PeachStamp
All model research

Porsche 911 (997 Gen 1) 3.6 / 3.8buyer's guide

2004–2008

The faults this model is known for, what they tend to cost in the UK, and what a service history has to show before any of them counts as covered. Researched from the sources shown, and open to correction.

Looking at one?

Type the registration for a free summary of its MOT history, read against the faults below. No sign-up.

If none of it were documented

£12,250–£21,500

Across 6 costed items, plus 2 named with no published figure and therefore not counted.

This is a ceiling, not a forecast. A report on a specific car counts only the items its paperwork does not already cover - a folder showing the clutch done last year removes that line entirely.

Known faults

Bore scoring

£8,000–£12,000

Vertical scoring of the cylinder walls, where the piston wears through the liner and lets oil into the combustion chamber. It shows up as heavy oil consumption, a darkened nearside exhaust tip and a tapping at idle. Markedly more common on the 3.8 than the 3.6, and absent from the later DFI engine. The only real fix is new liners and pistons, at which point the engine is usually refreshed throughout.

Reported at around £10,000 for the engine work once it is done properly. This is the one item on this car that can cost more than the car is worth if it is ignored.

Typically bites

Not tied to a mileage. Cars used gently and warmed up briefly are described as being at more risk than cars driven properly, so a low-mileage example is not automatically safer.

Verified against 3 published sources.

IMS (intermediate shaft) bearing

No published figure

The bearing at the end of the intermediate shaft stops self-lubricating and fails, and when it does it can put debris through the engine. Its reputation is larger than its incidence, and the important detail is the date: Porsche revised the design during 2006, so only the earlier cars carry the original bearing.

No single reliable published UK figure - the range depends entirely on whether it is done preventatively with the gearbox out or after a failure. A failure is an engine rebuild.

Typically bites

Not mileage-driven. Failures are reported across the range, which is why a retrofit is treated as insurance rather than as work due at a set interval.

Verified against 3 published sources.

PASM dampers

£1,500–£3,000

The rubber gaskets perish with age and the dampers weep. They are replaced in pairs, because fitting one new damper against one tired one gives the car a diagonal it will never settle on.

From around £1,500 fitted, and always in pairs.

Typically bites

Age-driven rather than mileage-driven. Any car of this era on original dampers.

Verified against 1 published source.

Coil springs

No published figure

They crack with age and corrosion, usually with no warning and no symptom until one goes. Replaced in pairs for the same reason as the dampers.

No published figure found separately from damper work.

Typically bites

Corrosion-driven, so winter use and outdoor storage matter more than mileage.

Verified against 1 published source.

Coolant pipes and crossover connections

£400–£1,200

The crossover pipes corrode internally where aluminium meets steel, and the rubber sections perish with age and heat. The first sign is a sweet smell around the front of the car rather than a puddle, which is why it is often missed until a hose lets go. Described by specialists as very common on these.

Independent work in the mid hundreds, and considerably more if a failure has let the engine run hot.

Typically bites

Driven by age and heat cycling rather than mileage.

Verified against 2 published sources.

Rear main seal and clutch contamination

£900–£2,200

The seal between engine and gearbox weeps, and on this engine the oil lands on the flywheel and clutch. That turns a slow leak into a clutch job: the symptom owners notice is a judder on take-up rather than a drip. The labour overlaps with the clutch and the IMS, which is why all three tend to appear on one invoice or none.

Little on its own beyond labour, and the reason it matters is what it takes with it. A contaminated clutch is a four-figure job on this car.

Typically bites

Age-driven. A judder on clutch take-up on a car with no recent invoice is the thing to chase.

Verified against 2 published sources.

Air-oil separator

£250–£600

The separator regulates crankcase pressure and returns oil to the sump. When it fails, the car pours blue smoke at startup, idles roughly and drinks oil, and it looks far more serious than it is. Described as essentially a maintenance item rather than a fault on cars past a hundred thousand miles.

A few hundred pounds fitted at an independent, and one of the cheapest jobs on the car relative to how alarming the symptom is.

Typically bites

Commonly by 100,000 miles, and treated as due rather than as a surprise.

Verified against 2 published sources.

Clutch, and the jobs that share its labour

£1,200–£2,500

The clutch itself is ordinary wear. It is carried here because of what sits behind it: the gearbox has to come out, which is also the access for the IMS bearing and the rear main seal. A clutch invoice that does not mention either is a missed opportunity a specialist would have raised, and a car with a recent clutch and no IMS work has had the expensive access paid for and not used.

A clutch on these is a four-figure job at an independent. The marginal cost of doing the seal and the bearing at the same time is small, which is the entire argument for reading the invoice rather than the mileage.

Typically bites

Reported around every 50,000 miles, and heavily dependent on use.

Verified against 2 published sources.

Service interval

Every 20,000 miles or 24 months, whichever comes first.

Porsche moved these to a 20,000-mile, two-year interval. Independent specialists openly disagree with it on older cars and recommend an annual oil change alongside the MOT instead - so a folder showing services exactly to the factory schedule is not the same as a well-maintained car, and a folder showing annual oil changes is worth more than it looks.

The interval is here so a report can do arithmetic rather than form an opinion: last recorded service, current mileage, today's date. Whether a car is overdue is a sum, and sums should not be able to hallucinate.

What to check on the day

Written to be read standing next to the car, with the folder open.

  1. Ask for a borescope inspection, or evidence of a recent one. It is the only way to see bore scoring before it is expensive, and on a 3.8 it is the single most important check.

  2. Look at the nearside exhaust tip for heavy darkening, and listen at idle for a tap.

  3. Establish the build date, not just the model year - the IMS bearing question turns on whether the car predates the 2006 revision.

  4. Look behind each wheel for damper weeping, and ask whether any have been replaced and in which pairs.

  5. Count oil changes by date rather than mileage. Two years between services is the factory schedule and most specialists think it is too long.

What this research does not know

  • Scoped to 997.1 only. The 997.2 from 2009 uses the DFI engine, which has no intermediate shaft and does not carry the bore scoring reputation - applying this profile to one would materially overstate its risk.
  • Two of the four items have no published cost and are excluded from the total rather than estimated. The bore scoring band is the dominant figure and is drawn from specialist commentary rather than a price list.
  • DVSA model strings for the 911 vary and may not read exactly "911". Until live data arrives it is not possible to confirm this profile matches on a real lookup - see the note in research-key.ts about why an unmatched key fails silently.

Last reviewed on 16 August 2026. Cost bands are indicative UK independent-specialist rates, not quotes.

Now find out which of these the car you're looking at can actually prove.

Photograph or upload the seller's folder and every invoice is read, checked against the official mileage record, and matched against the list above - so the total you get is what this car has no evidence for, not what the model might cost in general.

See a full worked example