I’ve lost count of how many times a customer has walked in holding an ECU, saying: “I already had this repaired once. It failed again in a week.”

And nearly every single time, the story is the same. The ECU got fixed. The injector that killed it didn’t.

That’s the whole lesson of this article, and if you take nothing else away, take this: the ECU output stage is almost never the villain. It’s the victim.

What an “output stage” actually is (in plain English)

Inside your engine control unit there’s a low-power brain — the microprocessor — and a set of muscle. The muscle is the output stage: a group of power transistors (usually MOSFETs or IGBTs) whose only job is to switch heavy current on and off, thousands of times a minute.

Two big ones:

• Injector drivers — pull the injector’s negative side to ground so the solenoid opens. Peak current on a low-impedance injector can hit 4–8 amps in the opening phase.
• Ignition drivers — either switch the coil primary directly (older systems) or send a trigger signal to a smart coil / igniter module.

These transistors are tough, but they are not immortal. They’re sized for a known load. Give them a load that isn’t in the design spec — a short — and they die instantly. Not gradually. Instantly.

How a shorted actuator destroys the driver

Think of it like a fuse rated for the appliance you plug in.

A healthy high-impedance injector sits around 12–16 ohms. Low-impedance (peak-and-hold) injectors run roughly 1.5–3 ohms. An ignition coil primary is typically 0.4–1.5 ohms depending on the system.

Now imagine the injector’s internal winding rubs through its insulation and shorts turn-to-turn. Resistance drops to 4 ohms on a 14-ohm injector. Current doubles or triples. The driver transistor tries to sink it, heats up past its junction limit, and the silicon fails — usually shorted, sometimes open.

Worse case: the winding shorts to the injector body, which is grounded through the fuel rail. Now the driver is being asked to sink a dead short to chassis. Game over, and often it takes the flyback diode and a track on the board with it.

Same physics on the ignition side. A coil with a shorted primary, or a chafed harness wire touching the block, will take out the ignition driver — and on some ECUs, the drivers share a common substrate, so one failure can cascade to neighbouring channels.

This is why fitting a repaired ECU to an untested car is throwing money into a fire.

The pre-install test routine (do this every single time)

Print this. Tape it to the bench. It takes twenty minutes and saves entire rebuild jobs.

Step 1 — Disconnect the ECU. Fully.

Nothing gets tested with the module connected. You’re protecting the repaired unit and you’re removing parallel paths that will lie to your meter.

Step 2 — Measure every injector at the connector

Probe the two pins at the injector itself, not at the harness end. Compare all cylinders against each other and against the manufacturer’s spec.

• Any injector more than about 10% off its siblings is suspect.
• More than 20% off is condemned.
• Reading near zero ohms = dead short, replace it and inspect the driver again.
• Open circuit (OL) = broken winding, also replace.

Step 3 — Insulation test to body

Meter on its highest resistance range, one probe on an injector pin, the other on the injector’s metal body. You want infinity. Any reading at all — even a few hundred kilo-ohms — means insulation breakdown. That injector is a time bomb.

If you have a megohmmeter, better still: test at 500V and expect readings in the tens of megohms.

Step 4 — Coil primary and secondary

Primary across the two low-voltage pins, secondary through the tower (where applicable). Then insulation-test the primary against the coil body exactly like the injectors. Coils crack, and cracked coils leak to ground under heat.

Step 5 — Harness continuity and short-to-ground check

With the ECU unplugged and the injectors/coils unplugged too, test each driver wire from the ECU connector pin to ground. You want infinity on all of them. A short to ground here means chafed loom — very common where the harness crosses the cylinder head, the intake manifold bolts, or a stud near the firewall.

Also check driver wire to driver wire. Two channels shorted together is a classic loom-crush symptom and it’ll take out both drivers.

Step 6 — Voltage supply and ground quality

Key on, measure the injector/coil supply feed at the connector — should be within 0.5V of battery. Then do a voltage-drop test on the ECU grounds: under load you want less than 0.1V drop. Bad grounds raise the driver’s reference voltage and cook it slowly.

Step 7 — Current clamp on first start-up

Once everything checks out and the repaired ECU is in, clamp a low-amp probe around an injector wire before you drive anywhere. Compare peak and hold current against a known-good waveform. A scope here is worth its weight — a slow current rise or a peak that overshoots tells you something is still wrong.

Real-world causes we see over and over

Fuel contamination. Water or ethanol-heavy fuel corrodes the injector winding varnish. One tank of bad fuel, six weeks later, an ECU on the bench.

Deposits causing a stuck injector. The solenoid stays energised longer, heat builds, insulation degrades.

Over-fuelling from oversized injectors on a tune. A very common one in the tuning world. If you fit 1000cc low-impedance injectors to an ECU designed for saturated high-impedance units without a resistor pack or proper peak-and-hold drivers, you’re overloading the output stage by design. It might last a season. It won’t last two.

Heat-soak on coil-on-plug units. Turbo cars especially. Coils sit in a hot valley, insulation cooks, primary shorts.

DIY wiring. Aftermarket alarms, dashcams and boost controllers tapped into the wrong wire. I’ve seen an ignition trigger wire spliced as a “12V ignition source.” That car needed a new ECU and a new loom section.

Reversed jump-start. Rare but catastrophic. If it’s happened, assume every driver is compromised.

After the repair: what a proper rebuild looks like

If you’re sending a module out or doing it yourself, the standard should be:

1. Failed driver removed, board cleaned, tracks inspected under magnification and repaired where lifted.
2. Genuine or equivalent-spec replacement transistors — not the cheapest thing on the marketplace. Voltage rating, current rating and clamping voltage all matter.
3. Flyback/clamp diodes tested and replaced as a set on the affected channel.
4. Board bench-tested under simulated load, not just powered up.
5. Conformal coating restored where it was removed.

Ask your repairer whether they load-test. If the answer is vague, find another repairer.

The one-line rule

Fix the cause, then fix the ECU. In that order.

An output stage that failed once will fail again in seconds if the short is still sitting out there in the engine bay. Twenty minutes with a multimeter is the cheapest insurance in the whole job.

Quick FAQ

Can I test an injector with the ECU still connected? No. You’ll get misleading readings through the module’s internal circuitry, and you risk back-feeding the driver. Always unplug.

My injector reads correct resistance — can it still be the culprit? Yes. Cold resistance can look perfect while insulation breaks down at operating temperature. That’s why the body-insulation test and the hot current-clamp check matter.

Will a bad injector throw a fault code? Sometimes — “injector circuit open/short” or a cylinder misfire. But a driver that has already failed short may leave the injector permanently on, which usually shows as a fouled plug and a flooded cylinder rather than a clean code.

Is a repaired ECU as reliable as new? A properly rebuilt one with correct-spec parts, yes — often more reliable, since the original failure point gets upgraded components. A cheap repair with mismatched transistors, no.

How much current is too much for an injector driver? Depends entirely on the ECU, but as a rule: if measured peak current exceeds the injector’s rated peak by more than about 20%, stop and investigate before running the engine.