Get The Analogue
Diagnostics · E90, E91

BMW E90 N47 4B90 crank no start: 10 bar actual, 300 bar specified

A direct test order for an E90 318d N47 that cranks but will not start while actual rail pressure stays far below the requested value.

By Chase Waters · · 6 min read

A BMW E90 318d with an N47 that cranks but will not start, stores 4B90 and reports about 10 bar actual rail pressure against roughly 300 bar specified has a real pressure-control problem to diagnose, not a verdict on the high-pressure pump. Those figures come from the linked owner report. I would preserve the full fault record, watch one complete cranking capture, then split the system into electrical supply, low-pressure feed, pressure measurement, leakage and pump-capacity tests.

What the 10 bar and 300 bar readings prove

The two values prove only that the DDE's reported actual pressure did not follow its requested value during that attempt. They do not show where pressure was lost. A scan tool reads the rail-pressure sensor through the control unit, so a low number can represent genuinely low hydraulic pressure, a sensor or circuit problem, or bad control-unit power and communication.

The request itself is useful. It shows that the controller was asking the high-pressure system to build pressure. Before interpreting it, I would confirm that engine speed is present and steady while cranking, battery voltage does not collapse, the DDE remains online, and the values are labelled in bar rather than another unit. Save a graph of:

  • engine speed
  • battery voltage at the DDE or the closest available supply channel
  • rail pressure specified
  • rail pressure actual
  • low-pressure feed, if the vehicle exposes it
  • fuel quantity-control or pressure-control command, if available
  • cam and crank synchronization status

Do not turn the owner's 10 bar actual and roughly 250 to 300 bar requested into universal N47 thresholds. BMW's vehicle-specific test plan and technical data should decide pass or fail for the exact build. The gap matters here because it persisted through cranking and the engine did not start.

Check cranking voltage and low-pressure supply first

Start with the battery fully charged and supported by equipment suitable for diagnosis. Measure voltage at the battery posts during cranking, then voltage drop from battery positive to the relevant engine supply and from engine ground to battery negative. A starter can still turn while voltage loss disturbs DDE communication, sensor readings or actuator control. The owner thread included communication faults and a CAS connector repair, which is a reason to check supply and network behavior rather than treating every stored code as a separate failed module.

Next inspect the low-pressure side without opening a high-pressure line. Check fuel level, the correct filter installation, damaged or kinked supply hoses, air ingress and visible leaks. Command the in-tank pump with the diagnostic test plan where supported. Measure feed pressure at the approved low-pressure test point with a suitable gauge, and compare pressure and delivered volume with BMW data for the exact car. Pressure without enough flow can mislead, especially through a restricted filter or supply line.

BMW's N47 Product Information lays out the engine's common-rail fuel-system architecture. The high-pressure pump, rail, pressure sensor, pressure-control hardware and injectors operate as one controlled system. That architecture is why replacing the pump from a low scan value alone skips several possible failure paths.

Take a clean fuel sample from the low-pressure side in an approved container. Look for water, dirt or metallic debris under good light. If contamination is present, stop repeated cranking and follow the manufacturer procedure for the whole system. Fitting one clean component into a contaminated system risks another failure.

Separate a false reading from a real pressure loss

With ignition on and the engine stopped, look at the rail-pressure signal for plausibility and stability. Wiggle only the sensor harness and nearby loom while watching the value. Inspect the connector for damaged locks, spread or backed-out terminals, corrosion, fuel or oil ingress, and rubbing. Do not force test probes into terminal faces.

If the reading jumps when the harness moves, test the sensor circuit from the correct wiring diagram. Check reference supply, ground and signal with high-impedance equipment and approved back-probing methods. A stable electrical value still needs comparison with an independent, properly rated pressure test or BMW's guided test before condemning the sensor. Never substitute a general-purpose gauge on the rail.

If the signal is plausible but pressure remains low, test leakage before pump capacity. A common-rail injector can return too much fuel internally, preventing the rail from building pressure. Use the correct closed leak-back equipment, identical containers and the BMW procedure. Compare all four injectors during the same controlled crank period. Delphi's injector guidance also treats back leakage and pump output as measured tests, not reasons to replace a complete set by guesswork.

A basic leak-back test has limits. Very low rail pressure may prevent a useful comparison, and return volume does not prove that an injector nozzle is sealing into its cylinder. Check for a cylinder that appears wet, an unexplained rise in engine-oil level or oil that smells strongly of diesel. If fuel dilution is suspected, stop cranking, do not run the engine, and have the injectors tested by a diesel specialist. Do not place fingers near injector openings or search for leaks by touch.

Test the pump only after the rest of the circuit

Once feed pressure and flow are in specification, the rail signal is credible, contamination is absent and injector leakage has been assessed, run the BMW high-pressure system test for the exact N47 variant. A specialist can isolate pump capacity with sealed, pressure-rated equipment and the correct adapters. That test separates a pump that cannot generate pressure from a rail-side leak or control problem.

Do not clamp return hoses, loosen injector pipes or improvise plugs. Common-rail diesel pressure can penetrate skin, and a tiny leak may be nearly invisible. If the guided test calls for blanking or substituting a component, use the specified tools and follow the depressurization procedure.

The owner's resolved thread is useful because the path was not a clean pump diagnosis. The report says a specialist found a CAN-bus issue and suspect sensor readings. The eventual work included an injector described as stuck open into cylinder 3, a sensor replacement and a repaired CAS connector, after which the engine started. Because several faults were repaired together, the thread cannot prove which single repair accounts for every symptom. It does show why a remanufactured pump and a simple return check did not end the diagnosis.

Prove the repair under the original conditions

After the identified fault is repaired, correct any fuel-diluted oil before running the engine and complete the specified fuel-system priming procedure. Clear faults only after saving the original record. Then repeat a cold or warm start under the condition that produced the failure while logging the same channels.

I want actual pressure to rise consistently toward the requested value, stable cranking voltage, a clean start and no returning 4B90 or companion supply, synchronization or communication faults. Inspect for leaks without touching pressurized components. One successful start is encouraging, but several matched starts and a clean rescan are better evidence that the fault is gone.

Questions owners ask

Why does my BMW E90 N47 show 4B90 with 10 bar actual rail pressure and 300 bar specified while cranking?

That large gap says the reported rail pressure is not following the request during this owner's start attempt. It does not identify one failed part. Test the cranking electrical supply, low-pressure feed, rail-pressure signal, injector leakage and high-pressure pump as separate branches.

Does 4B90 prove the N47 high-pressure pump has failed?

No. Low feed pressure or flow, air in the supply, a misleading pressure signal, excessive injector or rail-side leakage, a control-valve problem and a weak high-pressure pump can all leave actual pressure below the requested value. Isolate those paths before buying a pump.

Can an injector cause an N47 crank no start even if a leak-off test looks normal?

Yes, a basic return-volume check is useful but it is not a complete injector test. In the cited owner case, the reported repair included an injector that was stuck open into cylinder 3, plus a sensor replacement and CAS connector repair. The thread does not prove that every car with the same code has the same cause.

Should I loosen an injector pipe to see whether fuel arrives?

No. Common-rail fuel can cause serious injection injury. Keep the high-pressure circuit closed and use scan data, the manufacturer test plan and properly rated diesel test equipment.

Sources

  1. BimmerForums owner case: E90 N47 crank no start with low rail pressure, page 2
  2. BMW Aftersales Training: N47 engine Product Information
  3. Delphi Technologies: common-rail injector service and diagnostic guidance