Tuning Wiki

This is a historical revision. Read the current page or return to its history.

Page organisation in this revision
Page type
Technical article
Subject
Tuning
Parent
Bosch M3.8.3
Applies to
Bosch M3.8.3
Tool groups
None
Manufacturers
Audi, Škoda, Volkswagen

What the sample identifies

TracqiTechnology/MxT supplies an 06A906018CG XDF whose header describes Bosch M3.8.3, Audi A3 1.8T AGU 150 PS and Bosch 0261206518. The repository's companion binary is 262,144 bytes; the XDF describes a 0x40000-byte region. These are properties of the published example, not proof that an arbitrary read is complete or that every ECU with a similar label uses this calibration.

The XDF header does not give a software ID. The identification text inside MxT's companion binary contains 06A906018CG, 0261206518 and 1037352127, consistent with the identifiers discussed in an independent owner's report. Embedded text is an identification clue, not proof of the file's provenance, original state or identity with another read. Match the actual ECU and file identifiers.

What is and is not defined

The XDF contains a table labelled MLHFM, described there as hot-film airflow linearisation, and a constant labelled MLOFS. It does not define complete fuel, boost and ignition map sets. Labels and values are the project's claims and have not been independently validated here; no complete tuning definition should be inferred from this file.

Finding: 256 points is the better-supported interpretation for the inspected example. The 266-point definition appears to include ten preceding zero words. This is an inference from the file layout and a separate community report, not a firmware-code or vehicle-test confirmation.

EvidenceObservation for this sample
MxT XDFMLHFM begins at file offset 0x7540 and specifies 266 little-endian 16-bit entries. That span ends at 0x7753, inclusive. Its voltage labels cover 0–5 V over all 266 entries.
MxT companion binaryThe 20 bytes at 0x7540–0x7553 are zero. The next 256 little-endian words, 0x7554–0x7753, run from raw 1600 to 6987 without decreasing. This candidate 256-word span and the XDF's 266-word span have the same end.
Independent practitioner reportIn a 2014 discussion of a 06A906018CG / 352127 file, participants identify 0x7554 and report a 256-point M3.8.3 MLHFM. The forum attachment has not been independently compared byte-for-byte with the MxT binary.
MxT fixture testThe published test assumes 0x7540 and 266 entries, then checks monotonicity and expects initial zeros. It tests that assumed span; it does not independently establish where firmware begins the table.

For reproducibility, the inspected MxT binary at revision e80501a0000f is 262,144 bytes and has SHA-256 41f4ce4da0c6d38276aaca951ba51d297b475a6f5c094f5ab8cc991447e7ec03. File offsets refer to that exact file; do not transfer them to another read without matching it.

The sample definition also overlaps two entries. Its MLOFS constant points to 0x7554, the same two bytes as the eleventh entry of its 266-word MLHFM table—the first word of the candidate 256-word span. An edit to either definition therefore affects the same bytes. This overlap does not prove that the ECU has a separately editable MLOFS constant at that address.

The XDF gives those bytes the value 200 when its factor of 0.125 is applied, and uses the same factor without an offset for MLHFM. An AGU M3.8.3 owner discussion describes a minus-200 display offset, but does not independently establish the correct conversion for this particular sample. Verify the actual voltage indexing, scaling and offset against a trustworthy definition and ECU behaviour; simply deleting ten rows or applying an assumed offset can leave the curve incorrectly labelled.

The candidate table has not been traced to its firmware lookup code or tested on an ECU. The exact axis conversion, MLOFS role, and applicability to other software remain unverified. Treat the published XDF as a research example and do not use it uncritically for calibration.

Verification before calibration

  1. Match vehicle, ECU hardware, complete software identification and original read to the definition's documented target. Record the exact file and definition revisions.
  2. Verify each proposed map's address, dimensions, data type, axis values, units and conversion independently against reliable information for that software. Compare the resulting values with plausible original behaviour, rather than accepting a label in isolation.
  3. For airflow changes, compare measured sensor response and logged airflow across relevant operating points. Confirm fuel delivery, injectors and measurement equipment before using fuel trims to infer metering errors. Do not assume a corrected MAF curve also corrects every load, fuelling or boost function.
  4. Use a verified definition for any additional map and retain the original and change record. Seek software-specific evidence if map boundaries or roles remain uncertain.

Continue with AGU tuning and diagnosis. Example attribution: TracqiTechnology/MxT, source revision e80501a0000f, Apache-2.0 licence. No vehicle or ECU testing by this wiki is implied.