TD03 vs TD04
center sections

A factory N54 runs twin Mitsubishi TD03-frame turbos: 40 mm inducer turbine wheels on journal bearings with a single factory thrust assembly. Most upgrade paths move to a TD04-family turbine — bigger, and often paired with a different bearing system entirely. Below, every part of that swap is opened up and graded for where the number comes from — manufacturer-published, independently measured, calculated, modelled, inferred or unknown — because most of what gets repeated about “TD04 handles more boost” is a conclusion, not a spec.

Everything visible from outside a turbo — the compressor housing, the turbine housing, the wastegate can — is the part that gets photographed. The center housing between them is where the actual engineering happens, and it is invisible until it is cut open. This is a schematic cross-section, not a photograph: real geometry for most of these parts is not published for most of the vendors compared later on this page.

Representative turbocharger hardware · reference photo

Turbine wheel

Representative turbocharger hardware · reference photo

Turbine housing · empty

Representative turbocharger hardware · reference photo

Compressor housing · empty

Representative turbocharger hardware · reference photo

Internal wastegate flapper

The clearest, best-documented difference between a TD03 and a TD04L turbine is simply size. Using the factory TD03 turbine (community-measured) against VIV’s own published TD04L dimension:

The turbine wheel is the easy, well-documented part of this swap. The center section underneath it is where the real architecture change happens — and where most vendors publish the least.

Representative turbocharger hardware · reference photo

CHRA · assembled

Representative turbocharger hardware · reference photo

Oil-feed banjo fitting

Boost pushes back on the compressor wheel; exhaust pressure pushes back on the turbine wheel. The net of those two forces is an axial (fore-aft) load on the shaft, and the thrust bearing's entire job is to react it without letting the shaft creep far enough for a wheel to touch its housing. This is widely reported as the single most common failure point on a hard-boosted stock N54 turbo — more common than journal wear, overspeed or housing cracking.

Use the thrust-load mode in the cutaway above, or the simplified calculation here: a first-order estimate of axial force is F = ΔP × A, pressure difference times the wheel's back-disk area. It is not the manufacturer's certified net-thrust figure — that also depends on balance-piston and seal geometry no vendor in this comparison publishes — but it shows why the load scales directly with boost, not with some fixed “25 psi rating.”

Representative turbocharger hardware · reference photo

Sectioned shaft & bearings

A journal bearing is not lubricated the way a door hinge is — it rides on a pressurized wedge of oil thick enough that the shaft never actually touches the bushing under normal operation. That film depends on three things staying healthy together: adequate feed pressure, a drain that can keep up, and oil that has not overheated past the point of breaking down.

The oil-flow mode in the cutaway above traces the path: one pressurized feed splits to both journal bearings and the thrust bearing, then all of it has to find its way back out through a single gravity drain. That asymmetry is the important part — the feed side is pressurized and can push through a partial restriction, but the drain side cannot. A drain that is even partially blocked backs oil up into the housing long before the feed side shows any symptom, forcing oil past seals it should never reach. This is why more oil pressure is not automatically the fix for a bearing problem: pushing more oil into a system that cannot drain what it already has makes the backup worse, not better.

Bearing clearance interacts with all of this directly. Too tight, and the film can't build enough thickness to stay hydrodynamic under load — metal-to-metal contact follows. Too loose, and the shaft has room to move radially before the film can react, which shows up as audible shaft play and, eventually, wheel-to-housing contact. Neither OEM nor any vendor in this comparison publishes journal clearance figures for the N54, so treat specific clearance numbers you see online as unverified.

The same real N54 hardware from the 19T geometry article, re-read for center-section specifics. “Not stated” means the vendor does not publish it — that is itself information about how much you can verify before buying.

Modelled / illustrativeA note on thrust-collar width: some builders describe upgraded N54 CHRAs running a wider collar or spacer — sometimes cited around 14 mm versus a narrower OEM-style part — as one reason they tolerate higher axial load. No vendor above publishes an exact figure, so this article marks it unknown rather than repeating an unsourced number. The cutaway and 3D comparison above draw a wider collar only to illustrate the general pattern, never as a measured dimension.

None of these are scare tactics — they are the mechanical chain from cause to consequence, so you know what a symptom actually means when you see it in a log or on a lift.

Here is the calculation that makes this article's title true. Twenty-five psi of boost is a pressure ratio of about 2.7:1 — full stop, regardless of RPM. What is not fixed is how much air the engine needs moved to hold that ratio, and that scales almost directly with RPM:

Going from 3,000 to 5,000 to 7,000 rpm at the same 25 psi target means moving about 26.6 lb/min of air, then 44.3 lb/min, then 62 lb/min — roughly 2.3× more from bottom to top of that range. More intake mass flow means more exhaust mass flow behind it, which means the turbine has to do more work to hold the same pressure ratio. That additional work has to come from somewhere: higher shaft speed, higher drive pressure, or both.

A boost number without RPM, airflow and turbine-side conditions tells you very little about actual turbo stress. Two cars logging an identical 25 psi peak can be putting completely different loads on their thrust bearings, journals and oil film, depending entirely on where in the RPM range that boost is held.

This is the most commonly misunderstood number in turbo tuning. A boost gauge reads gauge pressure — pressure above atmosphere. A compressor's actual job is described by pressure ratio: absolute outlet pressure over absolute inlet pressure, atmosphere included on both sides.

At standard sea-level atmosphere (14.696 psi absolute), 25 psi of boost is a pressure ratio of 2.7:1 — not 25:1, and not even close. That distinction matters because compressor maps, efficiency and choke limits are all plotted against pressure ratio, not gauge boost. Two cars in different climates or altitudes chasing the same boost number are actually asking their compressors for slightly different pressure ratios.

Boost is only half the pressure story. On the other side of the engine, exhaust manifold pressure (EMP) — the pressure the engine has to push against to exhale — can matter just as much as intake manifold pressure (MAP) for how hard a turbo, and the engine behind it, is working.

As EMP rises relative to MAP: pumping losses increase, more hot residual exhaust stays trapped in the cylinder instead of being scavenged, exhaust gas temperature climbs, thrust load on the turbine wheel increases, and the engine has to work harder just to maintain the boost target it already reached. None of the specific ratio thresholds above are universal cutoffs — they are a general shape, not a certified limit, and no Synergy EMP-instrumented log exists in this article's sourcing to turn them into measured figures. If you have logged EMP against MAP on an N54 and want it considered for a future update, send it our way.

Get these in writing from any center-section vendor before you buy:

• Bearing type — journal or ball — and, if journal, the diameter if they'll state it.
• Thrust bearing design and whether it's a named upgrade or the OEM-pattern part in a new housing.
• Oil-port configuration: feed and drain sizing, and whether the OEM drain is reused unmodified.
• Turbine wheel inducer and exducer diameter, and its heat-shield height if changing turbine families.
• Turbine housing size, and whether it's quoted in A/R or cm², and by whom.
• What RPM range and boost level the power claim was actually made at.
• Whether independent EMP or drive-pressure data exists for the specific product, not just a dyno number.

A “TD04” badge is a starting point for that conversation, not the end of it — see the 19T geometry article for the same lesson applied to compressor sizing. If you want a second set of eyes on a specific turbo against a specific power and RPM goal, that's what a build consult is for.