Logic built-in self-test

Template:Wikify is deprecated. Please use a more specific cleanup template as listed in the documentation.

Logic built-in self-test, or LBIST, is a mechanism that permits an integrated circuit to test the integrity of its own digital logic structures. LBIST operates by stimulating the logic-based operations of the integrated circuit and then detecting if the logic behaved as intended. In practice, an integrated circuit with LBIST has special test circuits embedded for generating the stimulus and detecting the response.

LBIST is similar to memory built-in self-test (MBIST) in that both test digital structures on an integrated circuit. MBIST tends to be of much simpler construction because a memory consisted of many repetitions of a single (memory) structure. LBIST, on the other hand, typically tests a wide range of logic structures wired in a highly randomized fashion, and so requires more complex test circuitry.

There are many implementations of LBIST, but almost all depend on generation of a pseudo-random sequence as stimulus for the design, while the response to this stimulus is captured in a MISR (Multiple Input Shift Register). This MISR is used to generate a "signature”. This signature is unique in the sense that each failure in the device would lead to a different value at the end of the whole process.

The utility of LBIST has long been debated throughout the VLSI industry due to its random approach to generating stimulus. The utility of LBIST has been taken over in most cases by the Automatic Test Pattern Generation methodology where the stimulus can be tailored to exercise the circuitry with a degree of determinism.

References

This engineering-related article is a stub. You can help Wikipedia by expanding it.