The engine that fires fine at idle and falls on its face at redline. Three separate races the coil runs against RPM - and a disciplined sweep tells you which one it is losing.
Spark is not a constant-capability event. As engine speed climbs, the ignition system is asked to do more work in less time against harder conditions. Every firing event is a race against the clock, and a coil can lose that race in three independent ways.
Race one is the dwell window: the time available to charge the coil shrinks as RPM rises. Race two is demand versus delivery: the voltage required to jump the plug gap climbs with RPM while the coil's output may not keep up. Race three is insulation under stress: the secondary-side insulation that holds at idle voltage can flash over when demanded voltage climbs. An engine that runs clean at idle and breaks up at speed is losing one of these races - and each one condemns a different thing.
The dwell-window math is pure geometry of the architecture, and the three common designs lose race one at very different speeds.
The underlying physics is current-limited, not mysterious: the primary winding is a resistance and an inductance, and current takes time to build toward saturation. At idle the window is long and even a marginal coil saturates fully. At redline the window can shrink to a tenth of its idle length, and a coil that charges fine with time cannot be given time that does not exist. Strategy matters too - constant-dwell strategies fix the window and let current vary, while current-limiting drivers vary the window to hold current - but neither repeals the clock.
It is not the rev limiter. The most common impostor in this symptom space is strategy: a hard, repeatable cut at an exact RPM is usually the limiter doing its job - fuel cut or spark cut at a programmed ceiling. Spark at the cut point is healthy; it simply gets switched off. Breakdown is different in character: it starts soft, as occasional misses, gets worse as RPM climbs, and does not land on exactly the same RPM every time under different loads. Healthy spark measured at the cut point clears the ignition and points at strategy. Do not diagnose the limiter.
It is not automatically a "weak coil." A coil that fires correctly at idle is not weak in any general sense - it is losing a specific race at a specific RPM state. Condemning the coil without knowing which race it lost is how good coils get replaced while the real fault (a resistive feed starving the charge window, a worn gap raising demand, a tracked boot stealing the discharge) stays on the car.
It is not fuel starvation wearing an ignition costume. A lean condition at high RPM also breaks up under load, and the symptom overlap is real. The separation is measurement: ignition breakdown shows its signature in the secondary - firing voltage, burn time, pattern - while the ignition is being given correct command, feed, and trigger. If the secondary looks right and the engine still falls over, the investigation leaves the ignition.
The three races are three failure modes, and they fail in distinguishable ways:
The races can stack. A coil slightly short on charge time feeding a gap slightly wide through a boot slightly tracked can pass every isolated test and still break up at speed. When the symptom is worse than any single measurement explains, look for the stack.
The SpeedNeeds approach to high-RPM spark breakdown is a scoped RPM sweep that proves where the failure lives before any part is condemned. Reproduce the symptom at a defined RPM state - loaded, because unloaded RPM does not reproduce demand-driven failure - then isolate the race: verify the coil is being given its charge time and reaching saturation (dwell and current ramp against RPM, race one), verify demand is being met (secondary firing voltage against RPM, race two), verify the discharge is reaching the plug and not flashing over (pattern integrity, race three), and verify the trigger and the feed are clean at speed so a good coil is never blamed for a bad signal or a starved supply.
The verdict rides on measured output at the failing RPM state - what the coil actually delivers when the symptom is active - never on a resistance number and never on an idle-only test. A coil that delivers correctly at the failing RPM state is cleared across the tested range. A coil proved short on charge, short on output, or leaking its discharge at that state is a verdict, not a guess - and the sweep says which race it lost, so the repair targets the fault instead of the symptom.
Free-rev is a weak reproduction. An unloaded engine reaches RPM with low cylinder pressure, which masks demand-driven failure (race two) and can understate the whole symptom. Where it can be done safely, the sweep should include loaded operation - a drive pull or a controlled brake-torque - because the failing state is RPM plus load, not RPM alone.
Heat stacks with RPM. A coil that is marginal on charge time at high RPM can pass a cold high-RPM sweep and fail the same sweep hot, because primary resistance rises with temperature and slows the current buildup further. If the symptom only appears on a hot engine at speed, isolate the two variables - reproduce at temperature - rather than letting RPM take the blame for a thermal problem or vice versa.
Smart coils change the dwell math. Current-limiting drivers adjust the charge window to hold primary current, which means commanded dwell and actual charge time are not the same thing. The sweep reads the current ramp, not the command - the command can be perfect while the charge is short.
Misfire monitors can muddy the water. Some strategies pull timing, cut fuel, or set codes in response to detected misfire, and those reactions can look like ignition breakdown on a scan tool. Prove the ignition's own behavior on the scope first, then read the monitor's opinion - never the reverse. And near the rev limit, soft-cut strategies that drop cylinders approaching the ceiling can mimic breakdown just below the cutoff; confirm the ceiling before diagnosing the approach.
Idle is the easiest condition the ignition ever faces. Speed is the hardest. When spark dies with RPM, the coil is losing a race against the clock, against rising demand, or against its own insulation - and a scoped, loaded, at-temperature sweep names the race before any part is condemned.
Stop guessing. Prove the next move. Measure at the failing RPM state, and let the measurement - not the parts cannon - pick the loser.
Stay in the loop
New QTS builds, tools, and garage stories — straight to your inbox. No spam. Unsubscribe anytime.
Read the Privacy Policy.