Low-Cost, Radiation-Tolerant, On-Board Processing Solution

Advancements in payload technology are creating opportunities for the aerospace community. Commensurate with these opportunities are great challenges. Until recently, technological growth was accommodated by increased data storage. However, with today´s systems of terabyte capacities and gigabyte-per-second instruments, missions are being limited by the inability to download this data. The need to get data downloaded quickly is driving the need for on-board processing. One difficulty for on-board processing is how to provide reliable processing capability for the space radiation environment in a cost-effective manner. SEAKR Engineering´s approach to this problem was to use state-of-the-art commercial processors and support devices to provide the lowest cost with the highest processing performance capability, and then characterize these devices in a radiation environment. Radiation effects are then mitigated or recovered from by using support electronics implemented in radiation tolerant devices. The result is a low cost, high performance, on-board processor with a low upset rate. High performance single board computers require the integration of many different technologies including SDRAM, nonvolatile memory, interface devices, processors, custom logic, and high current power supplies. In order to achieve high performance processing capability, our architecture employs commercial devices. Each of these devices was characterized for radiation effects. Radiation effects are addressed with various mitigation techniques. From the characterization data and mitigation techniques, system upset rates for each element are calculated for the mission-specific environment. These effects and rates are rolled up into a board level upset rate. Board level thermal analysis is used to determine a reliability estimate. A custom board support package (BSP) has also been developed that allows user application code to interface to the SBC hardware peripherals. It is written in VxWorks- 5.5 and uses a Tornado 2.2 debugging environment. Performance benchmark tests results such as Dhrystone are used to determine the processing capability. The paper describes the approach for creating a cost competitive product for the space environment