LIN Transceiver: ISO 17987/J2602 Auto-Baud & Wake
← Back to: Automotive Fieldbuses: CAN / LIN / FlexRay
A LIN transceiver is the single-wire PHY that makes LIN reliable on real harnesses: it controls edge slew, auto-baud locking, sleep/wake behavior, and fault survivability so frames still decode across noise, capacitance, and temperature.
This page turns those knobs into measurable margins (sync/framing/checksum/false-wake) with a bring-up plan and checklist, without drifting into protocol encyclopedias.
What is a LIN Transceiver (and where it sits)
A LIN transceiver is the physical-layer interface that converts a microcontroller’s logic-level TXD/RXD into a robust single-wire bus waveform on a vehicle harness, while adding automotive-grade wake/sleep behavior, EMI-friendly slew control, and protection/diagnostics.
Where it sits in the system
- MCU side: LIN UART/SCI provides the digital TXD/RXD interface (protocol framing is handled above).
- Transceiver side: generates/receives the electrical bus levels (dominant/recessive) on a real harness.
- Harness side: connector, cable capacitance, EMI coupling, ESD/surge return paths, and node loading.
Responsibility split (avoids misdiagnosis)
LIN in one sentence + typical nodes
LIN is a single-master network where a master schedules communication and multiple low-cost nodes respond—ideal for body/comfort functions where cost and low-power behavior matter.
Compared to CAN, LIN targets lower speed and lower cost for these nodes—further network architecture discussion is out of scope here.
Scope guard
This chapter only establishes system boundaries. Electrical details (thresholds, slew, wake robustness, harness effects) are addressed in later chapters of this LIN transceiver page.
Minimal protocol facts you must know (ISO 17987 + J2602)
This section keeps only the protocol facts needed for hardware selection and bring-up debugging. It avoids protocol-stack depth and focuses on how fields influence auto-baud, timing margin, slew/EMI trade-offs, and false wakes.
Frame skeleton (what shows up on the wire)
Classic vs enhanced checksum (one sentence each)
- Classic: checksum primarily protects the data bytes (useful when isolating payload corruption vs timing).
- Enhanced: checksum coverage expands conceptually to include identifier context, improving error visibility when IDs/payload associations drift.
In practice, repeated checksum issues that correlate with temperature/harness/EMI are often physical-layer margin problems, not “bad math.”
J2602 positioning (why it matters to transceiver choice)
SAE J2602 is commonly treated as a tighter, automotive-oriented constraint set that sits within LIN compatibility expectations. For a transceiver page, the actionable point is simple:
- Confirm whether the device is specified/qualified for J2602-oriented requirements in the target OEM/Tier-1 flow.
- Treat “LIN compatible” and “meets J2602 expectations” as not always identical in validation planning.
- If uncertainty exists, validation must stress auto-baud margin, slew/EMI settings, and wake robustness on real harnesses.
Field → impact map (hardware debugging lens)
Quick check: verify break detection stability across harness variants and with worst-case slew settings.
Quick check: measure edge-to-edge timing consistency (jitter of threshold crossings) rather than only amplitude.
Quick check: split failures by PID and node; if one PID dominates, suspect electrical margin at that node/harness branch.
Quick check: correlate error rate with payload length and temperature; “length-sensitive” failures point to edge/threshold margin.
Quick check: compare checksum-fail distribution against supply dips, wake events, and harness movement.
Electrical fundamentals: dominant/recessive, pull-up, thresholds, bus load
LIN reliability is primarily set by electrical margins: the dominant/recessive drive paths, the pull-up and total capacitance that shape edges, and the receiver’s threshold behavior under real harness noise and ground shifts. This chapter builds a compact model that explains EMI sensitivity, mis-detections, and waveform distortion without drifting into protocol-stack details.
Dominant vs recessive is a path story (not a logic story)
- Dominant typically comes from an active pull-down drive path inside the transceiver (bus is forced low).
- Recessive is often a release + pull-up process: the driver lets go and the bus returns via the master pull-up path.
- Because recessive recovery is commonly RC-limited, slow edges and threshold-crossing jitter usually show up there first.
Pull-up + total C sets edge shape and timing margin
Receiver threshold, noise window, and single-wire reference ground
- A receiver decides bits by threshold crossing. With slow edges, small disturbances cause large timing shifts at the crossing point.
- In a single-wire system, “ground” quality matters: ground bounce / offsets shift the receiver’s effective reference and move the crossing time.
- The real margin is a combination of threshold distance, edge slope, and crossing jitter—not just waveform amplitude.
Symptom → likely electrical cause (quick mapping)
Budget row: capacitance and rise-time sanity checks
- Capacitance budget: Ctotal = Ccable + ΣCnode + Cprot (parasitic).
- Time-constant lens: trise ~ Rpullup × Ctotal (conceptual).
- Pass criteria (placeholder): trise < X% of bit time across worst-case harness and temperature.
Auto-baud is fundamentally a timing measurement of threshold crossings, so this electrical model directly predicts auto-baud robustness.
Scope guard
This chapter builds the electrical model. Auto-baud failure modes are covered in the next chapter; detailed slew/EMI tuning is addressed later; protection selection and placement are treated in the protection and layout chapters.
Auto-baud detection: what it measures, what breaks it
Auto-baud is a timing measurement of threshold crossings during the Sync pattern—not a measurement of amplitude. When edges slow down or the receiver reference moves, the measured crossing time jitters and the computed baud can fall outside the valid window. This chapter explains the measurement chain and the most common break points seen on real harnesses.
What auto-baud measures (conceptual chain)
- Input: Sync edges on the LIN wire.
- Edge detect: locate threshold crossings (decision point).
- Timer window: measure edge-to-edge intervals.
- Compute: derive tBIT / baud estimate.
- Margin check: accept only if within the valid tolerance window (placeholder ±X%).
Slow slew: the hidden amplifier of crossing jitter
Threshold drift and reference movement (single-wire reality)
- Temperature and supply conditions can shift the receiver’s effective reference, moving the crossing point even if the harness is unchanged.
- Ground bounce and offsets change the local “zero,” so the same LIN wire waveform becomes a different crossing timeline at different ECUs or measurement locations.
- A stable auto-baud design focuses on crossing consistency under worst-case harness and ground conditions.
Ringing/reflection: multi-crossing traps
Measurement artifacts (why “scope looks OK” can be false)
- Probe ground lead inductance can exaggerate or hide ringing.
- Bandwidth limits can smooth edges and under-report crossing jitter.
- Measuring far from the receiver decision point can miss local ground noise and return-path effects.
Diagnostic matrix (symptom → quickest check → direction)
Direction: reduce Ctotal impact, review pull-up/slew settings, remove multi-crossing ringing.
Direction: confirm threshold/reference behavior and noise window under temperature drift.
Direction: treat as decision-point jitter, not amplitude; review ground return and parasitics.
Direction: re-balance protection parasitics and placement; suppress multi-crossing behavior.
Pass criteria (placeholder): computed baud stays within ±X% and auto-baud fail rate remains below X per 1k frames under worst-case harness and temperature.
Scope guard
This chapter explains the auto-baud measurement chain and failure modes. Detailed slew/drive tuning and layout return-path fixes are handled in later chapters; full bring-up sequencing is consolidated in the validation plan chapter.
Low-power states: sleep/standby, local vs bus wake
Low power is only meaningful when wake reliability is engineered and measurable. This chapter turns sleep/standby into an executable policy: clear mode definitions, wake-source attribution, false-wake root causes, and a data-oriented KPI set for validation and production monitoring.
Mode vocabulary: what “Normal / Standby / Sleep” typically means
Datasheet naming differs across vendors; the robust approach is to map each mode to which blocks are powered (Tx/Rx, wake detect, reference/threshold domain, diagnostic visibility).
Wake sources: Bus vs Local vs Timed (if supported)
A low-power design must ensure wake-source attribution: each wake event is tagged as bus/local/timed (or unknown) for field diagnostics and false-wake control.
False-wake root causes (practical classification)
KPIs to log and gate (data-oriented)
The core rule is simple: optimize Iq only after wake reliability is proven, and prove it with measurable KPIs rather than waveform screenshots.
Scope guard
This chapter focuses on transceiver power states, wake-source attribution, and measurable KPIs. Detailed EMC component selection and board-level return-path design are handled in dedicated protection and layout chapters.
EMI-friendly slew-rate control: knobs, tradeoffs, and pass/fail
Slew-rate control is a PHY knob with two competing goals: reduce high-frequency energy for lower EMI, while preserving timing margin and stable threshold crossings for sync/auto-baud and error robustness. This chapter defines the knobs, the trade-offs, and a measurable pass/fail loop without turning into an EMC handbook.
Two cause chains (why this is a trade-off)
- Faster slew → more high-frequency content → higher risk of radiated/conducted peaks.
- Slower slew → smoother edges → EMI peaks often decrease.
- Slower slew also increases time spent near Vth → greater sensitivity to noise and threshold drift → reduced timing margin.
LIN-relevant knobs (keep it actionable)
The validation goal is not “slowest possible,” but a setting that meets EMI targets while keeping auto-baud and frame robustness stable on the real harness.
Phase strategy: EVT → DVT → PVT (how to converge safely)
Pitfalls to avoid (common failure patterns)
- Only validating near-end waveforms; far-end threshold-crossing stability is what drives robustness.
- Choosing the slowest slew and triggering sporadic auto-baud failures on long harness or high Ctotal.
- Ignoring multi-crossing ringing after protection changes; slower slew can worsen threshold-region ambiguity.
- Using bandwidth limits or poor probe setup that masks crossing jitter and creates false confidence.
Pass/Fail loop (LIN-centric metrics, measurable and repeatable)
The correct acceptance is the intersection of EMI improvement and stable LIN robustness counters under controlled, repeatable conditions.
Scope guard
This chapter provides only LIN-relevant slew/drive knobs and a validation loop. EMC standards, detailed component selection (CMC/TVS), and full layout methodology are handled in dedicated EMC/protection and layout chapters.
Protection & diagnostics: shorts, thermal, undervoltage, fault reporting
Robustness is defined by fault behavior, not by a single headline rating. This chapter maps common abuse cases (shorts, thermal shutdown, undervoltage, bus stuck) to observable symptoms and loggable counters, and highlights how port protection parasitics can quietly break Sync/auto-baud margin.
How to read “robustness” without getting misled
- Survival: short-to-VBAT/GND capability and recovery rules (what resets, what latches).
- Power/thermal: undervoltage and over-temperature behaviors that can mimic bus issues.
- Bus abnormal: stuck dominant/recessive detection and safe output states.
- Reporting: flags/counters/timestamps that enable attribution and prevent “random” diagnosis.
The objective is a closed loop: fault → response → symptom → log.
Short-to-VBAT / GND: what matters and what to observe
- Is the short rating defined as continuous, time-limited, or repetitive pulses?
- Does recovery require mode cycling or only removal of the fault?
- Is the safe state deterministic (forced recessive, Tx off, or latched fault)?
- Bus level (dominant/recessive hold) and whether toggling is blocked.
- RxD/TXD behavior during and after the fault window.
- Fault flags/counters with timestamps for attribution.
A short event that “recovers” electrically can still degrade timing margin; regression counters after recovery are a practical gate.
Thermal shutdown & undervoltage (fail-safe semantics)
A clean fail-safe definition should include: entry condition, observable outputs, and exit/recovery rule.
Diagnostics signals to expect (common features)
Protection parasitics: LIN-specific side effects to watch
- TVS capacitance increases C_total → slows recessive rise → increases threshold-crossing sensitivity for Sync/auto-baud.
- Series R / RC can reduce ringing but may over-slow edges and shrink timing margin under worst-case harness load.
- Placement affects the effective stress path and ringing behavior; a “stronger” protector can still worsen stability if parasitics dominate.
The fastest sanity check for “protection change broke the link” is to correlate Sync fail and frame error counters with protection capacitance/placement changes.
Scope guard
This chapter defines fault behaviors, observables, and diagnostics mapping. EMC standards, detailed TVS/CMC selection rules, and full layout methodology are covered in dedicated EMC/protection and layout chapters.
Reference hardware: schematics that actually work (and why)
A reliable LIN port is mostly decided by a small set of placement and parasitic rules. This chapter provides a minimum viable reference (connector → protection → optional damping → LIN pin) and explains why each element exists, with the three most common mistakes that break auto-baud and robustness.
Minimum viable port topology (what “works” across harness variants)
- Connector → TVS → (optional R / RC) → LIN pin is the practical baseline.
- Power integrity matters: supply filtering and close decoupling stabilize threshold behavior.
- Return path is part of the circuit: port protection must have a short, direct ground return.
Why each element exists (cause chain, not folklore)
Connector and return path (what “short” means)
Port protection is only effective when its stress current returns through a short, direct ground path. Long return loops increase local ground movement and can amplify threshold-crossing jitter, causing Sync fail and intermittent frame errors.
The three mistakes that break stable LIN ports
- Over-capacitance TVS (or high leakage): increases C_total and narrows Sync/auto-baud timing margin.
- Protection too far from the connector: stress enters the PCB first; effective clamping and ringing control degrade.
- Ground return detour: long loops raise ground bounce and shift threshold crossings under burst conditions.
Scope guard
This chapter provides a block-level reference schematic and placement-driven rules. Detailed routing, stack-up, and EMC compliance methodology are handled in the layout and EMC chapters.
Layout & harness pitfalls: stubs, grounding, measurement traps
LIN failures are often caused by reference ground movement, harness branches, and measurement artifacts. This chapter lists real-world traps and provides fast checks that isolate threshold-crossing instability from “protocol-looking” symptoms.
Single-wire ground reference: why thresholds become unstable
- Receiver decisions depend on local ground. ESD return currents and load transients can move the local ground, shifting the effective threshold crossing time even when bus amplitude looks acceptable.
- Ground bounce increases crossing jitter around Vth, reducing Sync/auto-baud margin and increasing sporadic framing or checksum errors.
- The fastest confirmation is a measurement that compares connector-side and LIN-pin-side behavior under the same burst condition.
Harness branches (stubs): reflections that hit auto-baud and wake reliability
LIN is tolerant compared to faster buses, but long branches and heavy node capacitance can still create ringing and slow edges. The practical risk is not “eye closure” but multiple threshold crossings and longer time spent near Vth.
- Longer stubs increase ringing probability near the Sync region.
- More nodes raise C_total, slowing recessive rise and shrinking timing margin.
- Ringing-induced re-crossing can look like “random” Sync fail unless measured at the decision point.
Layout hot spots: connector, protection, transceiver, MCU
Measurement traps (why “scope looks OK” is not a pass)
Fast debug order (to prevent endless tuning)
- Validate measurement: probe, BW, trigger, decoder threshold, measurement point.
- Verify return paths: TVS ground return, loop area, sensitive ground separation.
- Check harness branches: stubs, node count, effective capacitance and ringing.
- Then tune knobs: slew/drive and protection damping, using counters as pass/fail gates.
Scope guard
This chapter focuses on LIN port layout/harness pitfalls and measurement correctness. EMC standards and component-level selection rules remain in dedicated EMC/protection chapters.
Bring-up & validation plan: what to test, in what order
Stable LIN bring-up requires a fixed order: establish static electrical correctness, prove basic traffic, stress auto-baud margin, validate sleep/wake reliability, then run EMC/ESD regression as a before/after gate. Each step uses measurable counters and clear pass/fail placeholders.
Test order principle (to prevent endless debugging)
- Static electrical (rails, idle level, fail-safe behaviors).
- Basic frames (stable traffic at nominal conditions).
- Auto-baud extremes (edge/threshold margin under worst harness).
- Low-power (sleep Iq, wake attribution, false-wake rate).
- EMC/ESD regression (repeat key steps and compare deltas).
Step 1 — Static electrical gate
- Check: VBAT/Vsup stability, idle level, undervoltage and thermal flags.
- Log: supply dip events, UV flag, thermal flag (if available).
- Pass: dips/flags remain within X over Y minutes (placeholders).
Step 2 — Basic traffic gate (nominal conditions)
- Run: fixed PID/length frames at nominal VBAT and a short harness.
- Log: checksum error, framing error, bus stuck status.
- Pass: error rates < X per 1k frames (placeholder).
Step 3 — Auto-baud margin sweep (worst-case harness)
- Harness: short / long / stub-heavy
- Voltage: VBAT min / nom / max
- Temperature: cold / room / hot
- sync fail count (primary)
- checksum / framing errors
- bus stuck indicators
Failures during this sweep should first re-check measurement point and thresholds (layout/harness chapter), then adjust slew/drive knobs.
Step 4 — Low-power and wake reliability gate
- Measure: sleep Iq, wake debounce, wake latency.
- Log: wake events and wake cause attribution (bus/local/timed/unknown).
- Pass: false-wake rate < X per 24h and Iq < X (placeholders).
Step 5 — EMC/ESD regression gate (before/after comparison)
- Repeat Step 2, Step 3, and Step 4 with identical scripts and logging.
- Compare deltas in sync fail, frame errors, and false-wake trends.
- Pass if deltas remain within X over Y repeats (placeholders).
Scope guard
This chapter provides an executable bring-up order with counters and regression gates. Detailed lab setup, equipment selection, and standard-specific EMC procedures are handled elsewhere.
Engineering checklist (Design → Bring-up → Production)
A stage-gated checklist that turns LIN stability into verifiable evidence: configuration intent, measurable margins, and production-ready diagnostics. Each item includes what to capture and what “pass” looks like (placeholders).
Design Gate
- J2602 requirement check: confirm whether the node must meet SAE J2602 constraints (timing/behavior limits). Evidence: requirement note + transceiver compliance statement. Pass: J2602 = Yes/No documented.
- Slew/drive plan: define at least two profiles (bring-up vs production) and the rule to select them. Evidence: register map / strap table. Pass: profile selection rule recorded.
- Protection parasitics budget: cap/leakage of TVS and any optional series R/RC are accounted for in edge and auto-baud margin. Evidence: BOM params + “C_total” budget. Pass: C_total ≤ X pF (placeholder).
- Wake strategy: bus wake + local wake policy (debounce/filter) and wake-cause attribution fields. Evidence: state table + log fields. Pass: wake cause = bus/local/unknown.
- Return-path intent: connector/TVS return kept short; protection currents do not traverse sensitive ground. Evidence: layout screenshot with marked loop. Pass: return length ≤ X mm (placeholder).
- Diagnostics visibility: define counters/flags required for serviceability (sync fail, framing, checksum, wake, UV/thermal). Evidence: log schema. Pass: all fields present in firmware build.
- LIN transceivers: TI TLIN1029-Q1, NXP TJA1021, Infineon TLE7258D
- LIN mini-SBC / LIN+LDO: TI TLIN1028S-Q1, TI TLIN1431-Q1, NXP TJA1128
- ESD/TVS for LIN: TI ESD1LIN24-Q1, ST ESDLIN03-1BWY, ST ESDLIN1524BJ, Littelfuse AQ24-01FTG
- Optional series resistor: Yageo RC0603FR-0733RL (33 Ω, 0603 example)
Bring-up Gate
- Waveform baseline: capture at two points (connector and LIN pin) with consistent probe and bandwidth settings. Evidence: waveform pack + scope setup snapshot. Pass: baseline stored.
- Basic frames stability: run fixed PID/length traffic for X minutes and quantify error rates. Evidence: checksum/framing counters. Pass: errors < X per 1k frames.
- Auto-baud margin scan: sweep temperature/voltage/harness; record sync-fail trends. Evidence: sync_fail_count vs (T, VBAT, harness). Pass: zero lock failures within envelope.
- Slew profile validation: compare fast/medium/slow against stability and emissions intent. Evidence: A/B counter deltas. Pass: chosen profile documented with reason.
- False-wake quantification: measure sleep Iq and false-wake rate; tag wake source when possible. Evidence: wake_events + wake_cause stats. Pass: false-wake < X per 24 h.
- Protection side-effects check: validate TVS/series R options do not reduce sync robustness. Evidence: sync-fail and edge timing comparison. Pass: Δmargin ≤ X% (placeholder).
Production Gate
- Temperature drift sampling: cold/room/hot sampling with the same traffic script and the same counters. Evidence: per-corner report. Pass: sync fail = 0 within spec corners.
- ESD/EMC regression: re-run core scripts after stress; focus on permanent deltas rather than one-off spikes. Evidence: before/after delta log. Pass: Δerrors ≤ X (placeholder).
- Fault injection visibility: verify short, bus-stuck, UV/thermal conditions are observable and logged. Evidence: fault→response→log matrix filled. Pass: no “silent failures”.
- Recovery behavior: confirm how the node exits faults (auto-retry vs reset vs mode change). Evidence: recovery steps. Pass: deterministic recovery defined.
- Test station consistency: lock probe, threshold, decode settings to avoid false correlation. Evidence: station setup file. Pass: identical settings across stations.
Applications (Body/Comfort nodes, small ECUs, sensors/actuators)
Application buckets focus on what actually constrains a LIN transceiver design: sleep current, false-wake robustness, bus-fault survivability, and measurable bring-up margins on real harnesses.
Bucket A · Body / Comfort (many small nodes)
Typical nodes: door modules, seat controls, lighting, mirrors.
- Top constraints: lowest Sleep Iq, strong false-wake immunity, clear wake-cause attribution.
- Watch metrics: sleep Iq, false-wake per 24 h, wake cause histogram.
- Example MPNs: LIN PHY: TLIN1029-Q1, TJA1021; LIN mini-SBC: TJA1128, TLIN1431-Q1; LIN ESD: ESD1LIN24-Q1, ESDLIN03-1BWY.
Bucket B · Small ECUs / Zone sub-nodes (harness variability)
Typical nodes: switch panels, HVAC flaps, small distributed controllers.
- Top constraints: wide auto-baud margin across harness variants, repeatable diagnostics, clean layout return paths.
- Watch metrics: sync fail trend vs (T, VBAT, harness), framing/checksum rates, bus-stuck flags.
- Example MPNs: LIN PHY: TJA1021, TLE7258D; LIN+LDO: TLIN1028S-Q1; LIN ESD: ESDLIN1524BJ, AQ24-01FTG.
Bucket C · Sensors (noisy edges, ground offsets)
Typical nodes: position sensors, rain/light sensors, small edge sensing modules.
- Top constraints: robust threshold behavior under ground noise, careful TVS capacitance/leakage, temperature drift awareness.
- Watch metrics: sync fail vs temperature, error rate vs VBAT dips, post-ESD delta on counters.
- Example MPNs: LIN PHY: TLIN1029-Q1, TJA1021; LIN ESD: ESD1LIN24-Q1, ESDLIN1524BJ, Infineon ESD24VS2U.
Bucket D · Actuators (bus faults and recovery matter)
Typical nodes: locks, small motors, valves, pumps.
- Top constraints: short-to-battery/ground survivability, thermal/UV behavior clarity, deterministic recovery after faults.
- Watch metrics: thermal/UV flags, bus-stuck detection, fault counters and recovery latency.
- Example MPNs: LIN PHY: TLIN1029-Q1, TLE7258D; LIN SBC (watchdog + wake policy): TLIN1431-Q1; LIN ESD: AQ24-01FTG, ESDLIN03-1BWY.
When a transceiver is enough vs when an SBC is warranted (conditions only)
- Transceiver-only is often enough when: the ECU already has power management and watchdog, wake policy is simple, and serviceability can be met with firmware logs.
- An SBC is often warranted when: unified wake/low-power policy, hardware watchdog/reset policy, or integrated high-side / limp-home behavior is required for the node.
- Example LIN SBC MPNs: TI TLIN1431-Q1, NXP MC33689, NXP UJA1069, NXP TJA1128.
Notes: part numbers are examples to anchor design discussions. Always verify automotive grade, package, suffix, and current lifecycle/availability before locking BOM.
Recommended topics you might also need
Request a Quote
FAQs (Troubleshooting, fixed 4-line answers + JSON-LD)
Long-tail troubleshooting only. Each answer is strictly four lines with measurable placeholders (X) so results can be logged, compared, and gated.
Auto-baud fails intermittently but the waveform “looks normal” — threshold crossing jitter or probing artifact first?
Slower slew passes EMI, but communication becomes unstable — check Sync edge or sampling margin first?
Checksum errors increase at low temperature — oscillator error or slower edges first?
Sleep Iq meets spec but false wakes are frequent — LIN pin coupling or pull-up leakage path first?
After ESD, communication still works but auto-baud margin is worse — fastest degradation check?
Harness length increase causes intermittent framing errors — cable capacitance or stub reflections first?
Problems appear after switching TVS vendor — capacitance difference or layout/return path first?
Master is stable, but one slave drops frames — node capacitance or ground reference offset first?
Bus occasionally gets stuck dominant — TXD dominant timeout or external short/moisture first?
Wake events are missed — wake-cause attribution/logging or gateway timed wake strategy first?
Claims J2602 compatibility, but field stability is poor — how to verify “constraints” first?
Dominant amplitude looks sufficient on the scope, yet communication is poor — receiver threshold/noise or sample window first?
Data note: keep metric names consistent across bring-up and production logs (sync_fail_count, framing_error, checksum_error, false_wake_rate, bus_stuck events) to enable trend and delta comparisons.
