Starter Drops Out and a Phantom Voltage on the Run Light

The drop-on-release symptom means the START path works and the parallel holding path does not. Everything in series ahead of the START button is proven good, because the coil picks up while START is held. That narrows the list to:

• the holding (auxiliary) contact itself - wrong contact (NC instead of NO), misadjusted arm, dirty or pitted tips, or landed on the wrong terminals;
• the two wires from the holding contact to the START terminals (open, broken strand, wrong terminal);
• a coil terminal connection that opens when the contactor seats (rare, but worth checking);
• a control transformer that cannot hold the coil once the seal-in path is the only supply - less likely because the coil holds fine while START is held.

Test sequence (starter energized, START held, contactor picked up): measure voltage across each series device - the closed STOP contact, the closed OL contact, and the closed aux contact. A closed contact should read approximately 0 V; the device showing close to 120 V across it is the open element. If both STOP and OL read 0 V and the aux contact reads 120 V, the aux contact is the fault. Back this up mechanically: de-energize, push the contactor in by hand, and check continuity across the aux terminals while watching the arm for over-travel or a missing spring. Confirm the print's contact designation before ordering a replacement.

Most likely not a fault - it is a high-impedance meter reading a coupled or leakage voltage.

A modern DMM has an input impedance around 10 MΩ. The 180 ft of control conductors share a raceway with 480 V motor leads, so capacitive coupling injects a small current into the floating control conductor. With no load to hold the voltage down, the DMM reads a real potential that cannot deliver any current. A neon pilot light in parallel adds its own high impedance and does not load the circuit, so the reading survives.

Prove it:

• Re-measure with a low-impedance (LoZ) function or a loading resistor of a few kΩ across the meter leads. If the reading collapses to near zero, the voltage is phantom.
• Or use a solenoid-type wiggle tester.

Do not conclude a fault exists from the DMM reading alone, and do not "fix" it by replacing the light. However, two things are still worth checking: the light is intentionally wired across the coil (so it goes out when the starter drops), and the light does not visibly glow when the starter is off - a visible glow means leakage current is large enough to matter, and the open STOP contact or wiring insulation should be investigated. If the glow is real, look at contamination across the STOP button and at cable damage in the shared raceway.

Wrong diagnosis, wrong fix.

The coil holds at 118 V while START is held, so the transformer is delivering adequate voltage and VA. Increasing transformer size does not close an open holding contact.

Reasoning to walk through with the tech:

• Voltage across the coil is the dependent variable. It is 118 V when the coil is energized and 0 V when it is not - that is exactly what an open circuit upstream does. A weak transformer would show a depressed voltage while energized, particularly at pickup, and would show elevated temperature or a blown primary fuse.
• The 500 VA transformer is already considerably oversized for a single NEMA-size coil (typically under 100 VA inrush, far less sealed). Upsizing increases available fault current on the 120 V control circuit and may exceed the interrupting rating of the control fuse and the wiring's short-circuit withstand.
• If inrush were genuinely marginal, the symptom would be chatter or failure to pick up, not a clean drop on release.

Fix the actual open circuit - the aux contact or its wiring - then re-test. If a transformer change is ever justified, size it from the coil's sealed and inrush VA and check the primary OCPD and secondary conductor protection accordingly.