A solid-state relay that switches one winding of a two-speed 230 VAC motor correctly and refuses to release the other is not a wiring fault and not a logic fault. The number that matters is the load current: roughly 50 mA on the low-speed winding. That is below the current the output device needs to commutate off at the zero crossing, and the output stays latched through the next half cycle whether or not the control input is energized.
Fixes That Do Not Move the Symptom
The first instinct is to suspect the command path, so the control signal gets scoped or a longer off-delay gets added in software. The SSR input is a DC-driven optocoupler; when the control voltage is removed the LED is dark within microseconds. Off-time in the program changes nothing because the output side is holding itself on, independent of the input side.
Second instinct: the SSR is undersized or oversized for the load. A device rated for several amps handles 50 mA fine on the conduction side, but rating headroom does not help at turn-off. Swapping in a larger SSR usually makes the latch-up worse, because larger output devices need more holding current, not less.
Third instinct: swap the SSR for a random-turn-on type instead of a zero-crossing type. Turn-on behavior is irrelevant here. Both types turn off the same way, by waiting for load current to fall below holding current near the current zero.
Fourth instinct: reverse the two speed leads or move the SSR to the neutral side. The high-speed winding works because it draws more current with a different inductive character. Moving the switching element does not change the winding's impedance.
Why the Low-Speed Winding Latches the Output
An SSR output is a thyristor structure. Once gated on, it stays on until the current through it drops below its holding current and stays there long enough for the device to recover blocking capability before voltage reappears across it. This is device physics, not logic.
Two things go wrong on a low-current inductive winding:
- The current magnitude is marginal. At 50 mA the peak load current is only a few times the holding current of a typical output device, so the window near zero crossing where current is genuinely below holding current is short.
- Motor current lags the line voltage. When the current finally crosses zero, the line voltage is already well away from zero, so the full instantaneous mains voltage is reapplied across the output device with a steep dv/dt. That dv/dt injects displacement current into the device junction and re-triggers it before it has recovered.
The high-speed winding presents a different current magnitude and phase angle, and it happens to fall inside the SSR's commutation envelope. The low-speed winding does not. Same relay, same wiring, different load physics.
A triac-output SSR fails this way readily because a triac conducts on both half cycles and must recover blocking state twice per mains cycle, with the reapplied voltage of the opposite polarity arriving immediately. Antiparallel SCR outputs, such as the SCR output stage in the Crydom DPA6111, give each device a full half cycle of reverse bias to recover, which is why an SCR-output device is the correct starting point for inductive duty. It does not, on its own, solve a load that is simply too small to commutate cleanly.
Quantities, Limits, and Where to Read Them
| Quantity | Value in this case | Where to read it |
|---|---|---|
| Load current, low-speed winding | ~50 mA | Clamp meter on the winding lead, running |
| SSR minimum load current | Device-specific | Datasheet electrical table, "minimum load current" or "minimum operating current" |
| Output holding current | Device-specific | Datasheet output characteristics |
| Off-state dv/dt withstand | Device-specific, V/µs | Datasheet, "critical rate of rise of off-state voltage" |
| Output stage type | SCR (DPA6111) | Datasheet block diagram / output description |
| Snubber capacitor that worked | 0.33 µF across the motor supply lines | Measured result on this installation |
If the measured load current is at or below the datasheet minimum load current, stop diagnosing the control system. That single comparison explains the entire symptom.
The Fix: Capacitive Load Across the Motor
Adding capacitance across the motor supply lines does two jobs at once. It draws leading current that adds to the winding current, pushing total SSR current above the minimum load threshold, and it slows the rate of voltage rise across the SSR output after commutation so the reapplied dv/dt stays inside the device's withstand rating.
- Measure running current on the low-speed lead with a clamp meter and compare it against the SSR's minimum load current in the datasheet.
- Select an AC-rated film capacitor. Use an X-class or motor-run type rated for continuous 230 VAC line service, not a DC-rated part. Start in the 0.22 to 0.47 µF range for a 50 mA load; 0.33 µF resolved this case.
- Wire the capacitor directly across the motor supply lines — between the switched winding lead and neutral, at the motor terminals or as close as the enclosure allows.
- Keep the capacitor leads short. Long leads add inductance that degrades the dv/dt clamping.
- Fuse or verify the branch protection. The capacitor is now permanently across the line whenever that winding is energized.
The added current is roughly I = 2πfCV. At 50 Hz and 230 V, 0.33 µF contributes on the order of 24 mA; at 60 Hz, about 29 mA. That is the same order as the motor current itself, which is why it flips the commutation behavior.
An RC snubber — capacitor in series with a resistor of tens of ohms — is the more conservative version and is preferred where a bare capacitor across a thyristor output raises concern about discharge current at turn-on. On a 50 mA load with an SCR-output SSR the series resistance is optional; on higher-current loads it is not.
Verification
- Command the low-speed winding on, then off, twenty times from the computer. Every off command must stop the motor within one mains cycle.
- Cycle the high-speed winding the same number of times. The capacitor must not have introduced a new failure on the winding that previously worked.
- Cycle low to high and high to low directly, if the control logic permits it, and confirm both transitions release cleanly.
- Run the motor on low speed for thirty minutes and check the capacitor body temperature by hand. A capacitor that is warm is undersized in voltage class or is a DC-rated part in an AC position.
- Measure voltage across the SSR output with the command off and the motor stopped. It should read full line voltage. A reading near zero means the output is still latched.
- Verify the motor is genuinely at rest and not turning slowly on leakage current through the snubber. Leakage through the capacitor is present continuously; on a very small motor it can produce audible hum or creep.
Recurring Pitfalls on This Class of Load
Oversizing the capacitor is the most common overcorrection. Too much capacitance produces enough leakage current through the off-state SSR to hold a small motor in a buzzing, partially energized state. If the motor hums when it should be off, step the capacitor down one value.
DC-rated capacitors fail here. A 400 VDC electrolytic or a DC film part placed across a 230 VAC line will overheat and can fail short. Do not fit a DC-rated capacitor across the mains — a shorted capacitor puts a bolted fault across the winding.
Snubbing at the SSR terminals instead of at the motor is less effective. The wiring inductance between the relay and the motor is part of the problem; place the capacitor where it sees the winding directly.
Two-speed motors also make the two windings look interchangeable in the diagnostic. They are not. The pole-count difference means different inductance, different current, and different phase angle, so a relay that commutates one winding reliably can fail on the other in the same machine at the same line voltage.
Where the machine cycles frequently or the winding current sits far below the SSR minimum load, a small parallel bleed resistor sized to raise the total current past that minimum is an alternative to capacitance. It dissipates continuous power and needs a wattage rating with margin, so it is the second choice, not the first.
FAQ
What happens if the load current is below the SSR minimum load current?
The output thyristor never falls below holding current long enough to recover blocking state, so the SSR stays conducting after the control signal is removed. The motor keeps running until the branch is opened upstream.
What happens if I use an oversized snubber capacitor?
Leakage current through the off-state SSR increases and can hold a small motor in a humming, partially energized condition instead of stopping it. Drop to the next smaller value, working down from 0.47 toward 0.22 µF on a 50 mA load.
What happens if I swap in a larger SSR instead of adding a snubber?
Turn-off gets worse. Higher-rated output devices generally require more holding current, so a marginal 50 mA load becomes further from the commutation threshold.
Escalate to the SSR manufacturer's application support when the measured running current is above the published minimum load current and the output still fails to release, or when adding capacitance in the 0.22 to 0.47 µF range does not change the behavior. Have the datasheet part number, the measured winding current, and the off-state voltage reading across the output ready. A device that latches with adequate load current and an in-spec dv/dt environment is a candidate for return under warranty rather than a field fix.