Automatic In-Flight Repair of FPGA Cosmic Ray Damage
NATIONAL AERONAUTICS AND SPACE ADMINISTRATION MOFFETT FIELD CA AMES RESEARCH CENTER
Pagination or Media Count:
FPGAs are finding an increasing number of applications within NASA in deep space probes, planetary rovers and manned vehicles. Like other silicon devices, FPGAs can be damaged by high energy cosmic ray impacts, resulting in permanent latch-up conditions that manifest as stuck-at faults. Traditionally, multiple redundancy and voting logic have been employed as a work-around, particularly for high reliability, extreme environment applications. However, reconfigurable FPGAs are becoming increasingly common in flight systems, offering a potentially valuable possibility for improved levels of fault recovery. After a fault is detected and localised within an FPGA, it is feasible to reprogram the device, in flight, with an alternative, equivalent circuit that does not depend upon the damaged portion of the chip. Designing such alternative chip layouts by hand is a valid option, though costly in terms of the man-hours of effort required a fully automated alternative would be far preferable. In this paper, a technique is presented that allows the automatic generation of FPGA configurations for fault recovery purposes by means of non-clausal SAT solver technology.
- Theoretical Mathematics
- Manned Spacecraft