How to Test a Soft Starter: Step-by-Step

How do you test a soft starter? A soft starter test sequence checks the power circuit insulation, the SCR (thyristor) pair on each phase, the bypass contactor, and the current-limit and ramp-time settings against IEC 60947-4-2 duty definitions, in that order, before the unit goes back into service. Skip one step and the starter often passes a quick power-up but fails under load a week later, usually from a shorted SCR or a bypass contactor that never picked up. This guide covers isolation and lockout, insulation resistance testing, SCR continuity and leakage checks, bypass contactor verification, current-limit and ramp-time functional tests, and protection relay confirmation, in the sequence a commissioning technician or troubleshooter should run them.

Before You Test: Isolate and Discharge

Disconnect and lock out the incoming line power before opening the starter enclosure. SCRs hold no charge once line power is removed; they are not capacitors. The control power supply and any keypad backup battery on units such as the ABB PSTX or Siemens 3RW55 can still carry voltage for several seconds after disconnection, though. Confirm zero energy with a multimeter across each line terminal to ground and phase to phase before touching any power terminal.

For readers new to the technology, how a soft starter reduces starting current covers the phase-angle control principle behind every step below.

What we see in the field: technicians re-energize a starter that "just tripped" without isolating first, because the panel looks dead. A control transformer fed from a separate circuit, or a PLC run-command input backed by a UPS, can leave part of the circuit live even with the main breaker open. Check the schematic, not just the breaker position.

Insulation Resistance Test on the Power Circuit

With the starter isolated and disconnected from the motor and line, measure insulation resistance between each power terminal and the enclosure ground using a megohmmeter, typically at 500 V DC. A healthy starter reads well above 1 MΩ. A reading in the low kilohms points to moisture ingress, carbon tracking on the heatsink, or a failed snubber network. Test each phase separately and record the readings — one weak phase out of three is easy to miss on a single spot check.

Do not run this test with the motor still wired in if the motor's own winding insulation has not been checked separately. A failed motor winding will mask or distort the starter's own reading. Disconnect the motor leads first when practical.

SCR (Thyristor) Continuity and Leakage Test

Each phase carries a back-to-back SCR pair, as covered in how SCR phase-angle control fires each phase. With power removed, use a multimeter's diode-check or ohmmeter function across the anode and cathode of each SCR. A healthy device reads open, very high resistance, in both directions when the gate is not triggered. A low reading in either direction with the gate untriggered means the SCR has failed shorted.

Check gate to cathode next. This junction behaves like a diode: low resistance one direction, open the other. A gate that reads open in both directions has a broken gate lead or a blown gate-cathode junction, and that SCR will never fire even though the anode-cathode path looks fine.

What we see in the field: a single shorted SCR rarely stops the motor outright. The other two phases still conduct, so the motor starts unevenly and the starter trips on phase imbalance or phase loss rather than raising a dedicated SCR fault — the imbalance protection reacts faster than the SCR diagnostic on some models. Do not stop at the first fault code. Check all three phases.

Verifying the Bypass Contactor

Once the motor reaches full speed, the bypass contactor should close and short out the SCRs. Confirm this with a clamp meter across one SCR at run speed: current through the SCR should drop close to zero once bypass closes, since load current now flows through the contactor's copper contacts instead. If the SCR still shows near-full load current at run speed, the bypass never picked up.

SCRs left in the circuit dissipate roughly 1-1.5 W per amp per phase while conducting. On a 200 A starter that is 200-300 W per phase, continuously, with no bypass — enough to trip a thermal protection function or cook the heatsink over an extended run. A starter that runs hot with no fault logged, on a motor that used to run cool, points here before anywhere else.

Verify the bypass contactor's own contacts too, with the starter isolated: check contact resistance across each pole, and confirm the coil pulls in at rated control voltage. A pitted contact can close but still drop enough voltage under load to leave the SCRs partially conducting.

Functional Test: Current Limit and Ramp Time

With the motor connected and running under normal load, measure actual start current with a clamp meter or the starter's own trend log and compare it against the programmed current-limit setting.

A 50 A FLC motor with the current limit set to 350% should draw close to 175 A during the ramp. A measured start current well above or below that figure means the current-limit parameter is set wrong, the FLC value entered into the starter does not match the nameplate, or a current-transformer ratio is misconfigured.

Checking for Undersized Starting Torque

Starting torque falls with the square of applied voltage: at 50% voltage the motor produces roughly 25% of its direct-on-line locked-rotor torque. If the motor stalls partway through the ramp instead of reaching full speed, the load's torque curve is likely crossing above the reduced motor torque curve somewhere in that ramp. Not a starter fault. Raise the current limit, or move to torque control instead of a plain voltage ramp, rather than assuming the starter itself is defective.

Time the ramp with a stopwatch against the programmed setting. A ramp that finishes early usually means the current limit was reached before the timer ran out, which is normal behavior on current-limited starters, not a fault.

Testing Protection and Fault Relay Outputs

Force or simulate each protection function in the starter's test or commissioning mode and confirm the fault relay operates: overload (I²t, class 10/20/30), phase loss, phase imbalance, phase sequence, locked rotor or stall, shorted-SCR detection, and undercurrent on pump duty. Cross-check each output against the panel wiring diagram, not just the starter's local display. A fault relay that trips locally but is wired to the wrong PLC input looks identical to a starter that "doesn't protect the motor" until someone traces the wire.

What we see in the field: the too-many-starts thermal model gets ignored during commissioning, because the motor is cold and every single test start passes. It surfaces months later when an operator bumps the start button repeatedly during a process upset, and the starter blocks a start the operator expected to work. Test this function deliberately, with several starts in quick succession, rather than assuming it works because one start does.

Undercurrent detection matters on pump duty: it catches a dry-running pump, a sheared coupling, or a broken belt where the motor still spins but does no useful work. Set the undercurrent threshold below normal running current and above no-load current, then verify it by briefly running the pump against a closed discharge valve or with the coupling loosened, whichever the process allows safely. A threshold set too close to normal running current trips on ordinary flow variation instead of a real fault.

Recording Results for the Next Technician

Log every reading against the starter's serial number and date: insulation resistance per phase, SCR anode-cathode and gate readings, bypass contact resistance, measured start current against the current-limit setpoint, and which protection functions were force-tested. A baseline from commissioning turns a future fault call into a comparison instead of a fresh investigation. Without it, the next technician re-tests everything from zero, on a unit that may already be years into service.

This matters more on starters built into OEM skids or panel assemblies shipped years apart, where the original settings sheet is the only record of what "normal" looked like for that specific motor and load.

Conclusion

Testing a soft starter in sequence — insulation first, then SCRs, then bypass, then current-limit and ramp settings, then protection outputs — catches the failure modes that a quick power-up check misses. Most nuisance trips trace back to one of these five areas rather than a defective starter overall. For sizing checks before a starter goes back in service, see the soft starter selection guide; for replacement units across ABB, Schneider Electric, and Siemens ranges, browse the current soft starters catalog at Stoklink.