Tuning Wiki
Applies to

Check the hardware and software limits stated in this article.

Scope. The direct injectors meter fuel into the cylinder, unlike port injectors which spray upstream of the intake valve. Injection phasing and pump capacity are separate but related topics. This page concerns injector characterisation, electrical timing and the feedback-based COSI path described in one community Simos 18.1 extraction. Equipment, software, definitions and enabled branches differ: match ECU ASW/CAL, injector part, fuel, pressure reference, indices and current profile before applying any name.

Why injector characterisation matters

The ECU must relate a requested per-cylinder fuel mass to an electrical injection command. The relationship depends on the actual injector and driver, fuel properties, pressure difference, temperature, pulse length and prior pulses. A single static flow rating does not define all these behaviours. An experimental GDI injector study measured the effect of pulse width, driver voltage, fuel pressure and current modulation on delivered quantity; it is general research on its test hardware, not calibration data for the EA888.

A direct injector upgrade requires matched injector and electrical-driver data and calibration support. A lambda reading averages gas reaching a sensor and does not establish that each cylinder receives the same fuel quantity or spray pattern. Direct injection also has a finite usable crank-angle window; see direct-injection timing and phasing. The corresponding MPI injector guide covers port injectors; do not transfer its pressure correction or battery dead-time data automatically to direct injection.

Injection-time maps and correction paths

The SC8S5031_C_OEM INJR final-injection reference lists the following examples. It is an extracted, community reference with some conflicting captions and incomplete axis metadata; representable ranges are not recommended injector limits.

  • ip_ti_4_l, ip_ti_4_h and ip_ti_4_over_prs are described as injection-time maps for low, high and over-pressure injector-current profiles. Each has a fuel-mass-labelled axis in mg/stroke and a pressure-difference-labelled axis in hPa, with an ms output. The axes are marked NO_INPUT_QUANTITY; extracted narrative and table captions disagree in places about profile labels. Verify actual lookup inputs and selected profile in the target implementation. They are separate maps, not generic low/high-pressure pump targets.
  • ip_ti_4_cor_ti_4_h and ip_ti_4_cor_ti_4_l are injection-time-dependent characteristic corrections for their named current profiles, with ms-labelled axes and outputs. The INJR injection-realisation reference has conflicting neighbouring captions and NO_INPUT_QUANTITY axis metadata; verify the equation and whether a branch is active before treating either as an offset or replacement time.
  • ip_rho_fuel_inj is a fuel-density map with high-side pressure fup_h in hPa and injector fuel temperature tfu_inj in °C as axes. ip_rho_fuel_inj_test is a distinct test-bench-density map. The extracted output unit is garbled; confirm it and whether temperature is measured or modelled before using the value. Fuel density is not stoichiometric AFR or ethanol percentage.

COSI effective-time and feedback path

In the same extraction, ip_ti_eff_mff_temp_l, ip_ti_eff_mff_temp_mid_1 and ip_ti_eff_mff_temp_h are temperature-specific effective-injection-time maps associated with COSI interpolation. The cited low and high map definitions have fuel-mass and pressure axes and outputs in microseconds, unlike the millisecond output of ip_ti_4_*. Their axis metadata does not specify the live inputs; verify the mid-map definition, temperature selection, units and actual COSI enablement. These maps are not interchangeable with injector dead time or a direct measurement of delivered mass. INJR final-injection reference.

The COSI control reference describes using feedback-related opening and closing timing to update an internal working curve. It lists ip_t_open_0 as a basic opening-time map and ip_ti_0 as a working-curve starting-time map; both have pressure- and temperature-labelled axes with NO_INPUT_QUANTITY metadata, and output microseconds. ip_fac_fil_cosi_ctl is a working-curve adaptation filter with a mass-difference-labelled axis in mg/stroke; establish the actual input and timing. The reference's effective-time calculation combines feedback-related timing values. This is not an independent laboratory measurement of injected fuel mass, and the existence of names in the extraction does not prove the vehicle uses or exposes COSI.

The INJR COSI application reference describes ip_mff_stnd_min_cosi_stat as a stationary minimum standard fuel mass curve for its COSI path, against a pressure-labelled axis, and separately lists ip_mff_stnd_min_l, ip_mff_stnd_min_h and ip_mff_stnd_min_over_prs for different current profiles. Their outputs are mg/stroke in this image and input metadata is incomplete. They are model/control constraints, not a guarantee of accurate operation below that quantity.

ip_mff_pls_delta_cor is labelled a fuel-mass correction for pressure and the time between consecutive injections: its second axis is t_delta_inj_prev[0][0] in ms. Verify the pressure input, pulse indexing and active branch. This correction differs from the crank-angle pulse-separation limits in the phasing guide. Do not disable constraints or feedback to mask poor idle, a misfire or unsupported injector hardware.

Practical validation

  1. Record injector and driver compatibility, ECU and calibration identification, fuel and expected density, measured low- and high-side pressures, and whether the vehicle has stock or modified MPI. Obtain characterisation data for the exact injector, fuel and operating conditions.
  2. Check charging voltage, injector circuits, mechanical condition, leaks and fuel pressure before editing characterisation. GDI systems contain high-pressure fuel: use the applicable manufacturer's depressurisation and repair procedures.
  3. Where a matched logger supports them, compare commanded and actual rail pressure, final lambda target and actual, fuel trims, injection mode, pulse times, SOI/EOI and range-limit flags, per-cylinder misfire and knock signals. Record whether values are requests, feedback-derived estimates or physical measurements. Check warm-up and idle, short-pulse and mode transitions as well as controlled high-load runs.
  4. Assess each cylinder and operating condition. A normal aggregate lambda, a representable map maximum or a stable rail-pressure log alone does not prove adequate injector delivery or even distribution. Compare the new data and hardware against the matched calibration before increasing demand.

Injector service and an owner-reported example

In a 2017 US-market Golf R owner discussion, the owner reported an EQT Stage 2 tune, intermittent idle or low-speed misfire and cylinder shutdown, with no reported wide-open-throttle symptom. They listed P0300, P0301, P0302, P130A, P02EE, P02F0, P2279 and P0068; scan-tool descriptions and exact software were not independently verified. After replacing injectors they reported finding an extra O-ring on cylinder 3 and suspected it had contributed; idle seemed improved after warm-up, while cold start remained somewhat uneven. The injector replacement and assembly changes occurred together, and a durable fix or single cause was not established.

The case makes injector installation and prior work relevant diagnostic context: it does not prove that one particular seal caused all those codes, or that a calibration change is required. Inspect and install the correct parts using the exact manufacturer's repair instructions for the engine and injectors; direct fuel systems require specified depressurisation and service procedures.

See also Simos 18 fuel temperature modelling and compensation for temperature-model maps, injection and pump corrections, and measured versus modelled channels.

Related: direct injection and fuel pressure · MPI injector calibration · high-pressure pump upgrades · misfire diagnosis. References reviewed 17 November 2026.

COBB examples of different injector-time strategies

In its MQB tuning guide, COBB identifies 5G0906259A and 5G0906259D among its 2015–2016 GTI calibrations as using “Injection Time (High, Low, Over) Pressure” display tables. For the later GTI and Golf R calibrations it discussed, COBB instead describes “Effective Injection Time (High, Low, Mid) Temperature” tables. These are COBB’s examples and display names, not proof of the active path in every vehicle with the same model year. Match the actual software and definition and verify selectors, axes and units; do not copy a time curve between the two approaches merely because the tables look similar.