P0191
- Powertrain
- Generic (SAE)
- Emissions
Get it looked at now
Wrong rail pressure information leads to hard starting, power limitation, stalling and limp mode, and the same code is set by a high-pressure pump that is on its way out. High-pressure fuel systems also demand real caution: injection-pressure spray penetrates skin.
- Safe to drive
- Drive only as far as needed to reach a shop. Stop immediately if power drops, it stumbles under load, or it stalls.
Fuel Rail Pressure Sensor Circuit Range / Performance
The fuel rail pressure sensor's signal is electrically normal but does not agree with the pressure the computer commanded. Either the sensor is reporting a pressure that is not real, or the pressure really is wrong — the code cannot tell you which.
This is a rationality fault, not a circuit fault. The ECM commands rail pressure through a metering or pressure control valve and closes the loop using the rail pressure sensor. P0191 sets when the measured value stays outside the expected window for the commanded pressure and the control valve's position, while remaining inside the sensor's valid voltage range. It appears mostly on common-rail diesel and gasoline direct injection systems, where rail pressures run from hundreds to tens of thousands of psi and the control loop has very little tolerance for error.
How the car detected itThe module compares three things: the pressure it asked for, what the control valve is doing to achieve it, and what the sensor reports. If the sensor value fails to move as expected when the valve moves, or if the correction needed to hold target keeps growing, the plausibility test fails. The critical limitation is that the module cannot distinguish a lying sensor from a genuine pressure shortfall — from inside the loop, both look exactly the same.
A trouble code records what a control module measured. It does not identify which part failed. Test before replacing anything.
- Safe to drive
- Drive only as far as needed to reach a shop. Stop immediately if power drops, it stumbles under load, or it stalls.
- Urgency
- Get it looked at now
- Will it pass emissions
- No
- If ignored
- Expect progressively harder starting, reduced power, and eventually a stall or a no-start. If the real cause is a wearing high-pressure pump rather than the sensor, continued driving risks pushing wear debris into the injectors and rail, multiplying the repair cost several times over. There is also a genuine stalling-in-traffic hazard once the fault matures.
- Check engine light
- Solid
- Clearing the light
- A completed drive cycle is normally required before the monitor re-runs and the light can clear.
These are the causes this code can have. For the order to work through them in, see what to check first.
| # | Cause | How often | Component |
|---|---|---|---|
| 1 | Pressure sensor drifted | Very common | — |
| 2 | Pressure regulator stuck | Common | — |
| 3 | High-pressure pump worn | Occasional | — |
Ordering reflects how often each cause is responsible in general, not a probability for your vehicle. Confirm by testing.
In order. Each of these is cheaper or faster than what follows it, and each one can make the rest unnecessary.
The reported rail pressure with the key on and the engine off
The rail should have decayed to the manufacturer's stated rest value once the system has bled down — but many common-rail and GDI rails hold real pressure for hours after shutdown, and some sensors report absolute pressure, so 'rest' is not always zero. A resting value well above the published figure points strongly at the sensor or its circuit; confirm it against service data before condemning anything.
Commanded versus actual rail pressure at idle and under load
Shows straight away whether the pressure loop is holding, drifting, or has given up entirely, and at what load it breaks down.
Fuel filter service history and current fuel level
A restricted filter or a starved supply produces a genuine pressure shortfall that the module reports as a sensor performance fault.
Any other fuel pressure, metering valve or injector codes stored with it
Additional codes usually name the subsystem outright and save hours of sequential testing.
The sensor connector and its harness for oil, coolant or corrosion
Rail pressure sensors sit in a hot, oily area. Contamination produces readings that look exactly like sensor drift but are repairable rather than replaceable.
Record freeze frame, then graph commanded against actual rail pressure
Tools Scan tool with live data graphing
Capture stored data first, then record both values across idle, cruise and heavy load. The shape of the divergence separates the two possible worlds: a sensor problem or a real pressure problem.
ExpectedActual tracks commanded closely at all loads. A gap that widens with load indicates a supply or pump limitation; a fixed offset at every load suggests sensor drift.
Read rail pressure with the key on and the engine off
Tools Scan tool, manufacturer service data
With the engine stopped and the system bled down, the rail sits at its rest pressure — a quick check that is often skipped. What 'rest' means is vehicle-specific: many common-rail and GDI rails hold real pressure for hours after shutdown, and some sensors report absolute pressure, so look the figure up rather than assuming zero.
ExpectedAt or near the manufacturer's rest value after bleed-down. A resting reading clearly above that, repeated across key cycles, indicates the sensor or its circuit rather than the pressure itself — but it is an indication, not a verdict; the independent pressure measurement in step 4 settles it.
Verify the sensor's reference voltage, ground and signal at the connector
Tools Multimeter, vehicle wiring diagram
Backprobe with the key on. Test the ground by voltage drop rather than continuity. Pin assignments vary by vehicle and even between engine variants — use the wiring diagram rather than a generic pinout.
ExpectedAbout 5 volts on the reference, under 0.1 volt of drop on the ground path, and a signal voltage corresponding to the system's rest pressure.
Measure actual rail pressure independently where the manufacturer provides a method
Tools Manufacturer-specified high-pressure test equipment
This is the test that settles the sensor-versus-reality question. Follow the manufacturer's procedure exactly. High-pressure common-rail fuel can exceed 20,000 psi, and injection spray will penetrate skin and cause a serious injury that looks minor at first — depressurise before opening any high-pressure fitting and never search for a leak by hand.
ExpectedIndependently measured pressure agrees with the sensor's reported value. If it does not, the sensor is the fault and the mechanical system is fine.
Test the low-pressure supply for both pressure and volume
Tools Fuel pressure gauge, graduated container, stopwatch
Measure inlet pressure while loaded, and measure delivered volume into a graduated container over a timed interval. Pressure alone will not reveal a supply pump that cannot flow.
ExpectedSupply holds up under load and meets the manufacturer's flow figure. Any sag under load explains the pressure shortfall outright.
Test pressure control and metering valve response with actuator commands
Tools Bidirectional scan tool
Command pressure changes and watch the rail follow. Also check the valve's own duty cycle — a duty sitting at a limit means the loop already exhausted its authority before the code set.
ExpectedRail pressure moves promptly and proportionally with the command, and valve duty sits inside its normal range.
Perform a rail leak-down and injector return test
Tools Graduated cylinders, return line adapters, scan tool
Check whether the rail holds pressure with the engine off, and measure each injector's return volume into its own container. Uneven return names a single leaking injector, which mimics both a bad sensor and a bad pump.
ExpectedRail holds pressure with only slow decay, and return volumes are even across all injectors.
Consider the high-pressure pump only when everything else is proven
Tools Scan tool, clean container, bright light
Supply volume, suction-side integrity, control valve response and injector return all need to be clean before the pump becomes the remaining explanation. Check the fuel and filter for metallic debris before fitting any new part.
ExpectedA pump that cannot reach commanded pressure with all of the above verified is the fault. Metallic debris anywhere means the whole fuel system needs evaluating.
| Repair | Only after you confirm | Typical cost | Who |
|---|---|---|---|
| Repair a corroded or contaminated sensor connector or its harness | Voltage drop on the reference or ground circuit, or a reading that changes during a wiggle test. | $100–$400 | DIY |
| Replace a restricted fuel filter and prime the system | Supply pressure or volume below spec ahead of the high-pressure pump, restored once the filter is changed. | $60–$300 | DIY |
| Replace the fuel rail pressure sensor | Reported pressure disagrees with an independently measured value, or the sensor reports a resting pressure well above the manufacturer's rest figure after bleed-down while its reference and ground both test good. | $150–$550 | Hard |
| Replace the pressure control or metering valve | Actuator commands produce no proportional pressure change with supply volume verified good. | $200–$800 | Hard |
| Replace an injector identified by excessive return flow | One injector's measured back-leak volume clearly exceeds the others in a timed return test. | $300–$1,400 | Shop |
| Replace the high-pressure fuel pump — gasoline direct-injection pumps sit near the bottom of this range, diesel common-rail pumps at the top | Commanded pressure unreachable with verified supply volume, a functioning control valve and normal injector return across all cylinders. | $400–$3,000 | Shop |
US independent-shop ranges, parts and labour, excluding diagnostic time — often 1.5-3 hours on this code. Sensor access varies enormously: some sit on top of the rail, others under the intake manifold. Diesel common-rail parts sit at the top of every range and some require coding after fitting.
- Replacing the rail pressure sensor because the code names it. The code names the signal that failed a plausibility test, not the part that failed.
- Skipping the key-on, engine-off pressure reading. It is quick and it points the diagnosis — as long as it is compared against the manufacturer's rest value rather than assumed to be zero.
- Assuming a rail at rest must read zero. Common-rail and GDI systems can legitimately hold real residual pressure long after shutdown, and some sensors report absolute pressure.
- Testing supply pressure without testing supply volume.
- Condemning the high-pressure pump before measuring injector return flow.
- Clearing the code and road testing before capturing freeze frame data, which throws away the conditions that produced the fault.
- Opening a high-pressure fitting without depressurising the system first.
- Fitting a new pump without checking the fuel and filter for debris from the old one.
- Fuel Rail Pressure Sensor
- High Pressure Pump
- Rail Pressure Regulator
- Definition
- Generic code set (SAE J2012 / ISO 15031-6) via the OBDexopen database, dedicated to the public domain under CC0-1.0.
- Diagnostic guidance
- Written and reviewed in-house. See our editorial policy.
- Review status
- Reviewed
Found something wrong? Tell us — we publish corrections.