Hubble Space Telescope mirror
| Launched | 24 April 1990, Space Shuttle Discovery |
|---|---|
| Flaw announced | 21 June 1990 |
| Error | spherical aberration, about ten times the specified tolerance |
| Template | the reflective null corrector, kept unchanged since manufacture |
| Spacing error | the field lens, 1.3 mm too far from the lower mirror |
| Other tests | an inverse null corrector, and a refractive one, both of which showed the error |
The primary mirror of the Hubble Space Telescope was polished to the wrong shape. The telescope was launched on 24 April 1990. On orbit it would not focus. Both high-resolution cameras showed the same distortion, spherical aberration, and on 21 June 1990 the project manager said so. A board chaired by Lew Allen found that the 2.4-metre mirror had been figured against an optical template, the reflective null corrector, whose field lens sat 1.3 mm too far from the lower mirror. That one measurement accounts for the blur.[1]
The mirror was not a bad piece of glass that had been polished carelessly. It was a very good piece of glass that had been polished, carefully, to the wrong instructions.
The cap[edit]
The position of the optics was set with metering rods of Invar, by reflecting a beam of light off a polished rod end and reading the interference. To keep the beam centred, Perkin-Elmer fitted a field cap over the end of the rod, with a small hole in the middle. The top of the cap was painted so that it would not reflect. Some of that paint had come off, in a small area around the hole. The operator, the board concluded, took the reflection from the bare metal of the cap rather than from the end of the rod. A test with the same equipment in 1990 found that this was quite easy to do, and even probable. The screws then had no adjustment left, spacers were added, and the lens was left 1.3 mm out.[1]
The corrector was preserved afterwards exactly as it had been during manufacture. When the board measured it, the error was still there.
The instruments that were right[edit]
Two other tests, made at the time, showed the same error. An inverse null corrector, built to imitate a perfect mirror, showed it in the template. A refractive null corrector, used to measure the finished mirror, showed it in the glass. Both results were set aside, on the view that those instruments were the ones that were flawed. The manufacturing plan had placed complete reliance on the reflective null corrector, and no check of its dimensions was made after it was first assembled. Staff inside the optical division were worried. The worry did not get out of the division.[1]
The telescope flew with a mirror that two of its own measuring instruments had already failed.
See also[edit]
References[edit]
- ^ Lew Allen and others, The Hubble Space Telescope Optical Systems Failure Report, NASA, November 1990.