The NanoSat
Revolution
Robert Sheldon
(with thanks to Jim Cantrell)
NASA/MSFC/NSSTC/VP62
April 20, 2007
Overview
1.
2.
Review of Jim Cantrell’s 2-day seminar
The NanoSat Problem
• Launcher cost mismatch  complexity
• Mission Creep  mass increase
• Bad Attitude  limited functionality
3.
The Revolutionary Technology
• CCR Laser -- Telemetry
• Elliptical Orbit + Spinner
• Laser-despun Fast Spinner
4.
Future Applications
Definition of “Sat”  kg.333
0.1kg 1kg
femto
pico
10kg
nano
100kg 1000kg
micro
small
“Smart dust”
BMDO
(Don’t blame an engineer, it was marketing!)
Jim Cantrell, Space Dev Inc, 2007
Smart Dust (Femto /Atto Sats)
Smallsats—the 30,000 ft view
1985: Smallsats were once considered to be toys
Small-sat conference in Utah in 1988 brought lots of skeptics
Mainly domain of universities and small companies
1990: Smallsats grew capability in power, pointing and
computational capability
1995: Smallsats become an increasing strong force in the
marketplace
1990-2000: Several small businesses (Spectrum, Ball, DSI) grew
capability into mainstream capability
2007: Smallsats have matured as the major spacecraft providers
have all built and flown them: They are the mainstream market
2008: Nanosats have the same market appeal today that
smallsats had 20 years ago
Jim Cantrell, Space Dev Inc, 2007
Small Sat Timeline
Small Satellite Launches By Size
(Excluding Iridium and Globalstar)
50
40
Pico
30
Nano
20
Micro
10
Mini
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
0
1990
Number of Small
Satellite Launches
60
Year
NASDAQ collapse
Jim Cantrell, Space Dev Inc, 2007
The Future of Satellite Types
(Excluding Iridium and Globalstar)
Mini
What happened to the
Nano/Pico?
Micro
2000
Nano
Pico
2004
The 2000 NASDAQ
bubble in telcomm
destroyed most R&D
for 7 years.
(Due to a slight
exponent error in a
telcomm report.)
Jim Cantrell, Space Dev Inc, 2007
NASA & DoD $/kg scaled
(DoD bus mass x 2, $ x 3)
Jim Cantrell, Space Dev Inc, 2007
Cost & Launch Impedence Match
The $/kg of a satellite is roughly constant above
some fixed cost floor.
2. The $/kg launch cost has not changed in 30 years
($22/g = Au spot price) in a discretized way-secondary & tertiary payloads are less expensive.
3. For primary payloads, the smallest launcher puts a
fixed cost on a launch (for OSC ~ $20M) though
smaller rockets would extend the linearity down
to ~100kg payloads.
Insurance couples 1) & 2) impedence matched =
But we aren’t launching gold, are we?
1.
Jim Cantrell, Space Dev Inc, 2007
Develop / Cost vs. Complexity
If cost is fixed, we must
maximize productivity =
complexity.
Time is linear, cost
quadratic (N!) with
complexity.
Failures occur when cost
& time are too limited
wrt the average.
The “gold” is in the
labor. No economies of
scale.
Jim Cantrell, Space Dev Inc, 2007
Mission Creep
What is the optimum level of complexity?
•
Requires linear programming, but real limits are not “hard”, extra
time, extra money is available. So we assume a linear or power-law
function near the limits, and estimate the optimum point by the first
derivative. Here’s a BOTE calculation:
$_profit = $_return - $_cost
$_returnfn(complexity)fn(telemetry+processing)F(power)
$_cost = fn(mass)G(size)
Optimum  d$_profit/dL = 0 = dF/dL - dG/dL
Some rules of thumb:
Density ~1g/cc (thermal, even in miniaturized components)
Mass~L3 (Since density is nearly constant)
Telemetry~Power (and dish size, but antennae are light)
Power~L2 body-mounted solar panels (~massL3 for 3-axis )
Optimization Condition
Spinner
$_profit = k1L2 - k2 L3
d$_profit=(2k1-3k2 L)LdL
Optimum SizeL0=2k1/3k2
Smaller is better.
3-axis stabilized
$_profit=(k1- k2) L3
d$_profit=(k1- k2) 3L2 dL
No Optimum Size!
Bigger is always better.
Iridium,
Teledesic
SDO…
Jim Cantrell, Space Dev Inc, 2007
Low-Complexity Spacecraft
Complexity index 0-0.33
High-Complexity Spacecraft
Complexity index 0.67-1
Small payload mass (~5-10 kg)
One payload instrument
Spin or gravity-gradient stabilized
Body-fixed solar cells
(Si or GaAs)
Short design life (~6-12 mo)
Single-string design
Aluminum structures
Coarse pointing accuracy (~1-5 deg)
No propulsion, no cold-gas
Low-frequency communications
Simple helix or patch low-gain antenna
Low data rate downlink (~1-10 kbps)
Low power requirements (~50-100 W)
No deployed or articulated mechanisms
Little or no data storage
No onboard processing ("bent pipe")
Passive thermal control with
coatings, insulation, etc
Large payload mass (~300-500 kg)
Many (5-10) payload instruments
Three-axis stabilized w/ reaction wheels
Deployed sun-tracking solar panels
(multi-junction or concentrators)
Long design life (~3-6 years)
Partially or fully redundant
Composite structures
Fine pointing accuracy (~0.01-0.1 deg)
Mono- or bipropellant w/ 4-12 thrusters
High-frequency communications
Deployed high-gain parabolic antennas
High data rate downlink (Mbps)
High power requirements (~0.5-2 kW)
Deployed and/or articulated mechanisms
Solid-state data recorders (5 GByte)
Onboard processing (30 MIPS)
Active thermal control with
heat pipes, radiators, etc
telemetry
shielding
lifetime
Jim Cantrell, Space Dev Inc, 2007
Attitude = Power
•
•
•
•
The problem of decreasing telemetry with distance
requiring larger amounts of power, is pointing a
large dish
Likewise data-gathering satellites improve with
pointing, even magnetometers.
This makes attitude control one of the important
requirements driving s/c selection, and the major
drawback to spinners.
3-axis nanosats have yet to be built, which means
nanosats have low capabilities.
Perform. Figure Of Merit
Small Small
Satellite
SatelliteCapabilities
Capabilities
120
0.0904x
y = 1.2352e
100
Insurgent satellite
providers occupy this
region of market
80
Market occupied by
small university
providers and others
60
40
20
0
0
10
20
30
40
Bus Cost (M$ FY01)
50
Jim Cantrell, Space Dev Inc, 2007
NanoSat Applications
•DoD
•NASA
–GEO SSA
–Atmospheric constellations
–High value asset inspection
–Station / Shuttle inspection
–ASAT defense [offense?]
–Exploration initiatives
–Space control
–Component test-beds
–Testbed for ORS components
•IC
•Commercial ??
–Researcher testbeds
–Distributed sensing
–High value asset escort
–Communications data exfiltration
–University testbeds
–Small commercial payloads
Applications for nanosats highly dependent upon propulsion and
other bus capabilities
Jim Cantrell, Space Dev Inc, 2007
Future of Small Sats
•Satellite market has followed “disruptive” path over past 20 years
•Market disruption is continual and tending towards smaller and
more powerful spacecraft
•Smallsats are now in mainstream of market
–This has been the case for 10 years & expected to continue 5-10 years
•Microsats are emerging as capable of addressing the mainstream
market
•Nanosats are still way below mainstream market demands for
capability but offer potential for insurgent provider entry points.
•Lower costs remain elusive since the cost of components [COTS
subsystems] remains high. [No economies of scale…]
Jim Cantrell, Space Dev Inc, 2007
Summary of Nanosat Bias
1.
2.
3.
•
•
Nanosats have limited pointing
Nanosats have limited power
Nanosats have limited telemetry
We can shrink a 3-axis s/c with miniaturized
components, but we can’t increase a spinner’s
capabilities by shrinking (though we do improve
it’s profitability).
The breakthroughs have to come in pointing,
power and telemetry.
Nanosat Solution #1: CCR
(low attitude requirement)
(1996) BU Space Constellation
Why do we
need a nanosat
constellation?
+
(1996) Space Physics Problem
•
“The effectiveness of the base-funded space
physics research program has decreased over the
past decade….The long-term trend that has led
to an ever increasing reliance on large programs
has decreased the productivity of space physics
research.”
A Space Physics Paradox, NRC
•
“Small”, despite being eliminated from the
“cheaper-faster-better” triumvirate, is even more
essential for survival, not just for Space Physics,
but NASA itself.
Rob Sheldon, 1996
(1996) Magnetosphere Problem
“The leap from static pictures of averaged structures and cartoon
sequences of processes to continuous sequences of global, 3-D, synoptic
images of the magnetosphere is widely recognized to be the logical next
move in magnetospheric science….The only known way to obtain
simultaneous, spatially comprehensive information on data fields of
invisible parameters…is through simultaneous multi-point in-situ
observations.”
G. Siscoe, 96 Roadmap whitepaper
Target Spatial Resolution: Ionosphere =>Magnetosphere
• Radial =10% dRe/Re from 5-15 Re --> 12
• Azimuthal = 10% dRe/Re all 360 degrees --> 63
• Vertical (Z) = 1 Re at plasmasheet 0-5 Re --> 6
TOTAL= 12 x 63 x 6 = 4536 Satellites!!
(1996) Cost: The Bottom Line
•
•
•
Maximum “solar probe” mission = $100 M
Minimum Constellation =
100 satellites
Maximum cost/satellite =
$1 M / sat
This is a “best case $” mission but not believable.
•
Typical cost/satellite
=
$100K / sat
Assume 50M$ and 500 satellites as “typical”
•
Astounding cost/satellite =
$10 K / sat
This would be a revolution akin to the Model-T (Exactly
how many copies to achieve this economy of scale?)
(1996) Launcher Considerations
•
•
Launcher
Pegasus XL
LEO Mass
Volume
Area
Cost
550 kg
5.8m3
1.3m2
$20M
# Satellites 100
Weight
Volume
Area
Cost/sat
Delta II
5,500 kg
63 m3
7.4 m2
$100M
500
5.5 1.1 kg
58 11.6 L
130 26 cm2
$200k $40k
100
500
55
630
740
$1M
11 kg
126 L
148 cm2
$200k
(1996) Miniaturization solution?
Detectors/Sensors
YES!
CPU/DPU
YES!
Solar Panels
Little, 20%
Bus/Thrusters
Some, 50%
Telemetry
NO! - 10,000%
Multiple satellites will require even faster/wider
bandwidth in order to dump the data in 1/100th of
the time, with less mass, less size, and less
power!
(1996) Solution
•
•
•
We MUST solve the telemetry problem before we
can fly a # > 100 constellation
Bandwidth problem was faced by telcomm industry
20 years ago—undersea cables were too limited:
Solution? LASERS
We propose to communicate with the satellites using
a ground-based laser system, so that the power is on
the ground, and the modulation is in space.
(1996) Nanosatellite
•
•
•
•
•
•
•
Weight: 1kg
Size:
10cm
Power: 1 watt
DPU Motorola603
Memory: 1 Gbyte
Telemetry : 10kHz
Sensors: 3-axis Mag
• 12 pinhole CCD
• SEU calibrated RAM
(1996)
Deployment
•
•
Bus with multiple
launchers, kick motors,
3-axis stabilized- “point
& shoot” capability
Solid fuel boosters, Spin
stabilized, separation
avionics, autonomous
deployment
(1996) Ground Segment
•
•
•
•
•
DSN already saturated
3 main laser ground
stations for 24hr coverage
Radar ground station in
Antarctic for polar sat’s.
~10 Institutional (<100K$)
downlinks at interested
sites
>90% temporal + spatial
coverage of >100 satellites
CCR Nanosat Conclusions
1.
2.
3.
We solved the attitude problem with corner-cube
reflectors, which need little attitude
We solved the telemetry problem with lasers.
We solved the power problem by sidestepping it:
miniaturize to reduce power, simplify to reduce
demand, and move telemetry power to the ground
Did it work? Our proposal was funded, and after study
we concluded (1998): Not for this application.
CCR 1/R4 return! So data from s/c at 8 Re took
too much laser power. But  MagCon Mission
(1998) DARPA grant to UCB
650nm laser pointer
2 day life full duty
Attitude control??
4 corner cubes
40% hemisphere
Long life
(1998) UCB Corner Cube
Top View of the Interrogator
Quarter-wave
Plate
Filter Polarizing
Beamsplitter
CCD Camera
Lens
0.25% reflectance
on each surface
YAG Green
Laser
Beam
Expander
45o mirror
(1999) 2nd Try: UCB Monolithic
300 um
(1999) U of AZ student project
Nanosat Solution #2: Spinner
attitude control
(1999) Elliptical Orbit Telemetry
•
•
•
•
Put the satellites into elliptical orbits with
perigee < 1Re away. Store data.
Use 10m dish on the ground to track the
satellite. Dump the data in burst mode at
perigee.
No real-time tracking, 5-10 hour delay in
updating the “magnetosphere” model.
Larger dishes (e.g., DSN) can provide realtime tracking of a few targets.
(1999)Why Nanosat Magnetometers?
•
Magnetic fields must be measured in situ
• (We too wish we could do it with imaging!)
•
Magnetic fields must be measured globally
• Fields are a global, not local, effect
•
Magnetic fields must be measured simultaneously
• Otherwise space/time ambiguities destroy the image
• Siscoe http://rbsp.info/rbs/RbS/PDF/wpaper.html
•
Magnetic fields must be measured densely
• Currents and structure are “narrow” boundary features
•
The economics of 100’s of satellites => nanosats
(1999) Magnetometer Accuracy
•
Need better than 1nT sensitivity at L>6 Re
This = 1/500 = 0.2% accuracy in magnitude
And = 1/2 degree accuracy in pointing
•
Science goals:
Alfven waves, current systems, global models
•
Nice to have 1nT at L<6 Re as well
Requires 1/30,000 = .003% accuracy in magnitude!
Requires 1/2 arc-minute accuracy in pointing!
•
Can nanosats get close to these requirements?
(1999) The Attitude Fix
1/4 wave antenna and
magnetometer boom
Inside: Batteries, power
system, fluxgate mag, CPU,
ADC, attitude determination,
Cell phone transmitter
10mm
aperature
100mm
focal length
sun-sensor
SPECS:
Mass
1 kg
Power
1 Watt
Spinrate
1 rps
Size
15x6 cm
2 watt
solar-panel
15cm
(1999) 3-Axis Spinners
•
•
In traditional satellites, only 2-axes are
Fourier sampled, leading to poor resolution
along the “z-axis”.
Manfred Boehm (1996) presented and
launched an offcenter (nutating) spinner that
was able to sample all 3-axes.
(2004) MagCon
Nanosat Solution #3: Laser scan
with drag chute positioning
Nanosat Solution
•
•
•
The spinner gives a nice attitude fix, but ruins most
of the benefits of a corner reflector except a the
poles. That means we’re back to omnidirectional,
low datarate radio. Can we recover the laser
advantage?
By similar reasoning, a CCR is 1/R4, whereas a
lasers at both ends give 1/R2. Can we remotely
direct a laser toward us with robust, autonomous
control? All laser communicators are feedback
stablilized telescopes on massive 3axis platforms.
Can nanosats do this?
The answer is a Scanner on a Spinner
(1998) LaserSatCom design
$50M10 Gbps cablein-the-sky. High altitude
balloons @10km altitude
relay radio through the
cloud deck.
Electro-Magnetic
drag chute
Spherical Fresnel Lens
Toroidal
Lens
Transmitted
signal
Laser
Scanner
Detector
Received
Signal
(1999) UCB Try 3: 2D scanning
AR coated dome
lens
Steering Mirror
laser
CMOS ASIC
(1999) UCB 8mm3 laser scanner
Two 4-bit mechanical DACs
control mirror scan angles.
~6 degrees azimuth, 3 elevation
(2004) Sensor Array Positioning
Because nanosats are small, they
lack stored energy, but can use
external forces more easily.
John Barker,
U Glasgow
2004
Applications
VLBA, radio/IR telescopes,
Planetary reconnaisance,
Military warspace-IFF, higher
resolution than GPS, unjammable,
disposable, quickly deployed…
Airborne Dust
UCB 2001
U Glasgow
Mapleseed solar cell
MEMS/Hexsil/SOI
“Smart Dust”, Planetary
exploration, Military
sensors
Rocket dust
MEMS/Hexsil/SOI
1-5 cm
Controlled auto-rotator
MEMS/Hexsil/SOI
Conclusions
•
•
•
•
Constellations of > 30 satellites will include
nanosats. Heterogenous constellations have many
advantages over homogeneous.
Nanosatellites will be cheap enough to be
“disposable” by being launched in quantity to
match launcher cost.
Nanosatellites will overcome the negative bias by
solving telemetry (w/laser), attitude (scan and/or
spin) and positioning (smart diffusion) challenges.
The next decade will be the Nanosat Decade