The Microsat Way in Canada

Abstract

According to a recent survey, over fourteen microsatellites have been launched per year on average in the last twelve years. These satellites are frequently built, launched and operated by university and amateur teams. Increasingly, however, small-and microsatellites are being used to fulfill commercial, military, remote sensing, and science missions as well. In the future, the use of microsatellites to achieve high-value missions at lower cost will become mainstream, and will occupy an increasing portion of the spacecraft applications market. The paper supports these conclusions by first examining the potential of microsatellites to reduce entry barriers to new space utilization, and their impact in changing the economics of space applications. An analysis of what makes traditional space programs expensive is undertaken, with a discussion of the upward cost spiral and how space programs can become trapped by it. Various ideas are put forward to energize a discussion on how things could be done better. The paper then discusses the economic drivers underlying microsatellites. It discusses why launch cost is a prime mover of spacecraft complexity and how this effect works in reverse for micro-and nano-satellites. It presents the impact of digital electronics in enabling microsatellite missions, and how a microsatellite program can take advantage of the rapid evolution of technology. Finally, it looks at the distribution of risk and return, and presents a proposal for a program blueprint to achieve successful, high-value missions at lower overall cost and lower risk.

Full text

The Microsat Way in Canada
Peter Stibrany Kieran A. Carroll
Manager, Microsatellite Systems Manager, Space Projects
Dynacon Enterprises Limited Dynacon Enterprises Limited
Tel: (905) 672-8828 x226 Tel: (905) 672-8828 x232
Email: ps@dynacon.ca Email: kac@dynacon.ca
Abstract
According to a recent survey, over fourteen microsatellites have been launched per year on average in
the last twelve years. These satellites are frequently built, launched and operated by university and
amateur teams.
Increasingly, however, small- and microsatellites are being used to fulfill commercial, military, remote
sensing, and science missions as well. In the future, the use of microsatellites to achieve high-value
missions at lower cost will become mainstream, and will occupy an increasing portion of the spacecraft
applications market. The paper supports these conclusions by first examining the potential of
microsatellites to reduce entry barriers to new space utilization, and their impact in changing the
economics of space applications.
An analysis of what makes traditional space programs expensive is undertaken, with a discussion of the
upward cost spiral and how space programs can become trapped by it. Various ideas are put forward to
energize a discussion on how things could be done better.
The paper then discusses the economic drivers underlying microsatellites. It discusses why launch cost is
a prime mover of spacecraft complexity and how this effect works in reverse for micro- and nano-
satellites. It presents the impact of digital electronics in enabling microsatellite missions, and how a
microsatellite program can take advantage of the rapid evolution of technology. Finally, it looks at the
distribution of risk and return, and presents a proposal for a program blueprint to achieve successful,
high-value missions at lower overall cost and lower risk.
Introduction
This paper sets out thoughts on the relevance of
microsatellites to Canada’s future activities in space. It
begins with a discussion of the microsatellite
phenomenon, moves to an analysis of space mission cost
drivers and concludes by laying out a “microsat way”
that could lead to the creation of more advanced and yet
less expensive Canadian space missions.
The key motivator in doing things differently is the
inevitable advance of silicon based electronics. Where
15 years ago, little practical science or engineering could
be done with a small payload that is not the case today.
Moore’s Law has given us electronics circuits with 1,000
time more processing power per dollar, enabling
missions such as MOST to be done in a small form factor
at a low overall cost. According the SSTL’s
compendium of small and microsatellite launches, an
average of 14 microsatellites have been launched in each
of the past 12 years. (1)
The “microsat way” as promoted in this paper, is to
design space missions and an overall corporate program
that maximises our ability to offer the most advanced
microsatellite equipment offerings at the lowest cost.
The best way to achieve this in the long term is to use
multiple, independently launched spacecraft, each using
the latest available technology to simplify its design and
maximise its utility, and each of which demonstrates the
next generation of technology. In this way, practical
experienced gained through in-space lessons can be
folded back into a (very rapid) design process that allows
the maximum benefit to be provided to the end user at
least cost.

The Past
The history of microsatellite development has not been
written, though one hopes that as the group of people
who pioneered this way of building spacecraft gets older,
their impulses may lead them to record their experiences
and recount to a more general audience what happened
and when. It would be unfortunate to lose a piece of the
history of technology that serves as an important
illustration and worked example of Kuhn's paradigm shift
hypothesis.
To say that microsatellites were initially, and to some
extent still, built by amateurs is to imply that
microsatellites can be built by anyone and don't require
an experienced technical team. This is far from the truth.
More accurately, microsatellites were initially built by
teams including experienced spacecraft builders who
were disillusioned with the way of building spacecraft as
practised by the NASA and military rule books over the
decades.
Rather than follow the methods of the earlier days of
exploration, when there was no such thing as an
approved parts list, NASA and others wrote weighty and
significant books of rules that captured the lessons of the
past.
The AMSAT pioneers saw the upward cost spiral that
this approach inevitably engendered, and were frustrated
by the slowness with which such this standard approach
took up new technology.
By successfully building, launching and operating
amateur packet radio satellite repeaters and store and
forward communications satellites, they demonstrated
that more could be done with less.
From this early legacy come the sometimes contradictory
characteristics of early microsatellites:
- they are built by amateurs and student teams
- they are built by experts in their spare time
- they are small, so that they can be launched
cheaply as secondary payloads
- they are not mass efficient
- they are made with donated spare parts, Radio
Shack parts, and equipment made in the
basements of hobbyists
- they are in many ways technically sophisticated
and have pioneered new approaches to radio
communications
- they have poor attitude control
- they sometimes fail to operate to their full design
requirements
- they last on the order of 4 - 6 years in low earth
orbit, with some living considerably longer
A microsatellite project can be defined as some
alchemical combination of small size, low cost, and
perhaps also a high technical innovation quotient. The
down-side, from the point of view of more traditional
approaches to building satellites, is that there is a lot less
paperwork and documentation generated, risks are taken
(and managed) rather than avoided, and real design
authority is passed to those in the best position to exercise
it.
As a welcome side-effect, microsatellite pioneers opened
the door to the use of digital architectures and software
on spacecraft long before such technologies were
acceptable to mainstream spacecraft programs. By
making it possible for a spacecraft to carry out more of its
functions through software, they also made it possible for
people who were not traditionally spacecraft engineers to
contribute successfully to the construction and operation
of spacecraft.
Microsatellite projects have demonstrated a much better
ability than other space endeavours to stay glued to the
technology curve. As a result, the relatively modest mass
and power available on a microsatellite now make it
possible to achieve science, governmental, and
commercial missions previously only possible with larger
and much more expensive approaches. Microsatellites
are not hobby spacecraft any more.
Crossing Paths with Moore’s Law
The rate of technology evolution is very fast; some say
it's exponential. Each new technology adds to the
possibilities and scope of previous ones. Gordon
Moore’s observation about the doubling of computer
power per dollar every 18 months has not only held for
the last two decades, but is expected to hold for at least
two more decades.
One problem for space equipment developers is that the
rate of launching new space projects operates on a
different time scale than the one required to track the
technology curve. The deployment of new space projects
is limited by the high investment required for space
activities and by the political processes involved in
mobilising that investment. Spacecraft are further
burdened by the enormous expense of deployment,
upgrades and servicing, leading to a natural reluctance on
the part of investors to take risks with technology
innovation.
The result is that space projects generally take a long time
to fund, develop, and fly.
This mismatch in rates of evolution means that spacecraft
companies that do not launch products regularly, perhaps

every year, do not have the means by which to maintain
technical relevance in their product lines.
This mismatch is not new. It was the case fifteen years
ago, when it was already obvious that the satellite
industry in general had significantly parted company with
the terrestrial technology boom. In fact, the satellite
industry is in many ways relatively low-tech.
The high-tech aspect of space programs frequently comes
from the brains needed to squeeze a prodigious amount
of function into wholly inadequate hardware.
In contrast, microsatellite designers are already done
flying their second generation of 386 computers and are
now upgrading to ridiculously more powerful
StrongARM and DSP computers. The recently launched
SNAP-1 microsatellite from SSTL boasts very capable
on-board computing hardware, as well as cameras, GPS
equipment, cold gas propulsion and attitude control – all
in a 6 kg package (2).
In comparison, the Canadian contribution to the Space
Station program is powered by Intel 186 and 386
computers. A good choice ten or fifteen years ago, but
now looking a bit dated.
Some observers argue that the reason for this
technological obsolescence in the space industry is that
projects are so large that they take very long to put
together. However, this does not take into account the
very large and complex projects started and completed in
much less time on Earth.
The authors believe that organisational, political, and
process factors are the most implicated as causing space
projects to overrun and underperform. These factors
cause missions to be expensive and infrequent, making it
difficult and expensive for their technology to be
relevant, and for their builders to keep their product lines
glued to the technology curve. This view is not
significantly at odds with that held by the upper echelons
at NASA who pioneered the Faster, Better, Cheaper
approach to space missions
In the next section, we will present the concept of the
"death spiral" and how it acts to trap space programs and
escalate their costs.
The Death Spiral
High cost is such an endemic part of space programs, that
it sometimes seems inevitable. Talking with various
colleagues over the years, we frequently said things such
as “you can’t develop a new unit for less than $5 million
to $10 million dollars.” A supplier from whom we were
recently attempting to source attitude sensor electronics
said “the lowest recurring cost for which we can give you
a digital interface and signal processing board is $100k.”
Initially, if a space program is perceived to have a high
cost, there are two distinct consequences. First, in order
for enough budget to be assembled to fund the project,
various promises have to be made to entice and secure
enough stakeholders. Typically, this leads to adding
mission objectives and/or payloads to the spacecraft,
reducing the number of spacecraft (typically to one) and
extending the mission life. It may also lead to various
restrictions on how the money may be spent.
In addition, stakeholders must be reassured that they have
a high probability of receiving the fruits of the payloads
they are sponsoring. Also, because of the perceived high
price of the mission, it acquires additional stature within
the customer’s agenda. This leads to an increase in risk
aversion on the part of the missions designers and
managers, because failure becomes less acceptable.
Instead of taking (calculated) risks, there is an impetus to
eliminate risk.
This causes cost growth because conservatism pervades
the design process. Designers begin to specify a cascade
of margins that radically overspecify the required
performance of the various parts and units of the
spacecraft. They begin to specify flight-proven parts (i.e.
old technology), or parts that have extensive reliability
estimates performed to quantify their performance.
As an example, recently, we had the experience of being
told that a JAN-S military specification diode was not to
be trusted, and that we should specify a JAN-TX part
instead. At this point, one moves into the territory of
paying as much for a handful of parts as one might for a
new car.
Designers also begin to add redundancy to the various
subsystems and functions of the spacecraft, radically
increasing its complexity and making assembly and
testing far more elaborate as a result. Instead of the
engineer being able to assure the performance of an
electronics board by inspecting and testing it themselves,
manufacturing is controlled by heavily documented
processes that are checked and cross-checked as they are
performed. An entire Product Assurance organisation is
frequently called into being to make sure that the
avalanche of requirements that are written down are
actually being implemented.
The second consequence of starting with the assumption
that the program is going to be expensive is that in order
to win what is perceived to be a high price program, a
consortium of suppliers must be put together.

Figure 1: The “Death Spiral” of Space Project
Cost Drivers
This guarantees a certain burden of documentation, staff,
miscommunication and lack of flexibility that
immediately pushes the program cost to a higher
category.
For example, in order to subcontract a unit or subsystem,
a contract must be put in place with a statement of work
and technical specifications. In practice, this means an
initial Design carried out by engineers who do not then
carry on with the detailed implementation, premature
freezing of the specifications, and a lack of flexibility in
changing requirements or eliminating units entirely.
When a program is perceived to have a high cost, then
the customer is naturally anxious to have an expanded
overview of the design. While this is valuable and
provides a better view into the evolution of the design, it
also creates another set of opportunities for
miscommunication and disagreement. The technical staff
from the customer are naturally frustrated if their
engineering judgement is overridden by the contractor’s
technical staff, leading to additional requests for
justification of the design and additional analyses to be
performed.
Finally, a large staff has a tendency to enlarge itself
through sheer organisational dynamics. As more people
appear, their individual utilisation tends to decrease as
they rub against each other’s technical and organisational
boundaries. Some things get done more than once, while
an increasing number of people is required to make sure
that everything is done at least once. More time gets
used in meetings, and more people are hired to streamline
operations and to solve problems as they come up.
These various consequences of a program initially
perceived as highly priced have the effect of igniting a
cost expansion bonfire so significant, that a round of cost
cutting is required during the very early stages of the
program. This cost cutting frequently has the effect of
reducing any margin for error in the cost estimates, and
makes the program more fragile to cost overruns.
The objectives of all high value-added endeavours is to
stay on the high side of the perceived value vs. cost curve
(Figure 2). This is true for government sponsored as well
as private ventures.
High Cost
More Payloads,
More Stakeholders,
Longer Mission Life,
Fewer Spacecraft,
More Risk Aversion
More
Analysis
More
Program
Staff
Stricter
Customer
Overview
Expanded
Document
Req’ts
Mission
Expectations
& Profile Launch Cost
Increased
Subcontracting
Tradition
Conservative Design
More Redundancy,
Stricter EEE Spec’s,
Older Parts and Designs
Stricter Process Spec’s
More Complex Testing,
More Design Margins
Technical Challenges
of Working in Space

Figure 2: The Problem of Diminishing Returns
The “death spiral” pushes the program along a curve in
which additional cost adds a diminishing return of value.
This devaluates the project as a whole and damages space
projects as not contributing value commensurate with
other ventures competing for the same resources. Not
only do space projects not run on “internet time”, they
also don’t contribute “internet value added”.
Getting On the Spiral
Given that there are many people in both the user and
supplier community that understand these cost driving
effects, how is it that the cost spiral is allowed to ignite?
We are able to trace several proximate causes of ignition
for this Spiral. The most important of these are as
follows:
- Launch cost
- Mission profile
- Traditions governing design and build
- The space environment
Launch Cost
Even supposing that someone were to build reliable
spacecraft for a price of $500/kg (in the neighbourhood
of the cost of high-performance communications and
computing gear today). The lowest possible price for
launching this satellite today is $10,000/kg (US) from the
Russian DNEPR launch vehicle.
For Government customers, the cost of a launch vehicle
for a large satellite is enough to make any space project
that requires such a launch vehicle to be considered
expensive. Immediately, the concept of launching two or
three spacecraft as part of one program is, if not ruled out
of the question, then at least significantly jeopardised
For commercial customers, gain is frequently more easy
to quantify than for research missions. The problem for
those who are looking to make commercial success using
satellites, becomes one of finding space applications that
add enough value to warrant the investment.
Finding commercial space applications that are
compatible with returning even the price of launching the
spacecraft, much less designing and building it, is
difficult.
Mission Profile
The price that government and scientific programs pay
due to the high ante in the spacecraft game is enormous.
In proportion to the amounts that taxpayers are willing to
pay the launch cost is high enough to put any large or
even mid-sized space project squarely in the political
limelight. Entering the limelight is another proximate
cause of death spiral ignition.
Being in the limelight means that no technical failure can
be tolerated that could lead to embarrassment of the
project stakeholders. When a project moves into the
limelight, its cost drivers migrate from managing the risk
of failure to proving that everything was done to prevent
failure. The cost of the latter is far higher than the cost of
the former. If, in addition, the project is so expensive that
all of its hopes are pinned on one spacecraft, the
conditions are set for extreme conservatism and risk
aversion as that project is carried out. Every attempt is
made to eliminate risk, rather than to manage risk.
When failure is not an option, success can be expensive.
Traditions Governing the Design and
Building of Spacecraft
It is probably a universal human desire to want to
duplicate successes. The problem with space flight is
that there are still unknowns. We frequently know what
leads to failure, but frequently don’t know what led to
success. We are in the old quandary that probably only
20% of what we do to build, analyse and qualify new
space equipment designs is all that is required to succeed.
Our problem is that we don’t know which 20%.
The Space Environment
Another cost ignition source is the technical difficulty of
operating in space. Spacecraft exist and operate in
nature's high-energy vacuum physics laboratory, where
many strange and non-intuitive things can happen.
Finding out what they are is expensive and uncertain, and
this leads to a great deal of superstition and conservatism
on the part of spacecraft designers. As discussed above,
it also leads to a body of tradition that makes innovation
difficult.
Mission Cost
Mission Value

There is a hard core of truth to the challenges posed by
making equipment for operation in space. However, it is
also true that the environmental challenge of space is also
frequently overstated. It was generally popular in some
quarters in the early 1980s to say how difficult it would
be to make solid state devices survive structural loads due
to launch and the wide temperature swings possible on
spacecraft. This was the case until the automotive
industry started to put significant amounts of silicon
processing in their cars, and insisting that it operate -40C
to +70C reliably. Today, components designed for
industrial and automotive use easily cope with the
thermal and structural requirements of spacecraft.
Spacecraft still must cope with ionising radiation,
charging effects and vacuum, however. If one had the
luxury of doing experiments and being able to learn from
one spacecraft and properly incorporate those lessons in
the next and in a short period of time, then this
environmental challenge is easily enough addressed. The
problem comes when, as happens in Canada and in many
countries, engineers and scientists can only design one
spacecraft every five years or more, and that spacecraft
has to succeed or else ruin the program in which they are
engaged.
No matter how enlightened the approach, however,
designing instruments and equipment to operate in space
costs more than designing them to operate on Earth.
Getting Off the Spiral – Worked Examples
Many of the activities undertaken as part of a standard
development scheme are done not as much to increase
mission reliability, though they may have that effect to
some extent. They are done more to provide the illusion
of control for those funding and those managing the
work.
Here is a thought experiment as an example. Suppose
one has a choice of two approaches to building an
electronics board. The first approach we'll call Design A
and the second Design B.
Design A uses 500 older, MIL-STD components
qualified (or proven on multiple flights) for space
applications whose reliability figures were well
established and which are built in low volumes on
controlled production lines. The components are a mix
of through-hole and Surface Mount Technology.
Processing is done in hard-wired logic (for example, with
fused FPGAs) and is not easily reconfigurable. The
board uses 4 Watts of power, masses 2.0 kg, and has a
calculated reliability of 0.995 for the mission. The board
costs $1,500k to design and qualify and $50k per board
to build in small quantities.
Design B uses 50 components, the centrepiece of which
is a reprogrammable Digital Signal Processor. All of the
components use Surface Mount Technology. The
components are designed for industrial and automotive
applications and are built in large volumes. Most of the
functions are implemented in software. The board uses
0.8 Watts of power and masses 0.5 kg, However, for this
board there are no reliability numbers associated with the
parts. This board costs $80k to design, and $5k to build
in small quantities.
Design A appears to offer completely contained and
quantified risk. Sure, it's more expensive but at least you
know what you are getting. Actually, a spacecraft full of
Design A choices is fully trapped by the death spiral and
a great deal more expensive than a spacecraft design full
of Design B choices. Looking just at the immediate
bottom line, however, the total cost of a 3 board
“program” for Design A costs $1,650k as compared to a
cost for Design B of $95k. Design A is therefore 17
times more expensive than Design B.
Detailing the various mechanisms that make it possible
for Design B to be so radically cheaper than Design A
would fill a large volume. One example may suffice to
illustrate the issues, however. In a time not longer than
15 minutes, an engineer can obtain the price, check
availability, fill in the paperwork and order an electronic
component for a microsat. That component will then
arrive the next day by courier.
Contrast this with the EEE parts procurement process for
a Design A program. The prices of these components are
not generally advertised. It may take several telephone
calls or even a face to face meeting and several days of
delay in order to obtain a quote for a MIL-STD part. The
amount of engineering time per part can easily stretch by
a factor of 10 or 20, given that many other specialists get
involved to review and sign off the request for purchase.
If qualified components are not available, and a COTS
solution is acceptable to that program, the timing and
work stretch out further. The engineer must now fill in a
Non-Standard Parts Application Request, and convince at
least one and possibly several other people that he has
selected an appropriate design. Depending on the
organisation, various inspection and screening processes
will then be discussed with that vendor at length, possibly
creating the need for customised handling and paperwork
to be done by the part vendor in order to meet the
upscreening requirements of the Design A program.
Anecdotally it has been said that doing your own
upscreening is just as expensive, if not more so, than just
specifying a MIL-STD part in the first place.

A recent paper (3) backs up these cost differences over
whole programs.
Table 1: Spacecraft Cost Comparisons
Spacecraft Cost Reduction
Ratio
Orbcomm prototypes
(communications) 24
RADCAL
(radar testing) 7
AMSAT OSCAR-13
(communications)
102
(assuming free labour)
~24
(correcting for labour)
Freja
(magnetospheric research) 11
Clementine
(BMDO test and lunar
science)
3
The paper calculates the cost of doing several programs
with costing models based on standard construction
approaches, and then compares these projections against
actual costs achieved when various versions of
microsatellite-style cost-reduction were actually used.
The different projects were carried out by different
groups. As Table 1 shows, the results are nothing short
of remarkable.
Designers of spacecraft have for many years been
pressured to reduce costs without adding significant risk.
The standard techniques that meet these somewhat
contradictory requirements and that have provided some
degree of cost reduction success are as follows:
1. Streamline Procurement and Project Oversight
2. Simplify Mission
3. Reduce Redundancy
4. Integrate Development Team
5. Improve Design and Manufacturing Systems
6. Selectively Reduce EEE Parts Req’ts
These steps are all evolutionary and can lower costs by a
significant factor. However, they do not allow costs to be
reduced by a factor of 24, as has apparently been
achieved by some programs. In order to achieve such a
dramatic cost reduction, a fundamentally different design
paradigm – the microsat way -- has to be implemented.
Mitigating the Risk
Designers who baseline the more expensive option do so
arguing that going the cheaper way introduces risk, and
this is quite true. The key is in managing and
overcoming the risks, rather than in avoiding them.
The additional risk in Design B comes in two flavours.
First is risk that the technical performance and durability
of the electronics will not match the need. The definitive
way that this risk can be abated by flying test payloads in
space. This means that if the program allows building
and flying only one spacecraft, no significant risk
abatement is possible.
The second risk is in the absence of quantification of the
failure rate of the electronics. However, for the low
failure rates associated with the vast majority of digital
components, the failure rate numbers are always small,
and offers only an illusion of control. Most projects only
build a few spacecraft. There is no law of large numbers
to fall back on.
The need to quantify reliability numbers comes from the
large-system realisation of the Apollo days that "if each
part were only 99.9% reliable, the rocket would fail many
times before it reached orbit". This argument may be
true for Apollo, but very few people are building Apollo-
sized systems these days. Those of us who are not do not
need to be prisoners of the reliability estimation
procedure and its associated parts requirements.
In fact, one quantification of the causes and rates of
failure for small spacecraft shows the following trend:
Table 2: Causes of Spacecraft Failure
Secondary
Payloads
Primary
Payloads
1 Launch Vehicle 25.0% 10.0%
2 Design 18.6% 22.3%
3 Environment 16.0% 19.2%
4 Unknown Cause 14.2% 17.0%
5 Parts (random) 12.2% 14.7%
6 Operations 3.5% 4.2%
7 Quality (random) 5.8% 6.9%
8 Other (random) 4.7% 5.7%
(Note: a launch failure rate was assumed at 25% percent
for secondary payloads and 10% for primary payloads.
According to unpublished research by Rick Fleeter of
AeroAstro, the success rate for launching secondary
payloads is actually lower. The rest of the information
comes from reference 4)
Of these causes, the first is largely out of the control of
the microsatellite designer. Beyond choosing a launch

vehicle with a good track record, little can be done by the
builder to ensure that the satellite will be launched
successfully into the right orbit.
The next two causes emerge from the inattention of
designers and insufficient testing and on-orbit experience.
Here, the death spiral reveals its malign power. By
making the spacecraft more complex, the attention of
designers to each element of the design is diluted and
opportunities for miscommunication multiplied. It has
been argued by AMSAT members with long experience
in building spacecraft, that redundancy can introduce
more opportunities for bad design and failures than they
eliminate.
Failures due to inadequate design or misunderstanding of
the environmental loads on the spacecraft can be
overcome through a multi-spacecraft program. Lessons
are learned and spacecraft designs improved as the
program proceeds.
The mitigation for avoidable failures is therefore the
same as the mitigation for the use of previously unknown
or unqualified parts: fly multiple spacecraft.
By flying spacecraft more frequently, there is more
accumulated experience both in terms of knowing how
well various designs work as well as building a well-
experienced cadre of technical and programmatic staff.
The factors that make Design B the cheaper alternative
in our example have a lot to do with the fact that the
electronics is much simpler, and thus easier to design
robustly. A second important factor is that many
functions are undertaken in software, which is easier to
change and adapt both on the ground and in orbit. In the
production process, the fact that Design B uses radically
fewer components means that parts procurement,
manufacturing, and testing are far simpler, cheaper,
easier, and that makes the end product more robust.
Next comes the set of failure causes that capture the
essence of why space is still a frontier. Every now and
then, spacecraft fail without providing a conclusive
indication of why.
Finally, spacecraft failure due to the failure of parts is
fifth in line, accounting for roughly 12% of failures. It
may seem strange, then, that such a small source of
failures is accorded such a large element of the typical
development and build cost.
The implication of this risk has frequently been that such
new parts are simply not considered or chosen. The
”microsat way” is to develop a process and program by
which such parts can be considered and qualified for use
on a regular basis. Beyond that, parts can be used even if
they are fundamentally of lower reliability, assuming that
the system design allows for this.
Rather than considering new technology as a threat,
feeding new technology into spacecraft design should be
the highest priority among spacecraft builders and the
spacecraft customer and user community.
The key in getting an operational system that costs less at
the same time as it delivers more and, and this is the
critical point, with higher system reliability and
robustness is to use a series of spacecraft. The details are
worked out in the next section.
Getting More For Less
In order to get more for less, we have to move more
generally from what we have called the “death spiral”
to its positive analogue: a “virtuous spiral” shown in
Figure XXX.
Figure 3: The Virtuous Spiral
The general idea is that faster and cheaper missions
mean more results more quickly, thus priming the
pump for more missions. Indeed, as more missions are
flown, a larger number of formerly COTS components
are qualified, thus bringing down the cost and
increasing the reliability of further missions.
Now, it remains to be seen whether the spacecraft
market is indeed price elastic. That is, with cheaper
missions, will there be a larger overall volume? One
argument that can be made is that if the market is not
elastic, then lowering the price of missions simply
deflates the value of the market and moves margins
from the spacecraft producers to the customers.
This would simply aggravate an already poor situation.
A recent study by Booz Allen & Hamilton (5) revealed
that return on sales in the space sector in the United
States has fallen by over 20% since the early 1980’s.
Experienced
Engineers & Techs
Lower Design and
Production Cost
Faster Production
Time
Cheaper Missions
More Missions
More Flight Parts and
Equipment
Happier Customers

(from around 8.5% to just over 6.5%). In addition, that
study concluded that “decreasing cycle times for design
and production of satellites has resulted in an excess
capacity of approximately 50%".
If the cost of spacecraft acquisition were to drop
significantly, then market volume and margins might
well take a further beating.
These observations provide one explanation why
existing spacecraft companies might find it difficult to
radically lower their costs using fundamentally
different approaches to spacecraft design, and why
these innovations are being pioneered by companies
such as SSTL and SpaceDev, who have no role in the
traditional spacecraft supplier infrastructure.
The authors believe that the market is indeed elastic.
Cheaper missions mean that science budgets that
previously were not large enough to afford spacecraft
would now be drawn into the market. Without new
science and exploration missions, it is not clear where a
significant enlargement of the spacecraft applications
market will come from. The world needs only so many
traditional-style communications and remote sensing
satellites.
Radically cheaper spacecraft mean that new business
cases that might have previously looked dubious or
risky may be more viable and therefore also be drawn
into the market. Only with successful new business
models can the commercial market enlarge.
A final, though perhaps the most critical, element in
this virtuous spiral is that faster and more frequent
missions lead to more excitement in the industry. This
will draw in more young, talented people both on the
technology and also on the science and business side.
The Booz Allen, & Hamilton study points to a
“greying” space workforce and the perception that
space is a “smokestack industry” as key challenges for
the industry. It finds that “in 1990 aerospace was the
third most desirable career field for science and
engineering graduates. By 1999, it had dropped to No.
7.”
The fact is that working on missions which take 15 years
from initial idea to flight is a highly resistible proposition
for young people used to the pace of “Internet time”. In
the next two decades, whole new fields will be created in
the areas such as functional genomics, wired and wireless
communications, and nanoscale technologies. Compared
to that, otherwise worthy large and long-term space
projects like the Next Generation Space Telescope may
well appear to exist in stasis.
The Future
Perhaps the most important objective for a spacecraft
component maker or integrator is to ride the technology
tiger. In the electronics arena, offering technology that is
more than one or at most two years old will mean being
squeezed out of the business.
Here are some observations and ideas that we believe
will shape and describe the future of the space industry:
1. Product innovation must be continuous, upgrades
must happen every year or two years at most. One
reason for this is that electronic parts become
obsolete very quickly. Unless one has pre-purchased
parts or is some other way guaranteed their
availability (driving up cost), electronic parts will
simply not be available for extended periods of time.
A second reason is that there are vast increases in
cost performance from one generation to the next. A
design that takes advantage of newer parts will
perform better, use fewer parts, and cost less.
2. The logical consequence of this observation is that
any supplier of space electronics must have a
continuous pipeline of hardware products
undergoing space qualification testing. There is a
role here for the Canadian Space Agency to serve as
a road to space, offering regular and frequent
technology flight opportunities. Project cycle times
must be reduced and the number of projects
increased, if new technology is to effectively be
injected into the spacecraft procurement cycle.
3. While this pipeline will generate a source of flight
proven designs, it will not eliminate risk. Mission
designers of missions based on the highest
performance to cost ratios must accept a certain
degree of risk in their missions, rather than attempt
to design all risk out. However, using a series of
cheaper satellites can provide an overall mission
reliability as high as required, even if each satellite is
significantly less reliable. The key is to design not
only the spacecraft procurement but also the
business case overall to take advantage of this
approach. Without an adjustment of the business
case, there will always be significant pressure to
revert to an expensive and high-rel approach to
building spacecraft.
4. Volumes will always be low, so development cost
must also be low. The days when we thought we
could reduce the cost of spacecraft units and
subsystems simply by increasing volumes are over.
Constellations of dozens of satellites are the
exception, not the rule. Development cost is best

reduced by having a cadre of technical staff
continuously employed and with access to frequent
opportunities to test designs. Building only one
spacecraft every five to ten years maximises
development cost, because those who designed the
initial spacecraft will be the “stick in the mud”
technical superiors who will hold back the next
generation from using the most advanced
technology.
5. Barriers to entry are dropping, and the number of
suppliers of units and subsystems for both micro and
larger satellite applications is set to surge as the
microsatellite industry develops and its methods
begin to infiltrate more traditional spacecraft
builders. Dynacon is among those companies
working to drop entry barriers for new spacecraft
organisations, by demystifying and providing
support in attitude control and other arcane
subsystems for which no analogue exists in ground
based equipment.
6. In the past, spacecraft operators have improved their
cost performance by demanding longer spacecraft
life. For many commercial communications
applications, longer-lived spacecraft are highly
desirable, because the stabilising effect of
entrenched standards and a large, installed user base
means that technology upgrades are not necessarily a
driving requirement.
This is not the answer for everyone, however. In
particular, it is not compatible with the explosion of
technology represented by Moore’s Law. In many
cases, a larger number of spacecraft, each with a
shorter life is a better answer. And even in the case
of the traditional applications, operators may come
to prefer upgrading their spacecraft every three or
four years rather than every fifteen years, assuming
their total costs and risks were the same. Newer
technology spacecraft would allow them to try new
product offerings and tap new revenue streams.
For science and remote sensing missions, and for
new services in which standards are not yet
solidified, technology can be a significant
discriminator and frequent upgrades to spacecraft
functionality is almost mandatory.
7. Although it is not a panacea, the use of multiple
smaller, cheaper, better spacecraft to implement
space mission objectives is not only likely to be the
best solution for the customer of the mission, but it is
by far the best alternative for the space industry. In
order to mitigate the risk of failure on launch, the
spacecraft should be launched on separate launch
vehicles. This total strategy will ultimately have the
effect of significantly adding to the value added
contribution of space missions, thereby expanding
the market base and increasing opportunities and
services for everyone.
References
1. University of Surrey Small Satellite Home Page, at
http://www.ee.surrey.ac.uk/SSC/SSHP/
2. “The SNAP-1 Machine Vision System”, paper
SSC00-III-6 by Richard Lancaster and Dr. Craig
Underwood, presented at the 14th Annual AIAA/Utah
State university Conference on Small Satellites.
3. “Microsatellites and Improved Acquisition of Space
Systems”, paper SSC00-IV-7, by Matt Bille, Robyn
Kane and Drew Cox; presented at the 14th Annual
AIAA/Utah State University conference on Small
Satellites
4. “Space Mission Analysis and Design, 2nd Edition,
edited by Larson and Wertz, Section 19.2, Microcosm,
1992.
5. “Space Industry Challenges and Changes”, by Thomas
S. Moorman, Jr. quoting a study by Booz Allen &
Hamilton, as reported in Space News, 23 October 2000.