The useful answer
An engine that idles smoothly but breaks up under acceleration needs to be diagnosed under the conditions that expose the fault. A clean idle is useful evidence. It is not proof that the ignition, fuel supply or mechanical condition will remain adequate when the engine is doing more work.
Start with two questions: Which cylinders are affected, and what changes as the symptom begins? Those answers help choose the next test. “It happens uphill” does not identify a failed coil, injector or fuel pump.
This guide covers gasoline spark-ignition engines. Diesel misfires require a different diagnostic approach.
Stop a damaging test before collecting more data
If the check-engine light flashes, reduce load and stop safely. Do not repeat hard acceleration to make the fault easier to record. Severe shaking, major power loss, overheating or a raw-fuel smell are reasons to end the test and arrange professional assessment.
Do not watch a scanner while driving, hold the vehicle against its brakes to create load, or disconnect ignition components on a running engine. A technician can decide whether controlled loaded testing is appropriate and how to perform it safely.
1. Describe the event before clearing anything
Save stored and pending codes and the available freeze-frame data. Record whether the engine was cold or warm, whether the symptom followed recent work, and whether it appeared during a hill climb, steady cruise or the first movement of the accelerator.
“At 2,000 RPM” is incomplete. The engine can reach that speed while parked with little work to do or while pulling the vehicle uphill. Those conditions are not interchangeable. Revving it in the driveway without a misfire does not reproduce the hill-climb complaint.
For an existing recording, identify the moment the symptom starts and examine a small set of useful channels:
- RPM, load and coolant temperature;
- cylinder-specific misfire information, where supported;
- fuel-system operating state and relevant mixture data;
- requested and measured fuel pressure, if available for the system;
- requested and measured boost when relevant.
Record unsupported data as unavailable. A scanner that does not display a parameter has not measured zero. Also record fuel-control mode: ordinary closed-loop trim interpretation may not apply during enrichment or fuel cut, and a fixed trim display during that interval does not establish correct mixture.
2. Confirm what the misfire data actually shows
A repeatable concentration on one cylinder supports investigating that cylinder’s spark, fuel delivery, circuit and mechanical condition. Activity across several cylinders makes shared causes worth considering, but it does not prove that one shared component has failed.
Make sure a displayed counter updates during the event. Stored monitor results and live counters are not interchangeable. Ford’s 2024–2025 gasoline OBD summary documents operating conditions that can suspend misfire monitoring, including certain rapid transients and rough-road conditions. That is an application-specific example of why an unchanged count cannot automatically rule out a misfire. Ford OBD operation summary, Misfire Monitor.
If the vehicle shudders but the available data does not establish combustion trouble, keep transmission and driveline causes in the investigation. Missing evidence is a reason to improve the diagnosis, not buy an engine part.
3. Check ignition suitability and condition
Spark-plug gap affects the voltage needed to fire the plug. Higher cylinder pressure also makes ignition more demanding. An ignition problem may therefore become apparent under load even when the engine runs smoothly at idle. NGK explains the relationship between gap, pressure and misfire in its gapping guidance.
Check the correct plug application and specified gap before assuming that a recently installed plug is suitable. Do not close the gap by guesswork to conceal a symptom. Modified engines need guidance appropriate to their actual configuration.
With the engine off and using the vehicle’s service procedure, inspection can reveal damaged insulation, carbon tracking, oil or water around the plug, or connector problems. DENSO identifies these as checks to make before condemning an ignition coil. DENSO ignition troubleshooting.
Where the design permits a controlled coil swap, change one variable and document both positions. If the fault reproducibly follows the moved assembly under comparable conditions, that is stronger evidence against that assembly. If the fault does not recur at all, the result is inconclusive. If it remains on the original cylinder, continue investigating rather than declaring every ignition component good.
Already replaced plugs and coils? Follow the deeper P0301 after plugs and coil guide.
4. Ask whether fuel delivery keeps up with demand
A pressure reading taken at idle answers an idle question. When fuel supply is suspected, the useful comparison is what the system requests and delivers as the complaint occurs, interpreted against the vehicle’s service information.
Gasoline direct-injection systems have both low- and high-pressure circuits. A high-side discrepancy alone does not identify which part of that supply chain is responsible. Bosch GDI system overview.
Before comparing pressure channels, confirm that requested and measured values refer to the same circuit, use compatible units and belong to the same part of the event. Low-side supply pressure and high-side rail pressure are different measurements. Treat a verified departure as the start of a branch: distinguish supply, electrical feed, control and mechanical delivery problems. Do not replace the high-pressure pump solely because a graph falls below its target. Check the measured-pressure signal for plausibility using the application procedure; the scanner display is evidence from a sensor and its circuit, not an independent pressure measurement. A slow-refreshing display can also miss a short event. Check the saved recording and the tool's supported update behaviour before interpreting a brief graph feature as a physical pressure loss. Snap-on explains that a smaller custom PID list can improve refresh rate; this does not guarantee that every scanner will capture the event.
Keep this stage to supported data interpretation unless you have the required training and equipment. Do not loosen a fuel fitting to check delivery; use the application’s pressure-release and test procedures.
5. Choose a test that separates the remaining causes
The following is a diagnostic decision aid, not a component verdict or a substitute for the vehicle’s test procedure.
| Evidence during the complaint | Useful next direction | What remains unproven |
|---|---|---|
| One cylinder repeatedly stands out | Controlled ignition comparison, followed by cylinder-specific fuel and mechanical testing as indicated | Which system caused that cylinder to lose contribution |
| Several cylinders misfire with a verified fuel-pressure departure | Investigate fuel supply and control under the relevant demand | Whether the pump, its supply, its control or another cause produced the departure |
| Symptom began immediately after plug or coil work | Verify application, installation and disturbed connections | That new parts are either good or bad merely because they are new |
| Boost performance is also abnormal | Compare the air-path and control evidence using the P0299 guide | That the turbocharger itself has failed |
| Available tests do not explain a repeatable cylinder fault | Select application-appropriate injector and mechanical tests | That a smooth idle has ruled out either branch |
Do not order every test at once. Write down what the next test is intended to separate and what each possible result would change.
6. Verify the repair against the original complaint
“Runs smoothly in the bay” is not enough when the original fault occurred under load. After repair, a technician should confirm operation under an appropriate, safely controlled version of the original conditions and review the relevant fault and monitor data.
If those conditions were not reached, record verification as incomplete. Clearing a code removes evidence; it does not demonstrate that the cause is gone.
Use the diagnostic worksheet to record the original condition, the evidence supporting the repair, and what happened afterward.
Common questions
Does a misfire under load always mean a bad coil?
No. Ignition deserves attention, but the symptom alone does not distinguish spark, fuel delivery, mechanical condition or even a non-combustion shudder.
Can new spark plugs still be involved?
Yes. Correct application, gap and installation still need verification. Replacement history is useful context, not a test result.
What if the car is smooth after I restart it?
Record that change. A restart is another operating condition; it does not establish a repair. Preserve the available codes and data before another attempt changes the evidence.
What should I bring to the shop?
Bring the complete code report, saved freeze frame, a precise description of when the symptom occurs, and a list of parts or work already performed. Ask what evidence supports the proposed repair and how the original complaint will be verified afterward.
This is general education, not a vehicle-specific repair procedure or hands-on test. Read the editorial policy.