Static flow meter or dynamic bench: how to test a variable-geometry turbo

It happens more and more often: a variable-geometry turbo overhauled by the book makes the engine run badly once fitted. Little boost, high consumption, sometimes the engine light. The reason is that today engine performance relies on the turbo far more than in the past, and the turbo has to stay within a narrow tolerance band. Where a visual check used to be enough, because the engine happily accepted a turbo slightly out of spec, today you need an instrumental measurement of its real characteristics.

Why the ECU has become so demanding

The ECU runs the engine with maps or mathematical models that link the compressor characteristics (flow and pressure ratio) to the exhaust gas flow and the geometry position. Any difference between the turbo the ECU expects and the one fitted misleads it: a geometry setting error, but also a wheel or housing with wrong dimensions.

It is no coincidence that the geometry lever, once moved by a simple pneumatic actuator, is now driven by an electronic actuator or a pneumatic one with position feedback: vane position has become so critical that it no longer tolerates the imprecision of the old system. For the same reason, empirical settings made by looking at the gap between the vanes at minimum opening or the vacuum in the actuator at end of travel are no longer acceptable. And the actuator must be set again every time the turbo is assembled or reassembled.

The static flow bench (flow meter)

The geometry is an obstruction to gas flow in the turbine housing: for a given pressure difference across the turbine, flow depends on the vane angle. Measuring flow at a known pressure difference therefore tells you the geometry opening, and you can set the actuator so that each command matches the right opening.

It is a valid tool and, for a turbo identical to the original equipment, a correct actuator setting is enough for it to respond as the ECU expects. The limit lies precisely in that “if”: a static bench guarantees only the consistency between actuator command and vane opening, and only when there is no leakage. It says nothing about how much air the turbo compresses.

The dynamic flow bench

A dynamic bench makes the turbo work as in the car: besides relating exhaust-side flow to the actuator command, it measures rotor speed and, above all, the pressure ratio achieved on the intake side, at different geometry positions. It is the only check that guarantees the turbo works correctly, because it measures what the ECU expects to get.

An extreme example makes the point. A turbo assembled with a smooth compressor wheel, with no blades, would compress nothing: the engine would run as naturally aspirated and the ECU would log a fault. Yet on a static flow meter it would be in spec, because the turbine side is perfect. A compressor wheel with the wrong profile has the same effect on a smaller scale: less pressure than expected, poor efficiency, high consumption. The same applies to turbine-side housings and wheels with wrong dimensions or poorly matched.

Leakage: the 0.1 mm test

A variable-geometry turbo performs so well because the geometry nozzles direct the gas onto the turbine at the right speed and angle. But not all the gas goes through the nozzles: high temperatures and dissimilar materials require clearances, and part of the flow leaks around the geometry ring and crosses the turbine without doing work. On new turbos this leakage is limited, and manufacturers increasingly complicate their designs precisely to reduce it (see for example IHI Engineering Review, vol. 44 no. 2, 2011). In a rebuilt turbo it remains a hard-to-predict variable.

To measure its effect, Cimat ran a test on its own Turbo Test dynamic bench. A turbo assembled with all the correct parts was set and measured: turbine-side flow, pressure ratio and speed were in spec. Then a 0.1 mm shim was placed under the geometry rollers of the same turbo, simulating a dimensional defect in the rollers alone, and the geometry was set again to spec.

Turbine-side flow is practically identical: to a static flow meter the two turbos are the same and both in spec. The dynamic bench shows instead that, with rollers a tenth of a millimetre higher, the pressure ratio drops to 1.225 against 1.31 at medium geometry and rotor speed falls by more than 12,000 rpm: values well beyond the tolerance the engine accepts.

What else a dynamic bench checks

• Oil sealing under conditions close to real operation.
• Whistling, which can be identified before the turbo reaches the car.
• Assembly errors that make the wheels rub on the housings.
• Electronic actuators, powered at 12 or 24 V and driven as the ECU would, with programming where applicable.

The comparison

How to choose

A static bench remains a good tool for setting the actuator correctly and should be considered the minimum for anyone working on variable-geometry turbos. But if you rebuild modern turbos, use aftermarket geometry or wheel parts, or want to be able to guarantee the result to your customer, the only complete check is the dynamic one: it measures the turbo the way the ECU will see it. Cimat offers it as a standalone bench, the Turbo Test, or combined with the high-speed balancer in the Turbo Test Duo stations: you will find them among the turbocharger equipment, for which we are Cimat’s exclusive dealer in Italy.