High-Traffic Technical Guide · Technical Perspective · Published: October 16, 2026
Table of Contents
Key Takeaways
- A 4-20 mA loop is a series circuit: the same current flows everywhere, so cable resistance can’t corrupt the reading — but it can starve the transmitter of voltage [1][3].
- Budget the voltage first. With a 24 V supply and a transmitter needing about 12 V of compliance, the loop can drive roughly (24 − 12) / 0.020 = 600 Ω of total resistance at 20 mA [3].
- Wire resistance is easy to overestimate: 400 ft of 20 AWG (about 800 ft round trip at ≈10.15 Ω/1000 ft) is ≈8.12 Ω, dropping only ≈0.162 V at 20 mA. Long runs usually break the budget through added devices, not wire alone [6].
- The live zero is a diagnostic tool: 4 mA is a valid 0% reading, 0 mA is a dead loop, and NAMUR NE 43 defines fault bands below 3.6 mA and above 21 mA [4][5].
- Ground the shield at one end only. Grounding both ends invites a ground-loop current that distorts the very signal the shield exists to protect [5].
- Before touching the transmitter, read the current at three points. The pattern of readings separates wiring faults from instrument faults faster than any swap-out [1][2].
Why the Humble Current Loop Still Rules
Every water treatment room runs on decisions — dose, regenerate, divert, alarm — and those decisions are only as good as the signals feeding them. The 4-20 mA loop has outlived decades of fieldbus and Ethernet challengers for one reason: current, not voltage, carries the measurement. A transmitter forces a defined current through a series circuit regardless of cable resistance, as long as the loop stays within its compliance range, so the value read at the panel equals the value emitted in the field [1][3]. Add a live zero — 4 mA means 0% of range — and a broken wire becomes distinguishable from a legitimate low reading [4][5].
That elegance hides the failure modes. Loops die of voltage starvation, ground potential, electromagnetic noise, sloppy terminations and misconfigured inputs. This article gives you the diagnostic model: how the loop is powered, how the voltage budget is computed, where noise and ground loops sneak in, what each current reading means, and how to walk a fault without guessing.
Loop Power Basics: Two-Wire, Four-Wire, and Who Powers Whom
In the classic two-wire loop, one pair does everything: the supply energizes the transmitter, and the transmitter regulates the current that represents the process value. Because the transmitter lives inside the loop’s voltage budget, its data sheet lists a minimum operating voltage — “compliance,” “loop burden,” or “drop” — that must be respected at 20 mA [1]. Four-wire transmitters get separate power and merely output the signal; they tolerate heavier burdens but bring their own grounding questions [1].
On the receiving end, the measurement is usually developed across a precision resistor — classically 250 Ω, which turns 4-20 mA into 1-5 V for the analog input [3][6]. Every element in series — safety barrier, isolator, indicator, long cable — eats part of the voltage budget. Add up each device’s power requirement, burden or impedance from the data sheet before you energize anything [1].
The Voltage Budget in One Calculation
Two facts make the arithmetic quick:
- Maximum loop resistance: R_max = (V_supply − V_compliance) / 0.020 A. A transmitter with 12 V compliance on a 24 V supply tolerates about 600 Ω total [3].
- Cable resistance: for 400 ft of 20 AWG, the round trip is about 800 ft at ≈10.15 Ω/1000 ft — roughly 8.12 Ω, or a drop of only ≈0.162 V at 20 mA [6].
Copper wire is rarely the budget-killer. The killers are 250 Ω input resistors, isolators, barriers and very long thin runs. When a loop reads fine at 4 mA but tops out near 18 mA and never reaches 20 mA, suspect a marginal voltage budget — the transmitter is starving exactly where it needs the most compliance [1].
Ground Loops and Shielding
The One-End Rule
A shield grounded at both ends connects two grounding points that are never at identical potential. That difference drives a current through the shield, which couples into the conductors and distorts the signal [5]. Standard practice for 4-20 mA is to ground the shield at one end only — typically the control-panel end — and insulate the field end so it can’t touch a junction-box wall [5]. The same logic applies to the loop’s own reference: a PLC input that quietly earths one side of the loop can create a second path for current, especially when field instruments sit on a different ground than the panel [1].
Symptoms of Ground Trouble
- Readings drift when nearby equipment switches on and off.
- A milliammeter at the transmitter and at the panel disagree by a few tenths of a milliamp.
- The error disappears when a shield drain is disconnected from a terminal it should never have touched.
When two grounds are actually required — long routes, high-noise areas — break the galvanic path deliberately with an isolated transmitter, an isolator, or an isolated input module. Don’t hope the two earths are equal. They aren’t.
EMI: Where Noise Enters and How to Keep It Out
Variable-frequency drives, contactors, motors and switching supplies all radiate and conduct noise that a 4-20 mA loop will happily sum with its signal — the receiver can’t tell induced current from measured current [2]. Field-proven containment, in order of priority:
- Physical separation. Never share a conduit or tray run between power and analog signal cables; separate trays or a grounded metal divider do most of the work.
- Instrumentation-grade cable. Twisted, shielded pairs for every analog loop; the twisting cancels magnetic coupling and the shield blocks capacitive coupling [2].
- Correct shield termination. One end only, with a short, low-impedance pigtail or a 360° termination — not a long “pigtail antenna” [5].
- Routing discipline. Where crossing is unavoidable, cross at 90°; keep parallel runs short.
- Filtered power. Feed loop supplies from a clean, dedicated DC source rather than sharing a rail with solenoid coils and contactors.
- Isolation as the last resort. A signal isolator between field and panel converts a coupled-noise problem into two quiet circuits.
Verification is quantitative: monitor the loop current while cycling nearby motors and drives. A healthy loop holds its value within a few tenths of a milliamp while the interference source switches.
What Each Current Reading Means
The live zero turns the loop into a self-diagnosing system — if the input card is configured to interpret it. NAMUR NE 43 formalizes the convention: a healthy transmitter stays within 4–20 mA, current below 3.6 mA or above 21 mA indicates an instrument or wiring fault rather than a measurement, and the control system should raise a fault instead of clamping the value to 0% or 100% [4]. Clamping a 3.4 mA fault to “0% scale” silently discards everything the live zero was designed to tell you [4].
| Loop current reading | Meaning | Where to look |
|---|---|---|
| 0 mA everywhere [1][2] | Open circuit: broken wire, dead supply, blown fuse, or open load | Supply output (full voltage with no current confirms an open), terminal screws, cable continuity [2] |
| Stable ≈4 mA, process should vary | Transmitter at zero-scale or failed low; loop healthy | Sensor wiring, process actually at lower range limit, transmitter status [4] |
| 0–3.6 mA with intermittent dropouts [2] | Loose terminal, corrosion, water in conduit, failing connection | Pull and re-terminate; check insulation resistance [2] |
| ≤3.6 mA or ≥21 mA steady [4] | NAMUR NE 43 fault band: transmitter reporting its own failure | Transmitter diagnostics; sensor burnout; hardware fault [4] |
| Reads fine low, cannot reach 20 mA | Voltage budget exhausted at high current | Recalculate burden: supply, compliance, 250 Ω input, barriers, cable [1][3] |
| Reading oscillates when motors/drives run | EMI coupling or ground loop | Separation, shield termination, isolation [2][5] |
| Panel reads higher than field meter | Parallel current path (ground loop) or instrument loading | Check for second ground; verify input isolation [1][5] |
| Full-scale pinned (20 mA) | Over-range, sensor fault, or input shorted | Disconnect the input; if current collapses to 0 mA, the transmitter was driving full scale [4] |
One symmetric truth is worth memorizing. A series loop reads the same everywhere, so identical readings at both ends mean the wiring is honest and the disagreement is interpretive — scaling, range, input configuration. Different readings mean physical trouble between the two test points [1].
HART: Diagnostics on the Same Wires
Smart transmitters overlay digital communication on the analog loop, so device status, secondary variables and diagnostics are available without extra wiring. HART superimposes its digital signal on the 4-20 mA current, and the analog value remains the primary measurement — which is exactly why a loop can be electronically healthy yet diagnostically rich. An over-range sensor, a failing temperature element, or an internal self-check all surface through the device’s status long before the current leaves its band. Practical rules: read diagnostics at the communicator or the HART-enabled input, remember that digital signaling needs adequate loop impedance to be seen at the receiver, and treat device status as evidence — a transmitter that says “sensor fault” is usually telling the truth.
A Step-by-Step Troubleshooting Sequence
- Read the current at the receiver. In band? Below 3.6 mA or above 21 mA is fault information, not process data [4].
- Compare field and panel readings. A clamp meter or loop calibrator at the transmitter end settles “wiring vs. instrument” in one measurement [1].
- Check the supply under load. Measure at the power supply and at the transmitter terminals at 20 mA demand; the difference is your wiring and series-device drop [1][6].
- Recompute the budget. Sum compliance, input resistor, barriers and cable resistance; verify against supply voltage at maximum signal [1][3].
- Inspect the commons. Look for shields grounded at both ends, input commons tied to earth, and junction-box drains touching metal [5].
- Hunt coupling. If the noise correlates with motor or drive operation, fix separation and shielding before touching calibration [2].
- Query the device. Pull HART diagnostics before deciding the transmitter is broken — or before deciding it’s fine.
- Document the loop. Record cable length, gauge, devices and voltage budget. The next fault becomes a lookup, not an expedition.
Design Notes for the Next Installation
Most loop faults are installed, not born. Specify shielded instrumentation cable from the start. Keep signal and power in separate zones from panel layout to field routing. Budget voltage at 20 mA including future isolators. Configure analog inputs to honor NE 43 fault bands instead of clamping, and terminate every shield to a labeled single-point ground [4][5]. Integrated instrumentation shrinks the fault surface further: a multi-parameter probe that carries several measurements on one loop means fewer terminals, fewer junctions, and fewer chances for a loose screw to impersonate a process upset. At Shanghai ChiMay we sum it up as “Four parameters. One probe. Zero alignment error.” — fewer connections is itself a reliability feature.
The loop is not complicated. It’s a single series circuit with a live zero and a voltage budget, and every symptom in the table above falls out of one of those three facts. Read the current, respect the budget, ground once — the signal will tell you the rest.
References
- Instrumentation Tools — “Troubleshooting a 4-20 mA Current Loop” — loop power-supply verification, burden/drop summation and the voltage-budget method for diagnosing starved loops. https://instrumentationtools.com/troubleshooting-a-4-20ma-current-loop/
- Industrial Monitor Direct — “4-20 mA Current Loop Voltage Drop Reference for PLCs” — field symptom table covering open circuits, intermittent low currents from loose terminals, and noise correlated with drive operation. https://industrialmonitordirect.com/ar/blogs/knowledgebase/understanding-4-20ma-current-loop-voltage-drops-in-plcs
- Electrical Engineering Center (ElectricNeutron) — “Current Loop Characteristics” — compliance-based maximum loop resistance calculation ((24 − 12) / 0.02 A = 600 Ω) and live-zero scaling. https://www.electricneutron.com/current-loop-characteristics/
- Tagsight — “NAMUR NE 43” — fault-state convention for 4-20 mA transmitters (below 3.6 mA / above 21 mA = instrument or wiring fault) and PLC/DCS alarm handling. https://tagsight.io/learn/what-is-namur-ne-43
- Zbotic — “4-20 mA Current Loop: Wiring Industrial Sensors Guide” — wiring practice for two-wire transmitters, shield termination at one end only, and ground-loop avoidance. https://zbotic.in/?p=599447
- Industrial Monitor Direct — “4-20 mA Two-Box Wiring: Long-Distance Loop Best Practices” — resistance and voltage-drop calculations for long runs (20 AWG, 10.15 Ω/1000 ft, 24 V supply with 12 V compliance and 250 Ω input). https://industrialmonitordirect.com/blogs/knowledgebase/4-20ma-passive-device-wiring-0v-bonding-for-remote-io
About the author: Written by the Shanghai ChiMay Technical Editorial Team — specialists in water quality analyzers and control-valve integration who support field teams with loop design, SCADA connectivity and signal-integrity troubleshooting.
