A pre-FAT code review finds ambiguous ownership, unsafe state transitions, hidden bypasses, weak diagnostics, and untestable interfaces before the factory floor becomes the test environment. The review should connect each finding to a requirement, revision, risk, owner, and FAT case.
The short answer
Freeze the review candidate and its build environment, then inspect structure, data ownership, modes, sequences, interlocks, alarms, communications, restart behavior, diagnostics, security-relevant settings, and test hooks against approved requirements. Record findings by severity and disposition. FAT should begin only after blocking findings are corrected in a new controlled revision and the test team can trace every critical requirement to an executable case.
- Review a reproducible source revision and exact target configuration, not an engineer’s live workstation.
- Trace safety and process requirements to code modules, interfaces, alarms, and planned tests.
- Distinguish correctness, maintainability, diagnosability, cybersecurity, and performance findings.
- Close every accepted, corrected, deferred, or rejected finding with an owner and rationale.
Decision guide
| Review area | Questions to ask | Release evidence |
|---|---|---|
| Architecture and data | Are responsibilities, interfaces, writers, types, units, limits and retained values explicit? | Module map, interface list, naming checks and dependency review |
| Modes and sequences | Are transitions, permissives, interlocks, timeouts, aborts, restart and recovery deterministic? | State model, cause-and-effect trace and abnormal-path tests |
| Operations and diagnostics | Can operators identify cause, affected scope and safe next action? | Alarm mapping, diagnostic paths, simulation and fault-injection results |
| Configuration and release | Are hardware, libraries, communications, access, clock and build versions controlled? | Export, dependency versions, checksums, review approval and FAT baseline |
A practical workflow
- 01
Define review scope, reviewers, severity rules, required evidence and the exact source, library, hardware and firmware baseline.
- 02
Map requirements and cause-and-effect entries to modules, data interfaces, modes, states, alarms and test cases.
- 03
Walk normal start, stop and production paths, then power loss, communication loss, sensor faults, actuator faults, restart and manual recovery.
- 04
Run static checks, simulation and selected bench tests; capture each finding with location, consequence, recommendation, owner and disposition.
- 05
Correct blocking findings in a new revision, repeat affected checks, approve the FAT baseline and prohibit uncontrolled online changes.
Evidence and acceptance
Retain the review plan, source and compiled checksums, hardware and software versions, architecture, requirement trace, code-review findings, severity and disposition, screenshots or exports, simulation results, corrected revisions, approvals, open risks, FAT cases, and the baseline released to the test team.
Common mistakes
- Reviewing style while skipping restart, fault, manual and recovery behavior.
- Inspecting code without the hardware configuration, libraries, symbols and requirements used to build it.
- Allowing findings to disappear in chat or email without ownership and retest evidence.
- Treating a successful download or clean compile as functional acceptance.
- Changing code during FAT without a new revision, impact review and repeated tests.
Engineering and security boundary
Primary sources
Technical claims and product boundaries were checked against these primary sources. Fact-checked
Frequently asked questions
When should the PLC code review happen?
Review the design early, then perform a formal review on the frozen pre-FAT candidate with its exact dependencies and configuration.
Is a compiler warning check enough?
No. The review must cover requirements, architecture, modes, abnormal behavior, diagnostics, interfaces, access and testability.
Can findings remain open at FAT?
Only explicitly accepted non-blocking findings with owners, risk rationale and closure dates should remain.
Does AutoNexit review PLC logic automatically?
No. Qualified reviewers perform the engineering review; AutoNexit keeps its inputs, findings, revisions and evidence connected.