Unit 12 Chiltern Business Village, Arundel Road, Uxbridge, Middlesex, UB8 2SN

Tel: +44 (0)1895 255699

[Rolling Road] [Turbochargers] [Exhaust Emissions] [Chips]

Rolling Road

I know what the rolling road is for but how does it work?

For four wheel drive cars it has two rows of rollers which support the car. The front roller of each set is connected to a huge electromagnet. Think of it like a massive disc brake but with magnets not brake pads.

With the car resting on the rollers and the engine running, 1st gear is selected, the clutch is engaged and the rollers start to turn. (The rear of the two rollers on each set measures wheel speed.) If at this point the accelerator is pressed and more power is transmitted through the tyres there is no resistance so the rollers accelerate rapidly. This is just like nailing the throttle in neutral.

So we can try to “brake” the power of the engine, hence the term Brake Horse Power or BHP. If we can match the braking force of the rollers to the power being transmitted through the tyres we will create a torque reaction in the electromagnetic brake (or brake caliper). The amount of energy required to balance the torque transmitted by the car is then measured. The computer then looks at torque output and revolutions of the wheel (recorded by the other roller) and displays power transmitted at the wheel. There is another set of roller output figures on four-wheel drive cars so both figures are added together. This gives us power at the wheels.

How do you come up with flywheel power then?

The test is carried out in fourth gear. (This is because 4th is usually direct on most cars so gives the least losses). We dial no load into the rollers and get the car rolling in 4th at 1000 rpm. Then the throttle is opened fully and resistance is applied to match engine power. As engine power increases (when the turbo comes in) the resistance is increased, just to provide a steady smooth rate of acceleration. When maximum power has been shown and power is decreasing the transmission is disengaged and allowed to free-wheel on the rollers. This means the huge amount of energy built up into the rollers is reversed and they drive the transmission up to the clutch.

There is then a negative torque reaction and the amount of power required to drive the transmission is added to the power previously measured at the wheels. Simple eh?

Err…kind of.

We are not claiming it is the world standard for measuring power but the German manufacturer of the equipment has been making chassis dynos since the war and it is fast, repeatable and accurate.

There is no tyre slip, because we can measure it if it happens, it is not hard on the engine, because the test only takes less than a minute and we have been doing it for years.

Give us a non-turbo car we have never seen before (turbo cars are affected much more by atmospheric conditions, heat build up and inaccurate boost settings) and we will give you a power output figure which we know will be within 1-2% of the manufacturers claimed power figures.

The only people who criticise rolling roads are tuning companies who don’t have one. These companies know that they cannot do as good work as people with them so they say they are unsafe, inaccurate or dangerous.

Would we have spent more than £120k on two of them if they were useless?

[Go to top of page]



A turbocharger is a gas driven gas pump — it utilises gases to pump around other gases.

It does not provide free power as has been claimed.

The idea is to use the exhaust gases to drive a compressor which compresses the air going into the engine — making it denser, which means there are more oxygen molecules per unit of volume — so it makes a bigger bang when ignited with more fuel. Hey presto! More power. One way of looking at it is that a turbocharger enables a 2 litre engine to burn as much air and fuel as a normally aspirated 3 litre engine, or even more.

Exhaust gas which is being pushed out of the cylinders is forced through a restriction and round a snail-shaped housing called the turbine housing. The snail shape causes a speed increase as the gas turns through 360°. At the centre of the turbine housing the gas spins a turbine wheel. This wheel is a high temperature steel alloy on 95% of all turbocharged cars, but can be a lightweight ceramic as used on the Nissan Skyline for example. This turbine wheel is connected via a shaft to the compressor side of the turbo. This is inside another housing called a compressor cover. The compressor’s job is to suck in air through the air filter and accelerate it and compress it before forcing it into the engine. This creates a positive pressure (i.e. a pressure greater than atmospheric pressure which in it’s natural state is 1 bar or 14.7 psi). However before the turbo spins up to speed the compressor acts as a slight hindrance to the flow of the air coming into the engine. The size and shape of the turbine housing, wheel, shaft and compressor cover dictate airflow in and out of the engine, and therefore power and response.

The shaft between the turbine and compressor runs in a bearing housing which is water cooled. There is not actually a bearing in the bearing housing but a guide channel through which oil passes. This pressurised oil becomes the bearing and takes up the clearance between the shaft and the housing. This is why when you feel the shaft movement on a turbo when the engine is not running it always feels as if the shaft bearing is worn out.

The bearing housing on the Impreza (and 98% of all turbo cars on the road today) is water cooled. Water cooling of the bearing housing is something developed in the ’70s when oil was not as good as it is now. It was developed to keep the bearing housing cooler after engine shutdown and avoid oil burning on the shaft. It is still used today by all manufacturers but these days modern oils do not burn on the shaft even after hot shutdowns. But it would be a brave manufacturer who admitted they had fitted a non-water cooled turbo to their latest super, quad turbo, five valve, V12, tarmac eating stonkermobile.

If you are using the car extremely hard then just allow a few minutes for it to cool down or use a turbo timer. It is not vital but it helps.

So this is a turbocharger.

If bigger turbos make more power, I want a bigger one.

Bigger turbos also mean more lag.

What is lag?

Lag is the time taken for the exhaust gas to spin the shaft fast enough for the compressor wheel to start to generate some positive pressure. Bigger turbos take longer to spin up so there’s more lag.

Great, now I want a smaller turbo.

Smaller turbos spin up faster so there’s less lag but they flow less air so deliver less power.

Now I’m completely confused, I think I’ll take the damn turbo off.

No don’t do that they are great fun. It is just that the rules governing them are not straight forward. The game engine designers play involves balancing exhaust gas flow with compressor air flow volume. Add to this equation shaft speed, shaft weight, compressor and turbine housing flows and turbine wheel tip speeds and you can see how difficult things are. (And I haven’t even mentioned inlet and exhaust cam design, exhaust primary and secondary pulsed extractor lengths, variable turbo inlets and wastegates!)

So what is a wastegate?

A turbocharger has a very limited shaft speed band during which it is efficient. But an engine works over a huge range of gas volume output (rev range to you sir).

Consider a diesel turbo for a truck. Most diesel turbos are matched to the engine flow. Therefore they get up to their efficient rpm and start pumping positive pressure. Then as engine revs rise so does boost pressure so roughly speaking max boost is at max revs. No problem on a diesel because they have such a limited rev range.

Take your Impreza. You want boost at low speed and boost at high rpm too. So the turbo is designed to spin up quickly and provide positive pressure at say 2500rpm. It will then make more boost until the turbo shaft reaches maximum speed at say 3500rpm. At that speed if the turbo goes any faster two things will happen very close together. Firstly the turbo will overspeed and start to explode (the tips of the blade will actually go supersonic and the resulting sound shock wave will start to break the tips off of the turbine wheel!), secondly it will generate rapidly rising boost of 2-3 bar (30-45 psi) at which point your engine explodes too.


So instead a valve opens inside the turbo which allows exhaust gas to escape without going around the turbine housing. This slows the turbine wheel and reduces boost. Then the valve closes and more gas goes through the turbine housing therefore increasing boost until the valve opens again, etc, etc.

This wastegate valve is controlled by a pneumatic actuator (although the Ford Focus World Rally Car has a hydraulic one) which itself is controlled by an electric solenoid valve which is controlled by the ECU which is controlled by the software which can be re-programmed by Power Engineering. Clever eh?

So do you understand about turbochargers now?

No, but please don’t explain any more.


[Go to top of page]


Exhaust Emissions

Love ’em or hate ’em we all make ’em.

When combustion occurs in an engine certain by-products are produced. Here are a few.

  • Carbon monoxide
  • Carbon dioxide
  • Oxides of nitrogen
  • Hydrocarbons
  • Oxygen
  • Water

There are some others too. Carbon monoxide is the one we are most interested in. Please no e-mails from physicists criticising:

Carbon Monoxide (CO) we measure as a percentage of total exhaust emissions to gauge the air fuel ratio. The higher the CO figure the more fuel is in the mixture and the lower the figure the more air is in the mix. Catalytic converters convert CO to CO2 by combining it with the oxygen not used in combustion.

If the CO percentage is too low the combustion burn temperature will rise and cause engine meltdown under certain conditions. If the CO is too high then power is lost very quickly. Broadly speaking 4% CO and below is too lean for safety on a turbo car and 8% and above is too rich and will lose power. An Impreza 22B we had on the rolling road ran over 10% CO (off the gauge). When we reduced it to 7% CO it increased power by 20 bhp.

With you so far. But if you rely on CO% to measure safe engine running and a catalytic converter converts CO to CO2 how do you know if an engine is running to lean?

Good question, clever dick! Now we introduce lambda sensors.

For a cat to work (light up) it has to see a correct and accurate mixture. It then triggers a chemical reaction and converts bad gases to nice clean friendly ones. So that the cat sees the correct mixtures a sensor measures the mixture as it comes out of the engine. If it is too weak the ECU adds a tad more fuel, if it is too rich then the ECU reduces fuel mixture. This is fine until we want to make some proper power. If we were to run the engine at full chat at the mixture that a cat requires to work it would go bang. Therefore when the loud pedal is at a certain position (usually mashed into the carpet) a command in the software says “lets put some proper fuel in and make some power”. Then the cat gives up and stops working and then we can measure CO%. All cars are like this. It is a fact that cats only work when the engine is on a constant throttle opening or idle. When you are driving like you damn well should the cat is doing nothing. True of a Nissan Micra or a Nissan Skyline.

Other exhaust gases can be used to tell us different things like combustion chamber efficiency, air leaks and cam profiles.

Is that clear?

I understand all that but why is an emission test carried out at idle and not under any load conditions when true bad emissions would show up?

Search me. Ask a politician.

[Go to top of page]



What is a chip anyway?

A chip is the common name given to a device within a car’s computer which stores data about how the car is designed to run. The verb “to chip” is new and is the common phrase used to mean “to make a car go faster”.

Almost anything electronic has chips in it. Your car has many chips but the ones we are interested in store data which is received from signals from the engine (water temperature, throttle position, revs etc) and outputs control signals to things like fuel injectors, ignition coil and wastegate solenoid valve.

You may have heard of an EPROM. This is a memory chip which is erasable and re-programmable. However Mitsubishi use a processor chip with the operating system and the data memory stored inside. These tend to be faster and have larger memory tables. They are also more expensive and extremely difficult to get inside.

OK, so how do you change what’s inside?

It is different for each version but we use an additional circuit board inside the ECU which we load the programming information into and modify information from there.

Why is the re-programmed ECU so expensive?

There are different models of Evo and they have slightly different programme requirements, not to mention the exhaust and other modifications they might be fitted with. We feel the best way is to individually map each car. This means setting up every car on the rolling road, measuring all the ignition and fuel parameters and making a custom programme for each car. Also the work involved in fitting and modifying the ECU is very difficult and requires specialist electronic equipment.

Why are Power Engineering’s ECUs better than anyone else’s in the world?

Because we individually set up each car. We set the car up for UK unleaded fuel and we really understand what is going on inside the software.

What do other companies offer for chipping?

The most common conversion carried out is a small additional board which interrupts the signal from the air flow meter and electronically “clamps” it at a value which is lower than the factory boost limit. This means the boost pressure can be raised using an air bleed valve on the turbo, and you get more power.

So what is wrong with that?

The ECU sees input signals saying the engine is making say 1 bar (14.7 psi) boost pressure, so it calculates the ignition and fuel requirements for 1 bar using the maps in the chip. But the engine is set up to make 1.2 bar (17.5 psi) boost so that’s 20% more air coming in. It also requires completely different ignition mapping for 1.2 bar than 1 bar because combustion chamber pressure will be considerable higher. Result: weak mixture and advanced timing.

Is that all?

What do you mean “is that all”? If that is not bad enough for you then I will explain more. Because the boost is controlled by a bleed valve and not by the ECU, the boost setting will alter with air temperature and pressure. So you will have more boost on some days and be even closer to engine failure. Also there are various safety features within the programme which under certain conditions will lower the boost. The ECU cannot do this if there is a bleed valve fitted.

So how does the Power Engineering re-programmed ECU make more power?

Within the programme is a sophisticated boost control system which looks at air temperature coming into the engine and outside air pressure. It then calculates the boost pressure required to match another map within the programme. If it cannot match the boost it will increase until the two match. We can alter these two maps to safely increase boost.

OK, but you said more boost means weak mixture and incorrect ignition timing and big bang.

That’s right. This is why we carefully re-map both fuel and ignition settings for each car. When you raise the boost, airflow increases and sometimes the airflow is greater than the scale of the maps used. So we often have to re-scale the maps to allow the engine to read completely from zero boost to the new maximum boost settings.

You’re losing me.

Sorry, but you did ask and it is not an easy subject. Do you want bull or facts?


[Go to top of page]