Fuel: Upgrade Considerations

From 3000GT/Stealth Wiki
Jump to: navigation, search

Electronic Fuel Injection Theory

Ultimately, the purpose of any internal combustion engine fuel system is to ensure that the proper amount of fuel and air are mixed, such that the intended Air-Fuel Ratio is present inside the combustion chamber at the point of ignition.

In a spark-ignition engine, the driver control is a throttle plate that varies the amount of air currently admitted into the engine. As such, the EFI system must measure the amount of air in the current cylinder charge, and then add the proper amount of fuel.

(A compression-ignition engine, like a diesel, has no throttle plate, and the user control is the amount of fuel injected)

As the reacting of fuel with air is effectively chemistry, the important measurements are the mass of fuel and the mass of air in the current cylinder charge.

Fuel, being a liquid, has a fairly consistant density, and so can be measured by volume with reasonable accuracy. Injectors can be rated by mass flow or by volume flow and the relationship is constant for fuel of known density, and the range of possible fuel densities is small enough to make little effective difference to the state of tune of a street-driven engine.

Air, being a gas, is much more sensitive to environmental differences, which means that an accurate air mass estimate must be made.

There are two main mechanisms for estimating charge air mass:

  • Speed/Density; and
  • Mass/Air.


In a Speed/Density system, the induction system is fitted with a Manifold Absolute Pressure (MAP) Sensor, and an Air Intake Temperature (AIT) Sensor. These two measurements provide the "density" term. As the displacement of the engine is known, the volume of the current air charge is directly related to the current engine RPM; the "speed" term. Thus, a measurement of air temperature and pressure, plus an engine speed measurement, give you air charge density and volume, which in turn, gives you mass.

In actuality however, the actual volume of air drawn (or forced) into the cylinder is never the same as the theoretical swept displacement. There are always flow losses inherent to induction design, such that the actual air inducted is other than the measured volume of the cylinder. The difference between actual and theoretical air volumes is called Volumetric Efficiency (VE) and must be accounted for in order to provide accurate air mass estimation.

Speed/Density systems thus have a lookup table of VE/RPM (often 3D VE/RPM/LOAD). VE cannot be calculated; it must be measured empirically, and if anything in the induction/exhaust system changes, so too will VE, and the VE calibration table will need to change with it.


In a Mass/Air system, a sensor is provided that can measure air mass directly; as such, the fueling table in the ECU need only store the correct injector on time for that particular air mass flow value. This makes Mass/Air systems "self healing" with respect to VE; if VE suffers (due to age, deposits, clogged filters etc) this registers simply as a lower mass flow at that RPM cell. This pays large dividends on the emissions front (air/fuel ratio will always be correct, even if VE changes) and so most modern EFI systems (3/S included) use mass/air systems.

System Comparison

  • Speed/Density systems do not place an obstructive mass airflow sensor in the airstream, so they typically offer slightly better throttle response and maximum power levels (all else being equal)
  • Speed/Density systems are tolerant of boost/vaccuum leaks, where mass/air systems will miscount the air reaching the cylinder if air is leaking in/out where the sensor cannot see it
  • Speed/Density systems measure absolute manifold pressure, which can have important ramifications on timing advance for forced induction motors. Mass/Air systems without a MAP sensor must infer MAP from empirical observations of MAP at a given mass flow per engine RPM.
  • Mass/Air systems are tolerant of modifications that change VE, and will correctly fuel the engine within the boundaries of its fuel map (and the underlying assumptions) Speed/Density systems must be recalibrated any time VE changes.


Once the current air mass is known, the engine is fueled by turning on the appropriate injector. Fuel injectors are electrically-operated on/off valves; there is no pump mechanism inside the injector. Instead, high fuel line pressure (~ 45 PSI) is used to force fuel into the engine. Fuel mass delivered into the engine is a function of the current line pressure, the volume fuel delivered per second of on time of that particular injector, and the amount of time the injector is actually open for.

For forced induction motors, fuel pressure is usually raised 1:1 for manifold pressure in order to keep the differential pressure at the injector valve constant. Were this not done, fuel delivered per second of on time would drop with rising boost pressure (with a line pressure of 45 PSI and a 20 PSI manifold pressure, effective fuel pressure at the injector nozzle is 25 PSI)

If either the base fuel pressure or the injector volume is changed, the mass of fuel delivered per second of injector on time will also change.

3000GT/Stealth Specifics

North-American-spec turbo models have 360cc injectors that are arguably ample for the airflow generated by the stock MHI TD04-9B turbos ([1]).

However, when the boost limit is increased to 1 kg/cm^2 (14.2 PSI), the injectors and turbos are maxed out from around 6000 RPM to redline. Additionally, the stock fuel pump runs at a reduced voltage under vacuum (i.e.- when boost is not measured in the plenum), which may contribute to detonation issues during voltage transition.

Thus, to operate at higher boost levels, a higher-capacity fuel pump ([2]) with hotwire kit that provides full voltage at all times is a worthwhile upgrade for even otherwise unmodified fuel systems.

Since the increase in flow can overrun the stock fuel pressure regulator, an adjustable FPR may be required ([3]).

These changes can result in noticeably smoother acceleration and set the stage for larger injectors to support larger turbos.

NOTE: When hotwiring the fuel pump, be sure to replace the stock voltage relay with a jumper so that noise is not backfed to the ECU, causing garbage data in logs (known to occur with some hybrid ECUs [4]) and potentially other anomalies!

Proper mixture is achieved by cycling the injectors by an amount appropriate for the measured airflow. Raising boost increases airflow which the ECU maps to higher duty cycles; excessive boost can lead to "fuel cut," where the ECU determines that injector flow at 100% duty cycle is insufficient and abruptly disables the injectors to prevent a lean condition that could damage the engine. The ECU also calculates ignition timing from airflow, advancing it to improve responsiveness and retarding it to prevent detonation ("knock").

The stock mass/air system has no concept of boost pressure; the ECU infers boost pressure from the mass flow at a particular RPM cell. This works for an unmodified engine, but a modified engine may flow different mass amounts for the same boost pressure (when compared to an unmodified engine). Optimum fuel mixture, as determined by injector pulsewidth, is a function of mass flow, so piggyback computers that modify the mass flow amount reported by the mass airflow sensor work well for compensating for larger than stock injectors. Optimum ignition timing, however, is sensitive to boost pressure, and when a piggyback computer is used to underreport the amount of mass airflow at a given RPM point, the ECU will underestimate the current boost level, and may attempt to use more ignition advance that is optimum (or even safe).

With stock 9B turbos limited to stock boost pressure (?? PSI) , maximum airflow at peak power (~ 320 HP) results in a peak Injector Duty Cycle (IDC) of about 80% and timing of around ?? degrees. With stock 9B turbos raised to 1 kg/cm^2 (14.2 PSI) pressure (which they cannot sustain to redline), maximum airflow at peak power (~ 360 HP) results in IDCs of approximately 97% and timing of around 30 degrees (figures vary with temperature, etc.).

Ideally, peak IDC should be limited to 80-90% to limit stress on the injector mechanism and to ensure maximum fuel flow (some injectors flow less at 100% duty cycle than at lesser duty cycles).

Note that IDC values reported by loggers tend to be estimates that should be viewed with suspicion ([5]); accuracy can be verified by computing duty cycle from pulse width as IDC = IPW * RPM / 1200.

Higher-capacity injectors deliver more fuel per cycle, rendering the rates computed by the ECU inappropriate and resulting in an extremely rich condition. A simple piggyback controller (e.g.- A'PEXi AFC NEO) effectively decreases the airflow signal reported by the mass airflow sensor by a user-controlled amount. If the airflow is reduced by an amount proportional to the increase in injector size, such that a lower duty cycle is mapped by the ECU, proper mixture ratio is restored. However, the low apparent load can result in timing too advanced for the high actual load, leading to detonation at otherwise acceptable boost pressures. Controllers with timing management are available (ITC, eManage, AEM) but tuning becomes more complex.

Injector capacity should be increased only enough to achieve the maximum fuel delivery needed for the maximum anticipated airflow ([6]). In particular, controlling mixture ratio at small fueling levels and short pulsewidths (such as at idle) can be difficult with very large injectors due to insufficient granularity.

Below is a hybrid ECU log of the stock fuel system accommodating 9B turbos limited to 1 kg/cm^2 (14.2 PSI) during full acceleration. Each trace is on a different scale for clarity. Peak Injector Duty Cycle is 96%, and while exact A/F ratio cannot be determined from the OEM narrowband O2 sensor, it can be seen that mixture never falls below the narrowband swtich point of 14.7:1 (as narrowbands read leaner with increasing temperature, the mixture is potentially very rich). The stock fuel system seems adequate only because the stock turbos are unable to sustain airflow at higher RPMs, as indicated by the falloff on the "Engine Load" trace.

9b boost falloff.gif

--DG 09:43, 5 July 2006 (EDT)

This article originated from a thread or post on 3000GT/Stealth International. (1928)