Choosing an Aftermarket ECU The fitment of an aftermarket ECU is a significant step in the process of building your project car. It marks the end of all limitations of the factory ECU and EFI system, and the beginning of your tuning freedom. When chosen correctly, the aftermarket ECU you select will stay with your build process and offer the flexibility to adapt to your latest engine changes. When chosen poorly, it can become a constraint which limits your future modifications. So it’s important to select the most appropriate one in the first instance. This article will advise you on what features you need to look for when choosing an ECU, and will inform you on how to make the best decision.

When do you need one? The first question, before even how to select the best ECU for your car, is, how do you know when you actually need one? And the answer is: when your factory ECU becomes a constraint in your engine modification project. How do you know when your factory ECU is a constraint in your engine modification project? The answer to this is: when it’s not doing what you want it to and you can’t adjust it! What do you want your factory ECU to do? Ideally it will give you the air-fuel ratios that you want, the ignition timing that you want, and so on. In practice, most factory ECUs are not adjustable, which means that once you increase the boost substantially above the factory level, the original EFI system is not able to cope. This is not only limited to the factory ECU, but will likely also include the factory fuel injectors, fuel pump, ECU and airflow meter. The symptom will be that the engine runs leaner than you would like at the high boost condition, as the fuel injectors are already letting through all the fuel they can. Unfortunately, simply fitting larger injectors or increasing the fuel pressure isn’t sufficient to solve the problem, because this will also increase the fuel delivered under idle and light load. This means terrible fuel consumption, lots of black smoke, fouling plugs, and perhaps most importantly on a street car, your car is now illegal as it certainly doesn’t meet emissions requirements. So fitting an aftermarket ECU allows you to control the larger injectors, so that the car can still meet emissions requirements, start and idle easily and so on. In many cases, the actual air-fuel ratio is not critical to power production, but running at stoichiometry (the theoretical ratio of air and fuel required for complete combustion) at idle and cruise is very important for emissions compliance. Furthermore, most factory ECUs do not allow you to adjust the ignition timing. Ignition timing adjustment is crucial to obtain the maximum torque (and therefore power) that the engine can produce, but advance the ignition too far and knocking (pinging) will likely become the limiting constraint, especially on a turbocharged engine. The amount of ignition timing that an engine needs to produce maximum torque (and how much timing

it can handle before knocking) varies with both engine speed and engine load, and to a lesser extent with octane rating of the fuel, and engine and air temperature. Therefore, an overall ignition timing adjustment such as rotating the distributor, which affects the timing by a fixed amount under all conditions, will result in as poor a result as increasing the fuel pressure to correct a lack of fuel at high boost. In some cases, we have seen manufacturers of aftermarket turbocharger kits recommend retarding the ignition overall by adjusting the position of the cam angle sensor, so counteract the fact that the factory ECU will not retard the ignition timing under boost. As a result, the off-boost performance and midrange torque is severely compromised due to poor ignition timing. In addition, having the ignition timing too advanced produces high NOx emissions, which is not good news for emissions compliance. In addition, there are many other features and benefits of aftermarket ECUs that may make an aftermarket ECU seem more favourable other than fuelling and ignition timing control. These can include: • Ability to control other functions, such as boost control, intercooler sprays and so on • Additional engine control functions such as launch control, traction control, geardependent boost and antilag • Ability to change other engine hardware, for example a different (larger) throttle body with a different idle control valve, changing from a distributor to coil packs or coil-on-plug, changing from batch or semi-sequential injection to fully sequential injection • Ability to log and analyse performance • Ability to add additional sensors to help diagnose weak spots in your system

Are there alternatives to aftermarket ECUs? Firstly, an aftermarket ECU is not the only way to solve the issue of fuel and ignition timing tuning. Several recent factory ECUs can be “reflashed” or reprogrammed by the diagnosis port. The level of control available for each OEM ECU depends on the ECU itself, including the market (for example, the Japanese version may be reflashable but the Australian version is not), and the product available for the reflashing. The GM LSx series engines, for example, have excellent aftermarket support in the form of VCM Suite. This system allows changing of injector size, removing the airflow meter and many other advanced features. A ROM emulator is a different solution. With this system, you must remove (or somehow disable) the ROM chip from the factory ECU. A new PCB is added into the factory ECU, which contains a circuit that emulates the ROM chip that you have previously removed. This new circuit then contains all the data which was on the original ROM chip, and a laptop software program allows you to adjust some of the settings such as fuel and ignition maps. In some cases they allow adjustment of other variables such as rev limiters, however in many cases they do not or are limited in what they do. The development of these is based on reverse engineering the original code in the factory ECU, and in many

cases not every function of the factory ECU is understood, meaning that some limitations remain. In particular, such systems often require use of original sensors on the engine. An “interceptor” is a third solution type. This solution also requires the factory ECU to be retained, and works by adjusting the sensor inputs to the factory ECU. This means that not only are the original sensors required, but often only a small range of tuning is available. In addition, often the factory ECU will “learn around” the changes that are made using an interceptor. In many cases for small modifications (such as an exhaust or a set of camshafts) they are adequate, but once there are major modifications such as conversion from NA to turbocharged, the limitation of the adjustment becomes a major constraint. Lastly, we have aftermarket ECUs. Whether wired in piggyback or standalone, they offer full control of all parameters on the engine, when chosen and installed correctly. The right ECU, correctly set up and tuned, will start and drive as well as a factory system. The compromises that are made with adjusting fuel pressure and ignition timing are nonexistent as the system can be tuned to what the engine needs under all its operating conditions. They also give unparallelled control over auxiliary outputs. However, because they do have such a large range of control, they do take longer to set up correctly than the above solutions.

What to look for in an aftermarket ECU There are three main areas to consider when selecting an aftermarket ECU. These are features, performance, and compliance. Features are really “what the ECU does”. You need to ensure that the ECU you select will do what you need it to do for the foreseeable future. I will later present a list of all the various features offered by the aftermarket ECUs, with a short explanation of how each one works and when you need it. Performance is really “how well the ECU does what it does”. This is an area that is a bit harder to determine what is required for your particular engine, but I’ll give some guidelines below. Compliance is all about legalities. If your car is street registered, it will need to comply with the local laws, whatever they are. The most critical of these (ie, most difficult to meet, and most difficult to demonstrate compliance) is emissions compliance. Legal compliance of the ECU product itself is also required for the ECU to be legally sold in many jurisdictions.

Features Below is a list of features of different aftermarket ECUs:

Trigger Compatibility: This is perhaps the most critical requirement of an aftermarket ECU. OEMs usually have good reasons for using the crank and cam angle sensors that they chose to on a particular engine. Many manufacturers have fairly standard systems duplicated across many engines. Great difficulty ensues when you try to change trigger systems on engines, because of the difficulty in mounting the trigger plates, building brackets for sensors, ensuring that they are reliable at high speed and so on. In most cases the OEM trigger solution will be all that you’ll need, and adding a different trigger system will add complexity and further variables. I’d always recommend using an ECU system that will work with the original trigger system on the engine where possible. Furthermore, some aftermarket ECUs require the sensor to be installed in a non-factory orientation, or on a different tooth, from the factory position. This is not recommended; you just know that when the engine needs to get rebuilt later with upgraded internals, and the engine builder puts the sensor back in the standard position, following the manufacturer’s recommendations, the ignition timing will be wrong when he goes to start the engine. Follow the standard systems; use an ECU that will handle the OEM trigger. See also Ignition Accuracy under Performance, below. Tuning modes: There are, by and large, three different load sources used for tuning EFI engines. They are airflow (from a mass airflow sensor), MAP (from a Manifold Absolute Pressure sensor), and TPS (from a Throttle Position Sensor). Airflow gives the most theoretically correct characteristic, as it measures directly the air being consumed by the engine. This, in combination with the desired air-fuel ratio, tells the ECU the amount of fuel that the engine needs. The ECU can then inject this amount of fuel. This automatically compensates for many variables, such as air temperature and changes to the engine’s volumetric efficiency such as a larger turbocharger or different camshafts. Downsides to using an airflow meter include the fact that it is a large physical constraint on the inlet piping geometry (it must be mounted between the air filter and the turbocharger, or throttle body on a NA or supercharged engine), they are expensive to buy when an upgrade is required (normally they are sized appropriately for the OEM engine solution, which means that when you increase the power output significantly, they are no longer adequate, much like fuel injectors), can give inaccurate results when dirty, and also give inaccurate results if there is a pressure gradient across the pipe (for example, if your intake piping requires a right angle bend just before the airflow meter). They also will read inaccurately if any air downstream of the airflow meter leaks out to the atmosphere (for example, via an atmospheric venting blow-off valve or boost controller), or if there is a vacuum intake leak. MAP sensor is the most conventional tuning mechanism for aftermarket ECUs. The placement of the MAP sensor pickup point on the manifold is important; often if the sensor is too close to the throttle body, the sensor’s reading will fluctuate under boost, potentially leading to fluctuations fuel and ignition timing delivery. This feels like a hesitation under boost, and will often give a fuzzy torque and air-fuel ratio line on a dyno printout. Advantages of MAP sensor over airflow sensor is that they are easy to mount, low in cost, and easy to source even for very high boost pressures. One downside is that retuning is required with modifications to the engine’s volumetric efficiency, for example by changing an exhaust system, camshafts or port sizes. Another downside is that unless

there is a large plenum after the throttle body, the air will not “settle” well enough to get a stable pressure reading. In addition, some engines (especially with large overlap camshafts or peripheral ported rotaries) produce so little vacuum that tuning such an engine on MAP is impractical. For these engines, TPS or TPS/MAP combined gives a more consistent fuelling result. TPS is the most primitive tuning method, and takes the engine’s load input from only the throttle position. This is necessarily difficult to tune, as the amount of air drawn in at a given throttle position is highly dependent on RPM, whereas at a given MAP value the amount of air drawn in is relatively constant with respect to RPM. In addition, any other mechanism that admits air to the manifold, without changing the throttle position, will change the amount of fuel the engine needs, for the same engine speed and throttle position. This is often not a problem for race cars, but for street cars with idle speed control this makes TPS-alone tuning completely unsuitable. TPS/MAP combined tuning is the most sophisticated tuning mode, and works excellently for individual throttle body engines, either forced induction or naturally aspirated. In this mode, the tune is essentially a TPS one, however compensation is made within the ECU for MAP. This means that when air gets to the manifold through an idle air controller, the amount of air “seen” by the ECU increases, and the ECU can increase the amount of fuel to suit. It also means that additional fuel will be added as the boost pressure increases on a turbocharged engine. This method also works preferably to the TPS-alone method on large cammed engines for the same reason. Ideally, you would choose an ECU that offers a tuning mode best for your engine. If you aren’t sure whether your engine will be too “cammy” or “porty” to run on MAP, then you would be best off selecting an ECU which can select between different tuning modes. If your engine runs individual throttle bodies, it is recommended to choose an ECU that offers TPS/MAP combined tuning. All the above tuning modes can be used in the traditional mode, or VE mode. VE mode is described below. VE Tuning: The traditional mode to enter fuel quantity in a fuel map is to enter the fuel injector duration, in milliseconds. For example, a “5” in the fuel map would instruct the ECU to hold the injector open for 5 ms, at that RPM and load point. The tuner adjusts this value to achieve the desired air-fuel ratio. The starting point (for a base map) can be estimated from knowing the injector size and engine capacity, or by “feel” if the tuner has done a lot of base maps before. When the ECU does its closed loop fuel control, it adjusts the fuel injector duration to give the desired air-fuel ratio. To change the air-fuel ratio in a closed loop system, the tuner needs to change the target AFR, as well as the value in the fuel map. VE tuning behaves differently, and is based on the volumetric efficiency of the engine. The volumetric efficiency is simply how efficient the engine is as an air pump. For example a 4 litre engine should, in theory, draw through 4 litres of air/fuel mixture for

every 2 crank rotations (or every 1 crank rotation for a rotary engine or other 2-stroke). In practice, it can’t flow 100% of this air, because of restrictions in port sizes, exhaust back pressure, and cam timing. If it can flow 80% of this air, then we say that the VE is 80%. With VE tuning, you simply enter this percentage in the fuel map instead of the fuel injector milliseconds. This allows you to start with 80% all over the map, which will allow the engine to run straight away, rather than having to guess the injector milliseconds (or do it by feel if you’ve already done hundreds of base maps). For the ECU to calculate the fuel injector duration, it must also know the engine capacity and the characteristics of the injector. Some ECUs will simply ask for the injector size and assume fairly linear behaviour from the injector, with the result that the fuel map deviates from the actual VE percentage at higher loads. Other ECUs have a large number of injectors fully characterised, meaning that the fuel map matches the VE closely. In practice, VE helps in that it makes the tuning process much faster than with the more traditional method, which is why most tuners who’ve been exposed to it really like it. However at the end of the day an ECU with a correctly tuned millisecond map will drive exactly as the same ECU with a correctly tuned VE map. Throttle pumps / acceleration enrichment: There are three things that can happen when the accelerator is depressed for the first time: the ECU can deliver a burst on all the injectors at once, the ECU can provide an enrichment for a duration after the throttle is opened, and the ECU can momentarily advance the ignition timing. Firstly, let’s investigate why this is necessary at all. Essentially there are two reasons. When the throttle is opened for the first time and the engine is idling, a large rush of air is admitted to the intake manifold. In some cases, this rush of air creates a positive pressure wave that is momentarily higher than the atmospheric pressure, so the engine needs even more air than it would do at wide-open-throttle at the same engine speed. However the main reason is that the ECU’s reading of the MAP sensor is slow to respond to this air being admitted to the plenum. The MAP sensor itself is usually quite fast; the response time is usually less than 10 ms. However the pressure fluctuations in the manifold mean that the ECU must provide filtering of the MAP signal to give consistent fuelling to the engine. This filtering necessarily slows down the response of the MAP reading. Taking a typical example where the throttle is opened to 15% from idle. On a fairly typical street engine, the MAP may be about 33 kPa at idle (ie, two-thirds of an atmosphere of vacuum). When the throttle is opened, even when only to 15%, this will present hardly any restriction to the airflow at idle speed because the engine is drawing through air so slowly. Therefore, the air pressure in the manifold rises rapidly to atmospheric pressure, or 100 kPa absolute. This is triple the value sensed by the MAP sensor, which means that the ECU must either look up the fuel map to the 100 kPa position, or provide an enrichment of 200% to allow for the additional air. Therefore, there are two ways that aftermarket ECUs handle this; one is through enrichments (such as providing an additional 200% fuel) or through having a table of

“Predicted MAP” against throttle position and engine speed, which allows the ECU to predict the MAP value of 100 kPa, knowing that the MAP sensor hasn’t caught up yet. Even with this accurate model of the pressure in the manifold, this doesn’t solve all transient throttle problems. The inlet valve on a piston engine is open for approximately 180 degrees. Therefore, on a 4-cylinder engine, one cylinder is performing its induction stroke at any given time. The injection is usually timed to complete just before the inlet valve opens, which means that if the throttle is subsequently opened, a large rush of air at higher pressure rushes in to fill up the cylinder. The fuel for that cylinder has already been delivered. Therefore that cylinder is guaranteed to have a lean stroke. To correct this, many ECUs have an asynchronous injection burst, which fires on all the injectors at once, when the throttle is opened. This can either be calculated automatically by the ECU, based on the difference between the “new” fuel quantity and the “old” fuel quantity, or it can be a table, or a set of tables, dependent on throttle position and throttle rate. Transient throttle based ignition advance is another feature which can assist with increasing “pick up”, or how quickly the engine revs up when the throttle is tapped in neutral. It can also be used to cover up inaccurate ignition timing due to changing RPM (see Ignition Tracking under Performance below). I would recommend choosing a system which uses a MAP prediction method rather than enrichments, unless you are in a position to mount a MAP sensor directly on the manifold (no hose) and therefore do not require filtering of the MAP signal. Transient throttle based ignition advance is also a bonus, but in general if the ignition tracking is accurate and fuelling is correct this usually is more of a “bonus” to performance rather than a requirement. Dual maps: Many aftermarket ECUs offer dual maps – that is, two fuel and two ignition maps with the ability to select between them with a separate switch. Often people say that they want this feature because they want “a power mode and an economy mode”. In general this is not required, because the part of the map that you use when you want power (ie, under wide open throttle conditions) is separate from the part of the map that you use when driving for fuel economy (ie, light throttle, cruise conditions). There is a slight benefit to having the light load part of the map tuned slightly richer, in that it can help with transient response slightly, but in general you don’t need two separate maps for power and economy, because you’ll be tuning these separately in your map anyway. There is a benefit to having two separate maps for maximum power and endurance, though. For endurance racing, often you will need more conservative tuning to maximise engine life, whereas for the occasional squirt on the street or drag racing, the tuning can be more aggressive. If you intend to run nitrous oxide, dual maps can be used to allow a “non-nitrous” and a “with nitrous” tune, if a simple ignition timing / fuel trim does not give you sufficient control. Dual maps are useful for dual fuel systems (eg, PULP and E85, ULP and LPG).

Dual maps is a feature that is handy to have in case you need it, but unless you need it, the absence of this feature should not be a deal-breaker. Boost Control: Many aftermarket ECUs can perform a boost control function. In its most basic form, this is simply a duty cycle control to drive a 3-port solenoid valve to set the boost. Ideally though, this duty cycle would be variable against RPM, so that the boost can come on sooner, boost spikes can be eliminated and boost taper at high RPM can be reduced. This can also be used for other tricks such as increasing the boost above the torque peak to flatten out the torque curve and increase top-end power. Ideally this would be variable against throttle position or gear, so that different boost settings can be chosen to maximise acceleration without compromising traction. Having closed loop boost control also assists with maintaining constant boost against other variables such as temperature. My recommendation is that if your engine is turbocharged, or it’s likely to turbocharged eventually, that you choose an ECU with boost control. General outputs: Aftermarket ECUs generally have multiple auxiliary outputs that can be programmed by the ECU installer or tuner. It is very important that you list the ancillary devices that you want to run from the ECU and select an ECU that will allow you to run all of these devices. These would include devices such as: • Fuel pump • Thermofan / condenser fan control • Idle speed (may need 1-4 outputs, depending on the idle valve or motor in use) • Variable valve control (either on/off or continuously variable) • Variable intake manifold flaps • Variable valve lift • Air conditioner control • Intercooler water spray • Water injection pump control • Shift light (s) • Electric water pump control • Drive by wire control (may require 2 outputs depending on the ECU) • ECCS relay (main ignition control) • Tachometer output • Automatic transmission control outputs For each of these outputs, make sure that the ECU can drive the output in terms of the electrical compatibility (current draw and whether it need to drive high or low), and that it can control the logic required. Various ECUs have different levels of configurability over their auxiliary outputs. It’s important to recognise the level of configurability that you need.

Gear detection: Gear detection is performed by correlating the engine speed with the gearbox output speed, either by measuring the tailshaft speed or the average speed of the two driven wheels. This is a relevant feature if you intend to run gear dependent boost, for example. Launch Control: Launch control is a special rev limiter which only applies below a certain road speed. This assists the driver in a take-off situation at the start of a race to give consistent launches by maintaining a certain RPM. This can be a useful feature. Antilag: This feature allows the driver to build boost on a turbocharged engine without applying load to the engine. This is a very useful feature, but can be very damaging to the engine if not configured correctly. This feature is almost essential for drag racing, turbocharged, manual cars. Traction Control: This feature allows the ECU to reduce engine power when it detects wheel slip. This feature requires individual wheel speed sensors to be fitted to the vehicle. Engine power is reduced by ignition retard at first, then for more severe power reductions the ignition can be cut to reduce one cylinder at a time. RPM based VVT / VVL, on-off: Many engines have an on/off switched system for changing the engine’s volumetric efficiency. Some engines have a “high power” set of camshaft lobes. An ECU controlled oil solenoid, when energized, engages the second set of lobes. In other engines, the solenoid, when engaged, advances the intake cam by a fixed amount. Other engines have variable valve lift, variable intake length runners, intake flaps between the runners or additional intake ports. Because the engine’s volumetric efficiency is RPM dependent, the effect of any one of these variables may improve the engine’s volumetric efficiency, but only within a specific RPM range. The easiest way to configure this is via an RPM based switch, but another option is to use a combined RPM and load (eg TPS) based switch. Ideally the switch points would occur where the VE is equal in either position, which means that there should be no discontinuity in the power delivery or in the resulting fuel map. If your engine has either on/off VVT, variable valve lift, variable intake runners or a similar VE optimisation device, ensure that the ECU either has outputs that you can switch on within a certain RPM range, or preferably within a certain RPM range and above a given throttle position. An alternative option is to use a 3D table of RPM vs load (TPS or MAP) to select whether the given output is on or off. Closed loop VVT: More modern engines have continuously variable valve timing, where the actual angle of the camshaft with respect to the crankshaft can be continuously adjusted by the ECU. This is given various names by different manufacturers, such as VVT-i (Toyota), iVTEC (Honda), SVT (Mazda). This can only be performed if the ECU can measure both the crankshaft position and the camshaft position, and perform closed loop control of the cam angle. If your engine has

continuously variable valve timing, ensure that your ECU can control continuous VVT with enough channels for the number of variable camshafts. In addition, you should ensure that you have either a table against RPM for the target cam angle, or a table against RPM and load. You must also ensure that the triggering mode of the engine is compatible so that the camshaft angle can be sensed accurately. Closed loop Idle control: Engines need differing amounts of idle air under different conditions. When an engine is cold, for example, the engine will require more air than when it is hot. Also additional loads on the engine such as the air conditioner, power steering and electrical loads such as headlights will all require additional amounts of idle air. Ideally the ECU would have variable amounts of air (either a number of steps of a stepper motor or additional duty cycle for a pulse-width modulated idle valve), and additional digital inputs so that the ECU can determine which of these loads are active. As well as the settings for basic idle air quantities, the ECU would ideally have a target idle speed and the ability to perform closed loop idle. To enable the closed loop function, the ECU needs to know either the road speed, so must have a vehicle speed input, or a clutch / neutral switch digital input. Closed loop fuel control: Closed loop fuel control allows the ECU to adjust the fuel delivery to meet the target air-fuel ratio, or target lambda. For this, the ECU needs a sensor that will give a reading over the range that you want to run the system in closed loop mode. For example, a capable ECU should be able to use a narrowband sensor to run closed loop at stoichiometry and then automatically run open-loop when the target air-fuel ratio goes outside the range of the narrowband sensor. The ECU should be able to run in closed loop mode at richer mixtures if a wideband lambda sensor is fitted. Many wideband oxygen sensors generate a serial data stream. Connecting this directly to the ECU’s serial input port is preferable to connecting the analogue output from the wideband sensor controller into an ECU analogue input, because the serial data stream includes other data such as whether the sensor is warmed up or has a fault. The analogue output from the wideband sensor controller, however, must sit at some voltage, which is still within the range of legal sensor values. Therefore, there is no way for the ECU to know whether the sensor is reading a sensible value when reading from the analogue voltage alone. Knock detection / closed loop ignition control: Some advanced aftermarket ECUs perform knock detection. This can be used to retard the ignition when knock is detected. Tuning an engine right on the edge may require the engine to be on the verge of knocking, and this may occur at different ignition angles on different cylinders. To get the absolute maximum power out of the engine would therefore require knock detection on a percylinder (or per-rotor) basis. The actual complexity and sophistication required depends on how close to the limit you intend to run. Sequential Injection: For standard engines, (eg, with OEM cams), sequential injection often isn’t necessary. However with higher overlap cams, or large ports, injection timing

is more critical. To achieve consistent injection timing on each cylinder (or rotor), the ECU must support sequential injection. This requires as many injector outputs on the ECU as the number of injectors you intend to run; for example to run sequential injection on a 4 cylinder requires 4 injector outputs from the ECU, and to run sequential injection with staged (secondary injection) on a 6 cylinder will require 12 injector outputs on the ECU. This also requires sufficient triggering information on the engine to give the ECU enough information for the engine angle. For a 4-stroke engine, the ECU needs the full 720 degrees of information, which requires a sensor mounted on the camshaft. A crankshaft only sensor will not provide enough information for fully sequential injection for a 4stroke engine. However a rotary engine fires each injector every 360 degrees, so a pattern that repeats every 360 degrees is fine. Spark Split: This is a feature required for rotary engines Plug-in compatibility: Datalogging: Live Visible outputs / hand controller: Drive By Wire Control: Flat shift:

Performance Main loop speed: Fuel resolution: Ignition resolution: Mapping resolution: Ignition jitter (steady state): Ignition tracking (changing RPM): Maximum RPM: 8/16/32-bit: Clock speed:

Andy Wyatt has a Bachelor’s of Computer Science, and a Bachelor’s degree in Mechatronics Engineering with a University Medal. He is also the founder of Adaptronic Engine Management, and has been developing ECUs since 1999. For more information on ECUs, visit www.adaptronic.com.au