Sensors

Sensors are where the control system meets the process, and they live
in the harshest part of the environment — bathed in coolant, buffeted by
vibration, and baked by heat. A large fraction of ‘PLC faults’ are
really sensor faults, so knowing the common types and their failure
modes pays off constantly.

Proximity and photoelectric sensors

Inductive proximity sensors detect metal at short range; capacitive
types detect a wider range of materials; photoelectric sensors use a
light beam broken or reflected by the target. All three usually present
a simple sinking or sourcing digital output, so they troubleshoot like a
digital input plus a sensing check. Confirm the sensor is powered, watch
its own indicator LED (most have one that lights on detection), verify
the target is within the sensing range and alignment, and check for a
fouled lens or a build-up of debris on the face. A photoelectric sensor
with a dirty lens or a drifting reflector is a classic intermittent
fault.

Encoders

Encoders report position or speed as pulse trains. A failing encoder
shows up as lost counts, erratic position, or a drive fault. Check the
mechanical coupling first — a slipping or cracked coupling mimics an
electrical fault perfectly — then the supply voltage, the signal wiring
and shield, and whether the channels are producing clean pulses. Because
encoders feed high-speed inputs or drive feedback, wiring integrity and
shielding matter more than with a slow digital input.

Temperature: RTDs and thermocouples

RTDs (such as Pt100) change resistance with temperature;
thermocouples generate a tiny voltage from a junction of two dissimilar
metals. Both are sensitive to wiring. An RTD reading that drifts high
often points to added resistance in the leads — a corroded or loose
terminal — because the card interprets extra resistance as higher
temperature. A thermocouple needs the correct extension wire and
polarity; reversed polarity or the wrong wire type produces readings
that are wrong in a characteristic, direction-dependent way. For both,
an open sensor typically drives the reading to an extreme that the card
flags as a fault.

The sensor’s own indicator is your friend

Most industrial sensors have an output indicator LED, and it is one
of the fastest diagnostics available because it tells you what the
sensor thinks it is doing, independent of any wiring back to the PLC.
Present a target to a proximity sensor and its LED should light; break a
photoelectric beam and its output LED should change. This splits the
sensing question cleanly from the signal-delivery question. If the
sensor’s LED responds correctly to the target but the PLC input does not
follow, the sensor is sensing and the fault lies in the wiring, the
input, or a sink/source mismatch. If the sensor’s LED does not respond
to the target, the sensor itself, its power, its alignment, or its range
is the problem. Two glances at an LED have divided the fault space in
half.

Alignment, range, and the physical world

Sensors fail as often from the physical world as from electrical
faults, and these causes are easy to miss if you only think
electrically. A proximity sensor mounted a hair too far from its target
works when the machine is cold and fails when thermal expansion nudges
the gap wider. A photoelectric sensor drifts out of alignment as
vibration works its mounting loose, or its lens fouls with dust, mist,
or coolant until the beam can no longer make it through. A
retroreflective sensor’s reflector gets splashed or knocked askew. None
of these show as an electrical fault — the sensor is powered and healthy
— yet the machine sees a missing signal. When a sensor’s LED is not
responding to a target you believe is present, check the physical
relationship: gap, alignment, cleanliness, and whether the target itself
is actually where and what the sensor expects.

Encoder faults in depth

Encoders reward a methodical approach because their faults can look
like drive faults, position errors, or erratic motion. Begin
mechanically: a coupling that slips, cracks, or loosens makes the
encoder report motion different from the actual shaft, and this mimics
an electrical fault perfectly while being purely mechanical. With the
mechanical coupling confirmed solid, check the supply voltage at the
encoder — a long cable can drop it below spec under load. Then consider
the signal wiring and shield, since encoders often run fast pulse trains
that are sensitive to noise and to marginal connections. A single
intermittent channel connection produces lost counts that accumulate
into position drift. Where the controller offers it, comparing commanded
position against fed-back position over a move reveals whether counts
are being lost and roughly where.

Temperature sensing pitfalls

Temperature measurement has failure modes peculiar to how RTDs and
thermocouples work, and knowing them prevents misdiagnosis. An RTD
infers temperature from resistance, so anything that adds resistance in
the sensing path — a corroded terminal, a loose connection, an
undersized or damaged lead — reads as higher temperature than reality.
This is why RTD readings that creep upward over months often trace to a
degrading connection rather than a real process change; three-wire and
four-wire RTD wiring exists specifically to compensate for lead
resistance, and wiring one incorrectly reintroduces the error.
Thermocouples generate a tiny voltage and are unforgiving about wire
type and polarity: the extension wire must match the thermocouple type,
and reversing polarity produces readings that move the wrong way or sit
at odd values. For both sensor types, an open circuit typically drives
the reading to a scale extreme that the card flags — a useful behavior,
since a temperature pinned at maximum or minimum often means a broken
sensor or wire rather than a genuine extreme.

A case file: the sensor blinded by its neighbor

Two photoelectric sensors mounted near each other on a machine
produce occasional false readings that defy explanation — each sensor
works perfectly when tested alone, yet in operation one intermittently
sees a target that is not there. The cause is optical crosstalk: under
certain conditions the light from one sensor reaches the receiver of the
other, which reads the stray light as its own beam. This is invisible to
electrical testing because nothing electrical is wrong; both sensors are
healthy. The fix is physical — repositioning, shielding, or using
sensors with different modulation so each ignores the other’s light. The
lesson generalizes: when sensors that test fine individually misbehave
together, suspect interaction between them — optical crosstalk for
photoelectric types, mutual influence for closely spaced inductive
sensors — rather than a fault in either one.

Matching the sensor to the job

Some recurring sensor faults are really application mismatches that
no amount of troubleshooting the individual failure will cure. A sensor
at the very edge of its sensing range works when everything is perfect
and fails with the slightest drift in target position, temperature, or
mounting. A sensor type poorly suited to its target — trying to detect a
dark or shiny or transparent object with a sensor that struggles with
such surfaces — produces marginal, unreliable detection that looks like
intermittent failure. A sensor mounted where it is regularly fouled by
the process fails repeatedly no matter how often it is cleaned. When a
particular sensor position fails again and again, step back from the
individual failure and ask whether the sensor is right for the job and
mounted in the right place. The durable fix is often a better-suited
sensor or a better location, not another replacement of the same part in
the same spot.

Preventive attention to sensors

Because sensors live in the harshest part of the machine and cause a
large share of faults, they reward preventive attention in a way that
pays back in reduced breakdowns. Sensors accumulate the conditions that
will eventually make them fail — a lens slowly fouling, a mounting
gradually loosening under vibration, a gap widening with thermal
cycling, corrosion creeping into a connection — and these developing
conditions are visible before they cause a failure if anyone looks. A
sensor whose lens is half-fouled still works but is on its way to
failing; a mounting that has begun to loosen still holds but will
eventually drift out of position. Attention to these developing
conditions during routine maintenance — cleaning fouled sensors,
checking mountings, inspecting connections for early corrosion — heads
off failures before they stop the machine. The same sensor failures that
appear as sudden breakdowns are often the end of a slow, visible
decline, and catching them in decline rather than at failure is a large
part of moving from reactive to preventive maintenance on the parts of
the machine that fail most.

A case file: the encoder coupling

A servo axis begins to show position errors and occasional faults,
and because the symptom is position-related the natural suspicion falls
on the encoder or the drive electronics. But following the discipline of
checking the mechanical path first, the technician examines the coupling
between the motor shaft and the encoder before condemning any
electronics. The coupling has developed a crack and is slipping slightly
under acceleration, so the encoder no longer faithfully reports the
shaft’s true position — a purely mechanical fault producing a symptom
that looks entirely electrical. Replacing the coupling resolves the
position errors completely, with no electronic part involved. The case
underscores a principle specific to encoders and any feedback device
tied to a shaft: because they report on mechanical motion, a mechanical
fault in how they are coupled to that motion mimics an electrical fault
perfectly, and checking the mechanical coupling first — before the
encoder, its wiring, or the drive — saves the wasted effort and expense
of replacing healthy electronics to chase a cracked coupling.

A structured summary of sensor faults

Sensor faults, spanning the many sensor types, still organize around
a common structure anchored by the sensor’s own indicator: is the sensor
sensing, and is its signal arriving at the controller? Sensing-side
faults, where the sensor does not correctly detect its target, include
power problems (the sensor lacks its supply), physical problems
(misalignment, excessive gap, a fouled lens, a target out of range or of
a type the sensor handles poorly), and sensor failures (the sensing
element itself degraded or dead). For the feedback and measurement
sensors, this side also includes mechanical faults for encoders (a
slipping or cracked coupling reporting false motion) and sensing-path
faults for temperature sensors (added lead resistance reading high on an
RTD, wrong wire or polarity on a thermocouple). Signal-side faults,
where the sensor senses correctly but the signal does not reach the
controller, are the wiring, convention, and input faults already
familiar from the input chapters. The sensor’s own indicator LED is the
pivot that separates the two sides: responding correctly to the target
means the sensor senses and the fault is in signal delivery, while
failing to respond means the fault is on the sensing side. And
underlying many recurring sensor faults is application mismatch — a
sensor poorly suited or poorly located for its job — which no individual
repair cures and which is addressed only by choosing a better-suited
sensor or a better position. This structure turns the diverse world of
sensor faults into an organized approach anchored by two glances at an
LED.

A case file: the thermocouple reading backwards

A newly installed thermocouple reads in a way that makes no sense —
the temperature appears to move in the wrong direction, dropping when
the process heats and rising when it cools. A reading that moves the
wrong way is a distinctive signature pointing at reversed polarity,
because a thermocouple generates a small polarity-dependent voltage and
reversing its connections inverts the sense of the reading. The
thermocouple, or its extension wire, has been connected with reversed
polarity, so the card interprets the signal backwards. Thermocouples are
unforgiving about this because they depend on the correct polarity and
the correct matching extension wire type; either a polarity reversal or
the wrong extension wire produces characteristic wrong readings.
Correcting the polarity — ensuring the correct conductors land on the
correct terminals, with matching thermocouple-type extension wire
throughout — restores a sensible reading. The lesson is specific to
thermocouples and worth holding: a thermocouple reading that moves in
the wrong direction or sits at an implausible value, especially just
after installation or rewiring, points at reversed polarity or wrong
extension wire rather than a failed sensor or process problem, and the
fix lies in getting the polarity and wire type right, a wiring detail
that thermocouples demand precisely because their tiny
polarity-dependent signal encodes temperature in a way that mis-wiring
corrupts in these telltale ways.

Part IV — Outputs