7
Importance of Integration
Integration Issues
Optical systems designed to operate on ground-based mounts endure as
exemplars of stability and precision. Building a very large, meticulously
crafted optical instrument into a package (that has passed thousands of tests
simulating launch and deep space temperatures and pressures) that will
then be accelerated to three times the acceleration of gravity in the Shuttle
Orbiter payload bay and ferried to orbit is a time-consuming, exacting systems engineering, engineering, and systems integration feat.
The types of subsystems on the Hubble Space Telescope include physical, electrical, optical, electronic, thermal control, power (generation, distribution, and
management), communications, computer processing and storage, pointing,
orbital stability, and operational/housekeeping configurations. Representative
performance issues focus on optical reflectivity and light baffling, pointing
accuracy, weight management, power consumption, and adhering to proper
temperature design parameters. Representative quality issues focus on the size
of the optical airy disk, pointing variances, power fluctuations, and temperature
excursions beyond design variances from specifications.
Integration Problems
There is a veritable rule for systems that are planned to launch into space and
expected to operate unattended for years—if they do not work on the ground,
they will not work in space. Unfortunately, these systems can work perfectly on
the ground, and still not work in space. The space environment is one that is
difficult to emulate on Earth. Specifically, large vacuum chambers that have full
instrumentation to test the completed system are extremely rare. Large-sized
vacuum chambers with thermal and vibration controls to mimic the intense
heat of the Sun and the deep cold of dark space are even rarer. The Hubble
Space Telescope was larger than existing chambers and thus not subjected to
the full range of tests that would cover launch conditions and the gravitational
anomalies that would occur over the course of its 83,000 orbits around Earth.
Smaller, unit-sized elements up to subsystems were tested instead. Additionally,
the practice had been to build two “payloads,” one a qualification unit for
launch and one a backup unit, both built to near the same specifications. In case
of a component or subsystem failure of one, the backup was a ready solution to
stay on schedule. However, for the Hubble Space Telescope, the systems integration approach was to “design–build–test–fix” in an iterative fashion that
conformed to the prime contractor’s previous work experience.
Compounding the systems integration problems, the project was over budget* and severely behind schedule, and subsystems were failing their acceptance tests. Everything from the scientific instruments to the ground and
control system needed considerable work to improve reliability to sustain
* Original budget estimates were $200 million, compared to a final cost near $2 billion (Mattice
2005).
Importance of Integration
Integration Issues
Optical systems designed to operate on ground-based mounts endure as
exemplars of stability and precision. Building a very large, meticulously
crafted optical instrument into a package (that has passed thousands of tests
simulating launch and deep space temperatures and pressures) that will
then be accelerated to three times the acceleration of gravity in the Shuttle
Orbiter payload bay and ferried to orbit is a time-consuming, exacting systems engineering, engineering, and systems integration feat.
The types of subsystems on the Hubble Space Telescope include physical, electrical, optical, electronic, thermal control, power (generation, distribution, and
management), communications, computer processing and storage, pointing,
orbital stability, and operational/housekeeping configurations. Representative
performance issues focus on optical reflectivity and light baffling, pointing
accuracy, weight management, power consumption, and adhering to proper
temperature design parameters. Representative quality issues focus on the size
of the optical airy disk, pointing variances, power fluctuations, and temperature
excursions beyond design variances from specifications.
Integration Problems
There is a veritable rule for systems that are planned to launch into space and
expected to operate unattended for years—if they do not work on the ground,
they will not work in space. Unfortunately, these systems can work perfectly on
the ground, and still not work in space. The space environment is one that is
difficult to emulate on Earth. Specifically, large vacuum chambers that have full
instrumentation to test the completed system are extremely rare. Large-sized
vacuum chambers with thermal and vibration controls to mimic the intense
heat of the Sun and the deep cold of dark space are even rarer. The Hubble
Space Telescope was larger than existing chambers and thus not subjected to
the full range of tests that would cover launch conditions and the gravitational
anomalies that would occur over the course of its 83,000 orbits around Earth.
Smaller, unit-sized elements up to subsystems were tested instead. Additionally,
the practice had been to build two “payloads,” one a qualification unit for
launch and one a backup unit, both built to near the same specifications. In case
of a component or subsystem failure of one, the backup was a ready solution to
stay on schedule. However, for the Hubble Space Telescope, the systems integration approach was to “design–build–test–fix” in an iterative fashion that
conformed to the prime contractor’s previous work experience.
Compounding the systems integration problems, the project was over budget* and severely behind schedule, and subsystems were failing their acceptance tests. Everything from the scientific instruments to the ground and
control system needed considerable work to improve reliability to sustain
* Original budget estimates were $200 million, compared to a final cost near $2 billion (Mattice
2005).