How to Test a Contactor Properly (Not the Way Most People Do It)

The standard test most technicians run on a contactor — ohms across the closed contacts, looking for something near zero — passes contacts that are actively destroying themselves. By the end of this article you’ll know why that test is unreliable, what to measure instead, and how to test the coil and contact circuits separately so you replace a contactor for a real, numbered reason.

Two Circuits, Two Failure Modes

A contactor contains two electrically separate circuits, and most diagnostic errors come from confusing them. The coil circuit is the control side — a low-power electromagnet drawing milliamps, switched by the thermostat, pressure switches and overloads. The contact circuit is the power side — heavy silver-alloy contacts driven mechanically by the armature, carrying full load current to the compressor or fan.

The coil can fail with the contacts perfect; the contacts can be destroyed with the coil perfect. Test both, separately, every time — a contactor that pulls in reliably can still be running a scorched pole underneath.

Testing the Coil

First, resistance, with the power off. Isolate, lock off, prove dead per the OHS Act 85 of 1993 sequence, and disconnect one coil lead so you aren’t reading a parallel path. Measure across A1 and A2.

These ranges are wide because resistance depends on the contactor’s size as much as its rated voltage — the manufacturer’s figure matters more than the table. The failure signature is unambiguous: an open coil reads OL and is scrap. A shorted coil reads near 0 Ω — under 2 Ω where it should read hundreds — and is scrap. A coil at roughly half its expected resistance has shorted turns: it pulls in weakly, runs hot, and needs replacing even though it still closes.

Second, voltage at the coil terminals when the circuit calls. Leave the coil connected and measure A1 to A2 while the circuit tries to close it. Full control voltage with the contactor not pulling in means a jammed mechanism or an open coil. No voltage means the fault sits upstream, and the contactor is not yet implicated.

Pull-In and Drop-Out: Two Thresholds, Not One

A contactor is not simply on or off. Under IEC 60947-4-1 it must close reliably between 85% and 110% of rated coil voltage, and drop out between 20% and 75%. A 230 V coil must pull in by 195 V, a 24 V coil by 20.4 V; drop-out on a 230 V coil falls between 46 V and 173 V.

The gap between those numbers is where intermittent faults live. A coil sitting at 180 V — 78% of rated — will often hold in once closed but will not reliably close from rest. The plant works after a manual push, then fails on the next automatic start, reported as “intermittent” because nothing looks wrong when you arrive. Measure coil voltage at the instant of pull-in, not at rest — the coil’s own inrush drags the supply down at exactly the moment it needs to close.

Chattering: Two Causes, One Measurement Apart

A chattering contactor buzzes rapidly open and closed, from two distinct causes separated by one measurement.

Low control voltage. A coil draws six to ten times its holding VA during pull-in. If the supply can’t deliver that surge, voltage collapses, the coil releases, voltage recovers, and the cycle repeats several times a second. Measure coil voltage during pull-in: below 85% of rated (under 195 V on 230 V, under 20.4 V on 24 V) confirms this cause. Look upstream for an undersized control transformer, a long control cable, or a high-resistance joint.

A cracked shading ring. An AC electromagnet’s flux crosses zero twice per cycle, which at 50 Hz would release the armature 100 times a second without correction. A copper ring in the pole face carries an induced current producing a phase-shifted flux so total flux never reaches zero. A cracked ring produces a loud 100 Hz buzz with correct coil voltage present. If voltage stays above 85% during pull-in and it still chatters, the magnet assembly is the fault, and the contactor is scrap regardless of coil resistance.

Contact Testing: The Measurement Most People Skip

This is the test that decides whether a contactor stays in service, and it is not resistance. Measure voltage drop across each closed pole with the load running at normal current — one probe on the line-side terminal, the other on the load-side terminal of the same pole, contactor closed, motor running.

Below 0.1 V per pole is healthy. 0.1–0.3 V means pitting has started — record and plan replacement. 0.3–0.5 V is deteriorating fast. Above 0.5 V, replace now. These thresholds look small because dissipation is drop multiplied by current: a healthy 0.05 V drop at 24 A dissipates 1.2 W, but 0.9 V on the same pole dissipates 21.6 W inside a plastic housing never built to shed it.

Safety note: this is a live measurement at energised load terminals. Use a CAT III-rated meter, keep one hand clear of earthed metalwork, and confirm the enclosure is otherwise safe before you probe inside it.

Why Resistance Across Closed Contacts Fails You

A multimeter’s ohms range pushes about 1 mA through the circuit and resolves to around 0.1 Ω. A healthy contact reads 1 to 5 mΩ — fifty times below that resolution — and the test leads alone contribute roughly 0.2 Ω. A pitted pole reading 0.2 Ω looks like lead resistance and passes, yet at 25 A it drops 5 V and dissipates 125 W. Voltage drop under load tests the contact at the current it actually carries; cold resistance never gets there.

Contact Welding

Welded contacts leave the load energised with the coil de-energised — the plant runs and will not stop. Confirm it by de-energising the coil and measuring load-side voltage: full voltage present with 0 V across A1–A2 means welded contacts. Isolate at the upstream isolator immediately.

Welding is always a symptom: check repeated locked-rotor current from short cycling, a contactor sized on its AC-1 resistive rating rather than AC-3 motor duty, low coil voltage causing slow closure and bounce, and chattering. Fit the replacement, then fix whatever welded the last one.

Table: Contactor Fault Matrix

Worked Example: A Cold Room Tripping Its Overload Every 20 to 40 Minutes

A 400 V three-phase, 13 kW scroll compressor ran 20 to 40 minutes, tripped its overload, and restarted fine on reset — a pattern pointing at heat building up, not a starting fault. Phase-to-phase voltage read 398 V, 400 V and 399 V, and running current read 24.1 A, 23.8 A and 24.0 A against a 24 A nameplate — balanced and correct. The compressor was not overloaded, so something else was tripping it.

The 230 V coil measured 624 Ω isolated, within band, and 224 V at pull-in — above the 195 V threshold, clearing the coil. Voltage drop across the three poles running told the real story: L1 0.04 V, L2 0.90 V, L3 0.05 V.

L2 sat at nine times the 0.5 V replacement threshold — a contact resistance of 0.9 ÷ 24 = 37.5 mΩ against a healthy 1–5 mΩ, dissipating 21.6 W where the other poles dissipated 1.2 W. The overload, mounted directly beneath the contactor, was tripping on that pole’s heat plume, not motor current — a thermocouple on L2 read 103 °C after 25 minutes against 44 °C on L1. The site log showed the unit restarting immediately after every load-shedding recovery with no anti-short-cycle protection, at up to fourteen starts an hour against a design figure of six — eroding that pole with roughly 140 A of locked-rotor current per restart, a stress covered in what load shedding does to refrigeration equipment.

Fitting a replacement rated for AC-3 motor duty and a 5-minute anti-short-cycle timer resolved it: rechecked, the three poles read 0.03 V, 0.05 V and 0.04 V, and the L2 terminal settled at 41 °C after 30 minutes. Condemned on “it trips the overload,” the compressor would have been replaced for nothing.

Final Thoughts

A contactor earns its place in service on two tests: coil resistance and pull-in voltage on the control side, voltage drop under load on the contact side. Cold resistance across a closed pole substitutes for neither — it cannot resolve a healthy contact from one about to fail. Before probing any live terminal, confirm your isolation and proving procedure as laid out in ghost voltage and why your multimeter reads a dead wire, and see the full sequence in our full electrical troubleshooting reference.