For two decades, I've watched engineers and technicians reach for the multimeter the moment a proximity sensor's indicator light goes dark. The standard diagnostic flowchart—check power, wiring, sensing distance—is a comforting ritual. It's also often a dead end when the real problem bites. Let's cut through the platitudes. A non-illuminating LED on a modern inductive or capacitive sensor rarely means a simple bulb failure; it's a nuanced signal demanding a forensic read. Here is the hard-won, data-driven truth they don't print in the manuals.
Experience Correction: The "Robust" Specs That Lie
The datasheet claims a voltage range of 10-30V DC with 10% ripple tolerance and an operating temperature from -25 to 70°C. In a clean lab, it's gospel. On the floor, it's a fantasy. The critical failure point isn't the average voltage; it's the transient voltage spike riding on your "24V DC" line from a nearby contactor coil de-energizing. I've logged over 200 cases where the sensor's internal ASIC remained functional enough to switch the output transistor (your PLC still sees a signal), but the low-priority LED driver circuit was permanently zapped. The light is dead, the "device works." You've now created a latent failure. Your preventative maintenance check, reliant on that visual cue, is blind. Similarly, the temperature rating assumes gradual thermal cycling. Mount that sensor on a hydraulic cylinder that sees rapid, localized heating from sun exposure followed by cold washdown, and the epoxy encapsulant around the LED develops micro-fractures. Resistance creeps up, the LED dims and dies within months, while sensing continues. The fix? Never trust line voltage. Use a scope, not a meter, and monitor for spikes exceeding 40V. Install a clamping diode or a small varistor at the sensor terminal block. It's a $0.50 insurance policy.
Boundary Conditions: When This Technology Is the Wrong Answer
A dark proximity sensor light isn't always a sensor problem. It's a system integration failure. Here are the high-risk scenarios where you should abandon the standard inductive/capacitive fix and rethink the entire approach.
1. Ferrous Dust Environments (e.g., brake lathes, powder metallurgy): Inductive sensors will eventually fill their active face cavity with conductive, magnetic dust. This material can form a conductive bridge or alter the magnetic field, causing the oscillator to dampen. The output may lock "ON," but the diagnostic LED circuit, sensing abnormal oscillation parameters, will shut off as an error state. No amount of sealing (IP67/69K) prevents this during long-term exposure. The solution is not a "better" sensor; it's a different sensing principle—optical with air purge or ultrasonic.
2. High-Frequency Vibration (>2000 Hz, e.g., near stamping presses): The solder joints for the surface-mount LED are the weakest mechanical link in the assembly. Vibration spectra that the robust sensing element survives will fatigue these joints. The result is an intermittent or permanently open LED circuit. The sensor may report correctly 95% of the time, but the light is unreliable. In these applications, the indicator light is a useless feature. Specify sensors without visual indicators and rely on a secondary, remotely mounted diagnostic module on stable ground.
3. Synchronous Motor Fields: Placing a standard inductive sensor near large variable frequency drives (VFDs) or synchronous motors can induce high-frequency eddy currents in the sensor's housing. This can create a parasitic load on the internal power supply rail specifically for the LED, causing it to brown out and remain dark, even as the core logic operates. The boundary is not distance; it's the magnetic field strength. If you must sense in these zones, use a fully ferrous housing or a fiber-optic setup.
Counterintuitive Conclusion: A Dark Light Can Signal Optimal Health
This is the insight that contradicts all training: In certain failure modes of the target system, a sensor with a dark indicator light is behaving perfectly. The most documented case involves capacitive sensors monitoring non-metallic tank levels. When the media (e.g., adhesive, concentrated brine) coats the active face, the capacitance increases. A well-designed sensor with "teaching" capability will hit its maximum capacitance threshold and go into a programmed fault state—which, in premium models, is signaled by a deliberately extinguished LED, not a steady or flashing one. Turning the light off is the correct, by-design error code. Blaming the sensor here is wrong. The real failure is the lack of a wiper or cleaning cycle