Turbo Overboost and Underboost: What P0234 and P0299 Actually Measure

A boost fault code does not describe a component. It describes a disagreement — between the pressure your engine management asked for and the pressure a sensor told it arrived. P0234 says the reported figure came in above target. P0299 says it came in below. That is the whole of the difference between them.

So these are not two subjects with two cause lists. They are one reading taken in two directions, and a single failed part regularly produces both.

What Overboost and Underboost Actually Are

The engine management holds a boost target for the conditions the engine is in — not a fixed number, but a figure moving with engine speed, load, temperature and fuelling strategy. Against it sits an actual reading from the manifold pressure sensor. The management compares the two continuously and works the boost-control hardware to close any gap.

A code stores when the gap persists beyond what it will tolerate. Both codes are therefore a verdict on a comparison, and neither says which of the parts between the request and the result went wrong. What the comparison does carry is a direction, and that is the useful part: it decides which suspects are still in the frame, and how urgent the car is.

Why One Fault Can Store Both Codes

Most modern diesels regulate boost with a ring of moving vanes around the turbine wheel: narrow the path the exhaust gas takes and the turbine is driven harder, widen it and it is driven less. Soot and oil vapour bake around the vane pivots until the ring stiffens, and where a stiff ring comes to rest decides which code you get.

Stopped toward the closed position, gas keeps being driven at the turbine after the management has asked for pressure to come down, and an overboost code stores. Stopped toward the open position, exhaust energy bypasses the job it was meant to do and the pressure never arrives. Same deposit, same pivots, opposite codes. A wastegate does the same: held shut, the turbo carries on past the point gas should have been bled away; held open, boost is never made.

Two things follow and neither is normally stated. A page listing the causes of overboost and a page listing the causes of underboost are largely listing the same hardware twice. And a mechanism that is stiff rather than seized can present as one fault and then the other — which is why specialists documenting P0299 list P0234 among the codes that keep it company, with a mechanism stuck closed given as the explanation for the pair.

The Five Suspects, and Which Ones Each Code Rules Out

Replace the question. Instead of asking what causes overboost, ask which link between the commanded pressure and the reported pressure has broken. There are five, and they are not equally available to each code.

Read the third column and the asymmetry is obvious. An underboost code leaves all five standing. An overboost code eliminates two outright, because neither can manufacture pressure that was never asked for: air escaping through a split hose is air the engine never receives, and wear in a rotating assembly costs flow rather than creating it. P0234 is the more informative of the two codes; P0299, far the more common, narrows nothing.

The Pressurised Path, and Why a Leak Only Ever Reads Low

Specialists who see these codes daily are consistent that the most frequent cause is not a failed turbocharger. It is a split intercooler pipe, a cracked charge hose, a perished coupler, or a boost-control mechanism stuck in one position. Air made by the compressor and lost before it reaches the engine gives an honest sensor reading of a pressure that genuinely is not there.

This link goes first for a practical reason: it is the only suspect that can be proven or eliminated without removing anything. The charge circuit is pressurised from a regulated source — never improvised, never beyond the system’s own limits — and the joints are checked for escaping air. Worth doing by hand as well, warm: a hairline split beside a hot exhaust manifold looks sound at rest and opens under load, which is how it survives a static inspection.

Since none of this can raise a reported pressure, an overboost code takes the whole link off the list. If your code is P0299 it stays at the top of it. Other faults imitate a failing turbocharger from the driver’s seat in the same way, and our guide to common turbocharger problems works through them alongside the symptoms they copy.

The Regulating Mechanism, and Which Way It Stuck

Two families of hardware do this job and they should not be blurred together.

A wastegate is a flap letting exhaust gas bypass the turbine once the target is reached, held shut by a spring in a pneumatic can. Its rod is not a fitting dimension — the length is a setting made to that turbocharger’s specification, so a sound actuator fitted at the wrong length gives the wrong boost with nothing broken. That is the subject of our wastegate actuator repair page.

A variable-geometry mechanism is the vane ring described above. Its failure is chemical before it is mechanical: deposits build around the pivots until the ring can no longer swing freely through its range. What a strip of one shows is set out on our Garrett GT2052V page.

Petrol engines carry a third device often mistaken for these, and it is neither. A diverter or recirculation valve vents charge air back toward the intake when the throttle closes; stuck open, it vents under load and boost never builds. It belongs on the underboost side only.

One caution saves a repeat failure. Whatever moves the mechanism is frequently the casualty rather than the cause: an actuator straining against a ring that will not swing freely wears its motor and gear train doing it, and a replacement fitted to an untouched mechanism goes the same way. Our pages on what a turbo actuator does and electronic actuator repairs cover that part on its own terms.

When Nothing Is Asking the Mechanism to Move

Between the engine management and the mechanism sits a command path. On a pneumatic system the management drives a solenoid which meters vacuum or pressure to the actuator can, so the solenoid, a perished control line, a failed check valve or a weak vacuum supply can each break the command. On an electronic system it drives a motor and reads a position signal back, where a corroded connector, water in a plug or an open winding does the same job.

Here is the difficulty, and it is why this fault is misdiagnosed so consistently. A mechanism that is free but receiving no command behaves identically to a mechanism that is seized while receiving a full one. Same error, same direction, same code, same drive. Nothing observable from the cabin separates them and neither does the fault code.

Which is why the electrical side is worth testing before anything is unbolted — most of it can be checked in place. Where the evidence is behavioural and electrical rather than physical, our guide to turbo actuator symptoms is the better starting point than this page.

The Instrument Reporting the Fault Is Also a Suspect

Every judgement described so far rests on one number, and one component produces it. The manifold pressure sensor is inside the loop, not standing outside it, and it can be the thing that is wrong.

A sensor reading below truth manufactures an underboost code on an engine making entirely correct boost; one reading above truth manufactures overboost on the same healthy hardware. A blocked reference port or a corroded connector does the same.

Here the fault sits in the reporting rather than the running, so the car may drive close to normally until the management acts on the bad number — at which point the capped torque is real even though the condition that triggered it was not. Fit a mechanism to cure a measurement fault and the code returns unchanged.

The Comparison That Tells the Three Apart

Three of the five suspects — mechanism, controller, instrument — produce the same code in the same direction and cannot be told apart by symptom. One comparison separates them. With the engine driven under load, log the boost being requested, the boost being reported, and what the management is doing to the hardware at that moment: solenoid duty on a pneumatic system, commanded against reported position on an electronic one.

• Actual pressure well short of requested while the controller is driven hard toward its limit. The command is being made and not obeyed — a stiff vane ring, a stuck flap, or an actuator that can no longer hold position under load.
• Pressure short while the controller is barely being asked for anything, or a command that does not change as load changes. The mechanism may be perfectly free; the fault is on the control side.
• A reported pressure that does not track load plausibly, or disagrees with the rest of the picture. Look hard at the instrument before anything is condemned.

The relationship between the traces is what matters, not any particular value. No figures are quoted here deliberately: targets, acceptable deviations and duty percentages belong to an engine and its calibration, and a threshold borrowed from another vehicle is worse than none.

The rule that falls out of it is worth stating plainly. A turbocharger should not be condemned on a boost code alone — not before the charge circuit has been pressure-tested, the control side checked electrically, and requested pressure compared against actual.

What Limp Mode Is, and Why the Car Seems to Fix Itself

Limp mode is not a symptom of the turbocharger. It is the engine management capping torque deliberately, because it can no longer trust the boost picture. That is why a car with a boost code often feels far worse than the underlying fault warrants.

It also explains the observation that confuses owners most. Switch off, start again, and the limiter lifts. Then the same conditions return — a long pull, a high gear under load — and it limps again. Nothing was repaired in between; the management re-tested the loop from a clean start and reached the same conclusion. Treating each recovery as the problem resolving itself is how an overboost condition gets left running.

Which Is More Urgent, Overboost or Underboost?

This is the one question the code answers well, because direction is the only thing it reliably reports — and the two are not equally serious.

Overboost is the more damaging. More air than the engine was designed to receive means combustion more aggressive than its components were specified for, and combustion-chamber temperatures rise with it. The recognised consequences run through detonation, heat damage to pistons and valves, and a cooling system pushed outside its design range. The turbocharger is not spared either: sustained overheating breaks down the oil film the bearing system depends on, and a shaft running without that film can seize. Overboost damages the very component people assume it proves is healthy.

Underboost is the less destructive and the more revealing. Lost power, poor acceleration, higher fuel consumption, soot or smoke on a diesel — unpleasant rather than dangerous in the short term, but often the visible end of something that does matter: a leak that has been growing, a mechanism progressively binding, or a rotating assembly already worn.

So the two deserve opposite responses. Act quickly on overboost, and do not load the engine. Do not act on underboost by buying parts, because five suspects are still standing and only one is the turbocharger. Metallic noise, heavy smoke or a hard cut shortly after full throttle are different territory again, covered under common turbocharger problems.

What a Bench Test Settles That a Scan Tool Cannot

The list divides cleanly. Charge path, controller and instrument are all testable on the vehicle. The mechanism and whatever drives it are not, because both questions are quantities rather than yes-or-no observations: how much resistance the vane ring is being swung against, and how much load the actuator can hold before it gives up position. A mechanism that moves when you watch it can still be stiff enough to store a code under exhaust-gas load.

That is the work our bench does. Rebuilding here is a mail-in service — the unit comes off the car and is posted to our Epsom workshop from anywhere in the UK or internationally, then comes back rebuilt, tested and ready to refit. It is stripped to component level and every part measured against original specification, with the resistance in the vane or wastegate hardware recorded rather than estimated.

Worn parts are replaced with genuine OE or OEM-grade components, a part with a known design weakness is uprated rather than reinstated, and the rotating assembly is balanced before the unit is run up. Electronic actuators are recalibrated to the vane sweep they will be working against, so they refit without programming. You get your own unit back rather than an exchange unit off a shelf, and the measurements taken during the strip are what finally name the link that failed. Every repair carries a lifetime, unlimited-mileage warranty; postage and packing are covered in our FAQs.

Turbo Overboost and Underboost FAQ

Final Thoughts

This subject is written up badly nearly everywhere because it is split in the wrong place. Divide it into overboost and underboost and you get two cause lists containing much the same hardware, no way of choosing between the entries on either, and no explanation of how one car produces both.

Divide it by the link that failed and it becomes tractable. The pipework either holds pressure or it does not. The control side is electrical and can be tested where it sits. The instrument can be judged against the plausibility of everything around it. What is left — the mechanism, and whatever moves it — is the part no roadside check settles, and the part most often replaced on the least evidence.

If your car is showing P0299, resist the conclusion the code invites: five suspects are still standing and only one of them is the turbocharger. If it is showing P0234, you have been handed a narrower field and a more urgent reason to use it.