal and maximizing reception. In the case of the UCubeSat signal being circularly polarized, it is still possible for signal reception to be maximized through rotating the Yagi Uda antenna around its boom axis; however, there will be a -3dB loss at best [17]. It is expected though that this loss can be absorbed in the margin of the link budget. Overall, UCubeSat signal polarization will be a problem that can be accounted for in the future process of designing the UCubeSat ground station. DISCUSSION The results of the CST measurements indicated that antennas chosen for the UCubeSat and the ground station do meet the gain requirements defined in the link budget. The combined gain of 13 dB allows for the received power to reach -109 dBm, which is 9 dBm above the estimated noise floor. Such a noise margin provides a considerable level of insurance the link between the UCubeSat and the ground station will not be unexpectedly interrupted, and link budgets with less than 4 dBm of margin have been reported as successful [6]. By detailing the process of quantifying expected signal losses in a link budget, estimating the noise floor, and testing probable antennas in CST, future UCubeSat club members would be provided with a viable solution to the problem of communicating with the UCubeSat. This would allow future UCubeSat members to focus on subsequent UCubeSat design challenges, such as design of the UCubeSat chassis and mounting the ground station antenna. Though the prototype of the Yagi Uda antenna has not been tested, the CST results are themselves strong indicators that the design of the Yagi Uda will have the anticipated gain. The need to test the Yagi Uda antenna also provides an immediate goal for future UCubeSat team members to pursue. 24 REFERENCES 1. Maini, K.Anil , and A.Varsha, Satellite Technology: Principles and Applications. 3rd ed. New York: Wiley, 2014. Web. [E-book] Available: ProQuest: Ebook Central 2. D. Dave, and S. Adam, The View from Above: Applications of Satellite Data in Economics. Nashville: American Economic Association, The Journal of economic perspectives, 2016-10-01, Vol.30 (4), p.171-198 3. C. Jamie, R. Coelho, J Foley, et all, CubeSat 101 Basic Concepts and Processes for First Time CubeSat Developers. NASA CubeSat Launch Initiative, October 2017. 4. Nanosats Database, “ figures: Nanosatellites by organizations,” Nanosats Database, 2014. [Online] Available: Nanosats Database | Constellations, companies, technologies and more [Accessed: November 15,2021]. 5. A. Miller. Tracking CubeSats. Aerospace America, May 2020 6. O. Popescu, "Power Budgets for CubeSat Radios to Support Ground Communications and Inter-Satellite Links," in IEEE Access, vol. 5, pp. 1261812625, 2017. 7. Aslan, A. R, H. B Yagci, M. E Umit, A. Sofyali, M. E Bas, M. S Uludag, O. E Ozen, M. D Aksulu, E. Yakut, C. Oran, M. Suer, I. A Akyol, A. B Ecevit, M. S Ersoz, I. Oz, S. Gulgonul, Baris Dinc, and Tahir Dengiz. "Development of a LEO Communication CubeSat." 2013 6th International Conference on Recent Advances in Space Technologies (RAST) (2013): 637-41. Web. 8. Krebs, Gunter D. “Tevel 1, ..., 8”. Gunter's Space Page. Retrieved April 05, 2022, from https://space.skyrocket.de/doc sdat/tevel-1.htm 9. L. Chang, C. Salinas, J. Wang, J. Su, Y. Duann, J. Hong, Yi. Chiu, S.Chen, A. Chandran, M.Mcgrath, D.Fritts, L.Gordley, J.Fischer. “A Preliminary Design for the INSPIRESat-1 Mission and Satellite Bus: Exploring the Middle and Upper Atmosphere with CubeSats.” AIAA/USA Conference on Small Satellites, August, 2016. 10. D.Barbaric, J. Vukovic, D. Babic, “Link Budget Analysis for Proposed CubeSat Earth Observation Mission,” Dissertation, Faculty of Electrical Engineering and Computing, University of Zagreb, Croatia, 2018. 11. M, Zubair, Z. Janjua, S. Khan, J. Nasir, “Atmospheric Influences on Satellite Communications,” Dissertation, COMSATS Institute of Information Technology, Islamabad, January 2011. 12. Radulow, Atanas. Dimitrov, Stanislav, members of EnduroSat CubeSat Product Company Outreach Branch, October 29, 2021 and November 2, 2021. 13. SatNOGS, “SatNOGS:dB:transmitters:spectrum, ” SatNOGS [Online] Available: SatNOGS DB - Transmitters [Accessed: November 15, 2021]. 14. W. Stutzman, and G.Thiele, Antenna Theory and Design. 3rd ed. Hoboken, NJ: Wiley, 2013. Print. pp 109-111, 153 15. Intersil, Appl. Note AN9804.1 pp.1-3 25 16. Endurosat, “UHF Antenna III,” Endurosat [Online] Available: UHF CubeSat Antenna Module | CubeSat by EnduroSat [Accessed: December 12, 2021] 17. Schurig, David. University of Utah Electromagnetics Professor, March 16, 2022. 18. K. Schraml, A. Narbudowicz, S. Chalermwisutkul, D. Heberling and M. J. Ammann, "Easy-to-deploy LC-loaded dipole and monopole antennas for cubesat," 2017 11th European Conference on Antennas and Propagation (EUCAP), 2017, pp. 2303-2306 19. H. Green, “Design Data for Short and Medium Length Yagi-Uda Arrays,” Institution of Engineers(Australia), Electrical Engineering Transactions, Vol. EE2, pp 1-8, March 1966. 26 Appendix A An Explanation for Why An Operating Frequency In S Band Was Not Chosen. The 2.0 Ghz to 2.3 Ghz range in the S band was the initial choice as a range for an operating frequency for the CubeSat. Though this approach did initially appear promising, further research revealed specifics of S band communications that are difficult to accommodate in 1U CubeSats. One advantage of selecting an operating frequency in the S band or a band of even higher frequency is that the available bandwidth per channel is increased, which is advantageous for communications that require a large amount of information to be transmitted in a short period of time [10]. In the case of the range of 2.0 to 2.3 Ghz, the bandwidth expected would allow for communications at a rate of 1 to 2 Mb/s [10]. The 2.0 to 2.3 Ghz range, while high in frequency is also below the threshold where attenuation due to atmospheric particles is not significant [11]. This makes S band ideal for when the CubeSat needs to send large amounts of data to the ground station. However, higher frequency operation does have some disadvantages. The main disadvantage is that higher frequency transceivers generally consume greater amounts of power and dissipate larger amounts of heat [12]. This in turn requires more volume within the CubeSat for both heat dissipation and storing of additional batteries [12]. Higher frequencies also experience higher amounts of path loss, which can be disadvantageous [6]. In the case of S band though, this can be made up for by using antennas of higher gain, such as the patch antenna like the one seen in Fig. A.1. Fig. A.1: Capture of 10cm by 10cm patch antenna in CST. 27 A patch antenna does have the advantage that for high frequencies a working patch antenna can be made small enough to fit inside a 10 cm by 10 cm area. This is the area of one face of the CubeSat, which means that a patch antenna can be incorporated into a CubeSat without any deployable parts or mechanisms. Patch antennas like the one seen in Fig. A.1 also have about 7dB to 9dB of gain if operating in the S band [10]. The patch antenna seen in Fig. A.1 produces a gain of 7.15 dB according to the results depicted in Fig. A.2. This is helpful for reducing the amount of needed power on the CubeSat antenna. However, as can also be seen in Fig. A.2, this results in the radiation being confined to a beam with an angle of 78.3 degrees. Fig. A.2: 2D CST capture of radiation pattern produced by patch antenna. Due to this narrow beamwidth, a CubeSat using this antenna would need to be able to point the antenna towards the earth [12]. This feat is possible but requires the presence of reaction wheels or magnetorquers within the CubeSat [12]. The directional nature of the patch antenna would also prevent communication with the CubeSat when the antenna is not facing the earth, making it difficult to communicate consistently with the CubeSat [12]. Consistent communication between the CubeSat and the ground station is essential, and many CubeSats that use an S band transmitter also incorporate an omnidirectional UHF communication system similar to the one planned for the UCubeSat [12]. Though it is possible to have an omnidirectional UHF communication system, an S band transmitter for high data rate downlink, and magnetorquers and/or reaction wheels for pointing within a CubeSat, the volume, weight, power supply, and heat dissipation requirements make this option viable only for CubeSats 3U and above in size [12]. Since designing and launching a 3U CubeSat is considerably more 28 complex and costly [12], it was decided to remain with a 1U CubeSat even if it meant a reduced data rate. A 3U CubeSat with the such hardware may be a project for future members of the UCubeSat club though. 29 Appendix B Explanation of Noise Floor Figure. To determine if the received power at the ground station is sufficient, it is vital to estimate the required noise floor [6][10]. The noise floor is the product of the Signal to Noise Ratio, (SNR) and the magnitude of the noise figure inherent to the system [15]. The noise floor represents the minimum possible power level that must be at the receiver for the incoming signal to have its information completely recovered [10]. The SNR is dependent on many parameters, as indicated in Fig. B.1. 𝐸𝐸𝑏𝑏 𝑅𝑅 𝑆𝑆𝑆𝑆𝑆𝑆 = ∗ (𝐵𝐵. 1) 𝑁𝑁0 𝐵𝐵𝑇𝑇 (B.1) calculates the minimum SNR in order to receive a signal with a particular data rate and bandwidth [10][15]. 𝐸𝐸𝑏𝑏 is the energy required per bit of information, 𝑁𝑁0 is the thermal noise in 1Hz of bandwidth, 𝑅𝑅 is the system data rate [bits/second], and 𝐵𝐵𝑇𝑇 is the system bandwidth [Hz]. The SNR is the ratio of the received signal power to power of the noise within the 𝐸𝐸 receiver [10][15]. The ratio of 𝑏𝑏 in equation B.1 is dependent on the modulation scheme 𝑁𝑁0 and desired bit error rate. A good rule of thumb is that a bit error rate of less than 1 out of a million. A modulation scheme defines how an RF signal is changed to carry information, such as digital bits. Fig. B.1 shows a chart of experimentally gathered values for the bit error probability vs the Energy per 1Hz bandwidth. Each line corresponds to a different modulation scheme. Fig. B.1: Bit error rate vs Eb/No . [15] 30 Following the line labelled DBPSK (Differential Binary Phase Shift Keyed) and 𝐸𝐸 DQPSK (Differential Quadrature Phase Shift Keyed), this chart indicates that the 𝑏𝑏 ratio 𝑁𝑁0 needs to be at least 11 dB. The line for DBPSK and DQPSK was used because these modulation schemes were closest to the intended modulation scheme of QPSK (Quadrature Phase Shift Keying). The 𝑅𝑅 𝐵𝐵𝑇𝑇 is also highly dependent on the modulation scheme. In the case of QPSK, the ratio of data rate in bits per second to bandwidth in hertz is two [10]. QPSK has the advantage of a constant signal amplitude, with the desired information carried exclusively in the phase. With the parameters for the SNR computed, the next step was to determine the noise figure. The main variety of noise that is focused on is Johnson noise, which is defined by (B.2). 𝑁𝑁 = 𝐾𝐾 ∗ 𝑇𝑇 ∗ 𝐵𝐵 (𝐵𝐵. 2) Where 𝑁𝑁 is the average noise power [watts], 𝐾𝐾 is the Boltzmann constant [Joules/Kelvin], 𝑇𝑇 is the temperature [Kelvin], and 𝐵𝐵 is receiver bandwidth [Hz]. (B.2) is often expressed with K multiplied with 290 K, which is approximately room temperature [15]. This results in a value of 4.002e-18 Joules, which is then multiplied by 1 hertz to get 4.002e-18 watts. By multiplying this value by the magnitude of the bandwidth of the receiver it is possible to determine the noise figure, which describes the power of the noise inherent to the receiver system. In the case of UHF, the maximum channel bandwidth is 20 kHz [10]. To systemize calculation of the final noise floor, all the required figures for the SNR and for the intrinsic noise level were converted to decibels. (B.3) describes the results of taking 10* log base 10 of both sides in order to perform this conversion. This causes the multiplication calculation for the total noise floor to become a sum of terms that can be found individually. This is similar reasoning to the process of creating a link budget. This derivation is not explicitly shown in the Intersil application note, but a similar procedure is used in the examples section [15]. 10 ∗ 𝑙𝑙𝑙𝑙𝑙𝑙10 (𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁𝑁 𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹) = 10 ∗ 𝑙𝑙𝑙𝑙𝑔𝑔10 𝐸𝐸𝑏𝑏 𝑅𝑅 𝐸𝐸𝑏𝑏 𝑅𝑅 𝑑𝑑𝑑𝑑𝑑𝑑 (𝑑𝑑𝑑𝑑) + (𝑑𝑑𝑑𝑑) + 𝐾𝐾 ∗ 𝑇𝑇 ∗ ∗ 𝐾𝐾 ∗ 𝑇𝑇 ∗ 𝐵𝐵 = + 𝐵𝐵(𝑑𝑑𝑑𝑑 ∗ 𝐻𝐻𝐻𝐻) 𝑁𝑁0 𝐵𝐵𝑇𝑇 𝑁𝑁0 𝐵𝐵𝑇𝑇 ℎ𝑧𝑧 The chart in Fig. B.1 already provided that is taken for the rest of the parameters, 𝑅𝑅 𝐵𝐵𝑇𝑇 𝐸𝐸𝑏𝑏 𝑁𝑁0 (𝐵𝐵. 3) ratio is 11 dB. When 10 log base 10 becomes 3 dB, the 4.002 e-18 watts becomes - 31 174 dBm/Hz, and the bandwidth becomes 43 dB Hz. Summing these figures together obtains the final noise floor. 32 Name of Candidate: Blayze Ashurst Date of Submission: May 11, 2022 33