SPIE_San_Diego_2010_Spitzer_Lessons_Learned_V12.5

jamaicaitalianΜηχανική

18 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

201 εμφανίσεις






1

Writing a success story: lessons learned from the
Spitzer

Space
Telescope



R. D. Gehrz
*
a
, T. L. Roellig
b
, and M. W. Werner
c


a

Department of Astronomy, School of Physics and Astronomy, 116 Church Street, S. E., University
of Minnesota, Minneapolis, MN, US
A 55455


b
NASA Ames Research

Center, MS 245
-
6, Moffett

Field, CA 94035
-
1000

c
Jet
Propulsion Laboratory, MS 264
-
767, 4800 Oak Grove Drive, Pasadena, CA 91109

ABSTRACT

A key to the success of the
Spitzer

Space Telescope (formerly SIRTF) Mission was a uniqu
e management structure that
promoted open communication and collaboration amon
g scientific, engineering, and c
ontractor personnel at all levels of
the
project
. This helped us to recruit and maintain the very best people to work on
Spitzer
. We describe the
management
concept that led to the success of the mission. Specific examples of how the
project benefit
ed from the communication
and reporting st
ructure, and lessons learned about technology are described
.

Keywords:
Infrared astronomy,
Spitzer

Space Telescope
,
Less
ons Learned
, NASA


1.

INTRODUCTION

NASA’s
Spitzer

Space Telescope (formerly SIRTF), launched on 25 August, 2003, is a cryogenic infrared (IR)
observatory that is studying the creation of the universe, the formation and evolution of pri
m
i
tive galaxies, the genesis of
stars and planets, and the chemical evolution of the universe. The cold (5.5K) phase of the mission operating at
wavelengths from 3.6
-
180

m ended in mid
-
May of 2009 when the liquid helium cryogen was exhausted, and the
mis
sion is now in a warm (30K) phase during which it will observe at 3.6 and 4.5

m until 2014. An expansive history
of
Spitzer

has been written by Rieke
1
. The success of the
Spitzer
Project has involved careful attention to lessons learned
not only within

the Project itself, but to lessons learned

from many other projects that were conducted

during the nearly







Figure
1
:
Spitzer
keeps co
ld

in its Earth
-
trailing
orbit by hiding from the Sun behind the solar panel
.
Reproduced by courtesy of NASA/JPL
-
Caltech










*

gehrz@astro.umn.edu
;

phone 1 612 624
-
7806; fax 1 612
-
626
-
2029






2

forty years since the development of the
Spitzer

mission began in the early 1970’s. Lessons learned that were examined
very closel
y during the design, development, and implementation of
Spitzer
included

those from the Infrared
Astronomical Satellite (IRAS)
2
, Hubble Space Telescope (HST)
3
, Infrared Space Observatory (ISO)
4
, Wide Field
Infrared Explorer (WIRE)
5
, Mars Climate Orbiter
6
,
and other space missions. In many cases, we benefitted greatly from
the fact that members of the
Spitzer

team had direct scientific, engineering, and management experience through their
work on these projects (another lesson learned). In this paper, we c
onfine our discussion to the design, development,
assembly, and testing of the flight hardware for the
Spitzer

Mission leading up to the launch.
Lessons learned during
post
-
launch operations are discussed elsewhere.
16



We describe the
Spitzer

Observatory
and examine the lessons learned
during the design, development, and early operational phases of the Project that led to a highly successful mission.



2.

SPITZER SPACE TELESC
OPE PROJECT OVERVIEW

T
he
Spitzer

Space Telescope Mission (Figure 1) has been descri
bed in detail by Werner et al.
7

and Gehrz et al.
8


Figure
2 shows a side elevation of the observatory. The Cryogenic Telescope Assembly (CTA), attached to a Spac
e
cr
aft Bus
(SCB), houses the cryogenic telescope (CTA), the
superfluid
-
liquid helium cryostat
, and the Multiple Instrument
Chamber (MIC) in which reside three Science Instruments (SIs) capable of imaging and spectroscopy from 3.5 to 180

m. The CTA’s 85
-
cm cryogenically cooled b
e
ryllium Ritchey
-
Chretien telescope operated at temperatures as low as

5
K
during the cryogenic phase of the mission. The unique Earth
-
trailing orbit (Figure 1) enables the telescope to be cooled
to 30K by radiation to space with the CTA hidden from the sun behind the solar panel (SP). The remainder of the
cooling power duri
ng the cold mission was provided by the boil
-
off gas from 350L of liquid helium in the cryostat
Dewar. Power is generated by solar cells mounted on the sunward side
of the S
olar
P
anel
. Spacecraft slewing, tracking,
and attitude control are accomplished b
y means of reaction wheels and dry nitrogen thruster jets on the SCB.

Field
acquisition is facilitated by star trackers mounted on the SCB.
Communications with Earth
are acco
m
plished using
fixed
low and high gain antennae located
at the base and bottom
o
f the SCB. Spitzer can be yawed and pitched to view a
37.5
o

wide annulus that sweeps the entire sky about twice a year, providing tempor
al viewing windows ranging from
about 40 days in the Ecliptic plane to contin
u
ous viewing in a 15
o

diameter circle at
the ecliptic poles.

















Figure
2
: Side elevation of the
Spitzer

Observatory facility
showing the major sub
-
assemblies. Repr
o
duced
courtesy of NASA/JPL
-
Caltech
.













3

The telescope was launched warm and cooled down on reaching orbit by a co
mbination of radiation and vapor cooling
using
boil
-
off gas from the superfluid liquid helium cryostat. JPL/Caltech was the
de facto
prime contractor for the
mission and provided the management team. The CTA was built by Ball Aerospace Technologies Corpo
ration (BATC).
SVG Tinsley provided the CTA optical
elements under contract

to BATC. Lockheed Martin Missiles and Space
(LMMS) built the SCB and integrated the observatory. The launch contractor was the Boeing Company.
Spitzer

has
three Scientific Instr
uments (SIs). The Infrared Array Camera (IRAC)
9
, a 4
-
channel IR imager with bandpasses at 3.6,
4.5, 5.8, and 8.0

m, was built by NASA Goddard Space Flight Center under management by the Smithsonian
Astrophysical Observatory. The Multiband Imaging Photome
ter for SIRTF (MIPS)
10
, a 24
-
160

m imager and low
resolution spectrometer, was built at BATC in collaboration with the University of Arizona. The Infrared Spectrograph
(IRS)
11
, operating from 5.2 to 37.2


m, was a collaboration between BATC and Cornell U
niversity.

3.

MANAGEMENT, TEAMING,

AND PROJECT DESIGN

Definition of the

Spitzer

mission and its hardware began in earnest in the fall of 1973. During the path forward from fall
of 1973, new mission design concepts occurred four times with each design phase t
aking nearly seven years before
Spitzer
’s

eventual launch at 01
h
:35
m
:39
s

EDT on August 2
5
, 2003. Many lessons were learned during these design
phases that enabled management, teaming, and design approaches to evolve over many years. When the
Spitzer

obser
vatory was finally given a new start by NASA in 1995, a highly efficient and successful project organization was in
place.

3.1


Four Mission Designs over 30 Years

Spitzer
(SIRTF
)
was first conceived to be a Shuttle Transportation System (STS) attached missio
n to be launched in
1979 at a cost of about $100M. By the early 1980’s, SIRTF proponents and the general IR scientific community had
begun to lobby NASA management vigorously with the notion that the STS
radiation
background and pointing
environments woul
d not permit a successful IR cryogenic STS attached mission. The Project was eventually directed to
re
-
design for a free
-
flying mission in a low earth orbit with an STS launch and periodic STS refurbishment
flights
similar to

those that have continually
revitalized
HST

over the years
. When the
shuttle
Challenger

(STS
-
51
-
L)
was lost
on January 28, 1986, NASA management concluded that liquid helium

was far too dangerous a cargo for the STS, and
the
Spitzer
Project was directed to re
-
design for a launch wit
h an expendable launch vehicle (ELV). The design that
resulted was to be launched by an Atlas
-
Centuar ELV into a high polar orbit, and the mission development cost
ballooned to more than $2 billion dollars. When the multi
-
billion dollar HST was launched
into orbit by STS
-
31 with
flawed optics on April 24, 1990, Congress soon mandated that no science emission be launched that cost more than
$500M. As described by Rieke
1
, Werner et al.
6

and Gehrz et al.
7
, this led the
Spitzer
Project to narrow its science
objectives, prune its SI capabilities, and adopt an Earth
-
trailing solar orbit that eliminated the Earth as a source of
parasitic background heat.


3.1.1 The
original
shuttle
attach
ed mission


Begun in fall of 1973, the

initial design phase of the Shuttle

Infrared Telescope Facility (SIRTF)

was marked by
highly
over optimistic

expec
tations about the breadth of

the
science that could be accomplished

and the frequency of Shuttle
flights. The concept included exceedingly complex SIs.


SIRTF’s 120
-
cm telesco
pe, to be attached to an arm extending
from the Shuttle bay, was designed to observe at wavelengths all the way from 2

m to 1 mm. Its SIs, accessed by a
rotating dichroic tertiary mirror that enabled on
-
axis guiding using an optical camera, included filte
r wheel photometers,
scanning grating spectrometers,

and very high resolution Fourier transform spectrometers (R =

/



~ 10
4



10
6
).
Background was to be removed using a chopping secondary mirror. As it became evident that the Shuttle
environment
would

interfere with the achievement of SIRTF’s

science objective
s, the IR community began to lobby NASA
Headquarters vigorously for a free
-
flying mission. During this phase, the SIRTF Project

began to learn

that team
building was important

and that mutual res
pect between scientists and engineers was crucial for success
.








4

3.1.2 Shuttle accessible low earth orbit


The 1980 Astronomy and Astrophysics Decadal Survey (The Field Committee)
12
, at NASA’s insistence, cited the STS
-
attached SIRTF mission as an approv
ed and continuing program
*
. However,

the IR community chaffed sorely at this
recommendation, and their unrelenting demand for a free
-
flying spacecraft caused NASA to place the project into the
regime of unapproved missions. At that point, the SIRTF miss
ion had to be re
-
assessed and prioritized by the next
Decadal Survey for the 1990’s (The Bahcall Committee)
13
, where it received top priority as a free
-
flying mission. In
retrospect, this was an important milestone. The STS environment implicitly advocat
ed by the Field Committee would
have doomed the mission to failure compared to the scientific milestones that were eventually achieved. T
he science
instrument
(SI)
teams and science wo
rking group (SWG) were selected in 1985, shortly after the decision to r
e
-
design
Spitzer
for an STS launched free
-
flying and STS refurbished low Earth orbit.
Overly broad scientific objectives and
complex scientific instrumentation continued to characterize this phase of the development

of the project, leading to cost
growth.


3.1.3 High
Earth
orbit

(Atlas
-
Centaur SIRTF)


The decision in 1987 to launch SIRTF with an ELV as a free flyer changed the “S” in the mission’s acronym from
“Shuttle” to “Space.” It was now the “Space Infrared Telescope Facility.” Free of the requiremen
t to be in the low
inclination Earth orbit dictated by a shuttle launch, the Project first aspired to the high polar Earth orbit used by IRAS
2

and COBE
14
. This orbit is free from day/night thermal effects and enables very efficient survey and scan strateg
ies.
From there
,
a little outside the box thinking led
to the idea that

Spitzer

be placed into a 100,000 Km high earth orbit,
which has substantial thermal and viewing advantages. This third version of
Spitzer
,

to be launched with an Atlas
-
Centaur ELV,

con
tinued to be designed around the scientific objectives and
SIs

from previous versions

of the mission
.
Because of the background radiation from the earth

(
even at
a distance of
100,000 km
)

and the desire to have a five
-
year
lifetime
, the
A
tlas
-
Centaur

vers
ion was quite
massive. The massive observatory and complex SIs combined to produce
a
development cost of well over $2 billion
, a budget rendered untenable by the Congressional fiscal constraints imposed
by the failure of the HST mission

and resulted in th
e decision by NASA to cancel the
Atlas
-
Centaur version of the
SIRTF mission
. The Atlas
-
Centaur version of
Spitzer

was endorsed by

the

Bahcall committee;
the Delta
-
SIRTF mission
was

endorsed
by a special peer review
by the NAS Committee on Astronomy and As
trophysics
following the de
-
scoping
described below
.



3.1.4 Warm launch, solar orbit
(Delta SIRTF, Spitzer)


The

final
version
of
Spitzer

was
designed

to realize

the primary

science objectives of the
Atlas
-
Centaur

mission at a
vastly reduced cost

utiliz
in
g

some important lessons learned during the design of the previous three versions. First, the
science objectives
of the mission were
limited to four major lines of investigation as a result of a

discussion at a

retreat
of Project personnel
held in Broomfie
ld
,
Colorado. This agreement became known as the
“Broomfield A
ccord.


Second, to
reduce

the launch weight of the spacecraft

so that a smaller, cheaper Delta ELV could be used,
a new way of
operating

had to be invented. T
his was done by launching the tele
scope warm into a
n earth
-
trailing heliocentric orbit

that eliminated the earth as a source of
heat
and cut the cost by
about $1 billion
.
This warm launch architecture (Figure
2) which makes substantial use of radiative cooling, is a departure from that use
d for IRAS, COBE, and ISO (and earlier
versions of
Spitzer
). These missions paid a significant weight penalty by launching the telescope cold within a massive
dewar.
Third,
the complexity
of the
SIs was reduced substantially

by limiting the wavelength co
verage and eliminating
almost all of the moving parts
through the use of large
-
format
detector arrays instea
d of filt
er wheels.

These actions



*

It should be noted that the Field Committee’s report
did

open the door for the possibility of a free
-
flying SIRTF for the
future by calling for the study of such a mission in a section on long
-
durat
ion cryogenic space infrared telescopes in
Chapter 7: “Programs for Study and Development.”






5

brought the cost of the Delta SIRTF down to the Congressional mandate of $500 million. The observatory was re
-
na
med
Spitzer



in honor of the eminent astrophysicist Lyman Spitzer
-

following its successful on
-
orbit checkout.

3.2

Major lessons Learned about Management and Teaming

The 30 year evolution of the development of the
Spitzer

Observatory was characterized by a n
umber of lessons learned
about management, teaming, and project design that helped the Project to refine the scientific objectives of the mission,
improve the performance of the observatory, and reduce mission cost:

Dealing with NASA Headquarters (HQ):

Es
tablishing and maintaining

an effective
working relationship with N
ASA
Headquarters through good communication was essential. Face
-
to
-
face meetings on key issues, about controversial
decisions, and at critical reviews were an important part of the communi
cation process. So were visits by HQ personnel
to contractor sites. Preparation of careful documentation to maintain corporate memory in an environment where players
were continually changing and where many missions were competing for attention proved to

be an important way of
keeping the Project moving forward over many years.

Have an expert, engaged Science Working Group (SWG):
A crucial element of the success of the
Spitzer
Mission was
the existence of a Science Working Group (SWG) that was scientifica
lly and technically skilled in
almost
all aspects of
the Project and whose members were willing to become deeply involved in Project oversight, the development process,
the testing program, and eventually the operation of the observatory. The Project made

the best use of all available
resources in a truly “badge
-
less” team. The original SWG was solicited in 1985 by a NASA AO that selected a Project
Scientist, a Facility Scientist, two Interdisciplinary Scientists, and three SI Principal Investigators. As

the Project
progressed, additional members were added as required to expand the portfolio of the SWG to include the expertise
needed to develop the observatory and its Science Center, and to engage the IR scientific community in preparing for its
use of
S
pitzer
. A clearly important incentive to the SWG was the award of a limited amount of Guaranteed Observation
(GO) time and monetary support in return for their involvement.

Organization of the Project:

The c
o
-
location of
key scientists, systems engineers
, and subcontracts managers

at JPL
during an early
Skunk Works
®
15

style design phase
was an important aspect of the success of the
Delta SIRTF
design
phase of the
project
. The tangible results of this
were
t
hat a good design was produced,
people got to kno
w each other,
mutual respect was built among the scientists and engineers,
members who were not team players

were rapidly identified
and weeded
-
out
, and team building occurred

through retreats
,

monthly face to face
management team
meetings
, and
frequent te
chnical meetings amongst engineers and scientists from the different development teams.
.

The
Project

is people:

People are the most important resource on a
large project

like
Spitzer
.
Project

management
learned to get the best people on the job, to empow
er them to do their work, and to keep them interested in their jobs.


It
was recognized that it
is important to remember and remind everyone engaged in a large
project

that the
Project

Off
ice is
not the entire project.
Every team member is part of the
pr
oject

and needs to be encouraged to feel personally invested
in the success of the mission.
Contractor personnel were encouraged to become invested in the Project by exposing them

to the excitement of
Spitzer
science
through

science colloquia
presented at
their places of business, at SWG meetings,
and at Project reviews. This investment paid off well by giving contractors the

incentive to make sure that the
project

maintain
ed

schedule
, stayed within

budget
,
and that
the observatory performed
to the
L
evel
-
O
ne specifications.

At
another level, many social events were held over the years where team members from the Project, the SWG, and the
contractors could co
-
mingle in an informal setting, get to know one another on a personal level, and discuss their mutu
al
interests in the Project in a productive way. These interactions had the very beneficial effect of facilitating good
communication among team members even at times when the project was undergoing extreme stress.

Promote open communication at all levels
:

Open communication
,

vertically and horizontally within all levels of the
project,

was a key to successfully identifying and solving problem areas before they were out of hand. SWG members
were encouraged to be assertive about their involvement in the

project

and communication between SWG members and
contractor personnel was encouraged. Open communication within the
project included the mantra “D
on’t treat a
dults
like children”
-

a
ll facets of the
project
, including technical, fiscal, and political pr
oblems encountered by the
project

were communicated openly at all levels of the
Project

to forestall the demoralizing effect of unsubstantiated rumors,
and
the opinions of all
who wished to be heard
were taken into account in making

decisions going forward
.


Celebrate your successes:
Project management recognized early that

it is important to celebrate successes as well as to
ac
knowledge failures.

In the 30 years of the development of
Spitzer
, there were many false starts and setbacks that were





6

demoraliz
ing. Ample attention to celebration of the victories


for example milestones in detector development or
encouraging optical test results
-

kept people engaged and muted the sting of the setbacks.

Frequently recite a few mantras that will keep the project

on track:
The
Spitzer

Project developed
several mantras that
we
re recited

frequently during the design
,

development
,

and testing phases of the mission
to keep people on track.
These included “Better i
s

the enemy of good enough,” “Keep it simple, stupid,”

“Don’t do anything stupid,” and “
I am
feeling a pinch about this.
” This last phrase could be used to express general uneasiness without requiring a very specific
concern. “You’re the boss” was frequently cited to team leaders to let them know that they w
ere encouraged to solve
problems and make critical decisions. Several of these mantras may seem obvious, but inattention to the obvious was
often the problem.


3.3

Major lessons Learned about Project Design

Define Limited Science Objectives:

P
erformance
was m
aximized and cost was
minimize
d

by

limit
ing the

science
objectives that
defined

the
L
evel
-
O
ne requirements of the mission. As noted above,
the
Spitzer

Science Working Group
defined

four major, but limited science objectives at the Broomfield, Colorado retr
eat. Spitzer has since done many
things not aligned with these objectives, but only the big 4 were allowed to drive the Level
-
One system requirements. In
addition, by restraining their appetites the SWG showed the rest of the Project team that they were
serious about
containing mission cost and scope.

Minimize the complexity of the Science Instruments (SIs):

The
number of moving parts
in the
Spitzer

SIs was minimized
by the eliminating all the f
ilt
er wheels and using the large detector arrays that had be
come available during the previous
21 years of development of the mission

to maximize spatial coverage of the imagers and spectral coverage of the fixed
grating spectrometers
.

Cross
-
dispersed Echelle gratings were used to increase spectral coverage on two
-
dimensional
array detectors. Large format arrays also facilitated background removal.
In the
final design, the only three

moving parts
in the instruments and telescope
we
re a shutter in the IRAC camera, a chopping/scanning mirror
in the MIPS camera
,
and
a cryogenic focus mechanism on the secondary mirror
. We note that the
IRAC shutter

failed
by stalling in the optical
path
during ground testing and was never used during flight
,

a reminder
that moving parts
that can cause a

single point
source of failure
s
hould be regarded with extreme caution
.

A further reduction in cost and complexity was achieved by
limiting
the
data acquisition

options to

seven primary
modes and by ruling that only one instrument could be on at a
time
.

These decisions limited both deve
lopment and testing/validation costs.
Several examples of how hardware and
software
were simplified
to improve the reliability of the
project

are worth mentioning. The IRS and the MIPS
instruments collaborated in several ways
.

T
he IRS
project

built the
detectors for the MIPS 24
µm

array since the two
project
s use
d

similar technologies and the IRS team also built
a
combined
set of
electronics that operated both
instruments. The IRS spectrometers
were designed to a particularly robust design so that

the det
ectors a
nd the optics
could all be bolted

in place and
in what is

call
ed

the

Bolt

and

go


technique where the optics
were shimm
ed during
ground based testing so that there were no
on
-
orbit

moving parts necessary to
get excellent

spectra
l data. The feasib
ility

of this technique was
proven

before
flight by ground
-
based demonstrations using the 200
-
inch

telescope
at

Palomar

Mountain.

Keep technology development alive:
The effort to downsize the number of moving parts involved the clever use of
technology to
simplify the system

while maintaining scientific capability

by
using multiple

large
-
area

arrays
instead of

filter wheels. In retrospect
,

one thing that ke
pt the Project

fresh
and alive
was that
it

maintained an active technology
development program even in

the years when the budget was lean. In this way,
there was continuing improvement of

the detector arrays
that eventually led to the “Keep it simple, stupid” concept of “no moving parts” in the cameras and
spectrometers,
and the concept of building a bery
llium telescope assembly

was developed
.

Interface design should be as simple as possible:
The Project capitalized on the “Keep it simple, stupid” mantra by
keeping all electrical and mechanical

subsystem interface
s
very clean
so as
to minimize integration
and test problems
.

4.

THE FABRICATION AND
TESTING PROGRAM

Following the
HST

optics disaster, NASA management began to encourage large projects to adopt the mantra “Test as
you fly and fly as you test.” In other words, conduct preflight testing of the flight
configuration in a flight
-
like
environment, and do nothing in flight that has not been proven in pre
-
flight testing. This mantra was difficult to follow





7

for
Spitzer
, because the observatory facility had to operate in an exceedingly

stringent thermal envir
onment
, the optics
had to operate at the low temperature

of 5K, and the telescope had to be launched in a heretofore completely unproven
warm configuration with no heritage. It did not appear to be fiscally
possible to
devise a ground
-
based test program t
hat
would faithfully
reproduce orbital
thermal
conditions
and that could provide images of real stars to test the optics. The
Project therefore
had to adopt an intelligently thought
-
out test program that was realistic within the resources available
.
The
tradeoffs between on
-
orbit performance

margin
and uncertainties introduced in the fabrication and testing process
were continually reviewed by Integrated Product Teams (IPTs) and other formal review teams to ascertain that Level
-
One requirements would be m
et with high confidence. We describe below some of the major lessons that were learned
in this process.

Use Integrated Product Teams (IPTs) to monitor the fabrication and testing process:

Integrated Product Teams (IPTs),
groups of experts drawn from all

project elements and made responsible for delivering a definable product required for
the success of a mission, were adopted as one mode of advancing the
Spitzer

Project.

They

proved to be a fundamental
mechanism for

general oversight, setting

of

require
ments, interpretation of test data, and
identifying and solving
problems
. The
Spitzer

IPTs
included teams on liquid helium management, optics development and testing, on
-
orbit
focusing of the CTA, pointing, payloa
d integration, aperture door
design, and t
hermal modeling and testing.

The IPTs
held regular telecon and face
-
to
-
face meetings to do their work. Many met weekly or even several times a week. The
“You’re the boss” mantra encouraged IPT leaders to lead their teams vigorously, operate independentl
y, and make
thoughtful decisions. Several examples of the importance of IPTs are relevant. The Helium Usage IPT conducted
studies showing how to maximize the lifetime of the
Spitzer

cryogenic mission by using a heater
in the helium tank
to
increase the bo
il
-
off rate in order
to cool the CTA to 5K only during MIPS observations. This methodology extended
the cold mission from about 4.5 to 5.75 years. When problems with the eject
a
ble aperture door assembly surfaced, the
Outer Door IPT was formed
to
devise t
ests and solutions. The Optics IPT uncovered an almost fatal flaw in the
BRUTUS optical test before it was conducted. The Focus IPT discovered how to evaluate focus with passive imaging so
that a risky focus sweep was not required on
-
orbit.

Have a well m
anaged data archive and transfer web site:

The design of a mission document archive system that is truly
useful can be quite tricky
-

SIRTF was fortunate to have an exceptionally talented person who led this development.

The SIRTF
Transfer, Archive, and Re
trieval System (STARS) password
-
protected
web site was a crucial asset to the
program dur
ing the design,
development
,

and operational phases of the mission.

It enabled the Project to

keep an
integrated record o
f all aspects of the design,
development
, tes
ting, and operation that could be accessed by all Project
personnel at any time. It was well maintained and had a very transparent architecture. The site also served as an ftp
i
nterface site for the transfer of documents

among IPTs and Project personnel.

Access to the site involved minimally
invasive password protection
.

There is no substitute for face
-
to
-
face
-
interactions:
Although the
Spitzer

Project made liberal use of telecons

and
webcasts, it was found that face
-
to
-
face reviews and periodic
manage
ment team and
SWG meetings were the most
important way of ensuring that the
project

was staying on track.

Nothing is simple


take nothing for granted:
Things that seemed simple and were taken for granted early in the
development of the mission turned out

not to be so. Problems with the
Spitzer
aperture door, a mechanism thought to be
quite simple at first, were uncovered by a sharp
-
eyed team member while double
-
checking some numbers in a review of
the design
.

A lesson learned from the failure of the WIR
E mission was that

the low thrust vents on the helium Dewar
had to be very carefully balanced in case of a premature door opening

or other violent venting event,

and a special
experiment
was conducted
to make sure that
the
Spitzer

vents were
,

in fact,
bala
nced.

Another mishap that could have
been avoided was an ice plug event during ground testing (described below) that could have been detected early by the
installation of an elementary alarm system or by following sound engineering practices. This mishap
occurred over the
Christmas holidays and resulted, in part, from lack of appropriate vigilance during periods when many Spitzer facilities
were closed or short
-
staffed.

Pay as much attention to the Ground Support Equipment as to the Flight Hardware:

Proje
ct personnel learned

that it
was important to pay as much attention to the ground support equipment as to the flight hardware. There were several
pointed

examples of this during the course of the
project. During
the ground t
esting of the cryostat at super
f
luid hel
ium
temperatures,

the
cryostat

nearly exploded
when an undetected ice plug formed

as a result of problems with

the GSE
pumping system. T
he
helium dewar
over
-
pressured before the
situation
was discovered.

In

retrospect more attention





8

should have

b
een paid to
devising
alarm systems that would have warned of the plug earlier.


Another example of a
problem that could have been detected earlier was a software stub that caused a critical
station
-
keeping
control gas valve
to be cycled
more than 10,000 ti
mes at low pressure.

The valve was meant to ope
rate only a few hundred times at

high
pressure on
-
orbit.

An extensive effort to test a witness valve to destruction and much soul
-
searching led to the
conclusion
that
a reasonable gamble was to launch with th
e
original hardware

rather than replacing it. Though the
gamble has paid off

thus far, this failure really

caused
the Project

personnel
to

feel the pinch.
” Another example of how
ordinarily routine things can go awry is that the

spare primary mirror
, tho
ught early during its fabrication to be the piece
part that would actually become flight hardware,
was bent when the manufacturer attempted to lift it off the polishing
machine
for an optical test
before it had been unbolted

from the machine bed
. The
elas
tic limit

of the mirror was
exceeded and it became impossible to use
it as flight hardware given the launch schedule. The moral of this story, a
quintessential example of “Don’t do anything stupid,” is “D
o not try to lift the polishing machine with
your

m
irror.


A
more successful example of paying adequate attention to GSE
is

the story of the
down
-
looking
OSCAR
test flat
mirror

used to test the optical performance of the assembled cryogenic telescope assembly (CTA) in double pass at its flight
operating t
emperature of 5K
8
.
The manufacturer of OSCAR was able to test the optic only at 300K and, instead of a
cryogenic test, provided an engineering analysis based on knowledge of the thermal properties of the glass showing that
it would remain flat at 5K.
The
Optics IPT

concluded it
would be judicious
to test
OSCAR at its intended use
temperature of 5K. The open communication structure of the Project facilitated a decision to find $2 million in funding
for this unanticipated test, and an experiment

was assembl
ed in

the RAMBO
cryogenic test chamber at Ball Aerospace
to perform a
skip test. In this test
,

OSCAR was

hanging
face
-
down
at 5K and a 4
-
inch diameter laser interferometer
beam
was reflected off the flat at a glancing angle. This produced
36

narrow data s
ausages tha
t had to be stitched
together t
o reconstruct the surface of the mirror
. The conclusion of this test, after interminable argument about the
reliability of the process of stitching together the data sausages, was

that OSCAR

actually became a conv
ex sphere with
a radius of curvature of 3

km at 5° Kelvin.

The test was repeated several times to establish the repeatability of the
deformation, and the problem it might have caused was nullified by moving the point
-
source glower (a “fake” star) used
in

the end
-
to
-
end optical test a sufficient distance below the focal plane to ensure success.

With the

radius of curvature
in OSCAR
unaccounted for
, the double pass

optical test
would have failed because the small range of motion in the
secondary mirror foc
us mechanism would have been insufficient to focus the point
-
source glower on the camera arrays.
It is estimated that the Project would have spent

$120
-

$150 million and

a
year’s delay to find out what had

done wrong
with this test and
then to
te
st the op
tical system correctly
.

Use the spacecraft and instrument flight software

and the flight operations system ad protocols

during pre
-
flight testing:
The
Spitzer

Project used the

flight
software
and systems
for ground
-
based testing. This g
ave a high degree o
f confidence
that things would work on orbit

and proved to minimize the problems that were encountered after launch.

Software is always a problem:
Software problems plagued the Project up until launch. One problem was
that software
developed by one organi
zation with
in our integration contractor

was being rewritten

by
another organization within the
same contractor (but located 1,000 miles away) with little communication between the two divisions. The open
communication policy helped to overcome the proble
m. The software stub over
-
exercising the control gas valve was
mentioned above. A solution to another software problem was devised by having m
uch of th
e pointing control software
designed and built at JPL and provided to the contractor who did the final d
esign and the integration of the spacecraft
and observatory.

You can’t conduct too many inspections:

N
umerous
analytical and
visual inspections of the instruments

were performed
as they were being installed in the multiple instrument chamber. A benefit o
f this

was the finding of an

unanticipated
reflection in the IRAC camera due to an un
blackened

surface. Throughout

the integration process, reflections off
steel
tooling balls
were used
to make
optical and interferometric
measurements of the relative posit
ions of instruments with
respect to the multiple instrument chamber backplane and the nominal focal plane. This worked w
ell enough to lead to
another Project mantra: “Success requires
balls of steel
.”


5.

LESSONS LEARNED IN R
ETROSPECT

Some of what mak
es a mi
ssion successful is having

good luck

and making good choices
.
The SIRTF Project tried to err
on the side of conservatism by employing

the principle of
“Better i
s the enemy of good enough.


This meant giving a bit
of ground in projected performance when a

search for perfection might have caused an outright retreat in terms of





9

schedule and budget.
A few examples will serv
e to illustrate the principle. When a slight amount of trefoil deformation
caused by strain in the secondary mirror mount was detected

d
uring

the optics test program, the Optics IPT determined
by extensive analysis that the problem would not affect the ability of the optics to meet the Level
-
One requirement for
diffraction
-
limited performance. The IPT recommended that the secondary be use
d “as
-
is”. D
egradation

of

one of the
IRS spectrograph

filt
ers caused concern about the ability of the instrument to reach
its desired

performance predictions.
The Project opted not to warm the cryostat to change the filter,

and on
-
orbit
it was

found tha
t the performance was good
enough. When the IRAC calibration shutter failed
, it was

decided to leave well enough alone and

find

a way to calibrate
without the shutt
er
. This solution proved to be valid

on
-
orbit.
When
1/8 of the detectors on the
MIPS

160

µm

array
f
ailed, it was decided to forge ahead and
change the data acquisition algorithm

to take into account the missing detectors
with a small hit in performance. Again,

the performa
nce on
-
orbit was satisfactory

enough to
achieve most of
the
mission’s
s
cience objectives. A
nother example is the value of

the

image analysis

activity
the Optics IPT

conducted to
determine
the

focus

quality of

the telescope

on
-
orbit
. Originally, it was predicted
that three
secondary mirror focus
mechanism
moves
would be requ
ired
to reach focus
. Bu
t after two moves
,

the
analysis
tools developed by the Focus
IPT
showed that the images
, although not as good as they could have been

at the shortest IRAC wavelength
,

met

their

specification
s,

so the
focus mechanism
was disabled and

never had to be moved again. We note that the focus position
of the telescope did not change after the cryogenics were exhausted and the telescope warmed up to 30° Kelvin
, as
predicted by the extensive pre
-
launch test data collected and analyzed by the O
ptics IPT
.
The mission did have its

share
of bad

luck
. For example on
-
orbit the
MIPS

instrument suffered the loss of one side of the 70

µm

array due to an
unanticipated problem with
its
cabling.


Although an attempt was made to achieve orbit
-
like
thermal

conditions

in the BRUTUS thermal
-
vacuum test, the
experiment failed abysmally.

The lowest parasitic heat load achieved during testing was 50 mW compared to the on
-
orbit value of 4 mW, and the test had to be aborted and repeated due to failure to thermall
y clamp mechanical interfaces
between the warm environment and the cryogenic shrouds. It was determined that this test would have benefited from
considerable review internal to the Project

but
external

to the contractor.

On the other hand, the 5K BRUTUS

optic
al
test went very well in proving that the CTA
would achieve
Level
-
One
optical performance on orbit
.


6.

SUMMARY

The
Spitzer

Space Telescope Observatory has been a highly successful mission. Much of that success owes to the
Project having taken heed of
the many lessons learned within the Project and from other missions.


ACKNOWLEDGEMENTS

We thank the entire
Spitzer

team for
their

tireless work on the
Spitzer

project
.
The Spitzer Space Telescope is operated
by the Jet Propulsion Laboratory (JPL), Californ
ia Institute of Technology (Caltech) under contract with the National
Aeronautics and Space Administration (NASA) under NASA Contract No. 1407


R
EFERENCES

[1]

Rieke, G. H., "The Last of the Great Observatories: Spitzer and the Era of Faster, Better, Cheaper at

NASA,"
University of Arizona Press:Tucson
,

AZ
(200
6
).

[2]

Neugebauer, G., Habing, H.J., van Duinen, R., et al.

The infrared astronomical satellite (IRAS) mission
,”

Astrophysical Journal 278, L1


L6,
(
1984
)
.

[3]

Hall, D.

B. ed.,

The Space Telescope Observatory,


NASA CP
-
2244, 1982.

[4]

Kessler, M.F., Steinz, J.A., Anderegg, M.E., et al.
The Infrared Space Observatory (ISO) mission. Astronomy
and Astrophysics 315, L27


L31, 1996.

[5]

WIRE:
http://spacese.spacegrant.org/Failure%20Reports/WIRE_Mishap_summary.pdf

[6]

Mars Climate Orbiter:
http://www.space.com/media/mco_report.pdf

[7]

Werner,
M.W.,

Roellig,
T. L.,
Low,
F. J.
et al.,

The

Spitzer Space Telescope Mission,” Astrophys. J. Supp.

Series 154, 1 (
200
4).

[8]

Gehrz,
R. D.,
Roellig,
T. L.,
Werner
, M. W.
, et al.,

The NASA Spitzer Space

Telescope,” Rev. Sci. Instrum.

78 , 011302 (
2007
).






10

[9]

Fazio, G. G., Hora, J. L., Allen, L. E., et al.
"Th
e Infrared Array Camera (IRAC) for the Spitzer Space
Telescope," Astrophysical Journal Supplement Series 154, 10 (2004).

[10]

Rieke, G. H., Young, E. T., Engelbracht, C. W., et al., "The Multi
-
band Imaging Photometer for Spitzer
(MIPS)" Astrophysical Journal Su
pplement Series 154, 25 (2004).

[11]

Houck, J. R.; Roellig, T. L.; van Cleve, J., et al.
"The Infrared Spectrograph (IRS) on the Spitzer Space
Telescope ," Astrophysical Journal Supplement Series 154, 18 (2004).

[12]

Field, G. B., Belton, M. J. S., Burbidge, E. M. ,

et al., “Astronomy and Astrophysics for the 1980’s Volume 1:
Report of the Astronomy Survey Committee,” National Academy Press: Washington, D. C. (1982).

[13]

Bahcall, J. N., Beichman, C. A., Canizares, C., et al., “The Decade of Discovery in Astronomy and
Ast
rophysics for the 1980’s,” National Academy Press: Washington, D. C. (1991).

[14]

Boggess, N. W. Mather, J. C., Weiss, R., et al.
,


The COBE mission
-

Its design and performance two years
after launch
,”

Astrophysical Journal 397,
420

(1992).

[15]

How the Skunk Works
® Got Its Name:
http://www.lockheedmartin.com/aeronautics/skunkworks/name.html

[16]

Laine, S.
, Silbermann
,
N. A.,
Rebulla,
L. M., Storrie
-
Lombardi, L. J. “
Spitzer Space Telescope p
roposal
process
,”
Proc. of SPIE

6270
,

627029
-
10 (2006)