Normally Open vs Normally Closed Sensors: Which Output Do You Need?

What do normally open and normally closed mean?

For a typical powered proximity sensor, normally open (NO) means the switching output is OFF without a detected target and ON when the target is detected. Normally closed (NC) reverses those states. “Normal” is the specified unactuated condition—not the condition in which your machine spends most of its time.

Here, ON means the output switching element conducts through its intended load path. It does not universally mean a high signal voltage, a lit LED or permission for the machine to run.

Is an electronic NC output the same as a closed contact?

No. An unactuated mechanical NC contact provides a conductive contact path; a mechanical NO contact does not. For a conventional relay, the reference is its de-energized coil. A transistor-output sensor instead needs power for its electronics. An NC label does not turn it into a passive closed contact when that power disappears.

For magnetic switches, pressure switches and configurable sensors, read the product’s stated reference condition and operating chart. Do not infer the meaning from the physical position of a door, cylinder or workpiece alone.

How is NO/NC different from PNP/NPN?

NO/NC tells you when the output conducts; PNP/NPN tells you which current path it switches. You can therefore have PNP NO, PNP NC, NPN NO or NPN NC. Changing one choice does not automatically change the other.

Conceptual conventional-current paths while the output is ON, not terminal assignments. Follow the exact sensor and input-module diagrams.

An active NPN output pulls its signal terminal toward 0 V, yet a correctly matched input can report an active bit. This is why “NC means high” is an unreliable wiring rule. Verify the input common, thresholds, output rating and connector viewing direction—not just wire colors.

What should the datasheet show separately?

Look for both output configuration and operating mode. For example, OMRON’s E2E NEXT DC 3-wire table identifies B-series output configurations as PNP and C-series as NPN, then distinguishes B1/C1 NO modes from B2/C2 NC modes. These are family-specific code elements, not complete order codes or a naming rule for other brands. The same documentation lists NO/NC switching through IO-Link for applicable models. Check the saved configuration as well as the nameplate.

What changes with two-wire or complementary outputs?

A two-wire electronic sensor carries its operating current through the load circuit. OFF-state leakage can keep a sensitive input active, while ON-state voltage drop reduces the voltage available to the load. Verify both against the input specifications; a continuity result alone cannot qualify the combination.

A complementary sensor offers opposite switching outputs on separate connections. Use only the documented pins and loading limits. Complementary outputs are not automatically two independent safety channels.

Does Light-ON mean normally open?

Not as a universal target-presence rule. Light-ON activates the output in the receiving-light state; Dark-ON activates it in the low-light state. What the target does to that light depends on the sensing method.

• Through-beam or retro-reflective interruption: an opaque target reduces the received light. Dark-ON is commonly used for an output that activates when the target blocks the beam.
• Conventional diffuse reflection: the target returns light to the receiver. Light-ON is commonly used for an output that activates when the target is detected.

For transparent targets, background suppression or specialized modes, use the exact operating chart and actual target/background test. First write “output ON when the part is detected” or “output ON while the path is clear”; then select the documented optical mode.

Which output fits your application?

Start with the physical condition you need to recognize and the signal you expect during that condition. A healthy energized signal can require NO or NC. The deciding factor is whether the healthy condition includes a detected target.

Part arrival or confirmed presence: start by considering NO

For a conventional proximity sensor counting metal parts, NO gives an active signal during detection. It is easy to interpret as “part detected.” If that signal becomes a process permissive, define what happens when it is missing, delayed or stuck active.

A clear sensing area: NC may fit the process logic

When the expected condition is “no target in this sensing area,” NC can provide an energized clear-state signal that disappears on detection. Losing that signal can also inhibit the process, provided the connected input and logic are designed to respond that way. It does not identify the cause of the loss.

Will NC detect a broken wire or make the machine safe?

NC alone provides neither complete wire-fault diagnosis nor a safety rating. An energized signal can be lost for more than one reason, and some faults can keep it active when it should change.

• Open signal wire: a basic input may see the same inactive state as a valid detection event on an NC sensor. It can react to that state without knowing which event occurred.
• Short circuit: a PNP signal shorted to positive supply can appear continuously active; a short to 0 V can force the opposite state. Evaluate the actual circuit, including NPN and shared-common arrangements.
• Sensor power lost or restored: the powered truth table no longer describes the full event. Input bias, backfeed paths, start-up delay and possible switching transients matter. Check model-specific power-on/off behavior.

Fault-specific information may require a monitored interface, device diagnostics or an evaluated complementary-signal check. Two opposite outputs can reveal some implausible states, but they do not inherently detect every internal, common-cause or wiring failure.

Why can the PLC read the opposite state?

Separate three observations: target condition, raw input bit and the program’s process tag. A correct sensor can appear inverted because the program intentionally reverses its input—or because someone reversed it twice.

Rockwell’s XIC instruction tests whether a bit is set; XIO tests whether it is cleared. Their contact-like symbols do not identify the physical sensor’s NO/NC mode. Record the raw bit for both target states before interpreting the ladder diagram.

Can an NO-to-NC replacement move the count event?

Yes. Consider a counter that increments on the raw input’s rising edge, with a compatible input that is active when the sensor output is ON. An NO sensor normally produces that edge on target entry; an NC replacement produces it on target exit. The part count may look plausible while the trigger occurs at the wrong position.

Check the configured edge, input filter, sensor response and the shortest real pulse and gap. Changing a tag’s polarity can restore the intended event, but it does not correct an incompatible output circuit or add missing fault diagnosis.

How should you test and approve a replacement?

Approve the complete device-to-controller behavior, not just a sensor LED or the letters NC on a quotation. Use the exact part number, connection drawing and current configuration as the starting evidence.

Use the right test for the output hardware

Dry contact: isolate the circuit, verify absence of voltage and remove misleading parallel paths before checking continuity in the rest and actuated states. Follow the meter instructions; a beeper is not proof that the contact meets its load rating.

Electronic output: use the rated supply and a compatible load or input, following the manufacturer’s circuit. Measure the loaded signal relative to the correct reference and compare it with the raw input bit. Do not use continuity mode on energized electronics or judge a transistor output solely by an unloaded voltage reading.

Electrical work and powered tests belong with qualified personnel. Isolate hazardous machine energy before changing connections; perform powered checks on a protected low-voltage bench or under an authorized commissioning procedure that prevents unintended motion. Do not improvise live short-circuit tests.

Record the states that the machine will actually use

1. Establish a healthy baseline. Record supply, output mode, input common, target geometry and all relevant sensor/PLC settings. Confirm the defined start-up behavior.
2. Test both target conditions. Capture the sensor indication, loaded output and raw bit with the target detected and not detected. Keep the processed process tag as a separate observation.
3. Verify the real event. Repeat at operating speed and the least favorable target position. Confirm the selected counting edge and that short pulses and gaps reach the controller.
4. Check the agreed loss-of-signal response. Under the authorized test plan, verify power interruption, the relevant open-circuit condition and recovery. Record the actual alarm or process inhibition; do not treat one successful test as coverage of every fault.
5. Release a reproducible configuration. Save the full sensor code/suffix, cable or connector pinout, input-module model, settings and observed state record with the replacement approval.

What should a supplier confirm before you order?

Request the complete order code and its datasheet, not only “M12 NC.” Confirm sensing method and usable distance for your target, NO/NC or Light-ON/Dark-ON behavior, PNP/NPN or other output technology, supply and load compatibility, connector pinout, timing and environmental limits. If the behavior is configurable, specify the delivered setting and how it is restored after replacement.

A useful requirement is concrete: “With this target detected, this input must be active; with it removed, it must be inactive.” Add the required response to signal loss separately. That gives purchasing, the supplier and the PLC engineer the same behavior to verify.