Siemens Extractor Internals

This document covers detection strategies, binary format details, and internal extractor structure. For the user-facing family reference, see Supported Siemens ECU Families.

Extractor directory structure

Each Siemens family lives in its own package under src/openremap/tuning/manufacturers/siemens/<family>/:

siemens/
├── simtec56/   — Simtec 56 (Opel/Vauxhall X18XE, X20XEV petrol)
├── simos/      — SIMOS / SIMOS2 / SIMOS3 (VAG 1.4–1.6L petrol)
├── ppd/        — PPD1.x (VAG 2.0 TDI Pumpe-Düse diesel)
├── sid801/     — SID 801 / SID 801A (PSA/Ford 2.0/2.2 HDi diesel)
├── sid803/     — SID 803 / SID 803A (PSA/Ford/JLR 2.0/2.2 HDi diesel)
└── ems2000/    — EMS2000 (Volvo S40/V40/S60/S70/V70 T4/T5 turbo petrol)

The registry in siemens/__init__.py lists all extractors in priority order — most specific first. When a new binary is submitted, the first extractor whose can_handle() returns True wins.

Registry ordering

The Siemens EXTRACTORS list is ordered by detection strength, strongest first:

Rule: EMS2000 must always be the last Siemens extractor. Moving it earlier would cause it to falsely claim binaries that a later, more specific extractor should handle.

Simtec 56

Overview

Opel/Vauxhall petrol ECUs (1995–2002) for X18XE and X20XEV engines. Vehicles include Vectra B, Astra F/G, Omega B, and Calibra. CPU is an Intel 8051 / Siemens C166 derivative.

Binary format

Detection strategy (5-phase, all must pass)

1. Size gate — exactly 131 072 bytes.
2. Exclusion — reject if any Bosch / Delphi / Marelli signature found (EDC17, MEDC17, ME7., BOSCH, 0261, MOTRONIC, PM3, PPD).
3. Detection signature — require 5WK9 (Siemens part prefix) in the binary.
4. Ident record — require RS or RT prefix followed by an 8-digit GM part number. This distinguishes Simtec 56 from other 5WK9-bearing families (SID801, etc.) that use different ident formats.
5. Header magic — first 3 bytes must be \x02\x00\xb0. Weak secondary confirmation — not sufficient on its own.

Ident record format

The RS/RT ident record is the single most reliable identifier. Format:

R[ST]<8-digit GM part> <12-16 digit serial>[a-z]5WK9<4-6 digits>

Example: RS90506365 0106577255425b5WK907302
         ││││││││││ │││││││││││││││││││││││
         ││││││││││ │││││││││││││││5WK90730 ← Siemens part (core 8 chars)
         ││││││││││ │0106577255425b ← serial (12–16 digits + lowercase letter)
         │90506365 ← GM part number (8 digits)
         RS ← prefix (RS = standard track, RT = alternate/test track)

The trailing digits after 5WK9 encode a 4-digit variant number plus a 1–2 digit checksum suffix. All confirmed samples show 5 trailing digits (4 variant + 1 check).

Extracted fields

Sub-field extraction

Sub-fields (GM part, serial, Siemens part) are split from the full ident record match using a compiled regex rather than separate lookbehind patterns — Python's re module does not support variable-length lookbehinds.

SIMOS

Overview

VAG (VW/Audi/Skoda/Seat) petrol ECUs from the late 1990s through mid-2000s, covering 1.4–1.6L engines. Three binary sub-types exist by file size.

Binary sub-types

Detection strategy (layered, fast-to-slow)

1. Exclusion — reject if any known Bosch / SID / PPD / Simtec signature found. Scans first 512 KB for speed. Exclusion list includes: EDC17, MEDC17, MED17, ME7., BOSCH, 0261, MOTRONIC, PM3, PPD, 5WS4, 5WK9, SID80.
2. Positive keywords — if any definitive keyword is found anywhere in the binary, accept immediately:
   • SIMOS — ECU family string
   • 5WP4 — Siemens SIMOS part number prefix
   • 111s21 — project code with leading sequence number
   • s21_ — project code with underscore separator
   • cas21 — calibration dataset prefix
3. Header magic + size gate — for "dark" bins with no ASCII:
   • 524 KB + first 2 bytes \xf0\x30 → accept
   • 262 KB + first 2 bytes \xc0\x64 or \xfa\x00 → accept
   • 131 KB + first byte \x02 → accept

"Dark" bins: Most SIMOS binaries contain no readable ASCII strings at all. Detection relies heavily on header magic + exclusion for these files. When strings ARE present, they are definitive identifiers.

Patterns (when strings are present)

Family resolution

The ecu_family field is resolved with cascading priority:

1. Specific SIMOS label from binary (e.g. SIMOS 2441 → normalised to SIMOS 2441).
2. Generic SIMOS keyword + size inference (262 KB → SIMOS2, 524 KB → SIMOS3).
3. Size-only inference when no SIMOS string exists.
4. Fallback: SIMOS.

PPD1.x

Overview

Siemens/VDO diesel ECUs for VAG 2.0 TDI PD (Pumpe-Düse / unit injector) engines, approximately 2003–2008. Three sub-variants: PPD1.1, PPD1.2, PPD1.5. Succeeded by Continental SID families when common-rail replaced unit injection.

Binary format

Detection strategy

1. Exclusion — reject if any Bosch / Marelli signature found in the first 512 KB (EDC17, MEDC17, MED17, ME7., BOSCH, PM3).
2. Positive detection — accept if at least one PPD signature is present anywhere in the full binary:
   • PPD1. — definitive PPD family identifier
   • 111SN — repeated SN project code blocks
   • CASN — calibration dataset prefix

No file-size gate is applied — detection is purely signature-based (after exclusion).

Ident record format

<serial>--    111SN<project>111SN<project>111SN<project>CASN<cal>.DAT    <VAG part> R4 <disp> PPD1.<ver>

Example:
  6576286135--    111SN100K5400000111SN100K5400000111SN100K5400000CASN1K54.DAT    03G906018DT R4 2.0l PPD1.2
  ││││││││││      │││││││││││││││                                ││││││││││││    │││││││││││ ││ ││││ │││││││
  │                │SN project codes (repeated 3×)                │CASN1K54.DAT   │03G906018DT│  │2.0l│PPD1.2
  │6576286135 ← serial code (10 digits)                           │← calibration   │← OEM part│  │    │← family
                                                                                               │R4 ← config

Extracted fields

SID 801 / SID 801A

Overview

PSA (Peugeot/Citroën) and Ford HDi diesel ECUs (DW10/DW12 engines, 2.0 HDi and 2.2 HDi), approximately 2001–2006. SID801A is a minor hardware revision — binary format is identical.

Binary format

Two binary sub-types

Detection strategy (4-phase)

1. Size gate — exactly 524 288 bytes. Eliminates >99% of non-SID801 bins.
2. Exclusion — reject if any exclusion signature found in full binary: EDC17, MEDC17, MED17, ME7., SID803.
3. Positive signatures — search first 128 KB for 5WS4 (hardware part prefix) or PM3 (project code prefix). First hit → accept.
4. Header magic fallback — for "dark" Type A bins with no embedded signatures, accept if first 4 bytes match either known SID801 header.

Ident record format (Type B only)

5WS4xxxxX-T <9-digit serial> <date+serial>S2<version>

Example: 5WS40145A-T 244177913   04020028014941S220040001C0
         │││││││││││ │││││││││   ││││││││││││││││││││││││││
         │5WS40145A-T│244177913  │04020028014941 ← date/serial block
         │← hardware  │← software version (9-digit serial — PRIMARY MATCH KEY)
                                               S220040001C0 ← S-record version

The 9-digit serial is unique per software calibration release and serves as the primary matching key for recipe lookup.

Extracted fields

Project and S-record references

SID 803 / SID 803A

Overview

PSA (Peugeot/Citroën), Ford, and Jaguar/Land Rover diesel ECUs from the mid-2000s onward. Two sub-groups distinguished by file size: SID803 (smaller, 458–462 KB) and SID803A (2 MB full flash).

Binary sub-groups

Detection strategy (3-phase)

1. Size gate — file size must be one of 458 752, 462 848, or 2 097 152 bytes. Immediate rejection otherwise.
2. Exclusion — reject if any exclusion signature found in the first 512 KB:
   • EDC17, MEDC17, MED17, ME7. — Bosch families that may share file sizes
   • PM3 — strongest negative signal. If PM3 is present, the binary is SID801, never SID803. This is the single definitive discriminator between the two Siemens diesel families.
3. Detection signatures — accept if at least one is found in the first 512 KB:
   • 111PO — PO block marker (highest specificity)
   • PO2 — PO2xx project code prefix
   • PO3 — PO3xx project code prefix
   • S122 — S122-series S-record reference (SID803A specific)
   • SID803 — explicit family string (not always present)

SID801 vs SID803 — the PM3/PO discriminator

PM3 present → always SID801, never SID803
PO  present → always SID803, never SID801

SID801 excludes SID803; SID803 excludes PM3. The two are fully disjoint.

Ident record format (SID803A, 2 MB files only)

5WS4xxxxX-T  <14-17 digit serial>

Example: 5WS40262B-T  00012345678901234

The serial portion (14–17 digits) is extracted as software_version and used as the primary match key.

Extracted fields

S-record series comparison

The higher S122 series is a distinguishing marker for SID803A versus SID801's S118/S120 range.

EMS2000

Overview

Volvo S40/V40/S60/S70/V70 T4/T5 turbo petrol ECUs (1996–2004). Also known as Siemens EMS2000 / Fenix 5. This is a "dark" ECU family — the binary is essentially pure machine code and calibration data with almost no embedded ASCII metadata.

Binary format

Detection strategy (exclusion-only, 3-phase)

1. Size gate — exactly 262 144 bytes.
2. Exclusion — reject if ANY known manufacturer signature is found anywhere in the full binary. The exclusion list is comprehensive:
   • Bosch modern: EDC17, MEDC17, MED17, ME7., BOSCH, 0261, 0281, MOTRONIC, SB_V, Customer.
   • Bosch legacy: /M1., /M2., /M3., /M4., /M5. (DAMOS strings)
   • Siemens SIMOS: 5WP4, SIMOS, s21, cas21
   • Siemens SID: 5WS4, PM3, PO, SID80
   • Siemens PPD: PPD, SN1, CASN, 03G906
   • Siemens Simtec: 5WK9
   • Siemens internal: 111PM, 111PO, 111SN, 111s2, CAPM, CAPO
   • Delphi/Delco: DEL, DELCO, DELPHI
   • Marelli: MAG, MARELLI, IAW
   • Denso: DENSO
3. Header magic — accept only if first 4 bytes match \xc0\xf0\x68\xa6. This is based on a single confirmed sample — not guaranteed to be consistent across all EMS2000 variants.

Lowest confidence extractor. EMS2000 produces no software_version, no match_key, and minimal metadata. The extractor exists primarily to prevent EMS2000 bins from being classified as "Unknown."

Extracted fields

The only pattern searched is volvo_vin (YV1 + 14 alphanumeric characters), and most EMS2000 dumps do not contain the VIN.

Confidence scoring internals

Dark bin impact

Siemens families are disproportionately affected by the "dark bin" phenomenon — binaries with little or no readable ASCII content. This affects confidence scoring:

IDENT BLOCK MISSING warning

Fires when software_version is None for families where absence is abnormal. For Siemens, this applies to:

Simtec56 · SIMOS · SIMOS2 · SIMOS3 · PPD1.1 · PPD1.2 · PPD1.5 · SID801 · SID801A · SID803 · SID803A

Exception: EMS2000 — software version absence is normal and expected. No warning is emitted.

Cross-family disambiguation

Several Siemens families share file sizes or structural features. The table below summarises how each potential conflict is resolved:

← Back to family reference · Back to README