Fuel trims describe how the controller adjusts fueling relative to its base calculation. Positive trim generally adds fuel; negative trim generally removes it. The pattern across operating conditions is more useful than one isolated number.
Short-term and long-term are different views
Short-term fuel trim is the immediate correction in the current operating condition. Long-term fuel trim is learned compensation, commonly organized into operating regions. A value shown at idle may not represent the learned correction used under load.
Check that the engine is in the control mode relevant to the test. Cold operation, enrichment, deceleration fuel cut and other special modes can make a casual trim comparison misleading.
Three comparisons that give direction
Save the codes and freeze frame before clearing anything. For a conventional warm-engine comparison, confirm closed-loop operation and record coolant temperature, RPM, load, STFT and LTFT for each supported bank. Keep the operating conditions beside the numbers. A bank or PID that the vehicle does not support is missing information, not a zero reading.
| Observation | What it suggests | Useful next step |
|---|---|---|
| Positive correction is strongest at idle and falls as airflow increases | Unmetered air is one possibility | Inspect intake and PCV plumbing, then use the vehicle's leak-test procedure if needed |
| Positive correction persists or grows under load | Fuel delivery or airflow measurement deserves investigation | Compare the relevant pressure and airflow data with vehicle specifications; do not diagnose this from free-revving alone |
| One bank differs markedly from the other under the same conditions | A bank-specific cause may be present | Confirm the bank layout, then inspect that bank's intake, exhaust and combustion evidence |
| Both banks change similarly | A shared input or condition may be involved | Check common air, fuel and control inputs before treating the banks as separate failures |
These patterns narrow the investigation; none identifies a failed part by itself. Use only the operating conditions and test procedures that are safe for the vehicle. Do not watch a scanner while driving. Log data or have a passenger operate it, and stop rather than reproduce severe misfire, overheating or another unsafe condition.
For a parked comparison with a return-to-idle check and worked data, use fuel trims at idle versus 2,500 RPM. A free-revving check is not a loaded fuel-delivery test.
Before making a controlled change, write down the response you expect. Afterward, compare the same PIDs under the same conditions. If the result does not repeat, record it as inconclusive.
Do not turn rules of thumb into specifications
There is no single universal trim percentage that identifies a defective part on every engine. Use the vehicle's monitor criteria and diagnostic limits. Interpretation can also depend on fuel composition and the meaning of the PID being displayed.
Compare STFT and LTFT from the same bank and observation period; do not mix banks. Adding STFT and LTFT is sometimes used as a rough screening aid. It is not necessarily the controller's exact final fueling calculation. Treat it as context, not a precision measurement or universal pass/fail test.
A worked example: why STFT alone can mislead
Suppose a warm, closed-loop engine displays LTFT of +18% and STFT of −3% on one bank. These are invented teaching values, not a vehicle specification or a customer repair. The small negative short-term number does not mean the controller has stopped compensating for a lean tendency: the displayed long-term correction is still positive. Adding the two gives +15% as a rough screening figure, subject to the calculation limits above.
Record both values and the operating condition. If a later comparison uses a different learned operating region, it is not a simple before-and-after measurement of the same correction. Use the pattern to choose a test, then verify against the vehicle's diagnostic information.
Check the inputs behind the correction
Air measurement, fuel pressure, exhaust leaks, misfire and sensor bias can all affect the pattern. An oxygen sensor reporting lean is not proof that the sensor caused the lean condition.
Wideband air/fuel sensors do not follow narrowband voltage rules. Confirm PID definitions and scaling. Use a safe controlled response test and circuit checks when the evidence points toward the sensor.
For the common paired lean-code complaint, see P0171 and P0174.
References & scope
These background references explain selected concepts. The diagnostic sequence is an original, general educational guide; it is not a reproduced factory procedure. Confirm specifications, test limits, cylinder layout and circuit details in the service information for the actual vehicle.
- Innova: fuel-trim data and bank-to-bank comparisons
- Ford: 2024–2025 gasoline OBD operation summary, Fuel System Monitor (pages 22–24), illustrating application-specific fuel-control conditions and learned correction cells
- California Air Resources Board: what the OBD system monitors and stores
This is general education, not a vehicle-specific repair procedure or hands-on test. Read the editorial policy.