Firstly, we should clarify the term "chip". This is a bit of a misnomer as most replacement chips for the WRX are in fact entire computer circuit boards with one or more chips. Thus a more correct term should be ECU or electronic control unit. Some of these ECUs completely replace the Subaru factory ECU in a plug and play fashion, whilst others splice into the wiring loom and work in conjunction with the factory unit. The latter are known as piggyback style ECUs (at least in the context of this article). 

When considering an after market ECU, the question could be asked as to why Subaru don't optimise their factory ECU to gain maximum performance anyway ? One answer could be that Subaru sell the WRX across many markets including Asia, Europe, Australia and South Africa. Each of those countries has a different climate and varying fuel quality. These two factors can radically effect the performance of the WRX. As such, Subaru tune their ECU so that it runs reliably and safely no matter what the conditions are like. Thus it is possible to improve the cars performance by tuning a new ECU to suit the conditions of the country you live in. 

Tuning 
Tuning basically involves changing the ignition timing, air to fuel ratio and boost pressure. Each of these parameters has its own map, which is simply a grid of rpm vs throttle position (or load). Some ECUs have more map points than others. The Microtech ECU (for example) has 16 load positions and 16rpm positions. This gives a total map of 240 different positions that can be set for each map. The Unichip has 12 load and 17 rpm sites for a total of 204 map points. A good ECU has narrow zones down low (say every 200rpm) for good driveability whereas up high wider zones are fine (say every 500rpm). 

Getting the best from your engine is about adjusting the maps to optimise the power output across the rev range at each throttle position. This is why tuning a car requires a lot of time, energy and experience. Simply leaning the fuel off may not get more power and in fact, you may get a pinging car. Ignition timing can be varied by a few degrees so that if the car is detonating, timing can be pulled out to try and stop it. This is an important safety feature of any ECU. 

When selecting a tuner to install and tune an ECU in your car, it is important you find one that is both reputable and experienced. There are plenty of tuners out there who may be able to produce big kw gains for you but don’t expect your engine to last too long. It is also important to find a shop that already has a basic map of the ignition and air/fuel ratios for the model car you have when fitting an ECU. Otherwise the tuner will have to build these maps from scratch. This will require a lot of dyno time for your car which isn't good.  

Some ECU manufacturers sell hand control units allowing the user to modify the settings. Whilst this may appear cool to your mates, you can do some serious damage to your car by fiddling around. Even a well trained dyno technician with years of experience and a full workshop (and 4WD dyno) at his disposal can struggle to map and tune engines correctly. An end user really has no hope of getting a good result from DIY tuning. 

One final point to remember is don't believe the marketing guff issued by any ECU supplier. The tuner is the key to getting your car right. Any ECU is only as good as it is tuned and a good ECU can be tuned badly in the wrong hands.  

Air Fuel Ratios 
The AF ratio is an important factor affecting performance and it is this area most tuners spend their time on. Note this ratio can also be referred to as lambda. The higher the ratio, the leaner the car runs and more prone it is to pre-igniting (or detonating). Naturally aspirated cars can run ratios of 13 or 14:1 (and even higher) under full throttle but turbo powered cars need to run lower due to the higher air temperatures being generated by the turbo. The factory ECU is tuned to run a relatively rich mixture in order to promote cylinder cooling and longevity. Tuners can set the car to run slightly leaner however without compromising these factors too much. 

Running lean improves fuel consumption and, at part throttle or cruise conditions, ratios of up to 14:1 can be run on the WRX without danger of detonation. At full throttle however, the AF ratio should be decreased (or made richer) to avoid detonation. Discussion with several specialist WRX tuners in Victoria, NSW and the UK proved interesting in that they all agreed the optimal AF ratio for the WRX at full throttle is about 12:1 for boost over 14.5psi. For lower boost, an AF ratio of 12.5:1 is acceptable and for cruise mode 14:1 is fine.  

A snapshot of the dyno parameters for the MY99. You'll notice Subaru use relatively rich mixtures in order to stay well clear of any detonation problems. 

Snapshot parameters for the MY98. 

Detonation 
Also known as pinging, knocking and a host of other names. This occurs when the air/fuel mixture in the piston chamber detonates before the spark plug fires and before the piston has reached it’s max compression (or upward) cycle. This results in the piston head being hammered back down causing major stress on the bearings and rings. It is this stress that results in the pinging sound. 

The air /fuel mixture detonates early because the air in the chamber is too hot or the fuel is too lean (and hence very volatile). The air may be too hot because of inadequate cooling or boost pressures that are simply too high (increased pressure means increased temperature). The fuel may be too lean because of incorrect mapping by the ECU or poor fuel quality (ie. low RON). The mixture may also detonate because ignition timing is too advanced. 

The knock sensors are designed to detect early detonation and retard timing if necessary. If the pinging is too severe however, they cannot retard enough and pinging will still occur which will eventually result in a blown piston. 

Dynos 
It is important to understand that dyno figures can be effected by many factors and it is important to ensure these factors are the same before comparing dyno charts with your friends. Bleating yours is bigger than mine is useless if we are comparing centimetres and inches. 

For example, variations in power can occur between supposedly identical dynamometers due to the use of different acceleration rates. The acceleration rate is adjustable because not all engines are powerful enough to allow the use of just one acceleration rate. A 60kw engine could not use the same rate of acceleration for graphing purposes as say a 220kw engine. The inertia factor can also be adjusted as different cars have different levels of inertia due to things such as wheel size and weight. Another important factor is the gear used during testing. A run in 4^th gear will not produce the same results as a run in 3rd (which will give a higher reading). The bottom line is that you use the same settings on the same type of vehicle if you wish to compare power graphs and have a degree of repeatability.  

It is also important to understand that dynos can only measure power and tractive effort (which is not equal to torque but similar) at full throttle. They cannot show gains made under part throttle. This can be quite important as part throttle response can change the character of the car significantly. Many people tend to get hung up on peak power dyno figures but how often would they drive with the foot to the floor ? Jokingly they will say "always", but in reality traffic will dictate otherwise. Thus it is the part throttle response that really shapes the character of the car. It is this power that must be assessed on the road as dynos are incapable of measuring this.  

Dyno curves can also vary by a few percent for the same car between runs. Indeed, my car was found to vary by a few kw from run to run. Things like temperature, atmospheric pressure, humidity and even wheel diameter (I had to change from 16 to 17" at once stage due to a puncture) can effect figures. Thus dyno figures should not be taken too seriously but merely used to judge relative improvements in the car only.  

Note : The cars tested within this article were only tuned and run on the APS dyno. This ensured all figures were comparable for each ECU. 

Warranties & EPA Regulations 
Whenever modifying your car from the standard factory build, you should always check the effect on your Subaru warranty and the legalities with the Road Traffic Authority (RTA). It is pretty much up to Subaru’s discretion when it comes to fixing modified cars. Basically if the mod killed the car, don’t expect them to fix it. Replacing the factory ECU with an after market one won't cause you problems if your door handle falls off, but might if you pop a piston. 

Changing the air/fuel ratios can effect tailpipe emissions. These emissions may breach the RTA regulations for your region. A difficult one to get caught for but you need to be aware of this all the same. Many tuners won't guarantee their ECUs for such offences so you should ask before purchasing. An emission test is in the order of $1,500 so is not something tuners can afford to do for every car they tune. All modern cars have an oxygen sensor so the ECU can monitor exhaust emissions and keep them within RTA spec. If your tuner is not using this sensor, you should be asking why.