Designing for fault-tolerance in the commercial environment

Fault-tolerant techniques have been successfully used for implementing highly reliable electronic systems. There is a range of fault-tolerant techniques to cater for any desired level of fault-tolerance. However, only cost-effective techniques can be used for commercial systems. The cost of implementing fault-tolerance can be estimated from the extra hardware required. The possible savings to the user can be estimated from the cost of "prevented failures". To estimate the cost of electronic failures to the user of computer numerically controlled (CNC) machines, maintenance records from several machine tools were analysed. The results of the study show that electronic failures constitute less than 10% of all failures of CNC machines. However, they top the list of average repair costs for different failure categories. Permanent electronic failures also have longest down-times and may result in substantial losses. Fault-tolerance techniques can be used to make machine controllers more reliable. The challenge is how to do it cost-effectively. Apart from the original investment, some fault-tolerant implementations increase maintenance expenses thus offsetting the benefits to the user. An example of the cost-benefit analysis for a double-redundant system shows how the major components of the life cycle cost change depending on the implementation.