Performance Level (PL a–e), ISO 13849-1
PL measures how reliable a safety function is, expressed as the average probability of a dangerous failure per hour (PFHD). PL a is the lowest and PL e the highest.
| PL | PFHD (dangerous failures per hour) | Roughly |
| a | 10⁻⁵ to < 10⁻⁴ | Lowest risk reduction |
| b | 3 × 10⁻⁶ to < 10⁻⁵ | Low |
| c | 10⁻⁶ to < 3 × 10⁻⁶ | Medium (minimum for an e-stop per ISO 13850) |
| d | 10⁻⁷ to < 10⁻⁶ | High; common on guarded machine access |
| e | 10⁻⁸ to < 10⁻⁷ | Highest |
Required PL (PLr) comes from the risk graph in ISO 13849-1, which asks three questions about each hazard:
- S (severity): S1 is a slight, reversible injury. S2 is a serious, irreversible injury or death.
- F (frequency/exposure): F1 is seldom or short exposure. F2 is frequent or long exposure.
- P (possibility of avoidance): P1 means the person can avoid or limit the harm. P2 means they scarcely can.
For example, S2 + F2 + P2 gives PLr e, and S1 + F1 + P1 gives PLr a. The PL you achieve is then calculated from four things: the Category (architecture), MTTFD (mean time to dangerous failure of each channel), DCavg (diagnostic coverage, meaning how many faults are detected), and CCF (measures against common-cause failures). The achieved PL must be at least the PLr.
Categories B, 1, 2, 3, 4, ISO 13849-1
The Category describes the architecture of the safety circuit: how many channels it has and what happens when a component fails.
| Category | What it means | On a fault | Typical machine use |
| B | Single channel built to basic safety principles | Safety function can be lost | Low-risk functions only; rarely the right answer for guarding |
| 1 | Single channel using well-tried components and principles | Can still be lost, but faults are less likely | A positive-opening guard switch on a low-risk cover |
| 2 | Single channel that is tested periodically | Loss is possible between tests; the test catches it | Type 2 light curtain with a start-up test on a low-energy transfer |
| 3 | Dual channel and single-fault tolerant | One fault doesn't lose the function; most faults are detected | Dual-channel e-stop and light curtain into a monitored safety relay |
| 4 | Dual channel with high diagnostics | Faults are detected in time; a build-up of faults still doesn't lose the function | Safety PLC or controller with OSSD inputs and monitored outputs, e.g. high-risk robot or palletizer cells |
Safety Integrity Level (SIL 1–3), IEC 62061
SIL is the IEC way to rate the same thing. For high-demand or continuous machinery functions, it uses the same PFHD bands:
| SIL | PFHD (dangerous failures per hour) |
| SIL 1 | 10⁻⁶ to < 10⁻⁵ |
| SIL 2 | 10⁻⁷ to < 10⁻⁶ |
| SIL 3 | 10⁻⁸ to < 10⁻⁷ |
SIL 4 exists in IEC 61508 (and the process-industry standard IEC 61511) but is not used for machinery. IEC 62061 stops at SIL 3.
PL vs SIL, approximate comparison
| PL (ISO 13849-1) | ≈ SIL (IEC 62061) |
| PL a | No SIL equivalent |
| PL b | SIL 1 |
| PL c | SIL 1 |
| PL d | SIL 2 |
| PL e | SIL 3 |
The comparison is approximate and based on the PFHD ranges. The two standards calculate differently, so pick one method per safety function and stay with it.
"CAT" means two different things. A Category (B, 1–4) in ISO 13849-1 describes circuit architecture. A Stop Category (0, 1, 2) in IEC 60204-1 and NFPA 79 describes how the machine stops:
- Stop Cat 0: Power is removed immediately, so the machine coasts to a stop.
- Stop Cat 1: A controlled stop (the drive ramps down), then power is removed.
- Stop Cat 2: A controlled stop with power left on, so it can't be used for an e-stop.
A machine e-stop is often "Category 3, PLd" for the circuit and "Stop Category 0 or 1" for the stop. Both can be correct at the same time.
Wiring basics
- OSSD test pulses: Safety outputs from curtains, scanners, and electronic interlocks drop out briefly on a staggered schedule. That lets the device catch shorts to 24 V, to 0 V, or between the two channels. The receiving safety input has to tolerate those pulses.
- PNP vs NPN: Safety I/O in North America is typically PNP (sourcing 24 V). Some devices, such as the Keyence GL-R and SZ-V, come in NPN versions too, so match the cable or setting to the controller.
- EDM / feedback loop: The NC auxiliary contacts of the two force-guided contactors run back to the safety device. If either contactor is welded, the system refuses to reset.
- Monitored manual reset: The reset only takes effect when the button is pressed and released, so a stuck or tied-down button can't restart the machine. Mount it outside the hazard zone with a view of the area.
- Power-to-release vs power-to-lock: A power-to-release lock stays locked when power is lost and suits hazards that coast to a stop. A power-to-lock lock opens when power is lost, so use it only where the hazard stops before anyone can reach it.
Device-by-device detail is in Safety hardware we design with above.
Worked machine example (typical, not a design)
Take a guarded transfer section with an operator access opening, a light curtain across it, and e-stops along the line. A risk assessment for this setup may come out at PLr d. A common way to meet it is a Category 3 design:
- Dual-channel e-stops and a Type 4 light curtain (e.g. a Keyence GL-R or Banner EZ-SCREEN-class unit, OSSD outputs)
- A monitored safety relay, or a safety PLC/controller when there are several zones, muting, or sequenced restarts
- Output through Safe Torque Off (STO) on the VFD, or redundant monitored contactors, with a Stop Category 0 or 1 stop
- The curtain placed at the ISO 13855 safety distance, based on the measured stopping time, with a manual reset outside the hazard zone
Then the design is verified (PL calculation, e.g. with the IFA SISTEMA tool) and validated against ISO 13849-2 before sign-off.
Important: This section is general education, not engineering advice. Final required levels (PLr or SIL) come from your machine's risk assessment, and safety functions should be designed, verified, and validated by a qualified safety engineer.