Walnut blasting and intake valve cleaning: the GDI carbon job, factory view

Direct injection moved the fuel nozzle into the cylinder, and the intake valves lost their only bath. On the N54 and N55 — turbocharged, DI, and often driven the way we drive them — intake-valve carbon is a design consequence, not a defect mystery. BMW's own answer is in the service data: a walnut-shell blasting procedure, per port, with the engine locked at TDC. Here is that factory procedure, what it says about how, and the honest version of when.

The physics, in one paragraph

Port injection sprayed fuel past the intake valves every cycle; the wash kept deposits from accumulating. Direct injection skips that trip entirely, so crankcase vapors and EGR residue bake onto the back of the hot intake valve with nothing to rinse them. The result is the classic symptom set: cold-start roughness, misfires under load, and a car that "needs a tune-up" that plugs do not fix. The factory's remedy is mechanical removal — media blasting — because the deposits are baked varnish, not loose soot.

The factory procedure (N54, `11 12 670`)

The document is unambiguous about the rule everything else hangs on: the intake valves of the cylinder being cleaned must be closed. BMW enforces it mechanically:

1. 01Preliminaries: cylinder head cover off, intake plenum off — the same teardown as our valve-cover guide, which is why these jobs get booked together.
2. 02Rotate the crank at the central bolt to firing TDC of cylinder 1 and lock it with `11 0 300`. With cylinder 1 at ignition TDC, the intake valves of cylinder 1 are closed — the document states this as the verification, not an assumption.
3. 03Fit the extraction adapter (`81 29 2 208 037`) over the port, insert the probe (`81 29 2 208 033`), connect compressed air and the vacuum to the adapter.
4. 04Fire the peener (`81 29 2 208 034`, loaded with walnut shell granulate): roughly **30 seconds** of cleaning per port, moving the probe in a circular pattern, then another 10–15 seconds of compressed air alone to chase the media out.
5. 05Verify no granulate remains in the port — the factory note says check, not hope.
6. 06Rotate the crank **120° at a time** and repeat, cylinder by cylinder, **in firing order 1-5-3-6-2-4**. Six stops, six clean ports.

The N55 additions

The N54's sibling procedure adds a warning worth the whole paragraph: cover the blow-by ducts. Granular material must not enter the cylinder head cover's breather areas — if it does, the document's remedy is cover replacement. The N55 job also leads with the cleanliness discipline (seal every opening, protect every removed part). The tooling and the firing-order rotation carry over.

DIY vs shop: the honest translation

The factory rig (peener + extraction adapter + shop vacuum + regulated compressed air) exists because media control is the whole game: blast the valve, recover every granule, let nothing fall into an open port or the breather. Aftermarket kits reproduce this logic at hobby prices, and the translation rules are exactly the factory's rules:

• Valves closed or the port packed — no exceptions, verified mechanically, not by eyeball.
• Vacuum on the adapter *before* media flows — recovery first, blast second.
• Walnut shell media, not sand or glass — the factory chose a media that cleans varnish without attacking aluminium.
• Check the port for leftovers with a borescope before rotating the crank.

A shop with the factory rig does this faster; a careful owner with a kit does it just as clean. What neither can do is skip the closed-valve law.

Intervals: what the data supports

Here is the honest part: the factory service data carries no mileage interval for intake port cleaning. It is a diagnose-then-do procedure ("after vehicle diagnosis" is how the multi-engine version is titled), triggered by symptoms — cold-start roughness, misfire codes, airflow imbalance between banks — not by a CBS counter. Any "every 40k/60k/80k miles" number you read is community pattern-matching from symptom onset, which varies with driving style, oil consumption, and breather health. What the data does support: oil that never gets changed feeds the breather that feeds the deposits — the contamination chain runs through maintenance you control.

Coverage notes

What the data carries: complete walnut procedures for N54 and N55 with tool numbers, times, and the TDC/valve-closed discipline; a multi-engine cleaning procedure whose validity includes B58 for the G-chassis era. What it does not carry: intervals, deposit photography, before/after flow numbers, or any claim that B58 needs the job on the N54's schedule — younger fleet, same physics, fewer miles of history. The S55's port injection (secondary injectors washing the valves) makes it a different story entirely, and this page does not pretend otherwise.

Every procedural claim above traces to the ISTA document ids in the sources.

FAQ

• **How do I know my valves actually need it?** Symptoms plus elimination: plugs and coils ruled out, misfires or rough cold starts that survive them. A borescope look through the port is the only direct evidence.
• **Why walnut shells and not a chemical pour-in?** The factory procedure is mechanical blasting — baked DI deposits do not dissolve meaningfully in a tank additive. Chemical service exists in the market; the data behind this page is the blast procedure.
• **Do I need to remove the cylinder head cover?** The N54 procedure lists it as preliminary work, along with the plenum. Plan the weekend accordingly.
• **Does port injection eliminate this job?** It largely funds the valves' wash again — which is why S55 owners with factory port injection and aftermarket PI kits (see our port injection guide) stop scheduling it.