SSC00-V-5 - Aalborg UniversitetSSC00-V-5 CubeSat: A new Generation of Picosatellite for Education and Industry Low-Cost Space Experimentation Mr. Hank Heidt 1, Prof. Jordi Puig-Suari

Heidt 14TH Annual/USU Conference on Small Satellites1

SSC00-V-5

CubeSat: A new Generation of Picosatellite for Education and Industry Low-Cost SpaceExperimentation

Mr. Hank Heidt1, Prof. Jordi Puig-Suari2, Prof. Augustus S. Moore3, Prof. Shinichi Nakasuka4,Prof. Robert J. Twiggs5

1NetDecide Corporation, tel: 703-610-523, e-mail: [email protected] Polytechnic State University, tel: 805-756-6479, e-mail: [email protected]

3Dartmouth College, e-mail: [email protected] of Tokyo, tel: (033) 812-2111 EX. 6590, e-mail: [email protected]

5Stanford University, tel: 650-723-8651, e-mail: [email protected]

ABSTRACT

The launch and deployment of picosatellites from theStanford University OPAL microsatellite in February2000 demonstrate the feasibility and practicability ofa new age of space experimentation. Two of the sixpicosatellites deployed from OPAL were built by TheAerospace Corporation in El Segundo, CA anddemonstrated new space testing of MEMS RFswitches and intersatellite and ground communicationwith low power wireless radios. These picosatellitesweighting less than one kilogram with dimensions of4x3x1 inch were built as test platforms for DARPAand were constructed and delivered for flight in lessthan nine months.

From this experience, a new generation of picosatscalled CubeSat is being developed by a number oforganizations and universities to accelerateopportunities with small, low construction cost, lowlaunch cost space experiment platforms. CaliforniaPolytechnic State University at San Luis Obispo, CAis developing launcher tubes that can be part of asatellite or attached to any orbiting platform to launchfrom 1-3 CubeSats per tube. These tubes will containCubeSats of 1-2 kilograms weight and approximately4-inch cube shape. This size as compared to thepicosatellites launched on OPAL provide bettersurfaces for practical solar power generation,physical size for components and a shape thatprovides better space thermal stability.

A consortium of potential CubeSat developers is nowwide ranging with universities from Japan, NewZealand, the US, amateur radio clubs and industryparticipants. Potential launch opportunities exist withthe Russian Dnepr (SS-18) about twice/year, with theOSP (Minotaur) every 18 months and possible 100

km altitude orbits from the second stage of Deltalaunches.

This paper will review the OPAL picosatellite launchand performance, the launcher being built for theCubeSat, the development and payloads of CubeSatdevelopers and cost and timing of launchopportunities.

TABLE OF CONTENTS

1. Introduction2. Stanford University OPAL Program3. CubeSat Development Program4. Participants in CubeSat Program5. Launch Opportunities6. Conclusions7. Acknowledgements8. References

1. INTRODUCTION

The current trend in satellites to do more for less costhas led to the “Smaller, Cheaper, Faster, Better”space missions. It is generally resulting in manycases of doing things the same way as before in asomewhat smaller format for less cost, but this hasbeen cited as the basis for some of NASA’s recentdeep space mission failures.

To support some science missions and in proposingnew missions, the trend is now to see what can bedone by decreasing the spacecraft sizes by orders ofmagnitude. This is becoming more practical with therapid advances in decreasing electronics size, ingreatly increased capability and very low powerconsumption. This decrease in size also directlybenefits the mission cost in lower launch costs.


There are several programs now exploring micro,nano and pico satellite sizes. These sizes rangingfrom <100kg to <1kg are being developed and testedby combined programs such as the ones discussed inthis paper with Stanford University, The AerospaceCorporation, Santa Clara University, University ofTokyo, Dartmouth College, California PolytechnicState University and groups of private amateursatellite enthusiasts.

This paper will describe the experience of StanfordUniversity with a student built microsatellites andtheir experience in providing launch opportunities forsix small picosatellites in January 2000. It will alsodescribe a follow-on program for a picosatellitecalled CubeSat as a collaborative effort to continuedeveloping the picosatellite, provide a convenientlow cost launch interface and coordinate launchactivities.

2. STANFORD UNIVERSITY OPAL PROGRAM

2.1 Introduction

The Space Systems Development Laboratory at theDepartment of Aeronautics and Astronautics atStanford University was established in 1994 with thepurpose of providing project based learning programsfor engineering graduate and undergraduate studentsto gain experience in systems engineering. Toaccomplish this goal, the program was designed totake students through the life cycle of a project, inthis case the design, development, fabrication,testing, launch integration and space operations of amicrosatellite. Additional goals for the laboratorywere to build the facilities, curriculum and researchinfrastructure for future laboratory programs.

2.2 Microsatellite Development program plan

The program plan1-2 at the Space SystemsDevelopment Laboratory (SSDL) was to haveprograms with the following attributes:

• Program team student managed• Student designed• Design assisted by mentor group

and support of industry specialists• Complete from design to “ready for

launch” in one year• Out of pocket cost not to exceed

$50,000• Use of commercial off-the-shelf

components• Spacecraft operational life of one

year

The students were responsible for the selection of theprogram manager, systems engineer, subsystemmanagers and other tasks. This was to be like adevelopment team within industry. Students weregiven general design parameters for the satellite suchas size, weight, expected launch orbits, some possibleexperiments to be flown, communication frequencybands and an approximate budget for each of thesubsystems. The mentors and industry experts wouldprovide design counseling, but could not do thedesigns or fabricate components for students.

The students could select experiments that theywanted to fly, but outside paying experiments weresolicited to help cover the materials and parts costs ofthe program as well as for laboratory developmentand student travel. These outside experiments wouldprovide the students with experience for dealing withan outside customer and valuable interaction withindustry and government customers.

It was expected that if a microsatellite could becompleted in nearly one year, that a new satellitedesign would be initiated each year with each newgroup of students entering the graduate program3-4.


2.3 The First Satellite SAPPHIRE

The first student built satellite, SAPPHIRE, wasstarted at Stanford in April 1994. This project wasstarted without any opportunity to launch, but on thefaith that a free or low cost launch would becomeavailable when the satellite was completed.

The initial concepts for SAPPHIRE5 are shown inFigures 2.1 and 2.2. A stacked modular approachwas selected with hexagonal shape. The initialparameters were for a 20kg weight, 18 inch diameterand 12 inch high structure.

The structure was a stack of modular shelves madefrom aluminum honeycomb. There were fourlongitudinal bolts with spacers that formed thestacked structure. The outer solar panels providedonly the surface for mounting the solar cells andexternal attachments such as antennas and the launchinterface.

The final SAPPHIRE contained two studentexperiments, a B/W Logitech CCD camera, a voicesynthesizer and the one government sponsoredpayload by JPL was as set of MEMS infraredsensors. Thus the name SAPPHIRE (Stanford AudioPhonic Photographic InfRed Experiment) wasderived. SAPPHIRE used amateur radios and wasbuilt to operate in the amateur frequency J-mode (2m145MHz uplink and 70cm 437MHz downlink). Itused a Motorola 68332 microcontroller for theC&DH, NiCad Batteries and GaAs solar cells. Iflaunched into a polar noon-midnight orbit, theestimated average power would be 7-8 watts. Thepassive magnetic stabilization scheme is shown inFigure 2.3 and the finished microsatellite in Figure2.4.

SAPPHIRE was completed in June of 1998 and at thetime of writing the paper is scheduled for launch inlate 2002

Figure 2.1 SAPPHIRE Concept

Figure 2. 2 Stacked Module Concept

Figure 2.3 Magnetic Passive Stabilization

Figure 2.4 OPAL with SAPPHIRE shown inBackground


2.4 The Second Satellite OPAL

The second SSDL satellite called OPAL6 was startedin April 1995. The initial major payload concept forOPAL was to demonstrate Mother-Daughter satellitetechnology where a mother satellite would carry anumber of smaller satellites internally, then when inorbit could be commanded to release these Daughtersatellites. By a general size convention, the StanfordSAPPHIRE and OPAL are considered microsatellites(> 10kg, <100kg); the Daughter satellites areconsidered picosatellites or picosats (>0.1kg, ~<1kg).

An opportunity was found to launch the OPAL inearly 1998 with sponsorship from DARPA and incooperation to launch some Picosatellites for TheAerospace Corporation. This launch was on a newAir Force launch vehicle called the Minotaur (alsoOrbital-Suborbital Platform - OSP), which used thefirst two stages of a Minuteman II missile and the lasttwo stages of the Orbital Sciences CorporationPegasus launch vehicle

The OPAL (Orbiting Automated Picosat Launcher)final design ended up with two of the launcher typesshown in Figure 2.5. OPAL carried six picosats, fourwere 4x3x1 inch and 2 were 8x3x1 inch. The finalconfiguration of the launcher tubes is shown inFigure 2.6 while one picosat is being loaded. Theother two experiments were an experimentalmagnetometer being tested for the Gravity Probe-Bproject at Stanford University and several sets ofcommercial accelerometers being tested for JPL.

The launch had delays from September 1999 toJanuary 2000. OPAL was mounted on the multiplepayload adapter on the Air Force Minotaur atVandenberg AFB, CA where final check out isperformed in late December 1999. Initial launch wasscheduled for September 1999.

OPAL was successfully launched on January 22,2000 from Vandenberg AFB, CA on the Minotaur.OPAL was initially operated from the SRI large 30-meter antenna on the Stanford. The picosats werereleased about one week after the launch via radiocommands from the ground station. The AerospacePicosatellites operated for several days on batteriesand successfully accomplished their missionobjective.

One of the smaller picosats called Stensat was builtby a group of amateur radio operators from theWashington D.C. area. It was the only picosat withsolar cells and secondary batteries. No long-term

response was received from it after launch. Studentsat Santa Clara University built three other picosats.No response was heard from them either.

OPAL successfully accomplished its mission bylaunching the picosatellites and testing the secondaryexperiments. Solar panel current data is shown inFigure 2.7 before an after picosat launch. Initialtemperature excursions are shown in Figure 2.8.OPAL continues to operate at the time of this paper.

Figure 2.5 OPAL launcher concept

Figure 2.6 Loading a Picosat into OPAL


Figure 2.7 OPAL solar panel currents before after picosat launch

Figure 2.8 OPAL temperatures before and after picosat launch


3. CUBESAT DEVELOPMENT PROGRAM

After the experience gained with OPAL and thelaunch of the picosatellites, it was determined thatpicosatellites could be used for many useful spacemissions as well as a practical space educationprogram.

The nominal OPAL picosatellite was 4x3x1 inch. Toget sufficient power from body mounted solar cells, itwas determined that a better size would be a cubethat had enough surface area to be able to generate atleast two watts with the present state-of-the-art solarcells. This was found to be practical with the new25% efficient triple-junction GaAs cells on a surfaceof 3.5x3.5 inches.

It was found that the OPAL tube launcher providedpractical means of launching picosats since it did notrequire a direct interface for each picosatellite. Thepicosatellite was held by four corners with two perlaunch tube. From this experience, it was decided todesign a cube that was 4 inches per side. Thisallowed room for the solar panels and room tocontain the cube on some rails in a launcher tube.With each cube being 4 inches long, it was decided todo an initial design for a launch tube that would hold3 cubes. Thus the new picosatellite called CubeSatand launcher tube was conceived.

Since the students at Stanford University were alloccupied building microsatellites, it was decided toform a collaborative relationship between Stanfordand California Polytechnic State University (CalPoly). Cal Poly also had a reputation as a veryhands-on undergraduate engineering university. Thiscollaborative effort was established in the fall of1999 where Cal Poly would complete the launcherdesign, build a prototype and evaluate forimprovements. The first prototype of this launch andmass model CubeSats is shown in Figure 3.1

The CubeSat program was announced to many of theorganizations, educational and amateur groups thatwere interested in building low cost picosatellites.This paper has descriptions of the launcher from CalPoly and CubeSat developments from an amateurradio group in Washington D.C. called The StensatGroup, University of Tokyo, and Dartmouth College.There are now approximately 20 organizations worldwide that are beginning development of the CubeSatpicosatellite.

Figure 3.1 Prototype CubeSat Launcher andMass CubeSat Models

4. PARTICIPANTS IN CUBESAT PROGRAM

4.1 The Stensat Group

4.1.1 Introduction

The Stensat Team is a group of engineers includingamateur radio operators who came together to designa satellite because we are engineers and have nothingbetter to do. We wanted an unusual hobby andfigured this can’t be too strange without the risk ofbeing put in a mental institute. Kevin Doherty, HankHeidt, Jim Bobbus, and Dave Nemai made up theoriginal group for the first Stensat amateur satellite.Joining the CubeSat program is Ivan Galysh and GilDutchover from the Naval Research Laboratory.Each member brings a unique skill to the group andhas the great desire to design and launch a successfulsatellite.

We were asked to develop a standardized genericvehicle. The purpose is to design a bus once andallow users to concentrate on the payload and not thespacecraft design. The CubeSat design will provide astandard bus to allow simple and easy control of thepayload while allowing flexibility in the design of thepayload.


4.1.2 CubeSat Design

The basic design of the satellite is comprised of acube structure with a stack of circuit boards inside.Each face of the satellite will be covered with solarcells. The center of the satellite will have tworechargeable batteries. The batteries split thefunctions of the satellite into two parts. One half ofthe satellite contains the satellite computer,communications electronics, and attitude controlsystem. The other half is available for a payload.

The CubeSat design is influenced by targetcharacteristics defined by the group and constraintsplaced by the CubeSat Program. The CubeSatprogram constraints are the dimensions and mass.The group defines all other design characteristics.

The CubeSat program put a couple of constraints onthe design of the satellite. The first constraint was thedimensions to be a ten-centimeter cube. The cubealso requires having rub rails for deployment out ofthe CubeSat launcher. The other requirement is themass of the satellite. The current constraint is onekilogram. This constraint imposes significantlimitations and challenges in designing the structureof the satellite. A 10-centimeter cube of water has amass of one kilogram. Materials used in the satellitestructure have a significantly higher specific densitythan water. This limitation requires a creativeapproach to designing the structure. A 10-centimetercube will have a large empty volume inside. Thecurrent CubeSat design has a mass of about 800grams leaving 200 grams for the payload.

There are several group defined target characteristicsinfluencing the design of the CubeSat. Thereproduction of the CubeSat is to be less than onethousand dollars. The design is to be simple and usestandard commercial components. The criticalcomponents selected should have some radiation dataassociated with it. For example, Motorola using theMOSAIC process manufactures the transmitter andreceiver integrated circuits. This process is known toproduce radiation tolerant devices.

Another characteristic includes low powerconsumption. Available power is to be 1 watt. Busmean power is specified to be about 250 milliwatts.Redundancy is designed into portions of the powersystem. Most of the redundancy lies in the solar cellcircuitry and power distribution. Each face of thesatellite contains eight solar cells. Four solar cells areconnected in series. The two series of solar cells on

each face are connected in parallel. If a solar cell failson a face, only four solar cells are lost. The other fourstill contribute to the satellite power bus. Each face isconnected in parallel. The batteries are alsoredundant. Circuitry isolates the batteries from eachother while allowing them to contribute to the buspower.

To minimize components and structural mass, somecomponents have multiple purposes. The solar cellswill provide power and also be used as sun sensors.The printed circuit boards with the solar cells alsohave a coil pattern on the opposite side for use as themagnetorquer coils. The circuit board stackcontaining the satellite processor, communications,and payload will also be structural members.

Payload services are designed to provide flexibility inoperating the payload. Communications and controlfrom the satellite processor include an IIC bus anddigital control signals. The IIC bus was developed byPhilips Semiconductor and provides a simple serialinterface to the payload. Four digital bidirectionallines are available to the payload. Two analog todigital converter inputs are made available to thepayload. A power control signal is provided to allowthe satellite processor to power down the payload. Apayload reset signal is also provided. The receiveraudio signal is provided to the payload. This allowsthe payload to monitor received signals. The user cancommunicate directly with the payload through thereceiver. The payload also has access to thetransmitter modulation signal. When allowed, thepayload can generate its own modulation fortransmission.

4.1.3 Current Design

The current CubeSat design uses some of the designfrom the original Stensat picosatellite. The CubeSatwill use the same uplink and downlink frequenciesand modulations. The receiver uses a MotorolaMC13136 single chip receiver. It is designed toreceive on the 2-meter amateur radio band. The gainof the receiver is about 100 dBm. It can support FSKmodulation bit rates up to 9600 baud. The receiveroperates at a voltage range from 3 volts to 5 volts. Itconsumes a mere 15 milliamperes.

The transmitter uses the Motorola MC13176 singlechip transmitter. It operates in the 70-centimeteramateur radio band. It can support up to 9600-baudFSK modulation bit rates. The transmitter operates at3 volts. It is connected to an RF Microdevices


RF2104 single chop amplifier. At 3 volts, it canoutput up to .5 watts consuming 300 milliamperes.

The integrated circuits used in the receiver andtransmitter are radiation tolerant. Based on data andtesting from other satellite programs, the Motoroladevices are manufactured using the MOSAICprocess. This process happens to be very radiationtolerant. The RF Microdevices RF2104 amplifieruses bipolar-junction technology and is also fairlyradiation tolerant. These devices have been used inprevious commercial and military satellites.

The antennas are deployable antennas mounted onthe surface of the satellite. It uses the rub rails in thelauncher for deployment. The antennas areconstructed on a piece of printed circuit board as acopper trace. There are four such circuit boards, twofor the uplink and two for the downlink.

A typical amateur radio satellite ground station canbe used to communicate with the CubeSat. Allmodulation techniques and frequencies are typical.

The power system consists of the solar cells,batteries, and the power distribution and controlcircuitry. Each face of the satellite contains eightsolar cells. The solar cells are grouped into two setsof four solar cells. Each set of solar cells is connectedin series. The two sets on each face are connectedtogether in parallel. They are isolated with schotkydiodes. If one series fails, the other series can still beused. Each face is connected in parallel. Thisconfiguration maximizes redundancy and maximizespower efficiency for the circuitry.

The batteries selected are lithium-ion rechargeablecells. The cells need special handling. They operatenormally between 3.1 volts and 4.2 volts. If the cellsare discharged too much or overcharged, the cellsmay explode. The power circuitry contains circuitryto maintain safe operation of the cells. The powerconverters are charge pump circuits. The power busis unregulated. The voltage can range from 3 volts to4 volts.

The satellite controller contains a single chipcomputer. The processor selected is a PIC controllerfrom Microchip. It is a PIC16C77. The PIC containsan eight channel eight-bit analog to digital converter.It also contains several digital ports. The controlleraccepts commands from a DTMF decoder using adual tone sequence. The controller also generatesAX.25 telemetry packets that are APRS compatible.An MX614 FSK transceiver is used to generate the

AFSK tones. The telemetry packet containsenvironmental sensor data, payload sensor data, andoperation status. The controller monitors itsenvironment with the analog to digital converter. Theinputs of the analog to digital converter are connectedto a temperature sensor, bus voltage sensor, buscurrent sensor, and two payload analog signals. Thefour payload digital signals are connected directly tothe controller. Each digital signal can be commandedto operate as an input or an output. The state anddirection of each signal is included in the telemetrypacket.

The satellite controller board contains a latchupdetection circuit. Being in orbit, the processor isexposed to high energy particles. Eventually, theprocessor will be hit by a high energy particle andhave one or more transistors latched in an improperstate. This can cause the processor to draw a largeamount of current. The latchup detection circuit isdesigned to detect a sudden increase in current andcycle the power to the processor. Cycling powerclears up the latchup. Not all high energy particlescause a latchup. Some will cause a bit flip called asingle event upset (SEU). An SEU can cause data tobe corrupted or program execution error. The circuitdesigned uses a MAX890L high-side current switch.The MAX890L allows the current limit to be set withan external resistor. When the current exceeds thelimit, the device will cycle the power. TheMAX890L is used to detect current increases greaterthan 200 milliamperes. Even though the PICprocessor consumes less than 40 milliamperes duringnormal operations, the total dose radiation effectsneed to be considered. As the processor is exposed toradiation, the total dose accumulation will cause theprocessor to increase its power consumption overtime. Total dose radiation testing showed that theprocessor consumed 180 milliamperes at its end oflife. In the case that a small latchup occurs where thecurrent increase is less than the limit, the voltagefrom the MAX471 current sensor is monitored forchanges. It is possible for a latchup to occur in theprocessor and the processor to still operate. Softwarewould be used to detect the smaller current changes.If the current change is determined to be a latchup,the processor can stop pinging the external watchdogtimer. The watchdog timer is a MAX824. If it is notpinged within 1.6 seconds, it will cycle the powerwith a MOSFET transistor. If an SEU causes data orprogram execution to be corrupted, the processor willstop pinging the watchdog timer. All of these devicesare connected in series. The MAX890L is connectedto the power bus. Next, the MAX471 current sensoris connected to the output of the MAX890L. The


MOSFET transistor is connected to the output of theMAX471. The two Maxim parts can support of to 1ampere of current. The power through each Maximpart is through a very low impedance resistor. Thevoltage drop is insignificant. The combination ofhardware and software provide a robust latchup andSEU detection system.

The PIC controller underwent radiation testing at theNaval Research Laboratory. Exposing the die to alaser performed Latchup and SEU testing. The laseris used to approximate high energy particles. Latchupwas determined to occur at about 17 MeV. Whenexposing certain parts of the die, the processorcontinued to operate during a latchup event. Themaximum current measured during a latchup was 600milliamperes. The PIC controller was never damagedby a latchup. The SEUs occurred at about 2 to 3MeV. It took little energy to flip bits. Exposing theEEPROM to the laser actually induced programexecution errors. None of the EEPROM bits could beflipped by laser exposure. The maximum energylevel exposed was about 300 MeV. Even at this highenergy level, no damage was induced. When thewatchdog timer was exposed to the laser, it stoppedworking. It was not able to reboot the processor. Anexternal watchdog timer is needed. For total dosetesting, the PIC controller was exposed to an x-raysource. The x-ray source was calibrated so that itcould be correlated to a cobalt-60 radiation source.The PIC controller failed after being exposed to 20Krads. The failure was due to the EEPROM being setto all zeros. The PIC controller could bereprogrammed. For a second exposure, the PICcontroller was exposed to and additional 15 Kradsbefore complete failure. The PIC controller couldn’tbe reprogrammed after the second exposure.

The first payload will be a picosat attitude controlexperiment. The payload will control the spacecraftmagentorquers. It will have magnetic field sensorsfor each axis. The processor architecture will allowfor software uploads. The processor will need to befast enough to calculate orbital position from timeand uploaded keplarian elements. At this time, one ofthe newer faster versions of the 80C52 eight bitprocessors has been selected. It can access externalmemory and peripherals. If the experiment issuccessful, then a more compact version of the80C52 can be used. Radiation hardened versions ofthe 80C52 do exist for those who need longerduration missions. Subsequent satellite designs willincorporate the attitude control on the satellitecontroller.

4.1.4 Satellite Structure Evolution

When the group started designing the CubeSat, thegroup went through a few design evolutions. Theinitial design specification early in the program wasfor a 3.5 inch cube. The design was highly modular.There are only two different types of panels. All ofthe side panels are identical. Each panel is made of1/8-inch aluminum. The center area is machined outto allow for mounting of a solar panel andmagnetorquer coil. The inner sides of the panels haverails machined to allow mounting of the circuitboards. The top and bottom panels are designed todeploy antennas. The antennas are constructed ofpiano wire and wrap around a racetrack machined into the top and bottom panel. A solar panel can bemounted on the panels. The panels also provideshielding for the RF electronics. The RF circuit boardmounts to the panel. The components are orientedinto the panel cavity. The circuit boards are stackedand can slide into a partially assembled structureusing the rails on the side panels.

Assembly is simple. The RF circuit board hasconnectors for each face. The connectors are used toconnect the solar cells, magentorquers, and sensors tothe bus. The side panels contain the matingconnectors. The side panels are attached to the topand bottom. The connectors help in alignment andprovide the snap together assembly. Assemblingthree sides to the top and bottom panels allows thecircuit board stack with battery compartment to beslid into the structure. The last side panel is thensnapped in and all screws can be inserted.

4.2 Dartmouth College

4.2.1 Introduction

DARTSat is a CubeSat being designed and built bystudents and faculty at the Thayer School ofEngineering at Dartmouth College. This program isto take a modular approach to educational andresearch microsatellite design. Professor August S.Moore directs this program.

An initial structure design was based on the 3.5-inchcube. The main difference is increasing the size to a10-centimeter cube and adding rub rails. The 10 cmcube meets the CubeSat structure requirements. Thepanels were designed for solid panels 1/8 inch thick.Then the problem with the design was that the massexceeded the limit. The structure had a mass greaterthan 1.5 kilograms.


The latest design uses minimal aluminum. Fourslotted posts are connected together with thin sheetsof aluminum. The assembly requires epoxy. Noscrews are used. The circuit boards contribute to thestructural strength. The battery is not mounted in anenclosure. It is epoxied between two circuit boards.

4.2.2 Standardized Design

The Stanford / Calpoly CubeSat program haspresented a unique opportunity to develop a modulararchitecture for extremely small educational andresearch spacecraft. The vision is to develop astandardized satellite bus that provides basic power,control, and communications functionality to one ortwo miniaturized payloads. These payloads areconstructed as stand-alone components withstandardized interfacing that will allow them to be“snap-in” additions to the DARTSat bus. Servicemodules are manufactured and subsequently pairedwith unique payloads as specific launch opportunitiesare realized.

The first generation DARTSat is a testbed for thisconcept and will begin an ongoing program ofsatellite development and space-based research atDartmouth College’s Thayer School of Engineering.

4.2.3 CubeSat Design

The approximately 4-inch cube specified by theCubeSat program defines the overall configuration ofDARTSat. Of the 61 cubic inches available in theconfiguration, 75% are dedicated to the satellite busand structure. A payload bay, designed to accepteither several PCBs or a component box measuring 3by 3 by 1.5 inches, is situated on one side of the cube.Photovoltaic cells are installed on five of the six sidesof DARTSat. The side not covered is optimized as aradiator to regulate the satellite’s internal thermalenvironment and act as the deployment point for thesatellite’s four antennas. DARTSat is passivelyoriented to the earth’s magnetic field to provideconsistent pointing of communications antennas,experiment sensors, and photovoltaic panels.

DARTSat’s highly integrated MicroprocessorControl Unit (MCU) is extremely compact andoccupies a single PCB that is only 2.75 inches square.It is designed around the Intel 8051 family ofmicrocontrollers. The microcontroller’s firmwareoperating system performs all of the satellites majorfunctions, including burst transmission ofexperimental data, power management, and mode

switching for the satellite’s transmitter. Groundcontrol of microprocessor functions is accomplishedthrough simple strings of “touch-tones” interpretedby a Dial-Tone MultiFunction (DTMF) decoder. A9800 bps packet modem is installed to transfer bothresearch and satellite system data via groundstationsequipped with standard HAM radio packet datareceiving capability. And a simple Power Allocationand Monitoring (PAM) subsystem, commanded bythe MCU, will provide real-time control of the powerconsumption of individual components andsubsystems. Electrical power is distributed on a 3.6Vbus. Two, 3.6V Li-Ion battery packs are used toprovide redundancy and buffer the power generatedby the external GaAs photovoltaic panels.

Satellite communications is accomplished throughFM Amateur Radio Frequencies. Use of twofrequencies, the144 MHz band for transmitting and440 MHz band for receiving, maintains flexibility,decreases transceiver complexity, and allows thesatellite to be used as a HAM radio repeater.Deployable dipole antennas, one for each band,provide fairly efficient transmission strength and thepassive pointing provided by the earth’s magneticfield insures good ground reception as the satellitepasses over North America.

4.2.4 Conclusion

The DARTSat team has focused its efforts on thedevelopment of the satellite service bus. Payloads forDARTSat are under development here at Dartmouthand at other institutions. These include plasma andauroral mapping experiments and unique GPSapplications. Because the extremely modular natureof DARTSat allows for payloads to be conceived andproduced independently from the satellite services, awide variety of payloads from a large number ofsources are possible.

4.3 University of Tokyo

4.3.1 Introduction

Intelligent Space Systems Laboratory (ISSL), alaboratory directed by Prof. Nakasuka in theDepartment of Aeronautics and Astronautics,University of Tokyo, has been studying a largemembrane space structure for several years with anintention to apply it to huge solar cells,communication antenna, debris catcher or othermissions7-9. One way to deploy such large membranein space is to use centrifugal force. The mission of


ISSL's CUBESAT is to experiment the deployment oflarge thin film solar cells using centrifugal forcegenerated by spinning the satellite main body. Figure4.3.1 shows the image of the satellite with itsmembrane deployed. The results of the experimentswill be useful for studying the way of deploying ahuge membrane as well as finding out a simple wayto provide large power to small satellites withoutmuch complicated systems.

Figure.4.3.1 Image of Univ. of Tokyo CUBESAT on orbit withMembrane Deployed

In the CUBESAT mission, we specified the followingfour success levels, which also show the missionscenario.Level 1) Educational objectives to experiencesatellite design, fabrication, testing, launch, operationand analysis of results.Level 2) To receive some beacons from the satellite.This success means that the satellite could endure theenvironment of launch, and that the subsystemsneeded for downlink communication work properlyin space.Level 3) Successful experiment of deployment ofmembrane and receive of the experiment data byusing uplink command. All the operation until thislevel is performed using the power loaded on thebattery before launch.Level 4) The solar cell on the membrane suppliespower to the battery and the satellite will survive forsome period.

4.3.2 Brief Description of the Satellite andSubsystems

4.3.2.1 Weight and Power Budget

The design is under way and so some parts aresubject to change.

4.3.2.2 Membrane

Figure 4.3.2 shows the configuration of themembrane. Total 120 solar cells are glued to somepart of the Kapton membrane, which will produce27.9 W when the solar angle is maximum. The sizeof the membrane after deployment is 880 mm by 880mm, and it is before deployment folded around thecubic body of the satellite.

Several experiments to find out the requiredrotational speed to deploy the membrane successfullyhave been performed, which indicated that therotational speed of 20 rpm is required to deploy themembrane from the folded position, and that 5 rpm isrequired to keep the shape of the membrane afterdeployment.

Figure 4.3.2 Configuration of Membrane (afterdeployment)

4.3.3 Schedule

Figure 4.3.3 shows the current schedule.


Figure 4.3.3 Schedule of University of TokyoCubeSat


4.4 California Polytechnic Institute

4.4.1 Introductions

One of the main objectives of the CubeSat program isthe development of a new class of standardizedpicosatellites. The CubeSats' size and mass standards,100 mm cube and 1 kg respectively, are developed asan extension of the picosatellites deployed byStanford's OPAL spacecraft. The CubeSats'dimensions are large enough to provide significantpower through the use of body mounted solar cells.Moreover, the CubeSats' mass is large enough tocarry a significant payload given currentdevelopments in small and efficient sensors,processors and communications equipment.

Beyond the initial form factor and mass of theCubeSats, additional standard features are determinedby the need to interface the spacecraft with a standarddeployment system. The CubeSat deployer mustsatisfy a number of requirements:- The deployer must protect the launch vehicle andprimary payload from any mechanical, electrical orelectromagnetic interference from the CubeSats evenin the event of a catastrophic CubeSat failure.- The CubeSats must be released from the deployerwith minimum spin and a low probability of collisionwith the launch vehicle or other CubeSats.- The deployer must have the ability to interface witha variety of launch vehicles with minimummodifications and with no changes to the CubeSatstandard.- The mass of the deployer should be kept to aminimum.- The deployer should incorporate a modular designthat allows different numbers of CubeSats to belaunched on any given mission.- The resulting CubeSat standard should be easilymanufactured without using exotic materials andexpensive construction techniques

The Poly Picosatellite Orbital Deployer (P-POD)shown in Figure 4.4.1 is the result of a lengthydevelopment process by Cal Poly students to satisfythe requirements above. The tube design produces areliable linear course for the CubeSats withoutsignificant spin. This deployment method wassuccessfully demonstrated in the Opal mission. Thetube provides an enclosure strong enough to handlethe structural failure of one of the CubeSats whileproviding a Faraday cage to protect the primarypayload. During the deployment sequence theCubeSats ride on rails built into the corners of thetube (see Figure 4.4.1) and a simple spring provides

Figure 4.4.1: P-POD releasing CubeSats with raildetail.

the force to push the CubeSats out of the deployerwith a linear velocity of approximately 0.3m/s. Deployment is initiated by the release of the P-POD'sspring loaded door using a G & H Technologies cablerelease actuator. The standard P-POD deployercontains 3 CubeSats although the design could belengthened to fit a larger number of CubeSats. Inaddition, the P-POD's design allows a number ofdeployers to be mounted together on a launch vehicleas shown in Figure 4.4.2. The P-POD is constructedusing 7075-T6 Aluminum due to this material's highstrength, ease of manufacture and relative low cost.The deployer is designed to sustain 15g loadsresulting in a total P-POD mass of approximately 1.5kg.

Given the P-POD's design, a final set of standards forthe CubeSats can be developed. A CubeSatspecification drawing is shown in Figure 4.4.3. Inaddition to the basic size and mass requirementssome standard requirements are introduced by thedeployer.

* The CubeSat needs 8.5 mm clearance on the fourside edges, which will be used to slide along theinternal rails of the deployer (Figure 4.4.1). * Eight 7 mm standoffs on the top and bottom facesof the CubeSat are required to provide separationbetween CubeSats. * On top of each side, excluding space for the railsand standoffs, an additional 6.5 mm space isavailable to accommodate solar panels, antennas, orother components extending beyond the 100mmlimit.* A minimum of one kill switch is required in thestandoffs on the top plate (Figure 4.4.3) to insure thatnone of the CubeSats are active during launch. Along with the kill switches there is also arequirement for a "remove before flight" pin, todeactivate the CubeSat during shipping and loading.


An optional data port can be included in the design inorder to complete last minute check or to chargeinternal batteries after the CubeSat is loaded into thedeployer.

Figure 4.4.2 Four P-POD's in a 2x2 configuration

4.4.2 PolySat: Cal Ploy’s Prototype CubeSat

The Cal Poly prototype CubeSat, PolySat, meets thedeployer's constraints for size, mass, shape, andinterface. The purpose of our prototype is twofold:first, to validate the deployer standard design andsecond, to demonstrate that CubeSats provide aviable platform for basic experiments in a space

environment. PolySat is a minimum configurationsatellite, consisting of a simple set of componentsrequired to demonstrate low-Earth orbit operation.

PolySat's structure is consistent with the deployer'sstandard. A 100 mm cube has been designed with therequired guide rails and interface for the deployer. The structure is composed of six individual panels ofaluminum 7075-T6, totaling a mass of 0.2 kg andstrong enough to survive launch loads. Mount pointsfor antennas on the exterior and various internalcomponents are provided.

The communications system consists of a downlinktransmitter, an uplink receiver, and independentantennas for each. Downlink is provided by amodified on-board Alinco DJ-C4T 440 MHz amateurradio transmitter. This is commercially available for$70 retail, and offers 300 mW output, at a cost of1.11 W when transmitting. It is very robust andcompact. Further, it generates very little heat inoperation and has low idling power requirements.Both our own requirements and FCC amateurregulations demand an independent uplink receivercapable of providing a minimum of an "off"command for the downlink transmitter. Requirements analysis revealed that commonlyavailable amateur radio receiving equipment wouldlikely suffer from overload and seriousintermodulation distortion (IMD) from strong earthgenerated signals on nearby frequencies was solvedby using a MICRF004.


Figure 4.4.3 CubeSat specification drawing


This data receiver on a chip (commonly used forgarage door opener receivers) is from Micrel, Inc. This solution is quite cheap to implement at about$30. It is easily modified to receive in the 144 MHzamateur radio band with extremely low powerrequirements and low sensitivity. Low sensitivity,usually a disadvantage, affords the CubeSat uplinkreceiver relative immunity to spurious signals. Compensation is easily provided on earth byincreased amplification and antenna gain withcommonly available amateur radio components. Thisreceiver choice enables future enhancements for theuplink channel such as reset command or even simplereprogramming of the onboard processor.

Independent dipole antennas mounted on one face ofthe box provide the downlink and uplink capability. The antenna design mirrors the technique used onOPAL, namely metal "measuring tape" available atany hardware store. It is flexible, holds its shape welland has served as adequate material on previousmissions. The antennas are folded down against thesatellite body with monofilament held down by ashort length of nichrome wire. Upon deployment, acurrent is passed through the nichrome wire, whichwill heat the monofilament and release the antennas.

Figure 4.4.4. The PolySat with Deployed DipoleAntennas

Amateur radio frequencies (and licenses) are used forlow cost and a readily accessed base of experience. In addition to inexpensive, commonly availableequipment, there exists a large network of amateurradio operators with well-equipped ground stationsavailable to assist in our mission. This is atremendous advantage for a small startup group ofundergraduate students.

Command is provided by an Atmel BasicX PICmicrocontroller running custom software, written forthe project. This chip provides 8 digital and 8 analoglines for I/O. The processor activates the transmitterat required intervals for downlink tasks. The FCCrequired identification of the transmitter is given inMorse code using tone modulated FM under the callsign "N6CP" (Cal Poly's Amateur Radio Club)followed by mission specific data. The data isencoded using dual tone multi frequency (DTMF)tones by the processor. The transmitter is keyed bythe processor at specified intervals while the encodeddata is passed to the audio input for transmission. The use of DTMF encoding involves a relativelyslow rate of data transfer but increases the simplicityand reliability of our link. This can be important ifwe unexpectedly lose gain or experience increasednoise on the frequency. Currently, sensors beingconsidered for PolySat include thermistors tomeasure the temperature of the structure and keycomponents as well as a voltage sensor for thebatteries.

Power is provided by two on-board battery packsconsisting of two lithium-ion batteries each. Step-down converters are used to provide 5V to thecomputer and receiver and 3.7V for the transmitter. This battery pack provides 209 watt-hours using fourcells of 116g each. This represents approximately150 hours of transmitting time.

PolySat uses passive thermal control, using coatingsand paints to control heat radiation and absorption. Operating range is expected to be -40ß C to 70ß C.

As described, the PolySat spacecraft is well under the1kg mass constraint. Since the transmitter operatesonly for small periods of time at specified intervals, itis expected to maintain operation over a lifespan ofseveral weeks.


Figure 4.4.5 Disassembled CubeSat ShowingBatteries and Electronics Tray

4.4.3 Future Advances

The main objective of the initial CubeSat missionwill be to validate the design of the standardizeddeployer. Once the deployer is space qualified it willbecome a new option to place small payloads inspace with low development time and cost. Inaddition, initial CubeSat missions will validate thePolySat design and will qualify a number ofcomponents that the Cal Poly team plans to use infuture CubeSats. In particular, the Cal Poly team isinterested in validating the spacecraft structure, theantenna deployment system, the communicationspackage and the spacecraft thermal models. Spacequalification of these components will facilitate thedevelopment of more sophisticated CubeSats in thefuture.

The Cal Poly team is currently working on thedevelopment of a solar powered CubeSat. Using thelatest solar cell technology, body-mounted solarpanels should provide around 1 W of continuouspower to the spacecraft. A solar powered CubeSatwould require a much smaller battery and anadditional 0.2 to 0.25 kg of payload could beincorporated. New components that are currentlybeing considered are a more powerful computer withmore storage capacity, a small CCD camera and asmall magnet to provide some attitude controlcapabilities, in addition the Cal Poly team is seekingcommercial payloads to be flown in future CubeSatmissions.

4.4.4 Acknowledgements

Bob Twiggs at Stanford University SSDL providedadvice, knowledge, guidance, and funding in thedevelopment of Cal Poly's CubeSat program. EdEnglish (W6WYQ) and Cliff Buttschardt (K7RR),local amateur radio operators affiliated with AMSAT,provided technical assistance and materials. TRW,Inc, provided much needed funding for the project. G&H Technologies provided much needed hardwarefor the project. Palm, Inc. generously provided freehardware. Alinco Electronics provided technicalassistance and products. CAD Research Center, CalPoly, provided support for student projects.

5. LAUNCH OPPORTUNITIES AND COSTS

The launcher tube, P-POD, being developed by CalPoly to hold three CubeSats is designed to beattached to many different launch vehicles. The P-POD launcher will be rectangular tube about 4.5inches square by 15 inches long and weigh about 5kg. It would require only a small power activationsignal from the last stage of the launch vehicle toactivate the release mechanism and open the launcherdoor to release the CubeSats. Since the launchercompletely contains the picosatellites, can beattached in a very small space and is lightweight,many launch vehicles can accommodate thissecondary payload. Multiple P-PODs could beattached to one launch vehicle for secondarypayloads.

There are several flight opportunities for CubeSats inthe near future. Two launch opportunities are nowavailable from with Thiokol Corporation in a jointVenture with Kosmotras on the converted RussianSS-18 called the Dnepr from the Russian launch siteat Baikonour. The first flight is scheduled for March2001 with a second one in late 2001. Theseopportunities are being coordinated through One StopSatellite Services in Ogden, Utah. It is expected thatthese flights will continue to occur at least twice ayear.

The Aerospace Corporation, which was a partner onthe OPAL projected and provided picosats for launchon that mission, is continuing picosat developmentwith launches on the OSP-II (second Minotaur) fromVandenberg AFB, CA in 2001. Additional USlaunch opportunities are quite likely on many Boeingand Lockheed Martin launch vehicles.


The Air Force OSP-III (third Minotaur) in nowscheduled for launch in late 2002 in which as manyas 15 CubeSats are proposed to be launched from it.

5.1 Launch Costs and Procedure

The estimated launch cost per CubeSat that weightsone kilogram or less is $30,000. This is based on theknown launch and integration cost using theKosmotras Dnepr. The present arrangements beingmade for launches on the Dnepr is a collaborativeeffort between Stanford University, Cal Poly and OneStop Satellite Services in Ogden, Utah. Contractualarrangements will be made with Stanford University;Cal Poly will provide P-POD test fixture to thecustomer, provide the flight P-POD, integrate theCubeSat into the P-POD upon delivery to Cal Poly atSan Luis Obispo, CA, perform final thermal andvacuum testing; then the P-POD will be shipped toOne Stop Satellite Services at Ogden, Utah. OneStop Satellite Services will then be responsible for allexport licensing, shipping to Kosmotras andintegration onto the Dnepr.

6. CONCLUSIONS

The OPAL program demonstrated that a low costlaunch system could be used to launch picosatellites.The success of The Aerospace Corporationpicosatellites demonstrated the first use of thesepicosatellites. The new picosatellite, CubeSat, nowproposed as a standard that can be launched with thelauncher tube developed by Cal Poly, the P-POD willprovide low cost opportunities for a new era in spaceexperimentation.

These new CubeSats and the low cost make itpractical for even universities and private groupssuch as amateur radio clubs to have access to space.The future of these space devices based on theCubeSat design will now depend upon howinnovative the science and general community canbe.

7. ACKNOWLEDGEMENTS

The initiation of the mother/daughter picosatellitework started at Stanford was done through the workof Professor Tom Kenny in the Stanford mechanicalengineering department and Mr. Jim Randolf at theJet Propulsion Laboratory. The final push to meetOPAL’s mission was due to Ernie Robinson at The

Aerospace Corporation and Al Pisano, the director ofthe MEMS activity at DARPA.

Many students at Stanford worked very long and hardat completing SAPPHIRE, which was thepredecessor to OPAL, and OPAL. Foremost of thesestudents supporting SAPPHIRE over several yearswere Mike Swartwout and Chris Kitts. The OPALeffort was led initially by Brian Engberg, ClemTillier and Carlos Niederstrasser and finally by JamieCutler and Greg Hutchins. These two projects hadmore than one hundred students work on them over afive year period. SSDL has outstanding mentorsgroup to support students in their work with muchvolunteer effort from John Ellis, Lars Karlsson, DickKors, Ron Ross, Richard Anderson and David Josephto name a few.

Boeing, DARPA, NASA Ames Research Center,NASA Langley and Jet Propulsion Laboratoryprovided research support for SSDL. LockheedMartin, Space Systems/Loral, Deskin Research andThe Aerospace Corporation provided additionalfinancial and facilities support.

The CubeSat program as done in collaboration withProf. Jordi Puig-Suari at Cal Poly demonstrates themotivation that a space program can bring toeducation. The Cal Poly students like Ryan Connellyand Jeremy Schoos are examples of students excited,enthusiastic and motivated by their education. Thesestudents are the most valuable output of theseprograms.

8. REFERENCES

1 “Initial Developments in the Stanford SQUIRTProgram”, Christopher A. Kitts and Robert J.Twiggs, EUROPTO, European Symposium onSatellite Remote Sensing, Rome, Italy,September 26-30, 1994.

2 “The Satellite Quick Research Testbed(SQUIRT) Program”, Christopher A. Kitts andRobert J. Twiggs, 8th Annual AIAA/USUConference on Small Satellites, Logan Utah,August 29 - September 1, 1994.

3 “Initial Developments in the Stanford SQUIRTProgram”, Christopher A. Kitts and Robert J.Twiggs, EUROPTO, European Symposium onSatellite Remote Sensing, Rome, Italy,September 26-30, 1994.


4 “The Satellite Quick Research Testbed(SQUIRT) Program”, Christopher A. Kitts andRobert J. Twiggs, 8th Annual AIAA/USUConference on Small Satellites, Logan Utah,August 29 - September 1, 1994.

5 "SAPPHIRE, A University Student BuiltSatellite for Space Experimentation", Twiggs,Robert J., and Christopher A. Kitts, November 1,1995, The AMSAT Journal,November/December 1995.

6 "Picosat Free Flying MagnetometerExperiment", Clarke, D.S., M.T. Hicks, A.M.Fitzgerald, J.J. Suchman, R. Twiggs, T.W.Kenny, and J. Randolf, June 15, 1996,Proceedings of the Tenth Annual AIAA/USUSmall Satellite Conference, September 16-19,1996. Presented by Michael T. Hicks at theTenth Annual AIAA/USU Small SatelliteConference, New Missions I Session, Logan,UT, September 16, 1996

7 Motohashi, S., Nakasuka, S., Aoki, T.,Narusawa, Y., Nagashima, R., Kawakatsu, Y.and Kinoshita, T., On-orbit Dynamics and NewControl Scheme for Large Membrane"Furoshiki" Satellite. Proceedings of 21st ISTS,98-e-20, 1998

8 Narusawa, Y., Aoki, T., Nakasuka, S.,Motohashi, S., Nagashima, R., Kawakatsu, Y.and Kinoshita, T., Behavior of MembraneStructure under Microgravity Environment,Proceedings of 21st ISTS, 98-b-16, 1998

9 Motohashi, S., Nakasuka, S., On-orbit Dynamicsand Control of Large Scaled Membrane withSatellites at its Corners, Proceedings of IFACSymposium on Aerospace Control, pp. 146-151,1998

Documents

SSC00-V-5 - Aalborg UniversitetSSC00-V-5 CubeSat: A new Generation of Picosatellite for Education and Industry Low-Cost Space Experimentation Mr. Hank Heidt 1, Prof. Jordi Puig-Suari