Short-Term Fuel Trim: Read the Pattern Before Replacing Parts

Short-term fuel trim (STFT) is the engine computer’s immediate fuel correction. A positive number means the computer is adding fuel; a negative number means it is removing fuel. Zero percent means no correction is currently displayed—not that every component is necessarily working correctly.

Do not diagnose from one STFT snapshot. Read it with long-term fuel trim (LTFT), closed- or open-loop status, RPM, load and bank number. The pattern is more useful than the isolated percentage.

What STFT actually measures

In closed-loop operation, the powertrain control module uses upstream oxygen or air/fuel-ratio sensor feedback to adjust injector pulse width. Ford’s OBD system description explains that short-term correction continuously adds or subtracts fuel around the commanded mixture. Scan tools normally display the correction as a percentage, with no correction shown as 0%, although the internal strategy uses a multiplier of 1.0 (Ford OBD System Operation Summary).

That makes the sign convention straightforward:

• Positive STFT: The computer is lengthening injector on-time to add fuel.
• Negative STFT: The computer is shortening injector on-time to remove fuel.
• STFT moving above and below zero: Often normal once a gasoline engine is warm and in closed loop. The control system is supposed to make rapid corrections.

STFT is a correction, not a direct measurement of air/fuel ratio and not proof that an oxygen sensor is defective. The sensor may be accurately reporting a mixture problem caused elsewhere. See upstream versus downstream oxygen-sensor roles before treating rear-sensor data as primary fuel-control data.

What is a normal short-term fuel trim?

On a warm, stable engine in closed loop, STFT should generally work around zero rather than remain heavily positive or negative. There is no universal percentage that condemns a part across every make, engine and operating cell.

OEM thresholds illustrate why one internet specification is unreliable. A Ford OBD document gives a typical lean-monitor threshold of LTFT above 25% together with STFT above 1%. A Mitsubishi diagnostic article for its covered models identifies an LTFT value of 12.5% as the lean-code setting point. These are calibration-specific criteria, not universal limits (Ford; Mitsubishi). Use service information for the exact year, model and engine when a pass/fail threshold is required.

For diagnosis, compare total fuel correction for the same bank:

Bank 1 total correction ≈ STFT Bank 1 + LTFT Bank 1

Ford uses this addition in a lean-code service procedure—for example, +13% LTFT and +23% STFT equals +36% total correction. Do not add Bank 1 to Bank 2 or mix STFT from one bank with LTFT from the other (Ford TSB 04-17-4).

Run a useful fuel-trim test

Use this sequence on a gasoline engine. A diesel may label and calculate cylinder or fuel corrections differently, so follow its OEM data definitions.

1. Scan all modules and save the evidence. Record stored and pending codes, freeze-frame data and current STFT/LTFT values before clearing anything. A reset can erase learned trim data that may identify the operating condition where the fault occurred.
2. Warm the engine and confirm closed loop. STFT during cold-start enrichment, deceleration fuel cut or another open-loop mode is not a valid baseline. Turn off major accessories and stabilize the engine in Park or Neutral.
3. Record each bank at idle. Capture STFT, LTFT, RPM, coolant temperature, mass airflow or manifold pressure, upstream sensor data and purge command if available.
4. Repeat at approximately 2,500 rpm with no load. Keep clear of moving and hot components. Ford’s applicable bulletin limits each 2,500-rpm hold to 20 seconds, while Mitsubishi’s procedure for its covered vehicles uses two minutes. Follow the exact service procedure for the vehicle.
5. Compare idle with higher RPM, then compare banks. If the symptom occurs only under road load, record a safe road-test graph or have a second person operate the scan tool. A no-load rev test cannot prove fuel delivery is adequate at high load.

Interpret the pattern

Observed pattern	What it supports	Next test
Total correction strongly positive at idle but falls toward normal at 2,500 rpm	Unmetered air or another vacuum-side leak	Inspect PCV and vacuum hoses, isolate branches where permitted, then smoke-test the intake
Positive correction remains high at idle and higher airflow	Airflow measurement error, inadequate fuel delivery, fuel quality or another fault affecting both conditions	Compare MAF/MAP data with OEM references and measure commanded versus actual fuel pressure
Only one bank is substantially positive	Bank-specific intake or exhaust leak, injector problem, or biased upstream sensor	Smoke-test that bank’s intake path, inspect for an exhaust leak ahead of its sensor, and perform injector and sensor tests
Both banks are similarly positive	A shared air-metering, intake, purge or fuel-supply problem	Test shared systems before replacing bank-specific parts
Correction is persistently negative	Excess fuel, incorrect airflow reporting, purge vapor, leaking injector, excessive fuel pressure or fuel-contaminated oil on an applicable engine	Inspect oil level and odor, test purge sealing and fuel pressure, then check injector leakage

The idle-versus-RPM comparison has OEM support. Ford notes that a fixed unmetered-air leak is a larger percentage of total airflow at idle. In the applications covered by its bulletin, a drop in total correction of more than 15 percentage points from idle to 2,500 rpm makes a vacuum leak likely. That threshold belongs to those applications, but the underlying pattern is broadly useful (Ford TSB 04-17-4). Mitsubishi similarly directs technicians to identify which bank is affected and whether the lean condition occurs at idle, 2,500 rpm or both (Mitsubishi Tech Talk).

A change during isolation is stronger evidence than the original trim number. On specified GM 3.6-liter applications, for example, GM instructs technicians to plug the foul-air-tube openings and watch whether positive trims move to zero or negative before replacing the tube. If they do not, the procedure moves to a purge-valve test (GM PIP5549A). The principle is to alter one suspected path safely and verify a repeatable response.

For a persistent negative trim, avoid assuming “bad injector.” GM documents certain 2.0-liter turbo applications where frequent short trips can dilute the oil with fuel and produce LTFT from approximately −18% to −30%. Its procedure checks the oil rather than automatically replacing the high-pressure pump or injectors (GM Bulletin 22-NA-056). Applicability matters.

Stop when the test does not support the part

Do not replace an oxygen sensor, MAF sensor, fuel pump or injector because STFT is merely nonzero. First prove that the correction is abnormal under a repeatable condition, determine whether it affects one bank or both, and run the matching air, fuel-pressure, purge, exhaust-leak or injector test. If pressure is the next branch, use the vehicle’s specified connection method and threshold rather than guessing from trim alone; the fuel-pressure-regulator test sequence explains that distinction.

After a confirmed repair, reproduce the original coolant temperature, RPM and load. Verify that total correction improves on the affected bank and that the code does not return. That confirmation—not a parts replacement by itself—is the end of the diagnosis.