A2L and HEX files

The A2L is a description, not a second copy of the calibration image. It names CHARACTERISTIC objects such as constants, curves and maps, and defines how their raw memory representation becomes a value an engineer can use. It can also describe measurement signals, conversions and the ECU interface used by a measurement and calibration tool.

The image contains bytes. Intel HEX and Motorola S-record encode addresses and data in text records; a raw BIN is just bytes and needs an agreed address base. An image can contain code as well as calibration data, so an address alone does not tell you which value is safe to edit.

Start with the ECU software identity and project variant. Then compare the A2L memory segments and addresses with the image ranges. Decode several known objects of different types, including at least one map with axes, and check that the values and units are credible against the released dataset or measured behaviour.

A shifted address range, changed record layout or different conversion method can still produce numbers that look reasonable. That is why a visual spot check of one cell is not enough. Treat a mismatch as a stop condition, find the correct A2L and image pair, and repeat the checks before writing.

Open the matching A2L and image, select the named calibration object, inspect its axes and physical conversion, then change only the intended cells. Export a diff that identifies the object, the old and new engineering values, and the raw bytes affected. Validate bounds and dependencies before accepting the change.

Checksums are ECU- and image-specific. A tool may support the applicable checksum algorithm, but the result still needs verification against the target ECU or release procedure. Keep the original image, the edited image, the validation result and the test evidence together. The vehicle measurement that justified the edit is usually an MDF4 recording, which is a different file format from the calibration image.

XHandbook connects a documented function to the A2L label and its address, layout and conversion. Hexx opens the image with that A2L so the engineer can inspect and edit named objects in engineering units. MDX reads the MDF4 measurement before and after the change, and HELIX keeps the proposed dataset, checks and approval history together for release.

These are different jobs. Documentation explains the intent, the A2L explains the memory representation, the image holds the bytes, and the measurement shows what the ECU actually did. Keeping those roles separate makes a calibration decision easier to verify.