Bathroom Heater / Fan: Drive, Sensors, De-Mist Control
← Back to: Smart Home & Appliances
Bathroom heater/fan engineering is about keeping de-mist and heating stable under wet, noisy conditions—by controlling heater/fan drive, sensing (T/RH), and interlocks with evidence-based tuning.
This page focuses on device-level power/EMC hardening, robust control strategies, and field-debug evidence so the unit stays quiet, safe, and reliable without frequent trips, chatter, or false touches.
H2-1 System Boundary & Use Cases
This page treats a bathroom heater/fan as a self-contained electromechanical controller: it must survive wet/condensing air, switch noisy loads, and still keep control stable. The boundary is set to protect uniqueness and avoid overlap with whole-home HVAC topics.
What this page covers (the locked boundary)
- Load control: resistive/PTC heater switching or modulation; fan drive (AC motor or BLDC) with tach/FG feedback.
- Moisture-aware control: temp + RH evidence, dew-point/condensation risk logic, and anti-oscillation (hysteresis/anti-chatter).
- Rugged power & EMC: inrush, triac/relay dv/dt stress, ESD/surge/EFT, brownout behavior, and reset immunity.
- Safety chain: over-temp, thermal fuse, fan-stall interlock, sensor-fault degrade modes, safe-off behavior.
- UI & connectivity (device-level): touch/keys/knob + indicators, and local wireless provisioning/keep-alive evidence.
Product variants included (so readers don’t mix different devices)
- Integrated unit: heater + fan + sensing + UI in one enclosure (highest coupling: heat + moisture + EMI in one cavity).
- Split system: separated fan box and heater module (longer wiring → higher surge/ESD sensitivity and more ground reference issues).
- De-mist submodule: mirror/zone anti-fog heater element (smaller power, but control stability and sensing placement dominate outcomes).
Use cases mapped to engineering targets (each is evidence-driven)
- Boost de-mist: target is “fast reduction of condensation risk.” Evidence: RH/T trend + fan airflow confirmation.
- Timed exhaust: target is “predictable moisture removal.” Evidence: tach/FG + current (backpressure changes).
- Humidity/condensation hold: target is “prevent mold without chatter.” Evidence: hysteresis + minimum-on/off timers.
- Space heating: target is “comfort without nuisance trips.” Evidence: inrush/line sag + over-temp margin.
- Night light / local indication: target is “stable, non-intrusive.” Evidence: LED PWM return paths do not corrupt touch/ADC.
Explicitly out of scope (to prevent cross-page overlap)
Whole-home HVAC control, ERV/HRV subsystems, IAQ hubs (PM/VOC/CO₂ chains), HEMS/utility metering deep-dives, cloud backend architecture, router/mesh setup tutorials, and protocol-stack deep-dive content are intentionally excluded. This page only keeps what is required to engineer a bathroom heater/fan unit reliably.
H2-2 Functional Blocks
The fastest way to keep this topic “vertical and deep” is to anchor everything to two paths: Power/Load Path (how energy reaches heater/fan safely) and Sense/Control Path (how the controller decides and proves correctness). Every later chapter will map back to one of these blocks.
Power / Load Path (what must stay stable under stress)
- AC entry & protection: surge/ESD/EFT absorption + EMI filtering; layout must control return paths and dv/dt loops.
- Low-voltage PSU: isolated or non-isolated rails (typically 5V/3.3V) with reset/brownout immunity under inrush events.
- Heater drive stage: relay/SSR/triac selection determines EMI, thermal rise, and failure mode (stuck-on vs stuck-off).
- Fan drive stage: AC motor drive or BLDC driver; speed confirmation (tach/FG) is the key “truth signal” for airflow.
Evidence hooks for this path: inrush current line sag / brownout dv/dt spikes rail droop (3.3V/5V) heater/fan current signature
Sense / Control Path (what prevents “looks OK but fails”)
- Temp & RH front-end: sensor placement + condensation resilience dominate accuracy; include fault detection for stuck/short/open states.
- Dew-point / risk logic: avoid raw RH threshold-only control; use hysteresis + minimum-on/off timers to prevent chatter.
- State machine outputs: boost → maintain → cooldown, with safe interlocks (fan stalled ⇒ heater inhibited).
- UI and indicators: touch/keys + LED/buzzer should be immune to wet surfaces and switching noise coupling.
- Local radio (optional): pairing + keep-alive counters; the design focus is power integrity and interference evidence, not cloud design.
Evidence hooks for this path: RH/T trend tach/FG confirmation sensor health flags event log (reset reason) chatter counters
Minimal BOM paths (differences only; no platform expansion)
- Entry (stability-first): relay + AC fan, 1–2 NTC points, basic RH (optional), hard safety cutoff (thermal fuse). Goal: safe and predictable.
- Mid (control stability): RH+T with placement discipline, tach/FG confirmation, de-mist state machine (hysteresis/anti-chatter), richer event codes.
- High (quiet + connected): BLDC + speed loop, improved EMC partitioning, wet-surface robust UI, device-level radio with counters and reset immunity.
H2-3 Rugged Power Entry & EMC
Bathroom environments combine condensing moisture, long wiring, and hard-switched loads. The result is a predictable stress set: surge/EFT/ESD events, common-mode noise, and load switching back-injection. This chapter converts that stress into evidence-first measurements and design levers that keep the controller stable.
Threat model: why bathrooms are harsher than “typical indoor” loads
- Moisture & condensation: water films create new leakage paths and reduce effective insulation, making high-impedance nodes easier to disturb.
- Long wires & large loop area: more inductance and stronger common-mode coupling; surge energy sees longer current paths.
- Switching loads: relay contact bounce, triac dv/dt edges, and BLDC PWM create sharp di/dt and dv/dt that re-enter the low-voltage rails.
Typical symptom mapping: random reset touch false trigger wireless drop heater/fan misfire sensor drift spikes
Evidence first: minimum measurements that isolate the root cause
- Power-up + load-step waveforms: capture 5V/3.3V rails during fan start and heater switching; confirm if BOR/UVLO is triggered.
- Reset / reboot evidence: log reset reason (brownout vs watchdog vs external reset) plus a monotonic reboot counter.
- Relay/triac switching transient: observe the instant of coil energize / triac gate event; look for rail spikes and ground bounce.
- Ripple + spike fingerprint: quantify low-frequency ripple and high-frequency spikes on rails; correlate spikes with UI glitches or radio drops.
Design levers (component + layout + partitioning)
MOV / TVS placement
MOV absorbs surge energy at the entry; TVS protects fast edges on sensitive low-voltage nodes and long I/O runs. The effective protection is dominated by placement (short return path) and energy rating, not part count.
RC snubber strategy
RC snubbers reduce contact arcing and dv/dt from triac switching, but add loss and can introduce leakage-like behavior. Use snubbers to control the switching loop, not as a generic “EMI fix”.
X/Y capacitors and leakage awareness
X capacitors reduce differential-mode noise; Y capacitors reduce common-mode noise but increase leakage current. Bathrooms magnify leakage sensitivity due to moisture and protection devices; keep Y choices conservative and evidence-driven.
Ground / return-path control and partitioning
Keep high di/dt power returns (heater/fan switching loops) inside the power zone and prevent them from sharing impedance with MCU/sensor/UI returns. Use a clear partition between power zone and control zone with a controlled tie point.
Isolation, creepage, and condensation reality
Isolation strategy should protect the control/sensing domain from mains-referenced stress. Creepage/clearance and leakage need to be treated as “dynamic” under condensation; mechanical layout and conformal strategies should match the bathroom duty cycle.
Acceptance criteria (what “rugged enough” means on this page)
- Heater/fan switching does not trigger uncontrolled resets; any brownout event returns to a safe-off or safe-recovery state.
- UI (touch/keys) does not sustain false triggers during load switching; any glitch is self-clearing within a bounded time.
- Event logging differentiates brownout vs watchdog resets, enabling field triage without guesswork.
- Common-mode noise controls prevent persistent sensor spikes (RH/T) that would destabilize de-mist logic.
H2-4 Heater Drive (Drive Options & Safety Chain)
Heater control is not a “simple switch” problem. The engineering priority is safe behavior under faults: even if firmware stalls or a drive device fails, heating must not remain uncontrolled. This chapter compares relay/SSR/triac choices through measurable criteria and then locks the safety chain as an independent, layered defense.
Start with the heater load (why drive selection changes)
- Resistive wire / ceramic: cold-start current can be significantly higher than steady-state; switching edges can be harsh.
- PTC: self-limiting helps temperature runaway, but inrush and EMI still dominate drive stress; control stability still matters.
- De-mist film: smaller power but sensitive to leakage and control chatter; safe-off behavior must remain deterministic.
Primary evidence: cold-start current steady-state current switching transient device case temperature
Drive branches: Relay vs SSR vs Triac (zero-cross vs phase)
- Relay: low on-state loss and near-zero leakage, but contact bounce/arcing can generate spikes; failure mode includes welded contacts (stuck-on).
- SSR: no mechanical wear, but on-state dissipation raises temperature; off-state leakage can create micro-heating; surge capability must be proven.
- Triac: flexible control (burst/phase), but most sensitive to dv/dt and noise; phase control is EMI-intensive and demands strong partitioning/snubbing.
Key metrics (measurable, not “marketing claims”)
- Surge / inrush headroom: cold-start peak and repetition; confirm that protection and drive survive without nuisance trips.
- EMI footprint: switching edge severity and rail spike amplitude; correlate with UI glitches and radio drops.
- Thermal rise: SSR/triac case temperature and enclosure airflow dependence; validate margin at worst humidity/temperature.
- Failure modes: welded relay, shorted SSR/triac, dv/dt false trigger; each must map to a safe response.
Safety chain (layered defense independent of firmware)
- Thermal fuse (final cut): irreversible protection as the last line; placement must track the true hot spot.
- Dual NTC sensing: one near heater core, one in airflow/cavity; cross-check detects blocked airflow and sensor faults.
- Over-temp shutdown: hard cutoff logic with bounded recovery; avoid “auto-retry forever” behavior under persistent faults.
- Fan-confirm interlock: inhibit heater when tach/FG indicates stall or below-min airflow; prevents heat soak under fan failure.
- Abnormal self-check: detect stuck-on (heater still draws current when commanded off) and sensor failures (open/short/stuck).
Evidence patterns that close the loop
- Heater current waveform: cold-start peak and edge spikes reveal required drive margin and snubber effectiveness.
- Triac dv/dt false trigger: unintended turn-on under noisy edges; evidence is heater current pulses without a valid control command.
- SSR leakage “micro-heat”: off-state still passes small current; evidence is a measurable temperature rise despite an “OFF” state.
H2-5 Fan Drive (Speed Evidence & Stall Protection)
De-mist and exhaust performance depends on actual airflow, not just a “fan ON” command. This chapter treats the fan as an airflow actuator with evidence: command → speed proof (FG/Tach/Hall) → protection (stall/backpressure). It also contains the fan drive noise so it does not destabilize MCU, UI, sensors, or radio inside the same product.
AC PSC vs BLDC: selection boundaries for bathroom duty cycles
- AC PSC motor: simple drive path and strong tolerance to harsh environments, but limited speed control range and weaker closed-loop consistency.
- BLDC motor: wide speed range and better efficiency/noise potential, enabling quiet/night modes and steadier airflow, but increases EMI risk from PWM/commutation.
Practical discriminator: need wide speed range need speed proof quiet mode EMI headroom
Drive architecture: what must exist for reliable airflow control
AC PSC path
Speed control is usually coarse (multi-step or controlled switching). The strongest reliability lever is confirming “fan really spins” with a speed proxy (tach or current fingerprint) and preventing heater enable without airflow proof.
BLDC path
A driver stage (integrated or discrete) performs commutation and accepts a speed command. The control loop quality depends on a stable speed feedback signal (Hall/FG or a validated BEMF estimator) plus protection thresholds that handle stall/backpressure.
Speed / position evidence: FG, Tach, Hall, and what each proves
- Hall: position feedback for commutation; also provides a robust “spinning” proof under harsh EMI if routed cleanly.
- FG / Tach: speed pulse proportional to RPM; enables minimum-RPM gating and stall timing rules.
- BEMF (sensorless): cost and wiring savings, but requires clean measurement windows; validate immunity to switching noise before relying on it for safety.
Evidence patterns: minimum measurements that isolate “airflow loss” causes
- Start current: excessive peak or long inrush suggests backpressure, mechanical friction, or undervoltage during spin-up.
- FG frequency / BEMF build-up: if RPM never crosses a minimum threshold in a bounded time, classify as stall/lock.
- Backpressure fingerprint: RPM may hold while current rises; this indicates duct restriction, filter clogging, or damper issues.
Noise containment inside the product (MCU/UI/sensors/radio stability)
- Partitioning: keep fan power switching loops inside the power zone; do not share return impedance with MCU/sensors.
- Signal hygiene: route FG/Hall away from high dv/dt nodes; use Schmitt inputs or modest RC where appropriate (without masking real faults).
- Supply resilience: local decoupling and filtered rails for MCU/sensors prevent commutation spikes from becoming resets or sensor glitches.
- Radio robustness (if present): correlate drop/retry counters with fan PWM events; isolate the power rail and return paths accordingly.
H2-6 Temp/RH Sensing (Placement, Drift, and Health Checks)
In bathrooms, humidity and temperature sensors must be treated as control-grade evidence, not “nice-to-have readings”. Placement dominates usefulness: a sensor that tracks cavity bias rather than condensation risk will produce stable numbers but poor de-mist outcomes. This chapter defines placement trade-offs, drift mechanisms under cosmetics/steam, and simple health checks that keep control decisions safe.
What to measure for condensation-aware control
- RH trend: direction and response time matter more than absolute accuracy for de-mist actions.
- Temperature gradient: heater proximity and airflow create strong gradients; interpreting RH without local temperature context can mislead control.
- Condensation risk (concept): treat as a thresholded risk signal derived from RH + local temperature, with hysteresis in downstream logic.
Placement trade-offs: inlet vs outlet vs near-mirror
Near inlet (room air)
Tracks ambient bathroom air more faithfully but reacts slower to de-mist actions. Useful for long-term humidity control and post-shower purge.
Near outlet (actuation-biased)
Reacts quickly but can read “artificially dry/warm” because it is close to heater and forced airflow. Without compensation, control may terminate de-mist early while the mirror zone still condenses.
Near mirror / target zone (risk-relevant)
Best correlation to fogging risk but most exposed to water films and contamination. Requires protection against droplets and robust health checks to avoid drift-driven false decisions.
Drift and contamination: how bathroom reality breaks RH sensors
- Water film: condensation on the sensor surface causes slow recovery and biased high readings.
- Cosmetics / aerosol / soap mist: residue changes sensor response time and offset; symptoms appear as “stuck high” or long tails.
- Oil/smoke ingress: films degrade repeatability and can create hysteresis that the control loop misinterprets as real humidity change.
Evidence & health checks: prove sensors are usable before trusting control
- RH response curve: after fan boost or heater action, RH should move in a consistent direction within a bounded window.
- Temperature gradient check: compare two temperature points (core vs airflow/cavity) to detect blocked airflow and bias conditions.
- Stuck / freeze detection: if RH/T does not change while fan/heater state changes, flag as “not responsive”.
- Open/short plausibility: detect out-of-range readings and unrealistic step changes; convert to a fault state.
H2-7 De-mist Control (From RH Threshold to a State Machine)
A single RH threshold is structurally unreliable in bathrooms: shower steam creates fast transients, sensor placement adds delay and bias, and mirror fog risk depends on the coupled behavior of RH and local temperature. A robust design uses a condensation-risk signal and a small state machine that enforces hysteresis, debounce, limits, and a cooldown phase to avoid rapid toggling and noise.
Why “one RH threshold” misfires in real bathrooms
- Steam transient: RH rises quickly, but fogging risk depends on temperature gradients; “RH high” does not always equal “mirror fog now”.
- Sensor latency & bias: inlet/outlet/cavity placement can lag or look artificially dry/warm during active heating and airflow.
- Noise and drift: short bursts of steam and sensor tails can cause frequent on/off toggles without a debounce and hysteresis strategy.
Recommended signal: condensation risk (concept) + timed logic
Treat RH and temperature as inputs to a risk signal (Low / Medium / High). The risk signal may be derived from local RH and temperature trends, and optionally a temperature point closer to the target zone. The goal is not perfect physics, but a stable and relevant decision input that correlates with fogging outcomes.
Minimum inputs
RH(t), T(t), plus a risk thresholding rule that is resistant to short spikes and biased outlet readings.
Minimum outputs
Fan level and heater power commands, bound by limits and gated by safety interlocks.
State machine: BOOST → MAINTAIN → COOLDOWN
- BOOST: rapid risk reduction. Enter on sustained High risk (debounced). Exit on sustained Medium/Low risk (hysteresis) or max boost time.
- MAINTAIN: keep fog risk from returning while reducing noise/energy. Exit to BOOST on renewed High risk; exit to COOLDOWN on sustained Low risk.
- COOLDOWN: prevent rapid cycling and allow sensor/thermal bias to settle. After a cooldown timer, exit to OFF if risk remains Low.
Required control hygiene: Hysteresis Debounce Min ON/OFF Max BOOST Cooldown timer
Limits that keep the loop stable and hardware-friendly
- Power limit: cap heater power and rate-of-change to avoid large transients and thermal overshoot.
- Time limits: enforce minimum ON/OFF times and a maximum continuous BOOST duration.
- Exit discipline: require sustained low risk before turning off; otherwise maintain a low, quiet airflow to prevent rebound fog.
Evidence curves: how to validate behavior with time traces
- Inputs vs time: RH(t), T(t), and the derived Risk(t) should show consistent movement after actuation.
- Outputs vs time: FanLevel(t) and HeaterPower(t) must align with state transitions and respect limits.
- State band: a BOOST/MAINTAIN/COOLDOWN band reveals whether logic is cycling too frequently.
H2-8 Safety Interlocks (Gating Rules, Degrade Modes, Evidence Logs)
Safety interlocks ensure heating cannot proceed unless the product can prove airflow, temperature integrity, and safe mechanical conditions. Interlocks must be written as gating rules with explicit degrade modes, plus a minimal evidence set (fault code, beeper/LED pattern, event log fields) so field troubleshooting can isolate the root cause quickly.
Interlock inputs: the minimum signals that must exist
- Cover / service door: open state disables heating immediately; log the event.
- Filter / inlet restriction: detect via a switch or via airflow evidence + current/thermal fingerprint (implementation varies, behavior stays consistent).
- Over-temperature chain: thermal fuse + one or two NTC points; over-temp triggers a safe shutdown and may require manual reset depending on severity.
- Fan proof: FG/Tach/Hall evidence within a time window; without fan proof, heating is inhibited.
- Sensor validity: RH/T plausibility and stuck detection; invalid sensors force conservative modes.
Heater enable gating (expressed as a deterministic rule)
Heater enable must be computed from a strict AND chain: Cover OK AND OverTemp OK AND FanProof OK AND Power OK AND (Sensors OK OR Degrade Policy). Any false condition forces a safe output state and records evidence.
Degrade policy examples
If RH/T is invalid, switch de-mist to conservative timed purge and limit or inhibit high heater power. If fan proof is lost, inhibit heating immediately and retry fan start within a bounded attempt count.
RCD/leakage-protection “compatibility symptoms” (phenomenology only)
- Trips on heater start: correlate timestamp with heater command level, inrush peak, and whether a previous event occurred under high humidity.
- Trips on fan start: correlate with motor inrush and commutation PWM events; confirm if fan-only mode reproduces the issue.
- Random trips: correlate with condensation periods, repeated restarts, and any logged surge/ESD counters if available.
Fault codes, beeper/LED patterns, and event log fields (minimal set)
- Fault code class: OverTemp / FanStall / CoverOpen / SensorFault / PowerReset / CompatibilityTrip.
- Beeper/LED: short/long/spacing patterns for major classes (keep the set small and field-recognizable).
- Log fields: reset_reason, last_state, fan_rpm_or_proof, heater_level, temp_core, temp_cavity, rh_raw, fault_count, timestamp.
H2-9 UI & Human Factors (Wet-Touch Robustness and Experience Evidence)
Bathroom UI must be engineered as a reliability system: wet hands, water films, condensation and power switching noise can turn a touch surface into an unintended sensor. A robust UI defines failure modes, hardens inputs (touch/keys/knob), and keeps feedback coherent (LED/display/beeper), with measurable evidence such as false-touch counts, button bounce statistics, and night-mode brightness limits.
UI elements and what can fail in a humid environment
- Touch (capacitive): baseline drift, water-film bridging, humidity-induced leakage, and noise injection during load switching.
- Buttons / knob: bounce, corrosion-induced intermittency, ESD paths, and inconsistent actuation feel over time.
- Display / LEDs: night comfort, visibility through fogged optics, and PWM flicker sensitivity.
- Beeper: user comprehension of patterns, night-mode quietness, and rate-limits to avoid “alarm spam” during repeated mis-touches.
Reliability hygiene: Input validation Debounce Rate limit Night mode
Wet-touch failure modes (why false triggers happen)
- Water-film bridging: a conductive film changes effective capacitance and couples adjacent electrodes, causing multi-point false touches.
- Condensation drift: slow baseline movement can cross thresholds over time if tracking and hysteresis are not designed.
- Glove / wet hand variability: coupling changes can create “no response” or “over-sensitive” behavior without adaptive thresholds.
- Noise coincidence: motor PWM and heater switching transients can pollute sampling windows and spike touch deltas.
Hardening tactics: hardware + algorithm + enclosure (device-level)
Hardware intent
Partition touch reference/return from power switching currents; add shielding/guard concepts to reduce edge coupling and water-film effects; define ESD return paths that do not disturb touch ground.
Algorithm intent
Use baseline tracking with two time constants; detect “wet mode” signatures (multi-channel lift + elevated noise); apply stronger debounce and higher thresholds while wet.
Enclosure intent
Use water management (drain paths, gaskets, coatings) so water does not sit on sensing edges; reduce persistent films that cause slow drift.
Buttons/knob and feedback coherence (night-friendly behavior)
- Button bounce control: measure bounce duration/count and enforce deterministic acceptance windows for short/long press events.
- Night mode: cap brightness and ramp changes slowly; reduce beeper volume and rate-limit non-critical alerts.
- Coherent feedback: align LED/display/beeper patterns to the same fault classes so users can interpret the state consistently.
Evidence: what to log and what to quantify
- False-touch statistics: false_touch_count, wet_mode_time, consecutive_rejects.
- Button bounce: bounce_count, bounce_max_ms, long_press_misdetect_count.
- Night comfort: night_brightness_level, ramp_time_ms, beeper_events_per_hour.
- Correlation hooks: touch_false bursts vs fan/heater switching timestamps.
H2-10 Connectivity (Device-level) (Provisioning, Power, Dropouts, Coexistence)
Device-level connectivity must be robust without relying on cloud or gateway assumptions. Engineering focus is on provisioning success, bounded reconnect behavior, controlled RF peak currents, and coexistence so 2.4 GHz activity does not destabilize touch or sensors. Evidence should include RSSI, reconnect counters, and correlations between RF bursts and power-rail ripple/reset events.
Scope: what is covered (and what is not)
- Covered: provisioning/join, power impact, dropouts, bounded reconnect policy, coexistence with device sensors and UI.
- Not covered: gateway architecture, Matter network design, cloud topology, router configuration tutorials.
Provisioning reliability: classify failures with minimal evidence
- Scan / discovery: record scan_count, rssi_last, join_time_ms.
- Authentication / join: record join_fail_code, join_fail_count, timeout_stage.
- User action coupling: link UI events to provisioning state to separate user abort from RF failure.
Power and stability: RF bursts must not collapse rails
- RF peak: TX/scan peaks can raise rail ripple and cause brownouts if power partitioning and decoupling are insufficient.
- Bounded behavior: reconnect must use a capped retry budget with backoff to avoid “reconnect storms”.
- Reset clarity: store reset_reason (brownout / watchdog) to distinguish power instability from firmware logic faults.
Dropouts & reconnect policy: deterministic, bounded, user-safe
- Disconnect detect: link-quality threshold and heartbeat timeout, with a short confirmation window.
- Fast retry window: a few quick retries to recover transient fades.
- Backoff: exponential backoff with a maximum interval; stop after a retry budget and remain in local-safe mode.
Local-first requirement
De-mist, safety interlocks, and UI must remain functional when radios are unavailable. Connectivity is additive, not required for safety or basic operation.
2.4 GHz coexistence: correlate RF activity with ripple and false UI/sensor events
- Coupling path: RF burst peak → rail ripple → touch baseline movement / sensor jitter.
- Scheduling: avoid sampling sensitive channels at the most aggressive RF burst instants where possible.
- Evidence correlation: compare touch_false_count and sensor_invalid_count against tx_burst timestamps and rail_ripple_mv.
Minimum device-level log fields (for fast field isolation)
- RF: rssi_last, reconnect_count, join_fail_count, last_join_fail_code.
- Power: rail_ripple_mv (if available), brownout_count, reset_reason.
- Cross-domain: last_state (BOOST/MAINTAIN/COOLDOWN), touch_false_count, sensor_invalid_count, timestamp.
H2-11 Validation & Field Debug Playbook (SOP)
A device-level, evidence-first SOP for bathroom heater/fan units. Each case follows: Symptom → First 2 Measurements → Discriminator → First Fix → Log Fields → Pass Criteria.
Symptom A — Startup trip / nuisance RCD trip
Typical timing: at power-on, heater enable, fan spin-up, or mode transition (BOOST ↔ MAINTAIN).
- Line/DC-bus surge snapshot: capture the first 0–200 ms with a repeatable trigger (power-on edge or relay/triac gate event).
- Inrush current profile: clamp current at the input or at heater/fan branch; record peak and decay constant (0–500 ms).
- High voltage spike but modest current → surge path / snubber / wiring loop issue (conducted spike).
- Very high current peak aligned with command → inrush / soft-start / phase strategy issue (current-driven trip).
- Much worse only in high-humidity conditions → moisture-related leakage/creep path; treat as a “wet-state compatibility” failure (device-level symptoms only).
- Across-line surge clamp: MOV example TDK/EPCOS B72214S0271K101 (line surge suppression class).
- Across-line EMI cap: X2 safety film example KEMET R46 series (Class X2, 275 VAC) (pick C by conducted-EMI/inrush balance).
- Switch edge control: if triac/SSR is used, add an RC snubber and/or use a higher dv/dt margin path (see triac/driver examples below).
- Drive strategy: stagger loads (fan first, heater ramp), enforce minimum off-time, and limit step changes at mode boundaries.
Heater switching examples (when relevant to the failure): Triac: ST BTA16-600B; Zero-cross triac driver: Vishay VO3063 or onsemi MOC3063/MOC3163 family; Relay (mechanical alternative): Omron G5LE-1.
- ≥ 50 consecutive power-on + mode-transition cycles with no trip in both dry and “after-hot-shower” conditions.
- Peak surge/current reduced or bounded with a stable decay; no repeated brownout/reset signatures.
Symptom B — De-mist ineffective (mirror still fogs)
Common root cause: the control loop is “blind” (sensor placement/response) or airflow is not achieved.
- RH/T response vs time: log RH(t), T(t), and state output for 10 minutes (BOOST/MAINTAIN/COOLDOWN).
- Airflow evidence: fan RPM (FG/tach) or current-vs-RPM correlation; record before/after filter/duct inspection.
- BOOST active but RPM is low → airflow path issue (filter/duct backpressure) or weak drive.
- RH/T curve changes but fog persists → sensor is not representative of mirror boundary layer (placement/thermal gradient).
- Heater power high but RH/T barely moves → heat/air coupling broken (recirculation, leakage, or incorrect vent path).
- Sensor sanity: replace/validate RH sensor (example Sensirion SHT4x / SHT40) and add wet-state plausibility checks (stuck/slow response).
- Airflow enforcement: enforce minimum RPM before enabling heater; add “filter blocked” heuristic when current rises but RPM falls.
- State tuning: extend BOOST window, add exit hysteresis, and add COOLDOWN to avoid rapid toggling.
- Under a repeatable “hot shower” test, fog clears within a defined time window (set product target), with stable state transitions (no hunting).
- Sensor response shows consistent time constant; “stuck/slow” detection triggers safe fallback.
Symptom C — Loud noise / vibration / “buzzing”
Goal: separate mechanical resonance from electrical/commutation artifacts using RPM/PWM evidence.
- FG/RPM + PWM snapshot: log RPM and PWM frequency/duty (capture during the noisy window).
- Sweep test: ramp RPM slowly up and down; identify narrow “noise peak bands” (resonance windows).
- Narrow RPM band peaks → mechanical/duct resonance or assembly tolerance.
- Noise jumps with PWM strategy (even at same RPM) → drive waveform / commutation / switching frequency issue.
- Noise correlates with current ripple → abnormal load/backpressure or near-stall behavior.
- Avoid-band control: implement “notch zone” (skip resonance RPM band), use a smooth ramp (jerk-limited).
- Low-noise BLDC drive: consider a sinusoidal sensorless driver for fan/pump class loads (example TI DRV10983).
- Stall protection: tighten the stall criteria (RPM drop + current rise) and degrade gracefully (reduce duty cycle, raise an alarm code, disable the heater).
- Noise peak band is avoided or reduced (repeat sweep shows no sharp spikes).
- No sustained near-stall operation; stall events trigger deterministic downgrade.
Symptom D — Random reboot / wireless dropouts
Key question: is it power integrity (droop) or software scheduling/reconnect storms?
- 3.3 V / 5 V droop capture: trigger on reset or on heavy events (heater enable, fan ramp, RF reconnect).
- Event log correlation: record reset_reason + reconnect_count + last_state + rail_ripple_mv with timestamps.
- Droop visible + brownout reset → power tree / decoupling / peak-current issue.
- Rails stable but watchdog reset with reconnect storms → scheduling/stack starvation issue (device-level).
- Dropouts align with touch false events or switching edges → coexistence coupling (sampling windows vs bursts).
- Improve the low-voltage rail: add soft-start / stronger transient response buck (example TI TPS62133), tighten local decoupling at MCU/radio.
- Off-line auxiliary supply (if used): low-BOM flyback switcher example Power Integrations LNK364PN (design-dependent; validate thermal and surge margins).
- Reconnect discipline: exponential backoff, retry caps, and “local functionality first” policy during link loss.
- 100+ stress cycles (heater/fan transitions + forced RF reconnect) with no brownout resets.
- Reconnect storms bounded (retry caps), and user-visible functions remain deterministic.
Figure F11 — Field Debug Decision Map (device-level)
A one-page visual cheat-sheet mapping each symptom to the first 2 measurements, discriminator, first fix, and minimum log fields.
Request a Quote
H2-12 FAQs ×12
Each answer follows an evidence-first pattern: First 2 checks → discriminator → first fix. All content stays device-level (no whole-home HVAC or cloud topology).
1) Rated power is modest but it trips at startup: surge/inrush or wet-state leakage compatibility?
2) Relay design has occasional “sticking”: what conditions trigger it, and how to confirm with evidence?
3) SSR is “off” but there is slight heating or faint glow: leakage or measurement illusion?
4) Why does a triac design more easily disturb wireless/touch? Which two waveforms to check first?
5) Airflow suddenly drops: backpressure/clog or motor drive derating?
6) De-mist becomes slower in cold + high humidity: dew-risk logic or sensor placement?
7) RH readings drift over time: contamination drift or condensation intrusion? How to run a health check?
8) Touch panel goes crazy during showers: water film or power-noise injection?
9) Fan stops but heater stays on: what is the correct interlock and failure handling?
10) More reboots after networking is enabled: power droop or 2.4 GHz interference?
11) De-mist control hunts (frequent on/off): how to set hysteresis and debounce to stop “chatter”?
12) Users complain of “piercing noise”: electromagnetic motor noise or structural resonance?
Figure F12 — FAQ-to-Chapter Map (at a glance)
A compact map showing which chapter anchors each FAQ, designed for quick navigation and audit.
