The Spitzer Space Telescope has a two-phase downlink system. Recorded data are transmitted during one telecom session. Then commands are sent during the next session to delete those data that were received on the ground and to retransmit those data that were missed. We must build science sequences that are as efficient as possible to make the best use of our supply of liquid helium. One way to improve efficiency is to use only the minimum time needed during telecom sessions to transmit the

more »

... cted volume of data. But, we must also not fill the on-board storage and must allow enough time margin to retransmit missed data. We describe tools and procedures that allow us to build science sequences that are single-fault tolerant in this regard and that allow us to recover quickly and safely from anomalies that affect the receipt or acknowledgment (i.e. deletion) of data. 2 We discuss the software tool that we have developed to analyze the current and future state of the on-board storage. The tool calculates the size, type and urgency of any backlogs, and guides the choice of actions to recover from an anomaly during sequence execution. We also discuss the procedures themselves and the failure-mode analysis used to develop them. We discuss the future development of the tool. Generally, the tool assumes that the most aggressive approach for acknowledgement and retransmission of data will be taken. This is not always necessary. In future versions of the tool, we hope to relax some of these constraints and allow easier analysis of alternative approaches. We review actual problems that have affected the downlink system in flight. We review the consequences of those problems and the way we have either solved them or minimized their effect. We examine potential changes to Spitzer flight software that would avoid filling the on-board storage. We retrospectively identify some areas of observatory telemetry and telecommand that might have been designed differently to simplify operation of the downlink system. We finally examine alternative downlink system designs and how the same sort of analysis described here might be applied.