Why Stable PLC Logic Still Fails When Process Data Is Wrong

A PLC does not directly understand pressure, temperature, flow or tank level.

It understands numbers.

A pressure transmitter converts a physical condition into an electrical signal. An analog input module converts that signal into a digital value. The PLC program scales the value into engineering units and then uses it for alarms, interlocks and control decisions.

Every stage must be correct.

When the measurement is noisy, delayed, drifting or scaled incorrectly, perfectly valid PLC logic can produce the wrong machine response.

The controller is not necessarily malfunctioning. It may be acting correctly on bad information.

The Complete Analog Signal Chain

A typical measurement passes through several stages:

1. The physical process changes.
2. The sensor detects the condition.
3. The transmitter generates a 4–20 mA or voltage signal.
4. Field wiring carries the signal to the control panel.
5. The analog module converts it into a raw digital value.
6. The PLC scales it into engineering units.
7. Filtering or averaging modifies the value.
8. Logic compares it with limits and setpoints.
9. The HMI displays the result.
10. The controller adjusts an output.

A problem at any stage can make the final value inaccurate.

For example, a pressure transmitter may be working correctly while the PLC uses the wrong raw range. Alternatively, the scaling may be correct while electrical interference causes the input current to fluctuate.

Looking only at the final HMI value does not reveal where the problem originates.

Understanding 4–20 mA Signals

The 4–20 mA current loop is widely used because it is relatively resistant to electrical noise and can detect certain wiring failures.

A typical transmitter might be configured as:

• 4 mA = 0 bar
• 20 mA = 10 bar

The span is 16 mA.

At 12 mA, the measurement is halfway through the configured range, so the expected pressure is approximately 5 bar.

The 4 mA starting point is called a live zero. It allows the system to distinguish a genuine zero measurement from certain faults.

For example:

• Approximately 4 mA may represent a valid zero process value.
• Approximately 0 mA may indicate a broken wire, missing supply or disconnected transmitter.

Many industrial transmitters also use currents below or above the normal measuring range to indicate faults or overrange conditions. The exact values depend on the device configuration and manufacturer.

Incorrect Scaling

One of the most common analog problems is incorrect conversion from the raw PLC value into engineering units.

A generic scaling relationship is:

Engineering Value =
(Raw Value − Raw Minimum)
×
(Engineering Maximum − Engineering Minimum)
÷
(Raw Maximum − Raw Minimum)
+
Engineering Minimum

Suppose the analog module represents 4–20 mA as raw values from 0 to 27,648, while the transmitter is configured for 0–10 bar.

The programmer must use the module’s actual raw range.

If the code incorrectly assumes that 4–20 mA corresponds to 0–32,000, every calculated pressure will be wrong even though the ladder logic and comparisons compile normally.

Common scaling errors include:

• Using the wrong raw minimum
• Using the wrong raw maximum
• Confusing 0–20 mA with 4–20 mA
• Entering the wrong transmitter range
• Reversing minimum and maximum values
• Scaling the same signal twice
• Using integer arithmetic where decimal precision is required
• Writing engineering values into a raw-data tag

The Transmitter and PLC Must Use the Same Range

A transmitter may be configured locally, through HART or through asset-management software.

The PLC scaling must match that field configuration.

For example:

Transmitter:
4 mA = −20°C
20 mA = 80°C

If the PLC assumes:

4 mA = 0°C
20 mA = 100°C

the HMI may still display smooth and believable temperatures. They are simply incorrect.

This is particularly dangerous because the value may appear reasonable enough to avoid immediate suspicion.

Always verify:

• Lower range value
• Upper range value
• Signal type
• Measurement units
• Sensor configuration
• Analog module configuration
• PLC scaling parameters

Do not rely only on labels from an old electrical drawing. Field transmitters are often reconfigured during maintenance without corresponding PLC documentation updates.

A Stable Value Can Still Be Wrong

A smooth trend does not prove that the measurement is accurate.

The value may be stable because:

• The transmitter is incorrectly calibrated.
• The impulse line is blocked.
• The level sensor is installed in the wrong position.
• The temperature probe is not contacting the process properly.
• The signal has reached its upper or lower limit.
• Filtering is hiding real movement.
• Communication has frozen the last valid value.
• The scaling range is wrong.

A frozen pressure value of 5.2 bar can look completely normal.

The PLC only sees that the number remains at 5.2. It does not automatically know that the transmitter stopped updating.

Important measurements should therefore include quality and plausibility checks.

Signal Saturation

An analog value saturates when it reaches the end of its measurable range.

Suppose a transmitter is configured for 0–10 bar. If the real process pressure reaches 13 bar, the signal may remain near its maximum output.

The PLC continues displaying approximately 10 bar even though the real pressure is higher.

The trend looks flat and stable—not because the process is stable, but because the instrument cannot report anything beyond its configured range.

Warning signs include:

• Value remains exactly at the upper limit
• Value remains exactly at the lower limit
• Process behaviour suggests further movement
• Related measurements continue changing
• Controller output rises without a corresponding process response

Alarm logic should distinguish between:

• Normal high process value
• Instrument overrange
• Invalid signal
• Module overflow

Electrical Noise

Analog signals can be disturbed by:

• VFD motor cables
• Contactors and relay coils
• Poor shielding
• Incorrect grounding
• Shared cable trays
• Loose terminals
• Unstable 24 V supplies
• Ground loops
• Long field wiring
• Damaged sensor cables

Noise may appear as:

• Rapid value fluctuation
• Periodic spikes
• Sudden jumps
• A repeating 50 or 60 Hz pattern
• Changes when a motor starts
• Different readings when cabinet doors are opened or closed

Before adding heavy software filtering, inspect the installation.

Check:

• Cable type
• Shield termination
• Signal grounding
• Separation from power cables
• Terminal tightness
• Power-supply stability
• Transmitter wiring
• Analog module configuration

Filtering a wiring fault may make the trend look better while leaving the underlying problem unresolved.

Ground Loops

A ground loop occurs when two parts of an analog circuit are referenced to ground through different paths.

Small voltage differences can create unwanted current and measurement error.

Symptoms may include:

• Offset that changes when other equipment starts
• Different readings between a handheld meter and the PLC
• Increased noise after connecting a second device
• Fluctuation associated with VFD operation
• Stable but consistently incorrect values

The solution depends on the circuit design and may involve:

• Correct shield termination
• Signal isolation
• Isolated analog modules
• Galvanic isolators
• Improved equipotential bonding
• Removing unintended grounding paths

Grounding changes should be made carefully. Randomly disconnecting shields can reduce one symptom while creating another EMC problem.

Excessive Filtering

Filtering is useful when a signal contains small, rapid fluctuations that are irrelevant to the process.

However, a strong filter delays the measurement.

Consider a low-pressure interlock:

1. Actual pressure falls quickly.
2. The transmitter responds.
3. The PLC filter changes slowly.
4. The displayed pressure remains above the trip threshold.
5. The pump continues running.
6. The filtered value eventually reaches the shutdown limit.

The comparison instruction is correct. The shutdown is late because the value being compared is delayed.

Possible filtering stages include:

• Sensor damping
• Transmitter damping
• Analog-module integration
• PLC software averaging
• PID input filtering
• HMI smoothing

These delays accumulate.

Filtering should be selected based on process speed. A level measurement in a large storage tank can usually tolerate more filtering than a fast lubrication-pressure interlock.

Raw, Control and Display Values

One filtered value should not necessarily be used for every purpose.

A better design may maintain several versions:

Pressure_Raw
Pressure_Control
Pressure_Display

Possible uses include:

• Pressure_Raw for diagnostics and fast fault detection
• Pressure_Control for stable PID regulation
• Pressure_Display for a smooth HMI trend

This allows the control system to reduce nuisance fluctuations without hiding fast process changes from protection logic.

The raw signal must still be validated. A noisy, unfiltered input should not be used blindly for critical shutdowns.

Analog Drift

A drifting transmitter gradually reports a changing value even when the process remains stable.

Possible causes include:

• Sensor ageing
• Temperature effects
• Moisture
• Mechanical stress
• Contamination
• Calibration loss
• Damaged sensing elements
• Unstable power supply

Drift can be difficult to detect because it happens slowly.

The PLC may compensate automatically through its control loop. For example, a valve output may gradually increase to maintain a pressure that the faulty transmitter reports incorrectly.

The process appears controlled until another instrument or manual measurement reveals the discrepancy.

Useful drift detection methods include:

• Comparing redundant transmitters
• Comparing related process measurements
• Monitoring long-term calibration offset
• Checking controller output against normal operating patterns
• Performing scheduled calibration
• Using transmitter diagnostic data

Sensor Placement Problems

The analog signal can be electrically perfect and still fail to represent the process correctly.

Examples include:

• Temperature sensor installed too close to a heater
• Flowmeter installed without sufficient straight pipe
• Pressure impulse line blocked or leaking
• Level sensor affected by foam or turbulence
• Sensor mounted in a stagnant section of pipe
• Temperature probe not inserted deeply enough
• Differential-pressure lines connected incorrectly

The PLC cannot correct a measurement that does not physically represent the required process condition.

Troubleshooting should include field inspection—not only online program monitoring.

Overflow and Data-Type Errors

Analog calculations can fail because of variable limits.

Possible issues include:

• Integer multiplication exceeding the permitted range
• Signed values interpreted as unsigned
• Negative values stored in non-negative variables
• Real values converted to integers too early
• Loss of decimal precision
• Raw diagnostic codes treated as measurements
• Division by an incorrect span
• Scaling blocks receiving out-of-range inputs

For example, performing multiplication before division in integer arithmetic can create overflow:

Scaled := Raw × Engineering_Span ÷ Raw_Span;

A large intermediate result may exceed the integer range even when the final result would have been valid.

Using appropriate data types and checking calculation status can prevent these faults.

Frozen Communication Values

Analog values received from remote PLCs, drives or network devices may remain at their last valid state after communication fails.

The logic may continue operating from stale data unless validity is checked.

A robust measurement structure can contain:

Process_Value
Data_Valid
Communication_Healthy
Last_Update_Time
Signal_Quality

The measurement is accepted only when all required quality conditions are healthy.

A watchdog or update counter can confirm that new data continues to arrive.

For example:

Remote_Value_Valid :=
    Communication_Healthy
    AND Data_Valid
    AND Signal_Age < Maximum_Age;

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

Plausibility Checking

A value can remain within its configured range and still be physically impossible.

Plausibility checks compare the measurement with expected process behaviour.

Examples include:

• Flow cannot exist when the pump is stopped.
• Pressure should increase after a pump starts.
• Tank level cannot change by 30% in one second.
• Temperature should not jump by 50°C between scans.
• Valve position should correspond with command and feedback.
• Two redundant sensors should remain within an acceptable difference.

These checks can detect:

• Frozen values
• Incorrect scaling
• Reversed wiring
• Sensor drift
• Communication mapping errors
• Implausible process changes

Plausibility logic should generate a diagnostic condition rather than silently replacing the measurement without indication.

HMI Trends Can Hide Problems

HMI trends may apply:

• Slow sampling
• Deadbands
• Averaging
• Compression
• Rounding
• Delayed communication

A process value can oscillate rapidly inside the PLC while the HMI displays a smooth line.

For serious analog troubleshooting, use a PLC trace or high-speed data recorder.

Capture:

• Raw input
• Scaled input
• Filtered value
• Alarm state
• Interlock state
• Controller output
• Related process measurements

Recording these values on the same timeline shows whether the problem begins in the signal, calculation, filter or controller response.

A Practical Troubleshooting Workflow

When analog-based logic behaves incorrectly, avoid rewriting the control code immediately.

1. Verify the physical process

Use an independent calibrated instrument where possible.

Confirm that the actual pressure, temperature, flow or level matches the expected condition.

2. Measure the electrical signal

For a 4–20 mA loop, measure the current safely using appropriate procedures and equipment.

Determine whether the field signal matches the process value.

3. Check the transmitter

Verify:

• Sensor type
• Configured range
• Units
• Damping
• Fault-current behaviour
• Calibration
• Diagnostic status

4. Inspect field wiring

Check polarity, shield termination, grounding, cable routing, power supply and terminal condition.

5. Examine the raw PLC value

Monitor the unscaled value directly from the analog module.

Compare it with the expected raw count for the measured current.

6. Verify scaling

Confirm raw minimum, raw maximum, engineering minimum, engineering maximum and data type.

7. Compare raw and filtered values

Measure how much delay the filter introduces during a realistic process change.

8. Check signal quality

Review wire-break, overflow, underflow and module diagnostic states.

9. Verify communication age

For networked values, confirm that the data continues updating.

10. Review control response

Only after validating the measurement should the PID, interlock or sequence logic be modified.

Useful Analog Diagnostics

A well-designed PLC program can provide:

• Raw input count
• Scaled engineering value
• Filtered value
• Signal quality
• Wire-break status
• Overrange status
• Underrange status
• Communication validity
• Last update time
• Rate of change
• High and low alarms
• Sensor disagreement
• Calibration or maintenance warning

Displaying only the final engineering value removes valuable troubleshooting information.

Final Thoughts

Stable PLC logic can still produce incorrect control because the controller responds to data—not directly to the physical process.

A smooth HMI trend may come from excessive filtering. A stable value may be frozen. A believable number may be scaled incorrectly. A transmitter may report a clean signal from the wrong physical location.

Before changing the program, verify the complete measurement chain:

• Physical process
• Sensor
• Transmitter
• Wiring
• Analog module
• Raw value
• Scaling
• Filtering
• Data validity
• HMI display

The PLC can only make decisions from the information it receives.

When that information is wrong, even perfect logic produces the wrong result.