Installing a 4-20 mA loop break and fault detection system in a factory involves more than setting a low-current alarm. The fault may lie in a transmitter, 24 V supply, cable, terminal, isolator, PLC analog input, or display logic. This guide takes a Thai brownfield plant through point selection, NE43 interpretation, procurement, FAT, SAT, and maintenance handover. Its aim is to make a detected fault actionable without treating every unusual current as a broken wire.
Define the action before selecting a detector
A 4-20 mA signal uses 4 mA as its live zero and 20 mA as its full-scale value. A near-zero current can therefore be distinguished from a valid zero measurement. Yet near-zero current alone cannot prove that a cable is broken: a failed supply, fuse, terminal, transmitter, or input circuit may produce a similar symptom. A transmitter’s self-diagnostic fault may instead drive a configured low or high alarm current. Fluke’s current-loop guide describes the supply, transmitter, receiver, and wiring as one series circuit.
For each signal, decide what operators should see and what the process should do. Should the controller invalidate the measurement, stop a pump, switch to a redundant sensor, or merely notify maintenance? A tank level used in a protection function calls for a different response from a temperature point used only for a report. A retained last value must be identified as retained, never presented as a live reading. The implementation examples here concern general monitoring and control; changes to an instrumented safety function require the plant’s own hazard assessment and approval.
Write the required outcomes in four columns: operator notification, invalid-value handling, process action, and diagnostic record. Not every point needs an automatic stop. A nuisance trip also has a cost, so define the alarm delay and recovery behavior before buying hardware. The point list is the foundation of the specification.
Map the whole 4-20 mA loop
A typical two-wire loop runs from a power source through the transmitter and back through an analog input. An existing plant may also include a signal isolator, indicator, surge protector, and terminal blocks. Record which device supplies loop power, whether the input is active or passive, and how the return conductor is wired. A mismatch between the drawing and the channel configuration can make any software diagnosis meaningless.
For every candidate point, record process variable and range, transmitter model, two-wire or four-wire connection, supply voltage, input-card model, terminal numbers, cable route, shield and grounding, and any isolator. Capture raw current, scaled engineering value, quality bit, diagnostic buffer, and HMI value at the same time. “The screen reads zero” could mean actual 0 mA, a rounded scaled value, a substituted fault value, or a communication failure. Trace where each substitution occurs in PLC logic and visualization.
Check the voltage budget at maximum loop current. Wiring, input resistance, isolators, and protective devices create voltage drops. Too little voltage at the transmitter can cause the output to plateau below 20 mA at high demand without producing a permanent open-circuit fault. Compare supply voltage, required transmitter operating voltage, and total load with the actual device data sheets, then verify at the panel and field terminals. Fluke explains this resistance and voltage relationship.
Separate an open circuit from overrange and device failure
A low signal may represent a slightly underranged process, a transmitter fault, an open cable, or a lost supply. These conditions have different first responses. NE43 concerns standardized analog signal levels for failure information from digital transmitters; it does not make every vendor’s diagnostic thresholds identical. NAMUR identifies the recommendation and its revision. Read the actual transmitter and receiver configuration before fixing boundaries.
For example, an Endress+Hauser input manual describes one device’s enabled NE43 monitoring: low underrange, sensor error, and cable-open conditions have distinct boundaries. An Emerson Rosemount 4051S manual distinguishes saturation levels from high and low alarm levels. These are device examples, not universal PLC constants. A purchasing phrase such as “NE43 compatible” does not establish the configured failure direction or the PLC’s interpretation.
| Observed condition | What to check | First action |
|---|---|---|
| Almost no loop current | Supply, fuse, terminals, open conductor, input wiring | Measure panel and field voltage |
| Low alarm region | Transmitter failure direction and diagnostics | Compare with its manual and settings |
| High alarm region | Self-diagnostics versus process overrange | Compare measured current and HART status |
| Implausible value within 4-20 mA | Sensor, process, scaling, noise, load | Compare field reading with raw PLC value |
| Frozen value | HMI hold, PLC substitution, communication quality | Record current, quality, and display separately |
The analog value cannot by itself establish the root cause. HART can provide additional information, but a healthy-looking current does not guarantee that all sensor functions are healthy. An Emerson 3144S example can maintain the analog output after one sensor in a redundant configuration fails while communicating the status digitally. Define the faults that need to be detectable and test that each actually reaches the operator.
Validate PLC analog-input diagnostics by model
Many PLC analog modules have configurable wire-break diagnostics. Determine whether the function is available for the selected 4-20 mA range, whether it can be enabled by channel, what event appears in the CPU, and what value or quality indicator the application receives. A Siemens analog-processing manual explains live-zero wire-break detection and configurable limits for certain modules. A specific Siemens module manual shows that diagnostic behavior and resulting codes depend on range and configuration. Neither should be copied blindly to another card.
Represent raw input, channel diagnostics, quality, and communication status as separate variables. A threshold on raw current alone may miss a card fault. Assuming that the quality bit always exists may fail when an old PLC or gateway discards it. Draw a truth table for the information the installed hardware can actually provide. Specify display text, event code, alarm priority, and the state of downstream calculations for each row.
Debounce matters. A single brief crossing may be electrical noise or a power transfer; immediately tripping a line may be too sensitive. A long delay, however, leaves a bad measurement active. Set entry delay, recovery delay, and hysteresis from the measurement update rate, process dynamics, and allowable response time for each point. Record the reason. Obtain operations, maintenance, and controls approval for consequential interlocks.
Choose the first points by consequence and evidence
Prioritize signals with high shutdown impact, repeated unexplained faults, long troubleshooting time, or difficult field access. Important level, flow, pressure, cooling-water, and utility-temperature loops are common candidates. Distinguish monitoring improvements from changes to existing safety functions, which require separate engineering and approval.
Score each loop using downtime, quality loss, equipment damage, environmental or safety consequence, and time to locate a fault. Start with maintenance logs, operator handovers, and spare-part lead times. In a plant with outdated terminal drawings, identifying and updating the wiring may have more immediate value than installing a new monitor. A small proof of concept should include a stable loop, an intermittently troublesome loop, and a critical loop. Diversity of failure patterns matters more than a large channel count.
Plan tests that can be done without opening a live production loop separately from tests requiring a planned shutdown. Obtain work permits, isolate where necessary, notify operators, and define a restoration check. Never cause an unplanned stop simply to demonstrate an alarm.
Compare architectures on what they can prove
The first option is to use the existing PLC card’s diagnostic function. It usually needs less new hardware and can feed existing HMI and historian screens. Check channel features, spare program capacity, and online-change restrictions. If the card supplies only a raw value, software can classify some conditions, but it cannot infer every internal device failure or supply problem from current alone.
The second option is a loop monitor or isolator. It may add isolation, signal splitting, relay fault output, or HART pass-through. Each added device also affects the power and load budget. Some products distinguish open and short circuits; others only relay an abnormal current. Specify detection conditions, output and power-loss states, isolation, HART handling, and reset behavior. Our related flow-meter monitoring guide explains how signal choice connects to acceptance testing.
The third option is independent monitoring that leaves PLC logic intact. It can reduce changes to the control system, but needs a clearly assigned alarm owner, timestamp synchronization, data retention, and an explanation for discrepancies with the PLC display. Whichever architecture is chosen, monitor the monitor itself: a dead gateway must not appear as a healthy loop.
| Architecture | Suitable when | Common omission |
|---|---|---|
| Existing PLC diagnosis | Card exposes channel fault and quality | Enable conditions and HMI path |
| PLC threshold logic | Hardware changes are difficult | Limits of current-only classification |
| External monitor or isolator | Isolation or relay output is needed | Voltage budget, HART, power loss |
| Independent historian | Control changes are constrained | Clock, communication loss, ownership |
Write an RFP that enables a real comparison
Attach the loop list and site conditions. For each point include measurement range, normal operating values, response to fault, input module, power supply, transmitter, cable and terminals, and HART requirements. Specify panel space, environment, spare channels, spare-parts strategy, and survey items that remain unconfirmed.
Request an explicit matrix for open cable, short, low alarm, high alarm, out-of-range, and power loss. Ask what appears on PLC/HMI, what is logged, when the state resets, and what is still unknowable. Require thresholds to be checked against the actual transmitter configuration. A catalog checkbox labeled “fault detection” is not an acceptance criterion.
Compare hardware, panel work, wiring, PLC/HMI changes, production downtime, FAT/SAT, training, spares, and support separately. Require loop drawings, I/O list, configuration backups, software change history, tests, and operator and maintainer instructions as deliverables. A low device price may be outweighed by a long night shutdown or poor diagnostic support. Obtain site-specific prices instead of borrowing generic figures from an article.
FAT: inject known currents and faults
In factory acceptance testing, connect a calibrator to the real or representative input. Check 4, 8, 12, 16, and 20 mA against raw PLC values, engineering values, HMI, and history. Then inject the transmitter’s configured low/high fault currents, simulate an open connection, and test power-loss and card-diagnostic events where possible. A current source cannot reproduce every internal transmitter fault, so specify separate procedures for device self-diagnostics. Record applied value, expected result, actual result, verdict, and tester. Fluke’s loop-calibration guide uses the same stepped-current approach.
Verify display semantics. Does the screen blank the number, show a clearly marked retained value, or combine a warning icon with the last reading? Test acknowledgement, silence, recovery, and repeated alarm behavior separately. Confirm that the event includes the channel identifier for a technician working at night. Also test downstream trends, reports, MES data, and control calculations. A high fault current must not become a false high-pressure production record. Preserve raw current and fault status separately where the system permits.
SAT: test the installed field path
Site acceptance testing adds real cables, terminals, grounding, protection devices, isolators, PLC input, HMI, and notification path. Compare a field current measurement with the raw PLC reading. Measure voltage margin at the cabinet and field end, especially at high loop current. If the output cannot reach full scale, check loop loading and supply before recalibrating the transmitter. Fluke’s troubleshooting guidance separates supply, wiring, instrument, and input-card checks.
Inject faults only with a permit and operator coordination. Verify control and interlock effects before opening terminals. Identify the tested channel, then restore the wiring, confirm terminal tightness, clear alarms, and check the live measurement. If a production loop cannot be opened, distinguish the coverage of a simulated input test from the untested field cable. Passing a cabinet simulation does not prove that an intermittent field conductor is sound.
Test who receives the message and who acknowledges it. HMI, andon, and email notifications should not flood the same team with duplicates. Record the initial response and escalation route. A message delivered without a practical response process is not a complete fault-detection system.
Intermittent faults and electrical noise
Faults that appear only during a motor start or humid night require time-aligned evidence. Candidates include loose terminals, condensation, corrosion, cable flexing, electrical interference, and a sagging supply. Correlate event time, equipment state, shared supply, nearby drives, raw minimum/maximum current, module diagnostics, and maintenance activity. Sampling too slowly may miss a brief fault.
Noise control and wire-break diagnosis are distinct jobs. Shielding, grounding, routing, and isolation are physical design questions. Increasing software smoothing may hide the symptom without fixing the circuit. Compare measured current and supply voltage with module diagnostics before extending alarm delay. Outdoor loops also need appropriate signal-line protection and grounding; our control-panel SPD guide addresses protection of 4-20 mA lines.
For the proof of concept, log alarm start and end, raw current, quality, and process state against a common time source. If HART status is available, include it without making the analog control dependent on a new communication channel. A fault that clears before a technician arrives still needs enough history to diagnose.
Turn alarms into maintenance improvements
The value of detection is faster diagnosis and fewer repeated failures, not a higher alarm count. Track time to identify, time to restore, probable cause, repeat rate, and false-alarm rate. Initial cause categories can be simple: open wiring, power, transmitter, input card, and unknown. Do not relabel every unexplained event as a broken wire; a large unknown category is evidence that more diagnostic data or field observations are needed.
For MES or cloud data, carry the number together with unit, tag, timestamp, quality, and fault code. Mixing invalid values into a normal trend can distort analytics. Discarding all abnormal samples can also remove evidence of an emerging fault. Define separate paths or flags for control, visualization, and analysis, including retention periods.
Review the first month of operation. Identify noisy points, duplicate alarms, confusing names, and gaps in response. Log any threshold change and compare before and after. Lowering sensitivity everywhere to reduce alarm volume can hide the very event the project was meant to detect. Return to power, wiring, and device fixes for repeat offenders.
Implementation sequence and responsibilities
Use a sequence of point selection, site survey, concept design, RFP, detailed design, FAT, outage plan, SAT, handover, and initial review. Define exit evidence for each stage. Survey completion means the loop drawings, input configuration, and fault response have been verified. FAT completion means test cases and defect resolutions are approved. SAT completion means the field path and restoration were demonstrated and documented.
Operations owns alarm priority and the process response; maintenance owns fault isolation, inspection, and spares; controls owns PLC logic, input configuration, historian mapping, and change control; IT/OT owns clock synchronization, access, and storage. The supplier provides configuration files and evidence against the specification. A responsibility table prevents the first real alarm from becoming a debate over who should investigate.
Train staff to narrow causes rather than memorize a single cause for each alarm. For a low alarm, compare HMI raw value with the transmitter indication, check supply and terminals, and then inspect device diagnostics. Only authorized personnel should open or simulate an active loop under the site’s procedure. Hand over current drawings and backups with the fault guide.
FAQ: 4-20 mA fault diagnostics and FAT/SAT
What current should trigger 4-20 mA wire-break detection?
There is no universal value to copy. Confirm the transmitter’s configured fault current, the specific PLC card’s diagnostic conditions, the minimum valid process signal, and alarm delays. NE43-compatible equipment may still differ in the boundary between saturation, device fault, and cable-open classification. Record the manual revision so a future replacement triggers review.
Is PLC analog-input fault detection enough?
It can be enough for the specified fault set if the module reports the needed events and they reach the HMI and history. It cannot necessarily identify every transmitter-internal failure from current alone. A device may communicate a diagnostic over HART while its analog output remains within range. Test the exact coverage before declaring it complete.
What should a 4-20 mA FAT/SAT record contain?
Record loop ID, models, supply voltage, applied input, measured current, raw PLC value, displayed value, quality, alarm and recovery times, expected and actual actions, verdict, and tester. FAT focuses on representative hardware and logic; SAT includes installed wiring and the notification route. If calibration is performed, preserve As Found and As Left records separately. Beamex’s example illustrates this practice.
Should frequent alarms be solved by lowering the threshold?
First compare event times with supply, terminals, wiring, noise, transmitter history, PLC range, and communication quality. If current is healthy but the display reports a fault, check substitutions and quality handling. Change a threshold only with an engineering reason, logged configuration, and repeat test.
Should a detected fault automatically stop the line?
That depends on the signal’s purpose. A monitoring-only point may need notification and a maintenance work order. A control or protection point needs an approved failure response based on hazard, nuisance-trip consequences, and redundant information. Changes to automatic stopping must be covered by the plant’s review and acceptance tests.
Design and test details that are easy to miss
Two-wire transmitters share their power and signal conductors; four-wire devices have a separate supply. Test the output behavior after loss of the main device supply according to the actual wiring. The loop drawing should show power direction, terminal numbers, grounding, and meter connection points. Even a replacement of the same model may have a different range, failure direction, or damping setting. Compare its backup, then test normal current, fault current, and recovery after replacement.
Keep raw current, engineering value, fault state, and quality in separate tags. Overwriting a faulted value with zero makes a genuine process zero indistinguishable from a substituted zero. HMI text should complement color. After the current returns to range, the process value may jump relative to a retained last value. Define re-entry into automatic control separately from clearing an alarm.
Design calibrator access and a repeatable test procedure from the beginning. A cabinet current-injection test proves the PLC and HMI path but not the sensor or field cable. State the connection point, bypassed devices, and untested parts in the report. Keep a healthy baseline with loop current, field-end voltage, raw PLC value, HMI value, date, and operating load for comparison months later.
Separate routine checks from detailed investigations. Opening every loop weekly can increase handling risk. Review event counts, supply alarms, and repeat-offender tags first, then measure specific points. If most events labeled “suspected wire break” ultimately come from the power supply, improve the label and fault-isolation guide rather than merely weakening the threshold.
Conclusion
Effective factory 4-20 mA loop fault detection begins with a complete loop map and an agreed response for each signal. Treat NE43 values and PLC diagnostics as device- and configuration-dependent, verify logic with FAT, and prove the installed path with SAT. Then use event records to repair recurring power, wiring, and instrument problems rather than merely adjusting thresholds.
If you are selecting the first 4-20 mA points in a Thai plant, TOMAS TECH can help review loop drawings, input modules, fault history, and the proposed FAT/SAT criteria at the planning stage. Contact us.
