Sensors and Switches

Field devices are where the control system meets the physical process, and they live in the harshest part of the plant — exposed to vibration, heat, moisture, and contamination. A large share of what appears as control faults are really field device faults, so knowing the common types and their failure modes pays off constantly.

The common devices

Industrial sensing spans a range of devices with shared troubleshooting logic. Proximity sensors detect the presence of an object without contact — inductive types sense metal, capacitive types sense a wider range of materials. Photoelectric sensors use a light beam broken or reflected by a target. Limit switches detect physical position through mechanical contact. Pressure, level, temperature, and flow switches sense a process condition and change state at a setpoint. Most of these present a simple contact — open or closed, or a sinking or sourcing signal — so they troubleshoot much like a switch plus a sensing check: is the device sensing the condition, and is its signal reaching where it needs to go.

A method for a suspect device

The key to sensor troubleshooting is separating two questions that a single check often answers: is the device sensing, and is its signal arriving. Many sensors have an indicator LED that shows their output state, which reveals whether the device is responding to its target independent of any wiring. Present the target and watch the indicator: if it responds correctly but the control system does not see the change, the device is sensing and the fault is in the wiring or the input; if it does not respond, the device itself, its power, its alignment, or its adjustment is the problem. For devices without an indicator, a meter at the device’s output terminals accomplishes the same split. This one distinction — sensing versus signal delivery — organizes nearly every field device fault.

Physical causes and adjustment

Field devices fail from the physical world as often as from electrical causes, and these causes are easy to miss if you think only electrically. A proximity sensor mounted slightly too far from its target works when everything is perfect and fails when vibration or thermal expansion widens the gap. A photoelectric sensor drifts out of alignment or fouls its lens with dust or coolant until its beam no longer carries. A mechanical limit switch wears, sticks, or has its actuator bent out of position. A pressure or level switch drifts off its setpoint. When a device’s indicator does not respond to a condition you believe is present, check the physical relationship — gap, alignment, cleanliness, actuation, adjustment — before condemning the device electrically, because the device is often healthy and merely unable to sense correctly in its current physical state.

A case file: the sensor that failed when it got hot

A proximity sensor detecting a machine position works reliably at the start of a shift but begins to fail intermittently once the machine has been running for a while and the area has warmed up. The correlation with temperature points at a thermal effect: either the sensor itself is marginal when hot, or thermal expansion is changing the gap between the sensor and its target. Investigating, the technician finds the sensor mounted at the outer edge of its sensing range, and as the machine warms and its metal expands, the gap widens just enough that the sensor can no longer reliably detect the target. When cold, the gap is within range; when hot, it is not. The fix is to reposition the sensor for a reliable gap across the full temperature range, or to use a sensor with greater sensing range. The pattern illustrates that sensor faults correlating with temperature often come down to thermal expansion changing a marginal gap, and that a sensor set up at the edge of its range works cold and fails hot, a fault found not by replacing the sensor but by correcting the gap for all operating conditions.

The device’s own indicator

Most industrial sensors have an output indicator that is one of the fastest diagnostics available, because it shows what the sensor thinks it is doing independent of any wiring back to the control system. Present a target to a proximity or photoelectric sensor and its indicator should change; operate a limit switch and its state should change. This splits the two questions that sensor faults reduce to: is the device sensing, and is its signal arriving. If the indicator responds correctly to the target but the control system does not see the change, the device is sensing and the fault lies in the wiring or the input — a broken wire, a loose terminal, a convention mismatch. If the indicator does not respond to the target, the device itself, its power, its alignment, or its adjustment is the problem. Two glances at an indicator thus divide the fault space in half before any meter comes out, and for devices without an indicator, a meter at the output terminals accomplishes the same essential split between sensing and signal delivery.

Sinking, sourcing, and convention mismatches

Many industrial sensors provide their output by either sourcing current (often called PNP) or sinking current (often called NPN), and a mismatch between a sensor’s output type and the input it feeds is a common and confusing fault, because the sensor and the input can each be perfectly healthy while failing to work together. A sourcing sensor wired to an input expecting a sinking signal, or the reverse, may never register, may sit in an indeterminate state, or may behave erratically, all without any component being faulty. This is especially a trap when replacing a sensor, since a replacement of the wrong output type will not work in place of the original even though it is new and functional. When a sensor and input will not communicate despite both appearing healthy — the sensor’s indicator responding correctly to its target but the input never registering — a convention mismatch between sourcing and sinking is a prime suspect, resolved by confirming the sensor’s output type matches what the input expects rather than by replacing either healthy component.

Signal integrity from device to input

Once a sensor is confirmed to be sensing correctly — its indicator responding to its target — the remaining question is whether its signal reaches the input intact, and this signal path has its own failure modes worth understanding. A broken or loose wire between the sensor and the input opens the signal, a short compromises it, and for analog signals, noise picked up along the route can corrupt it, particularly where signal wiring runs too close to power or drive cables. A signal that leaves the sensor correctly but does not arrive at the input has a wiring fault between them, found by tracing the connection and checking the terminals along it, with the loose or corroded connection being, as everywhere, the most common culprit. For analog signals especially, the routing and shielding of the signal wiring matters, because a signal that reads correctly at the sensor but noisy or wrong at the input may be picking up interference along a poorly routed path. Confirming signal integrity from the device to the input, once the device itself is proven to be sensing, completes the sensor diagnosis by verifying the delivery half of the sensing-versus-delivery split.

A case file: the replacement that did not fix it

A sensor is suspected of being faulty, so it is replaced — and the problem persists, the new sensor apparently as faulty as the old. The temptation is to suspect a second bad sensor, but two identical failures in a row point elsewhere: at the installation, the adjustment, or the wiring rather than the device. Investigating, the technician finds the replacement was installed with the same misalignment as the original had drifted into, or set up with the wrong adjustment, or of a subtly wrong type for the target — so that the new sensor, though functional, could no more sense correctly in that installation than the old one could. The fault was never in the sensor itself but in how it was mounted and adjusted, which is why replacing the device changed nothing. The case teaches a valuable caution: when a replacement does not fix a problem, the fault is probably not the component but its installation, adjustment, or wiring, and continuing to swap components will not help. A replacement that fails just like the original is a strong signal to stop replacing and start examining the installation — the alignment, the gap, the adjustment, the type suitability, the wiring — because the repeated failure of a new component points at the context it was installed into rather than at a run of bad components.

Matching the device to the job

Many sensor problems trace not to a failed device but to a device poorly matched to its job, and understanding the common sensing types helps both in troubleshooting and in selecting a correct replacement. Inductive proximity sensors detect metal at short range and are robust in dirty environments but respond only to metal targets. Capacitive proximity sensors detect a wider range of materials but can be affected by their surroundings. Photoelectric sensors detect targets by a light beam over longer ranges but are vulnerable to lens fouling and misalignment. Mechanical limit switches detect physical position reliably but wear with actuation and can be damaged by mechanical abuse. Each type has a sensing range, an environment it suits, and targets it responds to, and a device applied outside its suitable range or environment, or asked to detect a target it responds to poorly, will give trouble regardless of its condition. This matters acutely when replacing a sensor: a replacement must match not only the electrical interface but the sensing type, range, and characteristics of the original, because a subtly wrong device — a different range, a different output type, a type suited to different targets — will fail to work even though it is new and functional. When troubleshooting a troublesome sensor, considering whether the device is well matched to its job — the right type for the target, within its range, suited to its environment — sometimes reveals that the real fix is a better-matched device rather than another replacement of the same unsuitable type, and it always guides the selection of a replacement that will actually work in the application.

A structured summary of field-device faults

Field devices resolve into a structured diagnostic organized around one central split: is the device sensing, and is its signal arriving? The device’s own indicator, or a meter at its output, answers the first question by showing whether the device responds to its target independent of the wiring. If the device is sensing correctly but the control system does not see the change, the fault is in signal delivery — the wiring, the terminations, a convention mismatch between sourcing and sinking outputs, or noise on an analog signal — found by tracing the connection and checking its integrity. If the device is not sensing, the fault is in the device or its physical situation — no power, a failed device, or, very commonly, a physical problem such as misalignment, an excessive gap, a fouled lens, a worn or misadjusted actuator, or a device poorly matched to its target. This central split — sensing versus delivery — organizes the whole approach, and layered onto it are the recurring lessons that many field-device faults are physical rather than electrical, that a replacement failing like the original points at the installation rather than the device, and that a device must be well matched to its job in type, range, and environment. Together these turn field-device troubleshooting into a systematic process: check whether the device senses, thereby splitting the fault into sensing or delivery, then within the indicated half apply the specific checks — physical situation and device health on the sensing side, wiring integrity and convention matching on the delivery side. This structure resolves the large share of apparent control faults that are really field-device faults, and it directs attention to the physical and installation causes that field devices, living in the harshest part of the plant, so often suffer.