What carries across EDC16 and EDC17
The useful common ground is a functional method: identify torque requests and restrictions, the relationship between torque and fuel, delivery of that fuel, the available air and the feedback that checks the result. Apply those questions to the identified software and hardware. This guide compares recurring principles; it is not a claim that every EDC16 or EDC17 has identical maps, units, selectors or programming methods.
Bosch describes torque-controlled EDC introduced with EDC16: driver and vehicle-system demands are coordinated, engine losses are considered, and fuel/air actuators produce the permitted output. Its schematic includes separate starting control and protective quantity limits. Bosch Electronic Diesel Control, 2001, printed pages 75–77. For a later example, BitEdit's BMW EDC17CP09 module lists requests, torque/fuel conversion, condition-dependent limits, monitoring, injection, air and rail-pressure functions. That is vendor coverage for a named variant, not its complete function diagram. BitEdit EDC17CP09 module.
Read the discussion evidence and applicability index for the additional NefMoto and ECUedit cases behind this review. Their comments are useful investigation leads; their settings and attachment names are not validated calibrations.
Use the same questions, then prove the implementation
| Recurring task | Question to carry to another ECU | What must be established locally |
|---|---|---|
| Driver/vehicle requests | What output is requested at this operating point? | Pedal/cruise/idle inputs, physical units, corrections and active mode. |
| Coordination and limits | Which request or restriction currently determines output? | Torque versus quantity limits, temperature/altitude/gear conditions, transmission intervention and monitoring. |
| Torque/fuel relationship | How is permitted output related to fuel? | Torque reference, quantity units, conversion range, corrections and any inverse relationship. |
| Injection delivery | What electrical/timing command delivers that quantity? | Injector hardware, event meaning, duration/timing units and pressure dependence where applicable. |
| Air system | Does actual air support the requested combustion? | Airflow/pressure/temperature signals, EGR context, actuator type and control response. |
| Modes and feedback | Is this the normal-load path or another strategy? | Starting, warm-up, regeneration, faults/substitute values and the data actually exposed for logging. |
This table is a review checklist inferred from the cited control descriptions and case discussions, not a universal serial execution order. A request, limit, model, actuator command and measurement are different objects even when an editor gives them similar names.
Start with an output objective and a healthy baseline
Specify the intended torque curve, speed range and drivability, including part pedal and transitions. Keep faults, original software identity, hardware configuration and repeatable baseline measurements with the calibration record. A pre-existing boost-control problem is a separate diagnostic task; the BMW E46 case in the discussion index illustrates why improved acceleration can coexist with unresolved overboost.
A useful physics check is P [kW] = T [Nm] × rpm / 9549.3. For an invented example, 120 kW at 4000 rpm corresponds to about 286.5 Nm; 400 Nm at 1800 rpm corresponds to about 75.4 kW. These are unit examples, not engine targets. This relationship does not calculate injection quantity or establish hardware capability. Bosch uses the power/speed/torque relationship to explain torque as the command variable. Bosch, printed page 75.
Distinguish indicated/internal, output, wheel and communicated torque before comparing a map with a specification or dynamometer. Bosch explicitly accounts for engine losses and auxiliary loads. Bosch, printed page 76. A difference between one table and a catalogue torque figure is not evidence for a fixed percentage correction at every speed/load.
Build a definition record before changing values
Record the exact software, function, axes, units, datatype, byte order, conversion, address basis and evidence for each relevant object. ASAM's A2L description format separates calibration parameters from measurement variables and specifies their layout and physical interpretation. ASAM MCD-2 MC overview.
ASAM describes table-specific, fixed/calculated and shared axes, including axes stored away from the table data. It also describes conversion methods such as a linear slope/offset. This is a vocabulary for checking a definition, not evidence that a given binary uses any particular structure. ASAM MCD-2 MC: AXIS_DESCR, AXIS_PTS_REF and COMPU_METHOD.
Invented scaling example: if physical = 0.1 × raw − 10, raw 500 represents 40. Increasing the raw value by 10% gives 550, representing 45; a 10% physical increase would be 44. Confirm whether the editor is changing raw data or a converted physical value. Do not transfer a factor, offset or address just because another ECU has the same family name.
The Mercedes E320 CDI discussion in the evidence index includes a table initially suspected to be a boost limiter which a reply identified as temperature-dependent MAP-signal correction. The exact binary was not verified here, but the distinction between a sensor transfer/correction and an output limit is essential before editing.
Check range and active mode before extending tables
Establish which map or combination is used at the intended operating point, whether its axes cover that point, and how the target software behaves outside that range. An axis extending far enough is one check; it does not prove every downstream conversion, selector and limit is suitable. Conversely, a table need not be changed merely because a generic checklist names it.
Where a validated model needs a wider range, interpolation between known data and extrapolation beyond known data are different operations. Redistributing a fixed number of breakpoints can reduce resolution elsewhere. Recheck every consumer of a shared axis and justify the new values from the fitted hardware and measured response. A smooth-looking graph does not characterise an injector or prove ECU boundary behaviour.
BitEdit's CP09 list separately names boost/actuator data for base, regeneration, double-injection and bypass conditions, plus idle and main rail-pressure functions. BitEdit CP09 operating-mode coverage. Practical implication: similar-looking copies should be assigned a function and selector before editing. A normal-load objective does not automatically justify applying the same changes during regeneration or fault operation.
Keep models, requests and controls coherent
Write a reason for each proposed change. A request change expresses an output objective; a limit change alters permission; a conversion or injector-characterisation change alters what the ECU believes about delivery. Identify which of these is actually required. Keep quantity and torque reporting representative so that vehicle coordination and monitoring remain meaningful.
On common rail, rail pressure and injector energising time interact with delivery; on a unit-injector installation the pressure-generation and timing mechanisms differ. Follow the EDC16 hardware comparison and the engine-specific guides below. The shared task is to verify delivered fuel and combustion conditions, not to apply a common rail-pressure or duration increase to both systems.
For air control, distinguish the boost target, its restrictions, actuator command and actual response. Treat a feedforward or control-gain change as control work justified by response data. An exhaust-temperature restriction calls for identifying the operating cause and suitable demand/hardware, rather than automatically applying a forum suggestion to increase timing, boost or rail pressure.
Compare changes and test a stated hypothesis
- Keep an identified reference: distinguish an actual read from a server-supplied original and record memory regions, software and checksum/recovery handling for the selected tool.
- Review the complete difference: check changed cells, axes, scalar limits, selectors, code and diagnostic alterations. Match functions and units when comparing different software layouts; a donor tune is not proof of compatibility.
- State the expected result: for example, whether a particular permitted-quantity restriction should change while the relevant conversion remains valid. Record what observation would disprove that explanation.
- Log requests and responses together: where available, include permitted/estimated torque, requested/limited fuel, rail and boost target/actual, air mass, temperatures, actuator feedback, transmission intervention, regeneration state and faults. Label calculated signals and diagnostic display limits.
- Compare equivalent conditions: use repeatable gear, speed/load, temperature and operating mode. Inspect transients, hot operation, part pedal, lift-off and shifts as well as peak measured output. Validate independent output and relevant physical measurements where available.
- Reconcile unexpected results: explain a discrepancy before adding more changes. The number of maps edited, a checksum pass, a lively pedal or absence of limp mode is insufficient evidence on its own.
EVC documents comparing imported versions against an original and marking differences in WinOLS. EVC WinOLS project/version FAQ. The checklist above is a proposed evidence record, not a requirement to use a particular editor or proof that comparison software validates a tune.
Application guides and evidence limits
- BMW EDC16 tuning and diagnostic checks
- BMW EDC17 torque checks and adaptations
- VAG EDC16 identification and transmission checks
- VAG EDC17 tuning and diagnostic checks
Reviewed 17 November 2026. Primary control/calibration documentation supports the concepts; the workflow and examples marked as implications are this review's synthesis. Bosch's 2001 text predates EDC17 and is not an EDC17 function definition. Engine/software applicability and unresolved outcomes are recorded in the linked discussion index. No forum binary, map pack, tune or vehicle was tested.