Glossary · Functional safety
Fault tolerance
Also known as: hardware fault tolerance, HFT
German: Fehlertoleranz
In functional safety, fault tolerance is the ability to continue performing a required function in the presence of specified faults. Hardware fault tolerance N means that N + 1 faults can cause loss of the function.
- Functional safety
- Systems engineering
In one sentence
Fault tolerance is a system's ability to keep performing a required function despite specified faults, often achieved through redundancy.
Example
A two-channel light curtain evaluation on a robot cell has a hardware fault tolerance of 1: one failed channel does not prevent the stop, but a second fault could.
How it applies
- Functional safety: IEC 61508 and IEC 62061 use hardware fault tolerance, together with the safe failure fraction, to limit the SIL a subsystem can claim. In ISO 13849-1, Categories 3 and 4 express single-fault tolerance.
- Systems engineering: State which faults are tolerated. "Fault-tolerant" without a fault list is not a verifiable claim.
- Integrated systems: Tolerating a fault does not mean ignoring it. The first fault must be detected and reported, otherwise a second fault can remove the protection unnoticed.
- Technical documentation: Explain to users what happens after a first fault, for example a degraded mode, and what they must do.
Fault tolerance vs. fail-safe behavior
Fault tolerance keeps the function working despite a fault. Fail-safe behavior accepts losing the function but ensures the system goes to a safe state.