Ignition Coil Supplier: Ninety Per Cent, Four Times Over

Alan XiaoFounder, Sourcing Pioneer

· Updated October 22, 2026 · 12 min read

The spark plug article in this series ended with a number that can only be produced on an engine. The coil is the opposite kind of part: everything about it can be measured on a bench, and the standard spends its length on the bench. What makes it interesting is the durability schedule — two hundred hours at seventy degrees, with an open circuit inserted four times, and a final pass mark of ninety per cent of what the coil did before.

An open circuit is the worst thing you can do to an ignition coil. With no gap to jump, the secondary voltage has nowhere to go and keeps rising until something gives: insulation, potting, or a winding. The test does it on purpose, four times, and then asks whether the coil still makes at least ninety per cent of its original output. A supplier who has never run that schedule is selling a coil that has been measured once, cold, on a good day.

1. Four Parameters, Two of Them in Time

85.2122(9) lists four emission-critical parameters for what it calls an inductive system coil: open circuit voltage output, dielectric strength, flashover and rise time. It then defines three of them, and the definitions are the useful part.

A coil is a device used to provide high voltage in an inductive ignition system. Flashover is the discharge of ignition voltage across the coil. Dielectric strength is the ability of the material of the coil to resist electrical breakdown. And rise time is the time required for the spark voltage to increase from 10% to 90% of its maximum value.

That last definition is the one buyers underestimate. Two coils can reach the same peak voltage and behave completely differently: one gets there in microseconds and the other labours. The plug fires when the voltage across the gap reaches breakdown, not when the coil reaches its maximum, so a slow rise is energy spent making heat in the coil instead of spark at the electrode.

It also explains why the parameters are paired the way they are. Voltage output is what the coil can deliver; rise time is how quickly it delivers it; flashover and dielectric strength describe whether it survives the attempt. Four numbers, two of which are measured in time and two of which are measured by whether something breaks.

2. The Circuit on the Bench

The procedure is Appendix I, paragraph I, and it begins with a circuit rather than a sentence: the coil is set up as Figure 1, triggered by an applicable distributor or equivalent, through an applicable primary resistor, into a 50 picofarad load, at 14.0 ± 0.50 volts DC, stabilised at 20 °C ± 3 °C.

Every one of those values matters to comparability. The load capacitance changes the shape of the secondary pulse; the input voltage changes the stored energy in the primary; the ambient temperature changes resistance in both windings. A bench report that omits any of the three is a number from a different experiment — and a bench is only as good as the sample it is handed, which is why a factory audit should pull the sample itself., which is why the standard fixes them before it fixes anything else.

Then the test becomes a sweep.

3. Ninety and a Hundred and Ten, at Every Speed

With the original equipment coil installed, the tester records the “predominant minimum peak voltage” and the rise time at 300 and 500 distributor rpm, and then at 500 rpm intervals up to the maximum intended operating speed, taking each measurement after four minutes of running at that speed. The replacement coil then repeats the identical sweep.

The criteria are two inequalities applied at each test speed:

• open-circuit output voltage at least 90 per cent of the original equipment coil's;
• rise time no more than 110 per cent of the original's.

Read “at each speed” as the whole point of the paragraph. A coil can be strong at idle and weak at high rpm, because the time available to build the field shrinks as the distributor turns faster. A supplier who reports one figure — usually the idling one — has reported the easiest point on the curve. The standard asks for the worst point, at every step, and both inequalities have to hold there.

4. Two Hundred Hours, Four Open Circuits

The first test asks whether the coil is as good as the original when new. The second asks whether it is still good after a life, and it is the most specific schedule in this whole series.

The coil is installed in the ignition system with the applicable rotor, cap, capacitor, breaker points and primary resistor, then run from a regulated supply at 14.0 ± 0.5 volts DC in an ambient of 70 °C (158 °F) at 1750 ± 50 distributor rpm for 200 hours.

Every 50 hours, the distributor is run for five minutes with one open-circuit spark gap instead of a 12 kV gap. Four shocks, at the four quarter marks of the run. Through all of it, the coil must show no failure that would lose the spark across the 12 kV gap — and at the end, the open-circuit output voltage must be at least 90 per cent of what the same coil measured before the 200 hours began.

That last clause is worth reading twice. The comparison is not against the original equipment coil; it is against the replacement coil's own starting figure. A coil that begins 30 per cent above the original and loses a third of its output still passes the first test and fails this one. The regulation is asking whether the part degrades, not just whether it starts ahead.

5. Why an Open Gap Is the Hard Part

The standard does not explain its choice of shock, so it is worth stating plainly what the schedule does. In normal service the secondary circuit has a gap to jump: voltage rises until the gap breaks down, and then the energy leaves. Remove the gap — a disconnected plug, a broken lead, a boot that has come loose — and there is nowhere for the energy to go before the voltage has risen far past the level the insulation was designed for.

That condition is not rare in the field. It is what happens every time somebody pulls a plug lead with the engine running, or fits a plug with the wrong gap, or leaves a connector loose. The standard takes the most common way a coil dies in service and makes it a scheduled event, four times, under thermal load. A coil that survives it has been tested against the failure it will actually meet.

For a buyer, that reframes what a sample is for. A cold bench run at room temperature says almost nothing about the part that fails at 70 °C after a hundred hours. The number to ask for is not the peak voltage; it is the retained percentage after the schedule.

6. Where the Price Difference Comes From

Two coils with the same connector and the same mounting holes differ in five places, and the last one is the reason the durability clause exists.

The winding. Copper against a cheaper conductor is a resistance difference, and resistance is heat. At 70 °C for 200 hours, a winding with less margin runs hotter and ages faster than the number on the box suggests.

The insulation and the potting. Epoxy that fills the case without voids resists breakdown; a void is a place where the voltage concentrates. This is the property the flashover and dielectric-strength parameters exist to describe, and it is invisible from outside.

The core. Laminated steel and its geometry set how quickly the field builds, which is what rise time measures. A cheaper core shifts the whole curve rather than a single point.

The boot and the connector. The interface is where moisture enters and where a poor contact raises the resistance the coil has to drive. It is also the cheapest part to get wrong, and the first to suffer in transit: a connector that takes on moisture before fitting is a resistance problem the bench never sees. The packaging checklist covers what the carton has to do.

The test itself. Two hundred hours of rig time, four open-circuit shocks and a repeat of the full speed sweep is an expensive way to qualify a part. That cost is why so few coils arrive with a schedule behind them, and why asking for one separates a factory that builds coils from one that assembles them.

7. The Word Belongs to a Switch

Method unchanged. The federal recall file holds 245,845 records across 15,305 campaigns, no header row, every field scanned, twenty-five terms each with its own pattern. Controls: brake 28,639, air bag 16,162.

Only two terms returned nothing, and both are parts of the coil rather than symptoms of it: winding and flashover. That second one is a parameter from 85.2122(9) — the standard names it, the field never uses it.

The rest of the counts are ordinary. coil appears in 217 records and 41 campaigns, coils in 52 and 6. epoxy 364, potting 352, cracked 1,155, resistance 1,160, dielectric 152. ignition looks enormous at 9,308 records and 879 campaigns, until the headings are read.

Of the 316 campaigns under the plain ignition heading, 312 are switches. The three largest campaigns that contain the word ignition are a Chevrolet switch recall covering 5,877,718 vehicles, a Cadillac one covering 3,141,731 and a Saturn one covering 2,191,014 — none of them a coil. The switch outnumbers the coil by more than thirty times under the same word.

So the coil is not absent from the record, the way the air cleaner is. It is crowded out. Anyone searching the file for “ignition” finds a switch; anyone narrowing to the coil heading finds nine campaigns, and anyone looking for the mechanics of failure finds winding at zero.

8. Seven Questions an Ignition Coil Quote Should Answer

A quotation in this category is a photograph, a cross-reference list and a price. Everything below can be answered from a bench report and a drawing, and the first four are the ones the appendix already defines.

1. What did the speed sweep show, speed by speed? 300 rpm, 500 rpm, then every 500 up to the maximum intended speed, four minutes at each. A single figure is the easiest point on the curve, not the curve.
2. Output and rise time at the worst speed. At least 90 per cent of the original's voltage and no more than 110 per cent of its rise time, at every point tested.
3. Was the 200-hour schedule run? 70 °C, 1750 rpm, four open-circuit shocks at the fifty-hour marks, and the retained output at the end.
4. What survived the open gap? No loss of spark across the 12 kV gap during the run, and the output percentage afterwards compared with the same coil beforehand.
5. Winding and insulation. Conductor, insulation class, and how the potting is filled — voids are where breakdown starts.
6. Which speeds were used for the rise-time figure? The definition runs from 10 per cent to 90 per cent of maximum; a figure without a speed is a figure without a test.
7. Two prices. One as quoted, one with the five items below written into it.

9. Six Lines to Put in the Order

The first four lines are the appendix, translated into an order. The fifth is the drawing, and the sixth is the one that decides whether the first five mean anything.

1. The sweep, by speed, from 300 rpm to the maximum intended operating speed.
2. Output at least 90 per cent of the original equipment coil at each speed.
3. Rise time no more than 110 per cent of the original at each speed.
4. The 200-hour result, with the retained percentage and the open-gap intervals stated.
5. Winding and potting, by material and by process.
6. Two prices: as quoted, and with the five items above written into it.

An ignition coil is certified on a curve, not on a peak. The parameters are measured at a stated voltage and temperature, across a dwell sweep, with an open-circuit event every fifty hours for two hundred hours.

A single spark voltage printed on a carton says nothing about that. The two figures that decide survival are the rise time and the behaviour after the open circuit, and neither is visible in a photograph of the part.

Frequently Asked Questions