Confirmed PIDs

Confirmed PIDs — Voge 900DSX (LX900-A / BMW F900 platform, Loncin 4M96001 engine)

Every request in this table has been verified to return a real, correctly-decoded value on this bike (not just "no error") via live on-bike testing with a Bluetooth ELM327 adapter. This should generalize to other bikes on the same BMW F900 platform (Voge/Loncin 4M96001), though exact values may vary by model year/ECU revision — verify against a real ride before trusting any of it.

This is the same list the app polls (BluetoothOBDManager.CONFIRMED_PID_MAP), and the app's own Bike Info tab shows a live version of it — a green/red/gray dot per row for whether the last request on the current connection actually matched what was expected, not just a static claim.

Headers used: 7E0 = engine ECU (Bosch ME17.8.10), 7E1 = ABS/IMU module.

Reference vehicle (via Mode 09, cross-checked with another OBD app — PID 0902/0904): VIN LLCVPX1A8RA150987, calibration ID 8814LX28502BLXMT. Worth checking against when cross-referencing another BMW F900-platform bike's CAN/DID data (e.g. the community CAN-ID spreadsheet) — different calibration revisions can shift layouts even on the same platform.

✅ Confirmed working

Standard OBD-II (SAE J1979, Mode 01/03)

These follow the standard and should work on most OBD-II-compliant vehicles, not just this one.

Header	PID	Signal
7E0	010C	RPM
7E0	010D	Speed
7E0	0111	Throttle position
7E0	0105	Coolant temp
7E0	0131	Distance since DTC clear (not the real lifetime odometer — resets on Mode 04 clear)
7E0	0142	Battery / control module voltage
7E0	010F	Intake air temp
7E0	0104	Engine load
7E0	0146	Ambient air temp
7E0	010B	Intake manifold pressure (MAP)
7E0	010E	Timing advance
7E0	011F	Engine run time
7E0	0121	Distance with MIL on
7E0	013C	Catalyst temp, Bank 1 Sensor 1
7E0	013D	Catalyst temp, Bank 2 Sensor 1
7E0	0106	Short-term fuel trim, Bank 1
7E0	0107	Long-term fuel trim, Bank 1
7E0	0108	Short-term fuel trim, Bank 2
7E0	0109	Long-term fuel trim, Bank 2
7E0	0114	O2 sensor voltage/trim, Bank 1 Sensor 1
7E0	0118	O2 sensor voltage/trim, Bank 2 Sensor 1
7E0	0143	Absolute load value
7E0	0144	Commanded equivalence ratio
7E0	0145	Relative throttle position
7E0	0147	Throttle position B
7E0	0149	Accelerator pedal position D
7E0	014A	Accelerator pedal position E
7E0	014C	Commanded throttle actuator
7E0	0101	MIL status / stored DTC count
7E0	03	Stored DTCs (Mode 03, decoded per SAE J2012)

Manufacturer UDS DIDs (7E1 header) are covered in Not working below — as of 2026-08-09 none of them have ever returned real data on this bike; see that section for how this was checked.

📡 Confirmed CAN broadcast frames (passive — no request needed)

Found 2026-08-15 in the first laptop-direct session (tools/obd_terminal.py over the adapter's Bluetooth SPP COM port, captures diffed with tools/can_diff.py against known vehicle states). All 11-bit IDs. Rates are what a clone ELM327 delivers filtered to one ID (ATCRA); unfiltered full-bus capture drops most frames past ~150 frames/s, so unfiltered rates read ~5x lower.

CAN ID	Bytes	Signal	Encoding	Verified how
126	b0 low nibble	Gear	0=N, 1-6 = gear	Live shift-through: N→1→...→6 each matched the dash exactly; ~50Hz filtered
126	b0 bit 0x20	Transient during shifts	Set while working through gears with the bike being rocked (engine off), clear once settled; NOT clutch (tested), NOT sidestand (tested), NOT wheel motion alone (tested). Exact meaning open — mask it off when reading gear.	Elimination tests, engine off
120	b0-1	RPM	16-bit BE, raw = RPM (idle read ~1250, fluctuation visible)	Engine idling vs off
120	b4 bit 0x40	Neutral flag	0x42 in N, 0x02 in gear	Same shift tests as 126
3FF	b0-2	Odometer	24-bit BE, km (00 2B 5C = 11100 km, exact dash match)	Constant across a whole stationary session; final proof = increments after a ride. ~0.2Hz

TPMS candidates (dash shows tire pressures, so they reach this bus): 3A1 (A9 C2 56 57 00 00 00 0A) and 3A2 (AD C1 AD AD C1 BB 00 00) — slow (~0.1Hz), payload constant all session, and 3A2's two three-byte groups look like a front/rear pair. Reference for decoding: dash read front 2.3 / rear 2.7 bar at capture time (2026-08-15, all captures in that session). No byte maps to those values under the common encodings (kPa, psi, 25-mbar steps) with one shared formula, so the decode needs a changed pressure to diff against — recheck after a ride (warm tires) or an inflate/deflate test. Both IDs consistently carry the clone adapter's <DATA ERROR checksum marker; payloads are nonetheless bit-identical across hours, so the data itself is stable.

Negative results from the same session, all with the bike stationary (each tested engine off AND idling where it could matter):

• Lean angle is not broadcast while stationary. Full-bus tilt tests (bike leaned side to side) engine-off and engine-running: no byte anywhere on the bus tracks the tilt. ID 092 — the bus's fastest frame (100Hz filtered) — is all-zeros in every state tested; if it's the IMU frame it only fills in motion. Next step: a riding capture (laptop in a backpack works).
• Brake pressure/brake light is not visible stationary — not as a broadcast (2s squeeze/release choreography, diffed) and the 7E1 DIDs stay NO DATA even engine-running (protocol locked via 0100 first — not the cause). Motion may wake the ABS module; unproven.
• ID 123 b0-1 carry steady state flags (values shift between ignition-on/engine-run states), b6 = 4-bit rolling counter in the high nibble, b7 = checksum over the frame. ID 191 b7 is a rolling counter (looked like a tilt hit in a naive variance diff — it isn't).
• The engine-off bus inventory (ignition on): 092/12B at ~20Hz observed unfiltered, 120/123/125/ 126/12F at ~10Hz, 191/192/220 at ~4Hz, 240/241/270/2A8/2D2 at ~2Hz, 320/341/3A1/3A2/3FD/3FF sparse. (Unfiltered = undercounted; see note above.)

❌ Not working

All seven rows below were listed as confirmed in an earlier pass of this doc. Cross-checked 2026-08-09 against every real (non-simulated) ride ever recorded in the app's own database (55k+ records going back to 2026-08-01, excluding the day's 3 known-simulated test rides) — every one of these columns is exactly zero in every single real record, never once a nonzero value, which the diagnostic log's failures for the same requests corroborate.

Header	PID/DID	Signal	Response
7E0	2243F7	Gear	7F 22 31, even with a UDS Extended Diagnostic Session actively granted (50 03)
7E0	222503	ECU odometer	7F 22 31, same as Gear
7E1	222B05	Front brake pressure	NO DATA
7E1	222B06	Rear brake pressure	NO DATA
7E1	22D10D	Lean angle (bike IMU)	NO DATA
7E0	012F	Fuel level	NO DATA
7E0	015E	Fuel rate	NO DATA
7E0	43FE	Unmapped	7F 22 31, even with the session granted
7E0	43FF	Unmapped	7F 22 31, same as 43FE

Checking the recorded DB instead of just grepping the log for failures is what caught fuel level/rate and the two brake DIDs — a live spot-check earlier the same day had wrongly reported "yes, fuel level and rate work" going only off this doc's stale confirmed table. The extended-session hypothesis for 2243F7/222503/43FE/43FF was separately disproved the same day (session granted, 50 03, all four still failed identically) — root cause for any row above is still open.

Re-verified 2026-08-15 from the laptop, stationary: 2243F7/222503 still 7F 22 31 and all three 7E1 DIDs still NO DATA, in every state available standing still (ignition-only AND engine idling, protocol pre-locked via 0100). Whatever these DIDs need, it isn't just "engine running" — motion is the last untested variable. Gear and odometer no longer matter here anyway: both are now read passively from broadcasts (see the broadcast section above).

Two related open leads, not request/response PIDs so they don't fit the table above:

• Lean angle broadcast on the raw CAN bus — 22D10D never answers on request (see above), and the 2026-08-15 stationary tilt tests (see the broadcast section) proved it is not broadcast while standing still either, engine off or idling. If it's on this bus at all it only flows in motion — a riding capture is the remaining test. (Gear, which was in this same bullet list until 2026-08-15, IS broadcast and is now solved — ID 126, see the broadcast section; the reason the 2026-08-09 shift test missed it is that unfiltered clone captures drop ~80% of frames.)
• A completely different channel for fuel/gear/odometer, unexplored so far — the bike's own dash shows all three despite the OBD/UDS requests above never working, so it's getting them some other way. MOTO-HUB (a screen-mirroring app for Voge/CFMOTO/Zontes dashboards, unrelated project) confirms these dashboards have a T-Box (telematics unit) reachable over Wi-Fi Direct, separate from the OBD port - not something this app has ever tried talking to. Whether the dash also exposes anything over a direct Bluetooth link (rather than only Wi-Fi Direct to the T-Box) is unknown. Nothing implemented here yet - just a lead for whoever picks this up next.

How this list was built

Combination of: SAE J1979 standard PID scanning (sweepStandardPidSupport() in the app), manual UDS DID sweeps across the two confirmed-responsive headers (7E0, 7E1), and cross-referencing another OBD app's raw sensor dumps against real dashboard readings taken at the same time. Since 2026-08-15 there's also a phone-free route: tools/obd_terminal.py drives the adapter directly from a laptop over its Bluetooth SPP COM port (probe / raw commands / a change-only watch / CAN-ID histogram / ATMA captures, full-bus or ATCRA-filtered to one ID), and tools/can_diff.py diffs captures taken in different known vehicle states — that pair is what found every entry in the broadcast section. PIDs that returned data but couldn't be correlated to a real, verifiable value on the bike are deliberately left out of this list rather than included as a guess. As of 2026-08-09, entries are also periodically cross-checked against the app's own recorded ride database rather than trusted from the original sweep alone - see Not working above for what that caught.