BMW F30 N20 P0172 rich fuel trim rough idle gasoline smell in oil: a test plan

A BMW F3X with an N20, P0172 or 118002, a rough idle, and oil that smells like gasoline needs a measured diagnosis, not a parts list. The odor raises the stakes because fuel may be diluting the oil, but it does not tell me whether the source is an injector, the high-pressure fuel pump, repeated cold running, or another rich-running fault.

A real 2012 F30 owner described this exact cluster: P0172, a cylinder 2 misfire, hesitation, occasional stalling, a fuel-fouled cylinder, and a noticeable gasoline smell in the oil. Injectors, coils, plugs, MAF work, and a smoke test had not solved it. Another long F30 forum thread includes several N20 owners who fixed similar symptoms with a high-pressure pump, along with others for whom that replacement did not work. That split is the reason I test before buying.

Chasing P0172 and P0302 (Fuel Rich Bank 1, Cylinder 2 Misfire) in 328i

P0172 is the generic rich-mixture code for bank 1. BMW fault 118002 describes mixture control with the fuel-air mixture too rich. On an inline four-cylinder N20 there is one bank, so this is an engine-wide mixture judgment, even when a companion misfire points toward one cylinder.

Fuel trim is the useful starting point. Negative correction means the DME is trying to remove fuel. I want the actual short-term and learned mixture values at warm idle, around 2,000 to 2,500 rpm with no load, and during a steady road load if the oil condition makes a road test safe. I also save freeze-frame data before clearing anything. A code after idle tells a different story from one set under boost or immediately after a cold start.

A single wet plug or cylinder-specific misfire moves a leaking or poorly spraying injector higher on my list. Strong negative correction across all cylinders with no single-cylinder evidence can fit excessive fuel pressure, vapor entering through the purge or crankcase system, incorrect air measurement, or biased exhaust feedback. The pattern narrows the job. It does not convict a component by itself.

Why gasoline can show up in N20 oil

The N20 is a direct-injection engine with solenoid injectors and a Bosch high-pressure pump. BMW's N20 technical training shows that the pump is driven by a three-lobe cam on the exhaust camshaft. That architecture makes two fuel-to-oil paths worth investigating.

A leaking injector can drip into a cylinder after shutdown or deliver fuel badly while running. Liquid fuel can wet the plug and cylinder wall, contribute to a misfire, and pass the rings into the sump. BMW issued N20 and N26 injector replacement guidance for defective injectors of a specific index, which confirms that injector identification matters. It does not mean every N20 rich code needs four injectors.

An internal leak at a cam-driven high-pressure pump is another plausible path because the pump is mounted at the engine and mechanically actuated. The owner reports make that failure mode worth testing, but a forum success is not a universal diagnosis. One Bimmerpost owner reported strongly negative trims and a repair after replacing the pump. Other contributors replaced the pump and still had the code.

Fuel can also accumulate through repeated rich cold operation and short trips that never heat the oil fully. The crankcase ventilation system then carries volatile vapor back into the intake. This can create a loop in which contaminated oil contributes more vapor as the engine warms. A rising oil level, obvious thinning, or a strong raw-fuel odor supports the concern, while laboratory oil analysis is a better confirmation than smell alone.

The test order I would use

I would keep the first pass non-invasive and preserve evidence.

1. Check the oil before extended running. I record the electronic oil level, inspect a sample for obvious thinning and fuel odor, and send a sample for analysis when the degree of dilution is uncertain. If the level is rising or dilution appears severe, I stop routine driving. An oil change protects the engine temporarily, but it does not fix the fuel source.
2. Scan every DME fault and save freeze-frame data. Misfire counters, injector activation faults, rail-pressure faults, purge faults, and temperature plausibility can change the direction. I record learned mixture values before resetting anything.
3. Log mixture correction from cold start to warm idle. I compare short-term response, learned correction, lambda feedback, coolant temperature, low-side supply if available, requested rail pressure, and actual rail pressure. The goal is to see when the DME begins pulling fuel and whether pressure follows its target.
4. Compare idle with a held engine speed. Extra unmetered fuel often has a larger proportional effect at idle. If correction improves as airflow rises, I prioritize leaks that add a relatively fixed amount of fuel or vapor, including an injector or purge path. If correction remains heavily negative, I widen the test to pressure control and air or exhaust measurement.
5. Use the cylinder evidence. I inspect plugs without letting the engine idle long enough to erase the clue. A wet or unusually clean plug, one cylinder's misfire count, and borescope evidence can justify an injector leak test. I verify the installed injector type and part identification before ordering anything.
6. Test purge and crankcase influence. An engineer can isolate the tank-purge path briefly and observe whether mixture correction moves toward normal. Crankcase pressure and ventilation faults also need checking, especially if the rich correction changes as the oil warms.
7. Verify sensors rather than cleaning or replacing them on suspicion. I compare the air-mass signal with expected engine behavior, check coolant temperature against a cold soak and warm engine, and confirm that exhaust feedback agrees with an independent mixture check when readings conflict.

Fuel systems operate at dangerous pressure. Injector and high-pressure pump leak tests belong with an engineer using BMW repair information and the correct pressure-relief procedure.

What I would not assume from fault 118002

A BMW bulletin often shared with fault 118002 describes an overly sensitive diagnostic, winter fuel, short-trip operation, crankcase vapor, and an oil service. The May 2019 version lists B38T, B46O, XB1H, and XB2H applications produced up to July 15, 2018. It does not list the N20.

I would not apply that bulletin's listed remedy to an N20 or present it as proof that an N20 only needs an oil change. The document does make one useful general point: BMW says other causes of a rich-mixture fault must not be overlooked. On the N20, the hardware and live data still have to be tested against the car's VIN, duty cycle, and companion faults.

I also would not condemn the front oxygen sensor merely because it reports rich. If an independent exhaust check and fuel-trim response agree that combustion is rich, the sensor may be reporting the problem correctly. Nor would I treat a smoke test as the end of the diagnosis. A smoke test helps find air-path leaks, but it does not clear injectors, pressure control, purge flow, crankcase vapor, or sensor bias.

Repair, oil service, and proof

Once testing identifies the fault, I repair that fault with the correct parts and procedures. If an injector is implicated, its index, installed type, sealing hardware, and any required calibration data matter. If the high-pressure pump is implicated, requested-versus-actual pressure and leak evidence should support the decision. If purge, temperature, air measurement, or exhaust feedback is wrong, I fix that circuit rather than adding fuel-system parts.

After stopping the source, I replace fuel-contaminated oil and the filter. Then I verify the repair on the same operating sequence that exposed the problem: cold start, warm idle, held rpm, and an appropriate road load. I want smooth idle, stable rail pressure, no growing cylinder misfire count, and mixture correction that no longer runs to the rich limit. A follow-up oil sample or at least a stable oil level and no returning fuel odor provides a second check.

P0172 or 118002 says the DME has run out of room to correct a rich condition. The rough idle, cylinder evidence, pressure behavior, and oil condition tell me where to look next. That distinction is what keeps an N20 diagnosis from turning into another round of plugs, sensors, injectors, and pumps bought on guesses.

FAQ

What does P0172 or BMW 118002 mean on an N20?

It means the DME has reached a rich-side mixture correction limit. It does not identify the failed component.

Can a bad N20 high-pressure fuel pump put gasoline in the oil?

It is a plausible failure path on this cam-driven pump, but odor alone is not proof. I would test pressure behavior, injectors, purge flow, fuel trims, and the oil before replacing it.

Should I keep driving when the oil smells like gasoline?

I would avoid normal driving until the oil is checked. Fuel dilution can reduce viscosity, while a rich misfire can damage the catalytic converter.

Does BMW's winter-fuel bulletin for 118002 apply to the N20?

The May 2019 bulletin commonly circulated for this fault does not list the N20. I would not use its software remedy as an N20 repair instruction.