Ignition leads: testing them, and the failure that only shows up in the wet

An ignition lead is not really a wire. It is a tube of insulation with a conductor down the middle, and almost everything that goes wrong with one is a failure of the tube rather than of the conductor.

That single fact organises the whole subject, and it is the reason a lead can measure perfectly on a meter and still be the fault. An ohmmeter interrogates the core. The core is rarely the problem.

Why it fails under load, and why it fails in the wet

A coil does not produce a fixed voltage. It produces whatever voltage the spark gap demands, rising until something breaks down, and then it dumps its energy through whatever path broke down first.

So the demand is not constant. It goes up when the cylinder is full and compressed — that is, under load — because a denser mixture is harder to ionise. It goes up as the plug's electrodes erode and the gap widens. It goes up when the mixture is lean.

Meanwhile the alternative path — a hairline crack in the insulation, a scorched track across a boot, a lead resting on the cylinder head — has a breakdown voltage of its own, and that one goes down when the surface is damp or dirty.

The two curves cross, and where they cross is the whole diagnosis:

A reader who has been told the car "only plays up when it rains" has been handed the answer. A misfire with a weather pattern is an insulation fault until proved otherwise — see P0300 for a random misfire and P0301 onward for a single cylinder, which is the more useful reading here because it names the lead.

The test the original teaches, which you must not do

It describes the traditional method: start the engine and pull the leads off the plugs one at a time. If the engine stumbles the cylinder was contributing; if nothing changes, that cylinder was already dead.

The logic is sound. The procedure is not, and it should not be done on any engine with a control module — which, on anything with a lead in this sense, means an engine with an ignition module at minimum.

The original does note that "some believe" this damages the ECU. It is not a belief. Pull a lead off and the coil's secondary circuit is open: the energy has to go somewhere, so the voltage climbs far above what it would ever reach across a plug gap, and it discharges through whatever gives way first — the coil's own internal insulation, the boot, or backwards through the driver stage that switches the coil. That is how a diagnostic test turns into a coil, or a module, or both. There is also several tens of kilovolts looking for a path to earth while your hand is on it, and the unburnt fuel from the disabled cylinder goes down the exhaust into the catalytic converter, which is where a misfire does its most expensive damage anyway.

This is the second time this archive has taught an ignition test that can destroy the ignition module — the other was checking for spark by holding a lead near the block. The underlying physics is identical each time: never leave the secondary circuit with nowhere to go.

The modern equivalent is free and safe. Ask the scan tool for a cylinder balance or power balance test, or read misfire counts per cylinder, which most modules maintain continuously. That is the same information, with the module cutting the injector rather than you opening the spark path.

The tests that do work

Look at them in the dark, with the engine running. This is the original's best advice and it costs nothing. Tracking shows as tiny blue-white threads jumping from the lead to the nearest metal, and once you have seen one you cannot unsee it. Do it dry first.

Then mist them with water. A hand spray bottle, a fine mist, engine idling, lights off. This is the wet-morning fault reproduced on demand, and it is the single most useful thing in this article. Two cautions the original does not give: do it on a cool engine, and keep water out of the plug wells — a spark plug sitting in a puddle at the bottom of a deep well will misfire on its own and send you after the wrong part.

Feel and look at the boots. Most leads fail at the spark plug end, because that is the hottest part and the most flexed. Pull one — engine off — and look inside the boot for a grey or black track running down the inside, and at the plug's ceramic insulator for a matching scorch mark. That mark is a photograph of where the spark has been going.

Measure the resistance, knowing what it is worth. It confirms the core is intact and not much else, since the failure mode above happens on the outside of the insulation and does not change the core reading at all. Still, an open core is a real failure and this finds it in seconds.

What the resistance figures actually mean

The original says a lead runs about 16 to 18 kΩ per metre, and that a lead should not exceed about 5 kΩ. Those numbers are internally consistent and they are in the right territory for one type of lead. But it then explains them wrongly, and the explanation is worth correcting because it is backwards:

"This resistance causes the current to flow in the lead, and if it is less than this amount the current is not transmitted well."

Resistance never helps current flow. The resistance is there to suppress radio-frequency interference, nothing else — a spark is a broadband radio transmitter, and without suppression it puts noise into the radio, the sensors and the module. The resistance is a deliberate cost, accepted to keep the rest of the car working.

Which means there is no single correct figure, and a low reading is not a fault:

• Carbon-core leads — a graphite-impregnated fibre core — suppress by resistance, and run roughly 10–40 kΩ per metre depending on the maker.
• Spiral-wound leads suppress with the magnetic field of the winding instead, and can be under 2 kΩ per metre while being entirely correct and generally better.
• Solid metal core has almost no resistance and no suppression at all. The original is right that these must not be fitted to a car with an engine module, and it is the one place in the piece where the reasoning is exactly right.

So do not test a lead against a remembered number. Test it against the figure in the service data, and against the other leads in the same set — one lead reading three times its neighbours is the finding, whatever the absolute value is. See using a multimeter for why the meter's own leads matter when you are measuring anything low, and for the wiggle test, which on an ignition lead will find a core that is broken but still touching.

Routing, and the fault that is nobody's part

Two leads running parallel and touching can induce a spark in the wrong one, because a lead carrying a rising 30 kV pulse is a transformer primary looking for a secondary. If the two cylinders happen to fire consecutively, the induced pulse arrives at roughly the right moment to light a cylinder on its exhaust or intake stroke, and the result is a misfire that moves around and defeats everything.

That is why leads come with separators and why the routing in the workshop manual is a specification rather than tidiness. After any work in that area, put the leads back in their clips. Also keep them off the exhaust manifold and the head — the insulation is heat-resistant, not heat-proof, and a lead that has been resting on hot metal is the classic source of a track.

Coil-on-plug did not remove this fault, it moved it

The original opens by pointing out that modern ignition systems delete the lead and sit the coil directly on the plug. True, and it is a real improvement — no lead, no voltage lost on the way, no radiated noise to suppress.

But the boot is still there: a rubber sleeve with a spring inside it, carrying the same tens of kilovolts down a deep well that fills with oil when the cam cover gasket weeps. It carbon-tracks in exactly the same way, produces exactly the same load-and-weather misfire, and is a cheap part that people replace an entire coil to fix. Look inside it for the same grey track.

So a reader with coil-on-plug should treat this whole article as being about their boots. The relevant codes are P0351 through P0354, and the same rule applies: a coil code names the circuit, not the culprit. And the commonest culprit is not the coil at all. A worn plug's gap widens, the coil is asked for more voltage every cycle to jump it, and it eventually fails from the overwork — so a coil fitted over a tired plug does not last. Spark plugs has that argument in full.

Two numbers to correct, and the practical rules that survive

"Replace them every year." No. A decent set lasts years and tens of thousands of kilometres, and most manufacturers specify replacement somewhere around 100,000 km or not at all. Replace them when they test bad, when they track, or when they are off with the plugs anyway — annual replacement is an invented interval.

"About 12,000 volts." Low, and it matters because the whole failure mode above is about voltage finding an easier path. A modern coil will deliver 25,000 to 40,000 volts and more if the gap demands it. The insulation is holding back rather more than the original suggests.

What the original gets right and is worth repeating:

• Replace the whole set. They aged together and a new lead beside four tired ones just relocates the weakest link.
• Never swap a suspect lead with its neighbour to test it. Firing order is not negotiable and you will have two problems. Note that this is the exact opposite of the advice for a coil: swapping a coil to the adjacent cylinder and clearing the code is the standard first move, and the corpus recommends it at P0301 and P0351. The difference is that a lead decides which plug fires and a coil does not.
• Pull by the boot, never by the cable. Twist the boot to break the seal first. Pulling on the cable separates the core from the terminal inside the boot, and produces a brand-new fault that looks exactly like the one you were chasing.
• Do them one at a time so nothing can end up in the wrong place.

And if you have the plugs out while you are there, spark plugs: heat range, iridium and platinum covers what to put back and, more usefully, what not to.

Video guides

Video, for the parts of this that are easier watched than read.