Occasionally an observation would fail due to an instrument glitch, or due to a Single Event Upset leaving an instrument in an unsuitable configuration, or even in Safe Mode. When this happened, a list of failed observations was compiled and, on receiving the authorisation of the Project Scientist to do so, the observations were automatically cloned and released for re-scheduling; it was not necessary for the observer to request re-observation. The observer would be informed of all failed observations when they were cloned and ready for re-observation.

All observations went through Quality Control checking as part of routine processing and have a Quality Control summary attached in the Archive (see Section 8.3, “Quality Control” for more details). As a final stage of Quality Control, all observations were subjected to a visual inspection by an instrument specialist, who would flag any observations of deficient quality. When such observations were found, a decision would be taken as to whether or not the observation was sufficiently compromised as to have failed to reach its objectives. Once analysed, a recommendation was made to the Project Scientist, who could decide that the observation should be repeated in full, or in part, or that it need not be repeated.