A relay that chatters, welds shut, refuses to drop out, or makes the PLC glitch when it opens is behaving the way the physics says it will. The usual list of relay advantages is wrong or overstated in four places. Contacts wear out. Standard coils draw power the whole time they are energized. Switching contacts are a noise source, not just a noise victim. And a relay is only fail-safe if you picked the right contact form and the contact can still open.
The fixes are straightforward. Size the contact for the real load type. Suppress the coil and the contact separately. Wire the circuit so the de-energized state is the safe state.
Match the relay symptom to its cause first
Your first check is the coil voltage, measured at the relay's coil terminals rather than at the PLC output, in both commanded states. That one reading tells you in under a minute whether the fault is on the control side or on the relay and load side.
| Symptom on the panel | Likely cause | First check |
|---|---|---|
| Relay won't drop out when the PLC output is off | Off-state leakage from a triac or solid-state output, or capacitive coupling on a long AC coil run, holds the coil above its release voltage | Coil voltage with the output off, compared to the must-release voltage on the datasheet |
| Contacts welded closed, load stays on | Inrush current from a lamp, capacitive, or motor load, or a DC arc beyond the contact rating | Load type compared to the matching rating category, not just the resistive amp figure |
| Load drops out intermittently, contacts look clean | Low-level signal switched through power-rated contacts; the surface film never breaks down | Voltage drop across the closed contact at the actual circuit current |
| PLC input, analog, or network glitch when the relay opens | Unsuppressed coil kickback, or an arc on an inductive load conducting and radiating into nearby wiring | Scope the coil terminals and the load side at turn-off |
| Relay chatters | Coil undervoltage from supply sag or wiring drop, a damaged AC shading ring, or vibration | Coil voltage under full panel load compared to the must-operate voltage |
| Coil runs hot, relay dies early | Monostable coil energized continuously in a hot enclosure, or coil overvoltage | Coil voltage and enclosure temperature compared to the coil rating |
| Contacts pitted or blackened well before expected life | Arc erosion from a DC or inductive load with no contact suppression | Operation count compared to the electrical life curve |
Correct the relay claims that lead to bad designs
- "No wearing parts in the switching mechanism" is wrong. Contacts erode on every break under load. The armature, return spring, and hinge wear mechanically. Datasheets give two life figures: mechanical life with no load, and electrical life at rated load. Electrical life is far lower. Plan replacement on electrical life at your actual load.
- "Holds its state without consuming power" applies only to latching (bistable) relays. A standard monostable relay draws coil current the entire time it is on. That current heats the enclosure, loads your control supply, and drains batteries in standby.
- "Inherently immune to noise" is half true. The coil needs a threshold voltage and current to pull in, so it ignores small induced signals better than a logic input does. The contact arc and the collapsing coil field, however, are broadband noise sources. In many panels the relay is the noise culprit.
- "Defaults to a safe state" depends on you. It holds only if you chose the contact form so that de-energized is safe, and only if the contact can open. A welded normally open contact keeps the load running after control power is lost. For safety functions, use relays with mechanically linked (force-guided) contacts and a monitored safety relay module, and design against the applicable machine safety standard. A general-purpose relay is the wrong part for that job.
- "Isolation protects the controller" has limits. Coil-to-contact isolation holds up to the isolation voltage rating on the datasheet. You defeat it by sharing commons between the control and load circuits, or by bundling coil wiring with load wiring.
- "Voltage and current amplification" is really switching. The PLC output still has to supply the coil current. Check the coil current against both the output point rating and the output common rating.
Understand why inductive and DC loads burn contacts
Opening a contact under current draws an arc across the gap.
- AC loads: the arc usually extinguishes at the next current zero crossing.
- DC loads: there is no zero crossing, so the arc sustains longer, and more so as DC voltage rises. That is why the DC rating on a relay is much lower than its AC rating.
- Inductive loads (solenoids, contactor coils, motors): the inductance drives current through the opening gap, stretching the arc and transferring contact material from one face to the other.
Closing has its own failure mode. Lamps, capacitive loads such as switch-mode supply inputs, and motors draw inrush current several times their steady-state value. Contact bounce during that inrush micro-welds the contacts. The big amp figure printed on the relay cover is usually the resistive AC rating. Read the datasheet for the category that matches your load:
- Resistive
- Inductive
- Motor (hp)
- Lamp or tungsten
- DC at your voltage
Low-level signals fail the opposite way. Silver-alloy power contacts need enough voltage and current to break through their surface film, known as wetting current. Switch a sensor signal or a PLC input through them and you get intermittent opens. Use gold-plated or bifurcated contacts for dry circuits. Once a gold contact has switched a power load, the gold is gone, and the contact is no longer reliable for signal duty.
Suppress the coil and the contact separately
Coil kickback and contact arcing are two different problems. Treat each one on its own terms.
Coil suppression
- DC coil, flyback diode: clamps the kickback almost to zero, but it slows release. The stored energy recirculates through the coil, so the armature opens slowly, the contacts part slowly, and the load side arcs more.
- DC coil, diode plus zener, diode plus resistor, or TVS: clamps at a higher voltage and releases faster. Use this when release time or contact life matters.
- AC coil: fit an RC snubber or a varistor across the coil. Many relay sockets accept plug-in suppression modules. Choose based on how much release delay the sequence can tolerate.
Contact suppression
- Inductive AC load: fit an RC snubber or a varistor, preferably across the load rather than across the contact.
- Inductive DC load: fit a diode across the load.
- Watch the RC leakage: an RC snubber across an open contact passes leakage current. That leakage can hold in a small load, such as another relay coil or a PLC input, while the contact is supposed to be open.
Troubleshoot a misbehaving relay circuit in this order
- Measure coil voltage at the coil terminals, commanded on and off. Compare the readings to the must-operate and must-release values on the datasheet. If the off-state voltage sits above must-release, you have a leakage problem. Fix it with a bleeder resistor across the coil, or by changing the output type. A new relay won't fix it.
- Measure the on-state coil voltage with the whole panel loaded. If it falls below must-operate, the relay will chatter. Check for supply sag, undersized conductors, and voltage drop on long runs.
- Measure voltage drop across the closed contact with the load running. Compare it to a new relay of the same type at the same current. A clearly higher drop means eroded or contaminated contacts. Thermal-scan the socket terminals at the same time.
- Identify the load type and its inrush. Match it to the correct rating category on the datasheet.
- Scope the coil at turn-off and the load side at contact opening. A large transient means suppression is missing or has failed.
- Count operations. A PLC counter on the output works. Compare the count to the electrical life at your load. A relay past its electrical life is worn, not defective.
- Replace the relay only after you fix the cause. Otherwise the replacement fails the same way.
These fixes waste time:
- Swapping the PLC output card when the coil voltage is correct.
- Replacing the same relay over and over with no suppression added.
- Moving to a relay with a higher AC rating for a DC load without reading its DC rating.
Prove the fix under real load
- Coil off-state voltage stays below must-release with every other load in the panel running.
- The relay drops out on command every time. If you added a flyback diode, measure the new release time with a scope or a high-speed PLC timestamp. Confirm it doesn't break interlock or sequence timing.
- Contact voltage drop stays stable after the load reaches operating temperature.
- Cycle the load repeatedly and check the diagnostic counters. PLC input faults, analog glitches, and network error counters should stay flat.
- Terminal and socket temperatures stabilize and don't creep upward.
- For fail-safe circuits, remove control power and confirm the load goes to the designed safe state. For safety circuits, simulate a welded contact and confirm the safety relay's monitoring detects it, following the module manual.
Avoid the relay mistakes that keep coming back
- Swapping to a solid-state relay to eliminate wear: an SSR has off-state leakage and needs heat sinking. It commonly fails shorted, and it provides no open air gap when off.
- Using one relay for power and signal: a contact that has carried a power load won't reliably switch a dry circuit afterward.
- Paralleling contacts for more current: the contacts never close at exactly the same instant, so one of them takes the full inrush and welds.
- Putting contacts in series to break DC: this lengthens the effective gap. Do it only when the datasheet publishes a series-connection rating.
- Ignoring enclosure heat: a continuously energized coil in a hot enclosure needs derating. Read the ambient temperature derating curve.
- Running AC coils over long cables: cable capacitance can hold the coil in after the output opens. Use DC coils, interposing relays near the load, or a bleeder resistor.
- Choosing NO or NC by habit: decide what the load must do on loss of control power, then pick the contact form to match.
FAQ
What happens if I remove the flyback diode from a DC relay coil?
The collapsing coil field produces a voltage spike many times the supply voltage. That spike can destroy a transistor output and couple noise into adjacent wiring. If you need faster release than a plain diode allows, use a diode plus zener or a TVS instead of removing suppression.
What happens if I use an AC-rated relay contact on a DC load?
The DC arc has no zero crossing, so it sustains longer and erodes or welds the contact far faster than the AC rating implies. Size the relay from the DC rating at your actual voltage, and suppress inductive DC loads with a diode at the load.
What happens if a power relay switches a low-current PLC input?
The silver-alloy contacts may not break through their surface film at signal-level current, so the input drops out intermittently even though the contacts look clean. Use gold-plated or bifurcated contacts rated for dry circuits, and never reuse them on power loads.
What happens if a relay coil stays energized 24/7?
A monostable coil draws current continuously, heats the enclosure, and ages faster at high ambient temperature. Confirm the coil is rated for continuous duty at your ambient temperature. If standby power matters, switch to a latching relay.
When should I stop troubleshooting a relay and contact the manufacturer?
Escalate when the coil voltage, load type, suppression, and operation count all check out against the datasheet and the relay still fails early or drops out erratically. Send the relay manufacturer's technical support the part number, the measured coil and contact readings, the load description, and the operation count. For safety relay modules, contact the manufacturer whenever the monitoring fails to detect a simulated fault.