Home Security Kit Hardware: Sensor AFEs, Low-Power Radios, Backup
← Back to: Consumer Electronics
A Home Security Kit is a hub + battery sensor network where reliability is proven by hardware evidence: clean sensor front-ends, multi-year low-power wake behavior, robust RF margin in real homes, and outage-safe backup power without losing events.
This page focuses on measurable design knobs and field-debug proof (waveforms, RSSI/LQI/retries, and sequence-consistent logs) to reduce false alarms, drops, and resets.
H2-1 — Page Boundary, Target Reader, and What This Page Solves
A Home Security Kit is an event-driven hardware system: multiple low-power sensor nodes convert physical signals into auditable events (open/close, motion, shock, tamper), and a hub/base station guarantees local logging + alarm actuation under real-world constraints (walls/interference/power outages).
- False-alarm reduction with measurable thresholds: define “event” using threshold + time window + drift control, then validate with repeatable stimuli.
- Multi-year sensor-node battery life: build a sleep/wake state machine and a current budget that survives coin-cell pulse limits.
- RF robustness in real homes: use link-budget and field evidence (RSSI/LQI/retries) to separate “coverage” from “interference.”
- Power-outage resilience: hub switchover to backup power without losing event integrity (sequence + timestamp + dedupe).
Hub/base station (radio concentrator + event queue + local storage), door/window nodes (reed/hall + debounce + tamper), PIR nodes (pyro + optics + AFE + decision window), vibration/shock nodes (piezo/MEMS + energy window), alarm actuators (siren driver, keypad), optional range extender.
Camera/ISP video pipelines, smart locks (motor/fingerprint), mobile app UX, cloud/backend architecture, and protocol-stack deep dives. Backhaul is treated as an interface box only.
H2-2 — Reference Architecture: Hub + Sensor Nodes + Alarm Path
A single canonical topology helps keep every later decision traceable: sensor signal chain → event decision → RF transport → hub aggregation → local log → alarm actuation, plus a power-fail chain that preserves event integrity during outages.
- Standardize the node pipeline: Sensor → AFE/Comparator/ADC → MCU (wake/decide) → Radio → Battery.
- Define “event integrity” at the hub: sequence counter + timestamp + node ID + retry/RSSI evidence, enabling dedupe and loss detection.
- Separate alarm actuation from logging: siren/keypad load transients must not corrupt the log pipeline.
- Lock down outage behavior: AC loss triggers power-path switchover; critical rails stay alive; events remain consistent.
The “event” is the unit of truth. Each event record should minimally carry: sequence counter (loss/duplicate proof), timestamp (ordering), node ID + event type, and RF evidence such as RSSI/LQI + retry count. During outage/WAN loss, the hub remains responsible for local buffering and alarm decisions.
H2-3 — Door/Window Sensor Front-End: Reed/Hall, Debounce, Tamper
A door/window node is “simple” only when its event definition is robust. The hardware goal is to convert a contact/magnetic state into a stable event under vibration, alignment drift, and ESD disturbances, while keeping wake-up power low and avoiding phantom open/close bursts.
- Alignment tolerance: large install variation or moving frames favor Hall to avoid “near-threshold” toggling.
- Vibration exposure: repeated micro-shocks can produce reed bounce that looks like multiple state changes.
- Magnet variability: unknown magnet strength or long-term drift benefits from Hall threshold margin and calibration options.
- EMI/ESD risk: long leads and touch points require input protection regardless of sensor choice.
- Hardware conditioning: pull-up/down plus an RC time constant to slow edges and reduce chatter energy at the GPIO input.
- Wake-up interrupt ≠ event: interrupts wake the MCU, but only a stable state after a confirmation window becomes an event.
- Confirmation window: choose a window that rejects bounce while preserving acceptable response time for alarm logic.
- Reset safety: after a reset/ESD hit, re-sample input state and avoid generating “transition events” without evidence.
- Dedicated tamper switch: mechanical switch with a defined pull and an interrupt path.
- Debounce applies to tamper too: switch bounce can cause repeated wakes and battery drain.
- Tamper event integrity: store tamper reason with a sequence counter and timestamp for auditability.
| Symptom | Likely root cause | Measurement evidence | Fix direction |
|---|---|---|---|
| Random open/close bursts at night | Vibration bounce near threshold, or reset-induced phantom transitions | GPIO waveform around wake; reset/brownout flag; sequence counter gaps | Increase confirmation window; improve input protection; tighten alignment margin |
| Door “closed” but occasional open glitch | Magnet distance near switching point; reed chatter; weak magnet | Distance sweep vs event rate; edge timing at GPIO; hysteresis behavior | Switch to Hall; adjust magnet placement; add hysteresis/firmware windowing |
| Touching enclosure triggers events | ESD injection into input path; poor return path; insufficient clamping | ESD test + captured input spike; reset flags; event spikes without mechanical change | Add TVS/series-R; shorten loop; improve grounding and routing |
| Tamper causes battery drain | Tamper bounce or unstable pull causing repeated wakes | Wake reason histogram; tamper pin toggling; average current increase | Debounce tamper; strengthen pull; revise switch mechanics |
Store at least: sequence counter, timestamp, node ID, state (open/close/tamper), plus wake reason and optional RSSI/retry evidence for transport correlation.
H2-4 — PIR Motion Node: Pyro + Optics + AFE + False-Alarm Control
PIR nodes generate the most “mysterious” alarms when the signal chain is treated as a black box. A robust design turns motion detection into a controlled pipeline: pyro + optics produce a small AC-like signal, the AFE shapes it (gain + band-pass), and the system declares an event only when waveform evidence matches a decision window rather than a single spike.
- Pyro element + Fresnel lens: converts spatial IR change into a small differential signal.
- Input buffer (JFET, if used): protects the sensor node from bias/leakage sensitivity.
- Band-pass + gain stages: rejects slow baseline drift and high-frequency noise while amplifying motion components.
- Comparator / ADC: threshold crossing or sampled evidence feeds a decision rule.
- Decision window: event requires sustained or patterned evidence within a time window (not a one-shot glitch).
- Gain: too high turns drift into triggers; evidence is baseline movement crossing thresholds.
- Band-pass corners: low corner too low allows HVAC/thermal drift; high corner too high increases noise triggers.
- Noise + offset drift: temperature and leakage shift the baseline; evidence is a slow trend correlated with temperature.
- Saturation recovery: strong IR bursts can saturate stages; evidence is long recovery tails that mimic motion patterns.
- Input bias sensitivity: bias/leakage errors reshape low-frequency behavior; evidence is repeatable drift under humidity/temperature steps.
- Sunlight sweep: slow drift plus occasional bursts; typically looks like a baseline ramp and a clipped segment.
- HVAC drafts: low-frequency periodic components that survive if the band-pass low corner is too low.
- Pets: smaller amplitude but repeated patterns; controlled by windowing and pattern requirements.
- Lens artifacts: angle-dependent sensitivity; repeatable hotspots when moved through the same path.
- Dual-threshold: trigger threshold and release threshold stabilize decisions around noise.
- Time window: require evidence persistence within window T to reject isolated spikes.
- Pattern vs spike: motion is a sequence; isolated impulses are treated as disturbances.
- Saturation guard: after saturation, apply a short cooldown window to prevent tail-triggered events.
First separate waveform types: slow drift crossing vs fast burst/spike. Drift points to band-pass low corner and offset drift; bursts point to saturation, sunlight transients, or injected disturbances. Then verify with thresholds and time-window evidence.
H2-5 — Vibration / Glass-Break / Shock: AFE Options and Robust Triggering
Shock and vibration nodes fail most often by becoming “always-triggering” in real homes, where doors slam, floors vibrate, and enclosures resonate. A robust design separates mechanical coupling from AFE shaping and defines a triggering rule based on an energy window rather than single spikes.
- Piezo film: high sensitivity to impacts; lowest complexity for “shock present” detection; strongly dependent on mounting and enclosure resonance.
- MEMS accelerometer: enables windowed features (energy, duration) and better false-trigger control; demands careful wake/ODR strategy for battery life.
- Mechanical ball switch: ultra-low-power vibration wake; limited repeatability and directionality; best for coarse “tamper/move” detection.
- Microphone glass-break (optional): mention-only; requires band-limited envelope + time-window logic; avoid deep audio pipeline discussion.
- Charge amp / envelope / comparator (typical for piezo): key knobs are HPF corner, envelope time constant, threshold hysteresis, cooldown after an event.
- Direct ADC / sampled features (typical for MEMS): key knobs are ODR, bandwidth, window length, energy metric (peak/RMS/integral), N-of-M within window.
- Always-trigger root causes usually come from a low HPF corner, slow envelope, missing hysteresis, or resonance peaks amplified by mounting.
- Mount location: thin panels amplify certain bands; screw posts couple structure-borne shock strongly.
- Resonance: enclosure modes can convert one tap into a long ring-down that keeps crossing thresholds.
- Adhesive vs screws: adhesives change stiffness over time; screws improve repeatability but can increase coupling to the host surface.
- Practical control: change mount stiffness, add damping pad, or move sensor away from resonance hotspots before tuning thresholds.
- Standardized stimulus: fixed tap/shock location, mass, and height (or simple spring tap tool) to compare builds.
- Evidence features: peak amplitude, integrated energy in window T, duration of ring-down, and false-trigger rate across benign scenarios.
- Acceptance criteria: trigger probability rises with stimulus strength, while benign disturbances stay below energy threshold.
| Sensor option | Sensitivity | Standby current | False-trigger risk | AFE complexity | Mechanical dependency |
|---|---|---|---|---|---|
| Piezo film | High (impact) | Very low | Medium–High | Low–Medium | High (mount/resonance) |
| MEMS accel | Medium–High | Low–Medium | Low–Medium | Medium | Medium (placement) |
| Ball switch | Low–Medium | Ultra low | Medium | Very low | High (direction/lot) |
| Mic (optional) | Event-specific | Medium | Medium | Medium–High | Medium (acoustics) |
Use an energy window (integrate envelope or |x| within time window T), apply dual thresholds (trigger/release), and add a short cooldown to prevent ring-down tails from re-triggering.
H2-6 — Ultra-Low-Power Strategy: Sleep, Wake, and Event Integrity
Multi-year battery life is achieved by controlling time spent in each state and by preventing retries, brownouts, and duplicate events from quietly multiplying energy cost. The engineering target is a clean state machine, a structured current budget, and an event record that remains auditable across resets.
- Deep sleep dominates lifetime; wake sources must be filtered before becoming events.
- Wake on interrupt → stabilize → sample → decide → transmit → return to sleep.
- Retry control is part of the state machine, not a “radio detail”.
- Fail-safe return: any abnormal branch must return to low-power to avoid stuck-high current.
Use the structure: I_sleep & T_sleep, I_wake & T_wake, I_sample & T_sample, I_tx & T_tx, N_event/day, N_retry/event, plus N_rejoin/day if the link is unstable. Lifetime is driven by “time in each state” and “how often TX/retry happens”.
- Pulse current limit: radio TX bursts can exceed what a cold/aged coin cell can supply.
- ESR rises with cold and aging; voltage droop increases under the same TX pulse.
- Brownout (BOR) creates hidden failures: missed events, duplicate events, and sequence counter anomalies.
- Mitigation boundary: small bulk capacitance, load staggering (avoid LED/buzzer with TX), and robust reset-aware event logic.
| Symptom | Evidence to capture | Most likely cause bucket |
|---|---|---|
| Event gaps (missing alarms) | Sequence counter jumps; wake reason without transmit; Vbat droop during TX | Brownout during TX or repeated retries |
| Duplicate events | Same seq repeats; reset flag set near the event time | Reset + re-send without dedupe |
| Clock/time anomalies | Timestamp resets/backward jumps; RTC status after reset | RTC domain power loss |
| Battery drains unexpectedly fast | Retry rate rises; rejoin count increases; average current higher than budget | Poor link margin or stuck high-power state |
- Sequence counter: the core of “missing vs duplicate” detection.
- Dedupe rule: hub must drop duplicates based on (node ID + seq) rather than timestamps alone.
- Reset-aware policy: after BOR/reset, re-sample and confirm before emitting transition events.
- Retry budget: cap retries and record retry count as evidence; uncontrolled retries destroy battery life.
H2-7 — Low-Power Radios in Real Homes: Link Budget, Antennas, and Coexistence
Real-home reliability is dominated by fade margin, antenna efficiency, and time-varying interference. This section stays at PHY and hardware robustness level: frequency choice, link budget essentials, enclosure interaction, and evidence that separates “distance” problems from “coexistence” problems.
- Through walls: Sub-GHz typically keeps more margin behind thick walls; 2.4 GHz is more sensitive to obstruction and body/hand effects.
- Antenna size & efficiency: Sub-GHz often needs more antenna volume; 2.4 GHz fits small products but can lose efficiency inside tight enclosures.
- Interference ecology: 2.4 GHz faces Wi-Fi/BLE crowding; Sub-GHz has fewer common interferers but still needs margin for neighbor devices and fading.
- Failure symptoms: time-of-day retry spikes often indicate coexistence, while “move 1 meter and it recovers” often indicates fade margin.
Confirm Tx power, Rx sensitivity, and antenna efficiency in enclosed condition. Estimate wall loss, keep a clear fade margin, and validate at the worst spot using distributions (RSSI/LQI histograms) rather than single readings.
- Ground plane: small ground planes and splits reduce efficiency and distort patterns.
- Plastic vs metal: metal, shielding cans, screws, and metallized coatings change matching and radiation.
- Placement: antenna near edges, batteries, or cables can detune; hub antenna placement often defines system coverage.
- Hand/wall effects: small nodes near walls or corners can lose margin abruptly (especially at 2.4 GHz).
- Wi-Fi congestion: packet loss/retries rise in evenings; RSSI may remain stable while PER worsens.
- Microwave interference: short bursts of loss near kitchen areas; time-correlated with appliance usage.
- BLE crowding: dense RF environments cause retry storms; link quality fluctuates without movement.
| Evidence item | How to capture | What it proves |
|---|---|---|
| RSSI/LQI histogram | Log per-packet RSSI/LQI; plot distribution per node | Fade margin and variability (not single-point) |
| Retry counts | Log retries per event and per hour | Coexistence pressure and energy cost multiplier |
| Packet loss vs time-of-day | Hourly PER / drop rate trend | Time-varying interference vs geometry issues |
| Worst-spot walk test | Floorplan points with consistent orientation | Coverage gaps and placement sensitivity |
H2-8 — Hub/Base Station Hardware: Aggregation Without Going “Cloud Architecture”
The hub/base station is a local event concentrator and alarm actuator. Reliability comes from a clean interrupt-to-queue pipeline, durable local storage for offline operation, and electrical interfaces that keep working under RF load and WAN loss—without turning into “cloud architecture” content.
- Radio concentrator: receives packets, exports RSSI/LQI/CRC results, and asserts IRQ for events.
- MCU/SoC: parses frames, timestamps, dedupes, manages queues, and controls local alarms.
- Local storage (flash): offline queue + audit trail (node ID, seq, time, link metrics).
- Siren/buzzer driver: PWM/driver stage for immediate local actuation.
- Keypad & tamper I/O: local input that must function even without internet.
- Backhaul interface: Ethernet/Wi-Fi/LTE as an electrical interface block only.
Radio IRQ → parse/validate → timestamp → event queue → dedupe (node + seq) → commit to flash offline queue → local alarm actuation → optional forwarding via backhaul when available.
- Receive + record events with node ID, sequence counter, and timestamps.
- Local alarm (siren/indicator) can trigger immediately from local rules.
- Dedupe prevents duplicates when nodes retry or hub reboots.
- Offline queue retains ordering evidence for later upload/inspection.
- Tamper + keypad remain responsive under RF bursts and WAN loss.
- SPI/UART to radio: IRQ latency and buffer handling must avoid packet drops under burst traffic.
- GPIO IRQ: debounce/throttle to prevent interrupt storms from saturating the event loop.
- I2C sensors: bus errors can block processing; monitor queue depth and service time as evidence.
- Alarm driver output: ensure alarm actuation does not corrupt RF performance (power/ground transients).
- node ID, sequence counter, timestamp
- RSSI/LQI (or equivalent link metrics) and CRC fail / retry counts
- offline queue depth and drop counters (if any)
H2-9 — Backup Power: Power-Path ORing, Charging, and Outage Behavior
Backup power is a designable subsystem: power-path ORing, charging priorities, and an explicit outage behavior spec. Reliability depends on controlled switchover (no brownout), defined critical rails, and evidence that can prove the root cause when “bench works” but real outages fail.
- ORing / ideal diode: prevents reverse current and enables seamless adapter-to-battery switchover.
- Load sharing: powers system while charging, avoiding adapter droop during alarm or radio bursts.
- Charge priority: define whether system availability dominates charging, especially under marginal adapters.
- Thermal derating: charger temperature limits can reduce charge current and create “not fully charged” outages.
| Option | Strength | Common trap | Best fit |
|---|---|---|---|
| Li-ion pack | Good energy density; supports longer outage operation | Charging thermals + aging; switchover droop if ESR rises | Siren duration + sustained local operation |
| Primary cells | Long shelf life; simple power-path | Pulse current limits and cold ESR; brownout during bursts | Low-duty outage behavior with strict power budget |
| Supercap | High pulse power; long cycle life | Low energy density + self-discharge; fast voltage decay | Short hold-up for logging + graceful shutdown |
MCU/SoC rail, radio rail (basic receive / beacon), flash rail (offline queue), brownout/reset domain. Add siren rail only if required by target behavior.
Reduce radio duty (cadence-based beacon), limit nonessential peripherals, duty-cycle siren if needed, preserve event ordering evidence.
- Switchover droop: ideal diode / power-path control causes a brief Vsys sag crossing BOR threshold → reset and event loss.
- Aging ESR: battery internal resistance rises over time and at cold → deeper droop during radio TX or siren bursts.
- Charger thermals: prolonged operation reduces charge current → actual available capacity is lower than expected.
- Write during power dip: flash commits during droop can corrupt the offline queue unless write/commit is power-aware.
Test at worst-case load (radio bursts + alarm activity). Unplug adapter and verify: (1) Vsys minimum stays above BOR, (2) no reset flags, (3) event sequence remains continuous, (4) offline queue retains ordering and later replay is clean.
H2-10 — Validation Plan and Field Debug Playbook (Evidence-First SOP)
A good security kit is defined by how failures are diagnosed and prevented. This playbook is evidence-first: isolate quickly, capture required proof, map likely causes, apply fixes, and re-test with the same acceptance criteria. The goal is repeatability in real homes, not theory-heavy narratives.
Symptom → Quick isolation steps → Required evidence → Likely causes → Fix / design change → Re-test. Keep each step measurable and comparable across builds.
- Door sensor leads: long conductors inject fast transients; watch for phantom events and resets.
- Hub power input: adapter plug/unplug and surge-like events can create Vsys dips and brownouts.
- Keypad/exposed conductors: contact discharge can upset GPIO IRQ logic or induce latchups.
- Walk test: floorplan points → RSSI/LQI histogram per node (installed/enclosed condition).
- Retry vs time: correlate retries/PER with time-of-day to separate coexistence from fade margin.
- Detune check: compare enclosed vs open enclosure; large shifts indicate antenna/enclosure coupling issues.
| Required evidence | How to capture | What it isolates |
|---|---|---|
| Vsys / battery rail | Waveform during radio TX bursts and alarm activity | Brownout, droop, insufficient hold-up |
| RESET / BOR flag | Reset line or status flags correlated to droop | Proof of reset-driven event loss |
| TX_EN / current pulse | TX enable + current sense / shunt waveform | Which load triggers droop |
- Sunlight / heaters: IR changes can create slow drifts and comparator excursions.
- HVAC drafts: temperature gradients and airflow can produce repeated near-threshold triggers.
- Pets / motion patterns: define acceptance windows and rejection rules using recorded evidence.
- Vibration sources: taps, door slams, and floor resonance should be tested as a repeatable scenario set.
Sequence counter gaps (loss), duplicates (retries/reboots), and time-order violations (power instability).
Retries, CRC fails, queue depth, and drop counters correlated to time and location.
H2-11 — IC Selection Checklist + Example BOM Blocks (Within Scope Only)
This section converts “spec sheets” into an engineering checklist. Each subsystem lists what to check, how to validate it with evidence, and a small example BOM block with concrete manufacturer part numbers (MPNs) to speed up sourcing and prototype bring-up.
- Noise + bandwidth: ensure trigger decisions are not dominated by noise near threshold.
- Offset + drift: confirm threshold stability across temperature and aging.
- Input bias / leakage: critical for high-impedance nodes (RC debounce, piezo, envelope nodes).
- Wake latency / settling: first sample after wake must be trustworthy (avoid “first-sample ghost events”).
- Comparator hysteresis: enough hysteresis to avoid slow-drift chatter, but not so large that sensitivity collapses.
- Recovery behavior: fast recovery after saturation/transients to prevent multi-trigger bursts.
- Near-threshold noise histogram (or repeated trials) → false-trigger probability.
- Temperature sweep → threshold drift (offset/drift dominating or not).
- Transient injection / strong stimulus → recovery time and re-trigger behavior.
| Function | Example MPNs | Why it’s used in kits |
|---|---|---|
| Hall switch (door/window) | TI DRV5032 · Allegro A3213 | Alignment-tolerant magnetic sensing; avoids reed wear-out; supports low-power wake logic. |
| Nano-power comparator | TI TLV3691 · Analog Devices/Linear Tech LTC1540 · Microchip MCP6541 | Wake-on-threshold designs; hysteresis options; low bias helps RC/envelope nodes. |
| Zero-drift / low-offset op amp | TI OPA333 · Analog Devices ADA4528-1 | Reduces temperature-driven threshold shifts and long-term drift in low-frequency sensor chains. |
| Low-power SAR ADC | TI ADS7042 · Analog Devices AD7091R | Burst-sample capability for short wake windows; supports evidence capture beyond simple comparators. |
Tip: For “wake-only” triggers, prioritize hysteresis + recovery. For “evidence capture”, add an ADC path (even low rate) to debug false alarms.
- True sleep current: include RAM retention and RTC domain (not just “datasheet deep sleep”).
- Wake time: interrupt-to-sample-to-transmit latency defines battery budget and missed-event risk.
- RTC accuracy: timestamp integrity for offline queues; verify drift over temperature.
- GPIO wake sources: door/tamper/shock IRQs must work in deepest sleep mode.
- BOR/BOD + reset reason flags: ability to prove brownout resets and correlate to event gaps.
- Current waveform for sleep → wake → sample → TX (time + amplitude).
- Reset reason / BOR flags + event sequence continuity (gap/duplicate correlation).
| Function | Example MPNs | Why it’s used in kits |
|---|---|---|
| ULP MCU (external radio) | ST STM32L072 · Microchip SAML21 · TI MSP430FR2433 | Deep sleep + flexible wake sources; FRAM/low-power features can help event integrity in harsh power conditions. |
| Radio SoC MCU (2.4GHz) | Nordic nRF52832 / nRF52840 · Silicon Labs EFR32MG21 | MCU + radio in one device reduces BOM and can simplify evidence counters for link behavior. |
| Radio SoC MCU (sub-GHz/dual-band) | TI CC1312R · TI CC1352R | Improved wall penetration options; integrated radio timing can reduce wake-window overhead. |
- Rx sensitivity and blocking: determines real-home reliability under interference.
- Tx current vs power: must match battery pulse capability and brownout margin.
- Evidence outputs: RSSI/LQI access and retry/error counters to build histograms.
- Antenna + enclosure coupling: verify enclosed vs open performance (detune risk).
- RSSI/LQI histogram per node + retry rate vs location/time-of-day.
- Enclosure detune A/B test: same node, installed vs uninstalled condition.
| Function | Example MPNs | Why it’s used in kits |
|---|---|---|
| 2.4GHz radio SoC | Nordic nRF52840 · TI CC2652R7 · Silicon Labs EFR32MG21 | Common ecosystem; internal counters enable coexistence evidence and retry tracking without protocol deep dive. |
| Sub-GHz radio SoC | TI CC1310 / CC1312R | Better wall penetration options; useful for larger homes while keeping hardware evidence measurable. |
| Sub-GHz transceiver | Semtech SX1262 · Microchip ATA8520E | Pairs with an ULP MCU; focus remains PHY robustness (sensitivity, current) and evidence counters. |
- Coin-cell capability: cold start, minimum input voltage, pulse current headroom.
- Load switches: hard power-gating for sensors/radio domains to prevent leakage.
- Power-path / ideal diode: controlled switchover without droop across BOR.
- Charging behavior (hub): priorities + thermals + aging tolerance.
- Protection: ESD/surge at hub input; reverse/overcurrent where needed.
- Waveforms: Vbat/Vsys droop during TX + RESET/BOR + TX_EN/current pulse.
- Hub switchover: adapter unplug/replug at worst-case load; verify no resets and no event gaps.
| Function | Example MPNs | Why it’s used in kits |
|---|---|---|
| Coin-cell buck | TI TPS62740 · TI TPS62840 | Low quiescent current for always-on rails; reduces average draw in sensor nodes. |
| Coin-cell buck-boost | TI TPS63900 · Analog Devices LTC3335 | Maintains stable rails across battery droop and aging; helps avoid brownout-triggered “ghost events”. |
| Load switch | TI TPS22910A · onsemi NCP45560 | Domain gating to stop leakage; supports evidence-friendly “known power states”. |
| Power mux / ideal diode ORing | TI TPS2121 · Analog Devices/Linear Tech LTC4412 | Hub switchover control; reduces Vsys droop risk during outages and adapter plug events. |
| Li-ion charger (ULP / small pack) | TI BQ25120A · TI BQ24075 | Charging priority + protection options; align charge behavior to outage spec and thermal constraints. |
- Key storage boundary: secure element for key storage/anti-clone needs; keep scope at hardware level.
- Tamper inputs: case-open / tamper switch must be a low-power wake source.
- Event integrity: tamper events must carry the same sequence/timestamp discipline as sensor events.
- Interface cost: secure-element transactions must not explode the wake window (battery budget impact).
- Tamper-trigger wake reliability in deepest sleep.
- Secure-element transaction time + current impact measured in the wake profile.
| Function | Example MPNs | Why it’s used in kits |
|---|---|---|
| Secure element | Microchip ATECC608A · NXP SE050 · ST STSAFE-A110 | Hardware key storage and authentication building block without expanding into cloud/app security architecture. |
| Tamper switch + protection | (mechanical) + ESD diode: Nexperia PESD5V family (example class) | Simple tamper loop with ESD robustness; ensure the input remains valid during ESD touch events. |
Keep security in scope: secure element + tamper handling + evidence continuity. Avoid protocol-stack narratives.
A single diagram linking subsystems → selection knobs → evidence outputs. Use it as a review gate before freezing the BOM.
Request a Quote
H2-12 — FAQs (Evidence-First, No Scope Creep)
Each FAQ stays within this page boundary: hardware signal chain, low-power behavior, RF link evidence, backup power switchover, and local event integrity. For deeper context, each answer maps back to the relevant chapter.
1) Why does a door sensor report random open/close at night?
Start by separating mechanical bounce/misalignment from power/reset artifacts. Capture the input pin waveform (or interrupt timestamps) and verify debounce windows. In parallel, log reset reason flags and event sequence gaps/duplicates. If events cluster with TX bursts or brownouts, suspect battery ESR droop or BOR thresholding rather than the reed/Hall front-end.
2) PIR false alarms spike when sunlight moves across the floor—what to measure first?
Measure the AFE output vs threshold during the sun-sweep condition and check whether it is a slow drift crossing the decision band or a recovery artifact after saturation. Confirm band-pass corners and hysteresis are appropriate for human motion, not slow thermal gradients. A short “false-alarm campaign” with waveform snippets makes the root cause testable and repeatable.
3) A vibration sensor triggers when the washing machine runs—hardware or threshold logic?
Treat it as a mechanical coupling + energy window problem first. Record the sensor signal (or envelope/energy accumulator) during wash cycles and compare its integrated energy to the event threshold. If the enclosure or mounting resonates, the energy window will stay elevated for seconds, causing repeated triggers. Fixes usually involve coupling, filtering, and time-window logic rather than “more sensitivity”.
4) Nodes near the hub are stable, far nodes drop—what proves link budget vs interference?
Collect RSSI/LQI histograms and retry counts for both near and far nodes, then correlate failures with time-of-day. A pure link-budget issue shows persistently weak RSSI and narrow margins across time. Interference shows retries spiking at certain hours while RSSI remains similar. Also compare “installed vs uninstalled” measurements to detect enclosure detune and antenna placement errors.
5) Coin-cell devices reboot only during transmit bursts—how to prove ESR droop?
Capture three signals together: Vbat/Vsys, TX_EN (or RF activity), and RESET/BOR indication. If Vbat droops below BOR only during TX bursts, the coin-cell ESR (aging/cold) or power-path impedance is the trigger. Verify by repeating at low temperature and with a fresh cell; the droop signature will shift predictably. Then tune burst length, peak current, and BOR thresholds.
6) The hub “has backup battery” but still resets during outage switchover—what rails to scope?
Scope the hub’s main system rail (Vsys), the always-on/RTC rail (if separate), and the radio/MCU core rail across adapter unplug/replug. Add the power-path selection node (ideal diode / ORing output) to identify switchover droop. Correlate droop with reset flags and local event queue continuity to confirm whether it is a power-path transient or a depleted/aged backup source.
7) Adding an external antenna worsens performance—what layout mistake is most likely?
The first suspect is ground reference and matching disruption: long feed lines, poor ground return, missing keepouts, or an external connector placed without a controlled reference plane. Validate by comparing RSSI/LQI and retries for (a) internal antenna, (b) external antenna open-air, and (c) external antenna with the enclosure installed. If performance collapses only when installed, the enclosure/cable routing is detuning the system.
8) A tamper switch causes phantom wakes—how to debounce without missing real events?
Use a two-stage approach: a minimal hardware filter (RC or Schmitt input) to suppress micro-bounce, then a short firmware window that confirms stability after wake. The key is matching the debounce window to the wake latency and the minimum valid tamper duration. Prove correctness by logging raw IRQ timestamps and the “confirmed” tamper events; phantom wakes will show dense IRQ clusters without stable confirmation.
9) How to design a siren driver so it doesn’t brownout the hub rail?
Treat the siren as a pulse load. Measure current steps and Vsys droop at siren enable, then isolate the siren supply domain with proper bulk capacitance, controlled ramp, and a dedicated switch/driver path. Verify ground return integrity to avoid ground bounce coupling into reset lines. A fast test is “siren enable vs Vsys droop vs reset flags”; brownouts will align tightly with enable edges.
10) Why do retries explode at certain times of day?
Time-of-day retry spikes usually indicate coexistence/interference, not a static link-budget deficit. Correlate retry counters with RSSI/LQI and with household activity windows (cooking, microwave use, peak Wi-Fi usage). If RSSI stays similar but retries jump, the noise floor or channel occupancy is changing. Confirm by running a short walk test at the same locations across two time windows and comparing retry distributions.
11) How to keep event logs consistent through resets/outages?
Enforce event identity with a per-node sequence counter and include timestamps that survive reboots (RTC or monotonic backup domain). Store events atomically (or in append-only records) so power loss cannot partially corrupt the queue. On reboot, dedupe based on {node ID, sequence} and preserve an “offline queue” that continues when WAN is absent. Proof is simple: no sequence gaps without a matching reset/brownout record.
12) What’s the fastest field test to separate sensor-side vs hub-side faults?
Use a two-step isolation test: (1) move the suspect node to a “known-good” location near the hub and compare RSSI/LQI and retries; (2) swap a known-good node into the suspect location. If the issue follows the location, it is RF margin/interference or installation detune. If it follows the device, it is sensor front-end, power droop, or firmware wake timing. Capture one short evidence set: Vbat droop + reset flags + retry counters.
Figure F11 — Evidence-First Triage Flow (FAQ → Measurable Proof)
A compact flow to decide whether an issue is sensor front-end, power integrity, RF margin/interference, hub aggregation, or backup switchover.
