Safety circuits are the most important circuits on a machine. When a safety circuit faults, the machine will not run — and the temptation under time pressure is to look for a way around it. This guide is deliberately high-level, and the overriding rule is simple:
Safety devices must never be bypassed, defeated or overridden as part of fault finding.
These resources are for general engineering information and do not replace site-specific risk assessments, safe isolation procedures, manufacturer instructions, competent engineering judgement or applicable legislation.
The components of a safety circuit
- Emergency stops — maintained (latched) buttons that break the safety chain when pressed; must be reset (twist/pull/key) to restore.
- Guard switches / interlocks — confirm guards and doors are closed; some are safety-rated with coded actuators.
- Safety relays / safety controllers — monitor the e-stops, guards and field devices and only allow a reset when conditions are healthy.
- Permissives — conditions the safety controller (or PLC safety task) requires before allowing motion.
Diagnostic logic (high level)
- Is a device actually triggered? An e-stop latched in, or a guard open, is the most common cause — not a fault at all.
- Reset conditions — safety relays typically require a healthy inputs state and a deliberate reset action (a button, not automatic). Is the reset action present?
- Which device? Most safety controllers indicate which input has dropped. Use the diagnostics rather than guessing.
- Wiring and terminations — loose connections on a safety chain cause nuisance trips; check M12/M8 connectors, terminal blocks and cable flex points.
- Failed device — if one device is genuinely faulty, replace it with an equivalent safety-rated part; never short across it.
What is never acceptable
- Bridging out an e-stop, guard or safety relay to "get running".
- Repeatedly resetting without finding the cause.
- Substituting a safety-rated device with a non-safety-rated one.
Key points
- Treat every safety trip as real until proven otherwise; the proof is a genuine cause, not a workaround.
- Use the safety controller's diagnostics to localise — modern units tell you which input dropped.
- Document the cause so a recurring nuisance trip is fixed at source, not worked around.