To
address the increasing O2 density as air temperature decreases, the
AFM chip scales fuel via a temperature compensation table. A MAF sensor already
has temperature compensation built into the signal, so this temperature
compensation table is flattened to prevent a ‘double compensation’ effect.
The
best place to accommodate larger injectors is in the Fuel Quality Switch (FQS)
table. This applies a global scaling factor to the injector duty cycle signal.
The DME box has an 8 position switch allowing selection across 4 fuel x 2
timing combinations. The FQS table can be customised to the injector size you
wish to run. I need to know your injector flow rate at 3 bar fuel pressure, eg
578 cc / min for
Note:
massaging the MAF sensor signal to scale for larger injectors is not
recommended as it corrupts the relationship between the sensor and the on-chip
transfer function, distorting part throttle timing.
The
DME chip relates the AFM voltage to air flow via the Transfer Function (Figure
1). As a MAF sensor has a different calibration to the
factory AFM sensor, the on-chip Transfer Function is reshaped to match the
targeted MAF sensor (Provided data is for the 85mm Lindsey
Racing Stage 2 MAF).

Figure 1 - Comparison of MAF chip transfer function to that of the stock AFM chip. At 5 volts, the MAF chip defines twice the airflow to that of the stock AFM. Above 3 volts, the shape of the transfer function is more aggressive than that actually flowed by the targeted MAF sensor. There is a compromise going on here between describing a smooth curve within the constraints of the transfer function tables (for easy of piggyback tuning) and matching the flow calibration of the targeted sensor. Above 3 volts we are beyond part throttle flow rates so is not a major issue.

Figure 2 - Calibration sheet for the 85mm Lindsey MAF sensor

Figure 3 - Piggyback tuning of the targeted MAF sensor voltage. Below 3 volts, there is an accurate calibration between sensor and transfer function and minimal signal massaging is required. Above 3 volts, the signal needs to be scaled back to compensate for the transfer function rising too rapidly in this region. Due to greater fluctuation in the MAF sensor signal at low flow (idle), a flat line is programmed in the 0 - 0.6v to smooth this out for a steady idle.

Figure 4 – Wideband datalog: 3rd
gear full throttle 15psi Turbonetics T04E50 stage 3 wheel,

Figure 5 - MAF chip fuel maps, the part throttle map has been fine tuned for smooth transition from idle, compensating for limitations on programmable granularity / control of the shape of the transfer function. The Full throttle map has been flattened to simplify piggyback tuning.
MAF
chip timing is consistent with that used by modern aftermarket AFM chips.

Figure 6 - MAF chip timing is consistent with modern aftermarket AFM chips.