A relay coil with its suppression diode reversed is a bolted short across the driver the instant that point turns on. The diode often survives long enough to crowbar the 24 V rail; the output transistor does not. What you find afterward is a point that stays energized with the processor in STOP, on a module that passes every other test you can run on it.
Skip the Fixes That Cannot Work
Everything below gets tried at 2 a.m. before somebody accepts the silicon is gone. Two of them are worth doing. Neither one brings the point back.
| Quick fix | Why it fails | What it does tell you |
|---|---|---|
| Cycle PLC power / field power | A fused drain-source channel is a metallic short. No power-on reset re-opens it. | If the load re-energizes with the PLC in STOP, the driver is shorted. |
| Re-download, force the coil off | Logic never reaches the load; the failed driver bypasses the gate entirely. | Confirms the fault is hardware, not a stuck force or duplicate coil. |
| Replace the diode, retest the same point | Fixes the cause, not the damage. | Do replace the diode — just land the load elsewhere. |
| Reseat the terminal block or the module | Loose contacts create open circuits, never a hard-on output. | Rules out a wiring fault in about a minute. |
| Swap in a new module, rewire nothing | Puts fresh silicon into the same short. The second module dies on first energize. | Nothing. This is how one failure becomes two. |
| Add an inline fast-acting fuse to save the driver | Fuse clearing time is milliseconds; the transistor leaves its safe operating area in microseconds. | Fuses protect wire and power supply. Size them to the conductor. |
| Add a PTC resettable fuse | Trip time against a low-impedance short is tens of milliseconds to seconds. | Useful against a stalled or pinched-wire overload. Not against a bolted short. |
Know What the Backwards Diode Did to the Driver
Wired correctly, a suppression diode sits reverse-biased across the coil the entire time the output is on. It conducts only at turn-off, carrying the collapsing field current and clamping the inductive kick to roughly one diode drop above the rail.
Reversed, it is forward-biased the moment the driver applies rail voltage. The fault loop is the transistor's on-resistance plus the field wiring plus the diode's forward resistance — tens of milliohms end to end. Fault current is set entirely by what the 24 V supply can push before it folds back or goes into hiccup mode, and the switching device is the smallest thermal mass in that loop. Die metallization or the drain-source junction fuses, and what is left is a resistive short from the output terminal to its supply. That is why the point reads energized regardless of the ladder.
Use the failure pattern as a diagnostic. One point stuck on, neighbors normal, means driver silicon. An entire common gone dead with no points working means a group fuse or a blown common — check the module datasheet for whether that fuse is internal and whether it is field-replaceable. Stop here if more than one point on the same common reads shorted: that is a rail event and you have another fault to find before you spend a module.
The crowbar also hits everything else fed from that supply. Walk the rail and check sensors, safety relays, HMIs and comms devices for latched faults or brownout-induced errors before you call the job done.
Confirm the Point Is Gone
- Lock out, then remove field power from the output group.
- Lift the field wire off the suspect terminal so the load cannot influence the reading.
- Meter from the output terminal to the point's supply terminal per the module wiring diagram, both polarities. Near 0 ohms in both directions is a fused channel. A healthy point reads high resistance one way and a body-diode or leakage reading the other.
- Restore power with the field wire still off. Command the point off, then measure terminal to common. Full field voltage present with the point commanded off confirms the driver is shorted on.
- Repeat on every remaining point sharing that common. Collateral damage is common after a crowbar and is easier to find now than during commissioning.
- Test the diode out of circuit — roughly 0.4 to 0.7 V forward, OL reverse. A part that has passed fault current can be leaky and still measure close to normal, so replace it rather than argue with it.
- Load the 24 V supply and verify it holds regulation without hiccupping. A supply that has been crowbarred repeatedly can come back degraded.
Get It Running Before Shift Change
- Verify the replacement diode's band orientation on the bench, before it goes anywhere near the panel. The band faces the coil terminal that is positive when the coil is energized — on a sourcing point that is the module output, on a sinking point that is the field +V rail.
- Move the field wire to a spare point, update the alias or tag mapping in the program, and mark up the drawing on the spot.
- No spare on that module? Use a spare on another module or another common, or land a small interposing relay or SSR driven from any free point and switch the load through its contacts.
- Tape, tag and land nothing on the dead terminal. A shorted-on point energizes its load the second field power comes up, so treat that terminal as live output regardless of program state.
- Confirm nothing safety-rated was riding on that point. Safety functions do not belong on standard transistor outputs.
- Order the replacement module now and run on the reassigned point until you can swap it during planned downtime. Get it running, then fix it properly.
Make the Next Short Survivable
- Pluggable interposing relay modules with integral coil suppression. The diode lives inside the module and has no orientation a tech can get wrong. This is the cheapest permanent fix for this exact failure.
- Relays ordered with factory coil suppression. Same benefit, no discrete parts on the terminal strip.
- Non-polarized suppression where a discrete part is unavoidable. An MOV, a bidirectional TVS or an RC snubber cannot be installed backwards. Pick a clamp voltage above the rail and below the module's off-state voltage rating from the datasheet, and expect faster relay dropout than a plain diode gives.
- Interposing relays on anything whose wiring leaves the panel. A crushed conduit run then kills a relay instead of a driver you cannot replace in the field.
- Group fusing sized to the wire. It protects the conductor and the supply, and keeps one fault from browning out the whole rail. It will not save the transistor.
- Current-limited protected output modules. Foldback-limited drivers with fault reporting are the only thing that survives a hard short and tells the processor about it. Protected DC output modules have been built for the Terminator I/O family but have never been high-volume items, so confirm current availability with AutomationDirect sales before you design a panel around one.
- Spare-point discipline. Pull two spare outputs per module to the terminal strip during the build. That is what turns this failure into a ten-minute recovery.
Catch Reversed Diodes at Pre-Power-Up
- Check every suppression diode band against the drawing, not against the neighboring circuit. Copying the adjacent wire is how a whole row goes in backwards.
- Apply field power with the PLC in STOP and note supply current. A reversed diode draws nothing until its output turns on, so this step alone will not find it — it only proves the wiring is clean at rest.
- Energize each output individually through a current limiter: a bench supply set slightly above coil current, or a 24 V lamp in series with the output supply feed. Correct wiring pulls the relay in and the lamp stays dim. A reversed diode holds the lamp at full brightness, the relay does not pull in, and the supply sits at current limit.
- Remove the limiter only after every point has picked up its load correctly.
- Log which points were proven and which are spares, and mark spares on the panel door.
Stop and call AutomationDirect technical support with the module part number and the fault description if a replacement module fails again on a load you have already proven, if points fail on more than one common after a single event, or if the base or backplane logs faults after the rail collapse. Do not keep pushing modules into a fault you have not isolated. Availability of protected or current-limited output options for a given I/O family is a sales and support question, not something to assume from a catalog page.
FAQ
Can I reset a shorted PLC transistor output by cycling power or reloading the program?
No. The failure is a fused drain-source channel — a metallic short inside the device — so no power-on reset, firmware reload or forced-off coil re-opens it. Move the load to a spare point and replace the module at the next planned outage.
Does a PTC resettable fuse protect a DC output module from a shorted load?
Not from a bolted short. A PTC trips in tens of milliseconds to seconds, while the output transistor exceeds its safe operating area in microseconds. PTCs and fuses earn their place against sustained overloads, pinched wiring and stalled loads, and they protect the conductor and the power supply — driver survival requires a current-limited protected output or an interposing relay.
Can I use an MOV or bidirectional TVS instead of a flyback diode on a relay coil?
Yes, and it removes the polarity mistake entirely since neither part has an orientation. Select a clamp voltage above the supply rail and below the output module's rated off-state voltage, and expect faster relay dropout and higher coil voltage stress than a plain freewheeling diode produces.