The PLC Does Not See the Process—It Sees Data

A PLC does not know whether a pipe is pressurized, a tank is full or a motor is overheating.

It only receives information representing those conditions:

Electrical currents and voltages
Digital input states
Raw analog counts
Scaled engineering values
Network data
Filtered measurements
Device status words

The controller assumes that this information accurately represents the physical process. When it does not, completely correct PLC logic can make the wrong decision.

A pressure transmitter may fail while the HMI continues displaying a believable value. A noisy signal may oscillate rapidly while heavy filtering makes the trend appear perfectly smooth. Incorrect scaling may produce realistic but inaccurate engineering units.

The PLC cannot recognize every measurement problem automatically. It must be given suitable diagnostics, validity checks and fault-handling logic.

The Complete Analog Signal Path

A typical process measurement follows this path:

The physical process changes.
A sensor detects the condition.
A transmitter converts it into an electrical signal.
Field wiring carries the signal to an analog input module.
The module converts the electrical signal into a raw numerical value.
PLC logic scales the raw number into engineering units.
Filtering may smooth the value.
Control logic uses it for decisions.
The HMI displays the final result.

Every stage introduces a possible failure point.

The value displayed on the HMI is therefore not the physical process itself. It is the final result of several conversions and calculations.

A Bad Transmitter Can Produce Good-Looking Data

Transmitters do not always fail by dropping immediately to zero.

A damaged pressure, temperature or level transmitter may:

Freeze at its last measurement
Develop a fixed offset
Drift gradually
Saturate at the top of its range
React more slowly than normal
Produce intermittent spikes
Output a believable but incorrect value

A pressure transmitter reading 5.1 bar may look completely normal, even when the real pressure is 3 bar.

This is why field verification matters. Compare the PLC value with:

A calibrated pressure gauge
A reference thermometer
A loop calibrator
A handheld process meter
A redundant instrument

Do not assume the PLC value is correct merely because it looks reasonable.

The 4–20 mA Signal

Many industrial transmitters use a 4–20 mA current loop.

For example:

4 mA = 0 bar
20 mA = 10 bar

A current near 12 mA represents approximately half of the configured range, or around 5 bar.

The 4 mA lower limit provides a live zero. This helps distinguish a valid zero process measurement from certain faults such as a disconnected transmitter or loss of loop power.

However, the PLC must be configured for the correct signal type.

Common mistakes include:

Configuring the input for 0–20 mA instead of 4–20 mA
Using the wrong channel mode
Incorrect transmitter polarity
Missing loop power
Excessive loop resistance
Connecting active and passive devices incorrectly
Ignoring underflow or overflow diagnostics

The input may still produce numbers, but they may not represent the real process correctly.

Electrical Noise and Grounding Problems

Analog signals can be affected by nearby electrical equipment.

Common interference sources include:

Variable-frequency drive motor cables
Contactors and relay coils
Poorly suppressed solenoids
High-current conductors
Welding equipment
Improper grounding
Damaged shielding
Unstable 24 V power supplies

Noise may appear as:

Rapid fluctuations
Sudden spikes
Periodic movement
Changes when a motor starts
Different readings depending on machine load
Oscillation at 50 or 60 Hz

Before adding software filtering, inspect the installation.

Check cable routing, shield termination, grounding, terminal tightness and power-supply stability. A wiring or EMC problem should be corrected rather than hidden behind increasingly strong filters.

Incorrect Scaling

The analog module normally provides a raw numerical value rather than engineering units.

The PLC must convert this value into meaningful units such as bar, litres per minute or degrees Celsius.

A typical scaling calculation is:

Engineering value =
(Raw value − Raw minimum)
×
(Engineering maximum − Engineering minimum)
÷
(Raw maximum − Raw minimum)
+
Engineering minimum

If the raw range or transmitter range is entered incorrectly, the displayed value will also be wrong.

Suppose the transmitter is configured for:

4 mA = −20°C
20 mA = 80°C

But the PLC assumes:

4 mA = 0°C
20 mA = 100°C

The displayed temperature may still move smoothly and appear believable. It will simply be inaccurate across the entire range.

Always verify:

Transmitter lower range value
Transmitter upper range value
Measurement units
Analog module raw range
Channel signal type
PLC scaling parameters
Data types used in calculations

Data-Type and Calculation Errors

Analog scaling can also fail because of incorrect PLC data types.

Possible problems include:

Integer overflow
Loss of decimal precision
Signed values interpreted as unsigned
Negative measurements stored in unsigned variables
Real values converted to integers too early
Division performed before multiplication
Raw diagnostic values treated as measurements

For example, a calculation may produce the correct mathematical result but overflow during an intermediate multiplication.

Use appropriate integer or floating-point data types and test the scaling at several points:

Minimum input
25% of range
50% of range
75% of range
Maximum input
Underflow
Overflow

Testing only one point is not enough.

Analog Input Module Problems

The analog input card is another possible failure source.

Potential issues include:

Incorrect channel configuration
Damaged input circuitry
Calibration offset
Common-mode voltage problems
Loose backplane connection
Intermittent module power
Firmware or communication faults
Channel overflow or underflow
Incorrect two-wire or four-wire connection

A failing module may affect one channel or several channels at once.

Compare the measured loop current with the raw PLC value. If the field current changes correctly but the raw input does not, the problem is likely between the input terminals and the PLC program.

Where possible, test the channel with a calibrated signal source.

Overfiltering and Delayed Process Visibility

Filtering reduces noise by slowing the response of the measurement.

This can make trends look clean, but excessive filtering introduces dead time.

Consider a low-pressure shutdown:

Real pressure falls quickly.
The transmitter detects the change.
The filtered PLC value decreases slowly.
The interlock remains healthy.
The equipment continues running.
The filtered value eventually crosses the trip limit.

The comparison instruction may be completely correct. The shutdown occurs late because the value reaching it is delayed.

Filtering may exist in several locations:

Sensor electronics
Transmitter damping
Analog module integration
PLC averaging logic
PID controller input
HMI trend smoothing

These delays accumulate.

During commissioning, compare the raw and filtered values on the same trace. Measure how long the filtered signal takes to follow a realistic process change.

False Stability

A perfectly smooth HMI trend is not always a good sign.

The actual process may be unstable while the displayed value appears calm because of:

Heavy filtering
Slow HMI sampling
Trend deadbands
Rounding
Data compression
Frozen communication values
Signal saturation

For example, a pressure signal may alternate rapidly between 4 and 6 bar. A long moving average may display a steady value near 5 bar.

The operator sees stability. The real process is oscillating.

Use separate signal paths where necessary:

Raw signal for diagnostics
Lightly filtered signal for fast logic
Moderately filtered signal for control
Smoothed signal for operator display

One value does not need to serve every purpose.

Networked Analog Data Can Become Stale

Analog values may also reach the PLC through PROFINET, EtherNet/IP, Modbus TCP or communication with another controller.

If communication fails, the last received value may remain in memory.

The HMI could continue showing:

Pressure = 4.8 bar

even though the remote device has not sent an update for several minutes.

A robust measurement structure should contain more than the process value:

Process_Value
Communication_Healthy
Data_Valid
Signal_Quality
Last_Update_Time
Update_Counter

The value should be accepted only when its quality information is healthy.

For example:

Pressure_Valid :=
    Communication_Healthy
    AND Data_Valid
    AND Signal_Age < Maximum_Age;

A believable number without a recent update should not be trusted.

Sensor Placement Matters

The signal can be electrically perfect and still represent the wrong process condition.

Examples include:

A temperature sensor installed too close to a heater
A pressure impulse line blocked by contamination
A flowmeter installed without sufficient straight pipe
A level sensor disturbed by foam or turbulence
A temperature probe inserted too shallowly
A pressure sensor installed in a stagnant pipe section
Differential-pressure lines connected in reverse

The PLC cannot correct poor instrument placement.

When the value disagrees with real process behaviour, inspect the physical installation—not only the PLC program.

Plausibility Checks

A measurement may remain within its normal range while still being physically impossible.

PLC logic can detect some of these situations through plausibility checks.

Examples include:

Flow reported while the pump is stopped
Pressure failing to rise after a pump starts
Tank level changing faster than physically possible
Temperature jumping by 50°C in one scan
Two redundant sensors disagreeing significantly
Valve position not matching command and feedback
Controller output reaching maximum with no process response

Plausibility checks should normally generate a diagnostic or invalid-quality state.

They should not silently replace the measurement without informing the operator.

Do Not Debug the Controller Before Verifying the Signal

When analog-controlled logic behaves incorrectly, engineers often begin changing:

PID settings
Alarm thresholds
Interlock delays
Sequence logic
Software filters

That may make the symptoms less visible without correcting the real problem.

A better troubleshooting sequence is:

Verify the actual physical process with an independent instrument.
Measure the transmitter output.
Confirm transmitter range and configuration.
Inspect wiring, shielding and grounding.
Monitor the raw analog input value.
Verify PLC scaling.
Compare raw and filtered measurements.
Check module diagnostics.
Confirm communication validity and data age.
Only then review the control logic.

This separates instrumentation problems from programming problems.

Useful Analog Diagnostics

A well-designed PLC project should expose:

Raw analog count
Scaled engineering value
Filtered value
Channel diagnostic status
Underflow and overflow
Wire-break status
Communication quality
Signal age
Rate of change
High and low alarms
Sensor disagreement
Controller output

Displaying only the final engineering value removes valuable troubleshooting information.

Final Thoughts

The PLC does not see pressure, temperature or flow.

It sees a numerical representation created by sensors, transmitters, wiring, analog modules, scaling calculations, filters and communication systems.

Every stage can fail while still producing a believable number.

That is why stable logic can produce incorrect control behaviour. The PLC may be performing exactly as programmed using information that no longer represents the real process.

Before modifying the program, verify the signal from the physical process all the way to the HMI.

The controller can only respond to the data it receives.

When the data lies, the PLC follows it.

#4–20 mA signal #analog filtering #analog input fault #analog signal troubleshooting #Industrial Automation #PLC process data #PLC Scaling #stale PLC data #transmitter problems