123 Main Street, New York, NY 10001

Submit Your BOM

Electronic Fence & Vibration Cable — AFE, Timing & Event Logging

Electronic Fence & Vibration Cable — AFE, Timing & Event Logging

← Back to: Security & Surveillance

This page explains an electronic fence & vibration cable as an evidence-based edge device: where to measure V/I, how the AFE avoids false alarms in rain/EMI, and how events are timestamped, logged, and reported reliably for field-debug and audit.

H2-1. Definition & Scope: What “Electronic Fence + Vibration Cable” Covers

This page is an end-device hardware guide for perimeter detection: (1) high-voltage pulse generation and fence-loop sensing, (2) vibration-cable analog front-end (AFE) and features, and (3) device-side event timing, logging, and reliable remote reporting. It is intentionally not a platform or “VMS/NVR ecosystem” tutorial.

What this page treats as the “engineering system”
  • HV pulse chain (measurable): pulse driver → fence loop/return path → V_FENCE / I_RETURN sensing → signatures for cut/short/leak/tamper.
  • Vibration chain (measurable): sensor cable → band-pass/AGC → ADC → features (energy/spectrum/envelope) → threshold and confidence.
  • Event chain (auditable): decision → EVENT_TS → ring-buffer log → uplink (RS-485/Ethernet/cellular) with sequence/retry/de-dup.
What this page does NOT cover
  • Video/recording platforms: VMS ingest, NVR storage, recording compliance/WORM (handled in dedicated pages).
  • Network infrastructure design: PoE switch/PSE internals, PTP grandmaster theory, fiber panel/OTDR systems.
  • Generic surge encyclopedia: only device ports and fence/cable coupling are discussed (not broad lightning textbooks).
Evidence-first Scope locked Repeatable field diagnosis
What to measure
  • V_FENCE_PEAK / pulse width / repetition
  • I_RETURN_PEAK / leakage estimate
  • Vibration RMS energy + band power
  • Supply state + brownout flags
Pass/Fail signatures
  • Open/cut: V peak shifts + return current collapses
  • Short: V peak clamps + current spikes
  • Leak/humidity: slow baseline drift + duty changes
  • Vibration false alarm: energy outside target band
First fix (fastest)
  • Confirm sensing points are not saturating/clipping
  • Stabilize return-path and ground health monitoring
  • Lock timestamp + log commit policy on power-loss
  • Re-tune band-pass + thresholds per zone environment
Key signals & log fields used throughout this page
Name Meaning (device-side) Why it matters (diagnosis)
V_FENCE_PEAK Peak sensed fence pulse voltage at the output sampling node. Distinguishes open/cut vs short vs leakage; reveals clamp/saturation events.
I_RETURN_PEAK Peak return/current-sense proxy during the pulse window. Separates true load changes from AFE artifacts; enables leakage estimation.
PULSE_WIDTH, REP_RATE Pulse timing parameters measured by timer/capture. Helps correlate alarms with pulse scheduling, EMI coupling, and power droops.
ZONE_ID Zone identifier for multi-zone or segmented loops. Prevents “global thresholds” mistakes; speeds isolation and on-site validation.
EVENT_TYPE CUT/SHORT/LEAK/TAMPER/VIB_ALARM (etc.). Enables consistent audit trails and remote triage without platform deep-dives.
EVENT_TS Timestamp (RTC or monotonic counter) at decision time. Solves “missing / out-of-order” reports; aligns field reports with on-site reality.
CONFIDENCE Confidence score from signatures/features. Separates environmental noise from true intrusions; drives threshold tuning.
BROWNOUT_FLAG Power-loss or undervoltage indicator during logging/uplink. Explains event loss/duplication; prevents misdiagnosing “network issues”.
Electronic Fence + Vibration Cable — End-Device Evidence Chain HV pulse + sensing AFE + vibration AFE + timestamped event log + remote reporting HV Pulse Chain Pulse Driver H-bridge / switch Energy Storage Cstore + charger HV Transformer + Output Network Pulse shaping • safety envelope • isolation barrier Fence Loop + Return Path cut / short / leak signatures happen here Sensing & Decision Pulse Sense AFE protect • attenuate • shape Vibration AFE band-pass • ADC MCU/SoC Decision + Timestamp features • thresholds • confidence • EVENT_TS Event Log ring buffer • NVM Remote Comms RS-485 • Ethernet • cellular V_FENCE I_RETURN EVENT_TS ZONE_ID / EVENT_TYPE Support Blocks (Device-Side) Tamper & Self-Test cover open/short • inject test Power Health Hooks BROWNOUT_FLAG • supply_state Remote Diagnostics counters • last signatures Design goal: every alarm maps to a measurable signature + a timestamped log record.
Figure F1. Layered end-device block diagram. Key measurable nodes are V_FENCE, I_RETURN, and device-side fields such as EVENT_TS and ZONE_ID / EVENT_TYPE.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F1 (End-Device Evidence Chain). Link

H2-2. System Topologies: Energizer, Zone Controller, Sensor Cable Deployment

Topology choices are not cosmetic. They determine whether an alarm can be localized, whether thresholds remain stable across environments, and whether field validation is fast. This section describes three common device-side topologies and the minimum interfaces needed to keep the evidence chain intact.

Topology A — Single Zone (fast deployment, weakest localization)
  • When it fits: short/medium loops, uniform environment, limited maintenance access.
  • Key interfaces: one fence loop (2-wire), optional dry-contact alarm out, optional RS-485 for diagnostics.
  • Engineering consequence: a single global threshold must tolerate humidity and ground variation; mis-tuning causes either false alarms (too sensitive) or missed cuts (too strict).
Topology B — Multi-Zone (best localization, per-zone thresholds)
  • When it fits: long perimeter, mixed environments (trees/roads/metal fence sections), strict service SLAs.
  • Key interfaces: multiple fence segments; per-zone sensing and ZONE_ID in every log entry.
  • Engineering consequence: thresholds can be tuned per zone; diagnostics become actionable because “which segment” is known immediately.
Topology C — Remote Sensing Node (cable-noise isolation, service-friendly)
  • When it fits: very long sensor runs, high EMI exposure, or when bringing the AFE closer to the cable improves SNR.
  • Key interfaces: remote node uplink (RS-485 or Ethernet at the device edge), sequence IDs for de-dup, offline buffering.
  • Engineering consequence: remote node must carry “evidence payload”: minimal feature snapshot + timestamp + health flags, not only an alarm bit.
Interface checklist Avoid platform creep
Fence loop (2-wire)
  • Carries HV pulse signatures
  • Primary source for V/I evidence
  • Drives cut/short/leak decisions
RS-485 (device-side)
  • Alarm + diagnostics fields
  • Sequence ID + retry policy
  • Remote read of counters/health
Dry contact output
  • Alarm state only (simple)
  • No deep diagnostics payload
  • Use when panel expects discrete IO
Deployment variables that most often decide topology
  • Loop length and segmentation: longer runs amplify coupling and baseline drift; segmentation enables per-zone baselines.
  • Ground health stability: return-path variability is a top false-alarm driver; design for measurable “ground health” signals.
  • Environmental mix: wind/traffic/metal mesh sections can dominate vibration features; per-zone band/threshold tuning helps.
  • Service workflow: if technicians must isolate faults quickly, multi-zone logs with ZONE_ID outperform “single alarm relay”.
  • Evidence transport: remote nodes should send timestamped evidence summaries (not just “ALARM=1”).
Deployment Topologies (Device-Side) Interfaces shown: fence loop • RS-485 • dry contact • optional Ethernet A) Single Zone B) Multi-Zone C) Remote Node Energizer + Sense one baseline Fence loop (2-wire) V_FENCE • I_RETURN RS-485 Dry contact Label: fast deploy, weak localization Zone Controller per-zone thresholds Zone 1 loop ZONE_ID=1 Zone 2 loop ZONE_ID=2 RS-485 / Ethernet (device-side) Label: best localization, per-zone baselines Main Controller logs + uplink Remote Sensing Node AFE + timestamp + buffer RS-485 Ethernet Evidence payload EVENT_TS • seq_id • confidence Label: isolates cable noise, service-friendly
Figure F2. Three device-side deployment topologies. The critical difference is whether evidence can be localized and whether logs carry ZONE_ID plus timestamped diagnostics (not only an alarm bit).
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F2 (Device-Side Topologies). Link

H2-3. Threat & Fault Model: What You Must Detect (Cut/Short/Leak/Climb/Tamper)

Perimeter “alarms” must be engineered as measurable signatures, not vague event labels. This section defines the minimum event set and the observable evidence each event should produce across three channels: HV waveform, vibration features, and device-side fields (timestamp/zone/confidence/power state).

Event classes (to prevent logic mixing)
  • Intrusion / tamper: climb/touch, cut/open, bypass/shield/interference, enclosure tamper.
  • Electrical faults: short, leakage (humidity/aging), unstable return path / ground health degradation.
  • Environment: wind/rain/traffic vibration, EMI coupling, temporary wet surfaces.
What counts as evidence (the three-channel rule)
  • HV waveform evidence: V_FENCE_PEAK, I_RETURN_PEAK, rise/decay shape, repetition consistency (REP_RATE), and window-aligned measurements.
  • Vibration evidence: band energy, duration, envelope shape, and pattern consistency across repeated bursts.
  • Device fields: ZONE_ID, EVENT_TS, CONFIDENCE, BROWNOUT_FLAG, and sequence/de-dup counters.
Evidence-first False alarm control Field-repeatable
What to measure first
  • V_FENCE_PEAK + rise/decay shape
  • I_RETURN_PEAK + window alignment
  • Vibration: band energy + duration
  • EVENT_TS + ZONE_ID + CONFIDENCE
Discriminators (most useful)
  • Step-change vs slow drift (cut vs leak)
  • Cycle-to-cycle consistency (fault vs transient noise)
  • HV-only vs HV+vibration concurrence (touch/climb)
  • Power flags explain “phantom” event loss/duplication
First fix (fastest)
  • Ensure sensing paths do not saturate or clip
  • Separate environmental vibration from target band
  • Log commit + de-dup policy tied to SEQ_ID
  • Per-zone baselines when environments differ
Event → signature mapping (concise, measurable)
Event HV waveform signature Vibration signature Device fields that confirm
Cut / Open Step-change in V_FENCE_PEAK with a collapse/shift in I_RETURN_PEAK; abrupt change is key. Often none; if present, short burst near event time (depends on installation). EVENT_TS aligns with step-change; stable REP_RATE suggests real loop change, not scheduler drift.
Short Clamp-like behavior in V + elevated current proxy; repeated across cycles (not one-off spikes). Usually none. High confidence with consistent repeats; check BROWNOUT_FLAG to avoid mistaking resets for “shorts”.
Leak (humidity/aging) Slow baseline drift over minutes/hours; changes accumulate rather than instant steps. Not required; may increase noise floor indirectly. Trend by ZONE_ID; confidence fluctuates with weather/time; long-term logs are decisive.
Climb / Touch Transient HV signature near contact; best when correlated with concurrent vibration evidence. Short-duration energy rise with recognizable envelope in target band. EVENT_TS concurrence across both channels; cross-zone correlation should be low (localized).
Bypass / Shield / Interference Waveform becomes “inconsistent” with normal load model; cycle-to-cycle anomalies or unnatural stability changes. May appear as non-physical patterns; use consistency tests, not single samples. Use confidence + anomaly counters; never rely on one metric alone.
Tamper (enclosure) May be absent; treat as a separate digital/IO evidence path. May be absent. TAMPER_STATE with EVENT_TS; compare with BROWNOUT_FLAG to avoid false tamper from power drop.
Event → Signature Map (Evidence Channels) Use short labels only: Peak/Rise/Decay/Rep • Energy/Band/Duration/Pattern • EVENT_TS/ZONE_ID/CONF HV waveform lane Peak • Rise • Decay • Rep Vibration lane Energy • Band • Duration • Pattern Device fields lane EVENT_TS • ZONE_ID • CONF • BROWNOUT Events Icons + short names Cut / Open step-change Short clamp + repeat Leak slow drift Climb / Touch HV + vibration Bypass / Shield anomaly + consistency Tamper IO + timestamp Peak Rise Decay Rep Energy Band Duration Pattern EVENT_TS ZONE_ID CONF BROWNOUT
Figure F3. Event-to-signature map. A robust design avoids single-metric decisions by combining HV waveform evidence, vibration features, and device-side fields (timestamp/zone/confidence/power flags).
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F3 (Event → Signature Map). Link

H2-4. High-Voltage Pulse Generation: Energy, Waveform, and Safety Envelope

HV pulse generation for perimeter sensing should be treated as a controlled, measurable energy packet with a defined time window and safety constraints. This section explains the block-level architecture and the indicators you can measure to keep detection signatures stable—without relying on unsafe parameter disclosures.

Block-level concept: energy packet + time window
  • Energy formation: input power charges an energy store, then a switch stage releases a bounded pulse through the HV path.
  • Waveform shaping: the output network and the fence loop determine rise/decay behavior and repeatability.
  • Detection relevance: stable and measurable waveforms produce reliable V_FENCE_PEAK / I_RETURN_PEAK signatures across weather and installation variance.
Measurable indicator types (no unsafe numeric values)
Indicator Where it comes from Why it matters for detection
REP_RATE (repetition) Timer/capture count per second or per window. Correlates false alarms with scheduling/EMI; reveals instability that can mimic intrusions.
PULSE_WIDTH (time window) Capture window aligned to the switching event. Misalignment creates “phantom” current/voltage artifacts and unstable classification.
V_FENCE_PEAK Output sense via attenuator + isolation-capable measurement chain. Primary signature dimension: step-change vs drift separates cut/short/leak patterns.
I_RETURN_PEAK (or proxy) Return-path current sense (shunt/CT/Hall proxy depending on design). Confirms whether changes are real load shifts or sensing saturation/clipping.
Delivered energy estimate (proxy) Derived from waveform features (windowed integral / modeled equivalent load). Helps explain why alarms rise when energy delivery becomes inconsistent across cycles.
Safety envelope (device-side checks that must exist)
  • Isolation barrier integrity: a design-time and production-time verification hook; treat barrier status as a health item.
  • Interlock status: an explicit gate that can inhibit pulse generation when required by safety policy.
  • Fail-safe behavior: faults must push the system into a safe state and leave a diagnosable log record.
Evidence-first Waveform stability Safety hooks
What to capture
  • V_FENCE_PEAK & shape (rise/decay)
  • I_RETURN_PEAK in the same window
  • PULSE_WIDTH / REP_RATE
  • INTERLOCK / barrier health flags
Pass/Fail signatures
  • Cycle-to-cycle repeatability within expected variance
  • No clipping/saturation in the sensing chain
  • Interlock transitions produce explicit, logged state changes
  • Power-loss flags explain missing or duplicated events
First fix (fastest)
  • Align sensing window to the actual pulse event
  • Verify sensing headroom (avoid clamp artifacts)
  • Stabilize return path / ground-health monitoring
  • Make “safe state + log” the default failure behavior
HV Pulse Generation — Block-Level Architecture Focus: measurable indicators + safety envelope (no unsafe numeric disclosure) Energy path (left → right) Input Power DC supply Charge Control bounded energy Cstore energy buffer Switch Bridge pulse window HV Tx isolation Output Shaping Network + Fence Terminal waveform shape • port protection • fence loop coupling V_FENCE_PEAK I_RETURN_PEAK PULSE_WIDTH REP_RATE Safety envelope (device-side gating + health) INTERLOCK explicit enable gate Isolation / Barrier Health verify integrity • log status Fail-safe + Log safe state on fault Measurement window aligned to pulse event
Figure F4. HV pulse generation at block level. The key is not “maximum output,” but stable, measurable waveforms (V_FENCE_PEAK, I_RETURN_PEAK, PULSE_WIDTH, REP_RATE) under a safety envelope (INTERLOCK, barrier health, fail-safe logging).
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F4 (HV Pulse Generation Block Diagram). Link

H2-5. Fence Output Stage & Sensing Points: Where to Measure V/I and Why

This chapter defines where measurements must be taken so that captured signatures represent the fence loop (not switch-node artifacts). It formalizes three evidence channels: V_SENSE, I_SENSE, and GROUND_HEALTH_NODE, and shows how they support stable discrimination of open, short, and leakage events.

5.1 Separate the energy path from the observation path
  • Energy path: HV isolation/transform → output shaping → fence terminal → return path.
  • Observation path: protected sense nodes + isolation-capable measurement chain → ADC/capture → feature/log fields.
  • Design rule: a sense node is valid only if its common-mode and transient behavior stays inside the measurement chain’s safe operating envelope.
5.2 V_FENCE sensing (divider + isolation + common-mode)
Sense node placement
  • Terminal-representative sensing targets the fence output behavior and improves event fidelity.
  • Switch-adjacent sensing often captures edge spikes/parasitics that reduce repeatability.
  • Define V_SENSE_RAW as a named node tied to layout coordinates and BOM constraints.
Divider chain requirements
  • Divider must survive transients and maintain ratio stability across environment and aging.
  • Protection elements belong to the divider strategy (not an afterthought).
  • Divider output should be bandwidth-shaped to preserve event features (peak/decay), not cosmetic waveform detail.
Isolation amplifier constraints
  • Common-mode handling is part of measurement integrity (avoid “false event” artifacts).
  • Isolation barrier health must be observable via a status field (health flag + timestamp).
  • Measurement chain must avoid saturation/clipping; include a CLIP_FLAG and RANGE_ID field.
5.3 I_RETURN sensing options (shunt vs CT vs Hall)
Method Strength Risk / limit Best evidence fields
Shunt Direct linear proxy for return current; supports windowed integration features. Requires robust protection and layout discipline; susceptible to transient stress and parasitic error. I_SENSE_RAW, I_PEAK, I_WINDOW_SUM, CLIP_FLAG
CT Inherent isolation tendency and strong pulse sensitivity; good for repeatability scoring. Limited low-frequency/DC content; waveform recovery depends on burden and loading behavior. I_PEAK, I_RINGING_SCORE, WINDOW_ID
Hall Galvanic separation from the main path; can cover broad dynamic conditions (implementation-dependent). Offset/temperature drift can imitate “slow leakage”; bandwidth limits may hide fast signatures. I_BASELINE, I_DELTA, TEMP_TAG, CONFIDENCE
5.4 Stable discrimination bundles (open / short / leakage)
multi-metric multi-pulse confirm per-zone baseline
Open / Cut
  • Use step-change in V_SENSE + concurrent change in I_SENSE.
  • Require cross-cycle consistency to reject single transient artifacts.
  • Log: EVENT_TS, ZONE_ID, WINDOW_ID.
Short
  • Combine clamp-like voltage behavior with repeatable return-current evidence.
  • Cross-check protection/status bits to avoid confusing measurement clipping with real shorts.
  • Log: CLIP_FLAG, RANGE_ID, CONFIDENCE.
Leakage (humidity/aging)
  • Prefer trend evidence over single-shot peaks; bind trend to ZONE_ID.
  • Use an energy proxy (I_WINDOW_SUM) + ground health to separate from vibration noise.
  • Log: EVENT_TS series, GROUND_HEALTH, TEMP_TAG.
5.5 Ground-health node: why a third node prevents false narratives
  • Problem: return paths and reference points can shift with installation and environment; this can look like “events”.
  • Solution: a GROUND_HEALTH_NODE provides a measurable reference for “return integrity” and supports per-zone baseline stability.
  • Log binding: ground health status must be timestamped and correlated with event bursts.
Fence Output Stage + Sensing Points Measure where the loop is represented: V_SENSE, I_SENSE, GROUND_HEALTH_NODE Energy path (thick) and observation path (thin) HV Tx isolation Output Shaping port protection Fence Terminal loop coupling Return Path I_RETURN V_SENSE (V_FENCE) I_SENSE (shunt/CT/Hall) GROUND_HEALTH_NODE Isolation boundary (measurement domain) Divider protected Iso Amp CM-safe ADC capture Fields bound to measurement EVENT_TS ZONE_ID RANGE_ID CLIP_FLAG Legend Energy path Observation path Isolation boundary Sense node
Figure F5. Output stage with recommended sensing nodes. A valid design defines where V/I are measured, preserves isolation/common-mode integrity, and binds measurements to log fields for repeatable classification.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F5 (Output Stage + Sensing Points). Link

H2-6. Pulse Detection AFE: Front-End Protection, Filtering, and Dynamic Range

The pulse-detection front end must survive a harsh interface while still extracting small, repeatable differences. The AFE is successful only if it delivers non-clipped, window-aligned features suitable for discrimination and logging: PEAK, DECAY, WINDOW_SUM, and CONF.

6.1 Input protection chain (survival without erasing signatures)
  • Limit the stress (series impedance concept) before sensitive nodes.
  • Clamp abnormal excursions into a measurable envelope (with explicit status fields).
  • Shape the bandwidth with RC elements to preserve event-relevant features (peak/decay), not to “beautify” the waveform.
  • Isolate the measurement domain so common-mode shifts do not appear as false events.
6.2 Dynamic range strategy (multi-path / ranging / peak-hold)
Multi-path sensing
  • Coarse path prevents clipping during strong disturbances.
  • Fine path keeps sensitivity for small deltas in normal operation.
  • Fuse into one feature set with a logged RANGE_ID.
Switched attenuation / auto-ranging
  • Use discrete ranges to keep ADC inputs inside safe headroom.
  • Record range transitions to prevent misclassification from “range jumps”.
  • Expose CLIP_FLAG when any stage saturates.
Peak-hold / envelope features
  • Convert fast pulses into a slower, sample-friendly representation.
  • Prefer stable features (peak/decay/windowed sum) over raw waveform storage.
  • Bind extraction to WINDOW_ID for repeatability.
6.3 False-alarm suppression: wet leakage vs real cut
trend vs step repeatability per-zone baseline
  • Leakage (wet/aging): typically appears as slow drift and accumulates over time; trend evidence must be tied to ZONE_ID.
  • Cut/open: appears as a step-change and stays consistent across multiple pulses; require multi-pulse confirmation.
  • AFE policy: reject one-shot anomalies when CLIP_FLAG is set or when WINDOW_ID is misaligned.
6.4 What must be logged (so field debug is deterministic)
Field Purpose
RANGE_ID + CLIP_FLAG Proves whether measurements were within valid headroom; prevents mislabeling “clipped artifacts” as real shorts/cuts.
WINDOW_ID + EVENT_TS Binds feature extraction to the correct pulse window; essential for repeatability across captures and firmware revisions.
PEAK / DECAY / WINDOW_SUM Minimal feature set for robust classification without storing raw, noise-sensitive waveforms.
NOISE_FLOOR (or proxy) Explains sensitivity loss and rising false negatives under environment/installation drift.
CONF + ZONE_ID Allows per-zone baselining and consistent alarm policy without platform-level dependence.
Pulse Detection AFE — Protect • Range • Features Goal: non-clipped, window-aligned features (PEAK/DECAY/WINDOW_SUM/CONF) Sense Input V_SENSE / I_SENSE WINDOW_ID PROTECT limit • clamp • RC Rlim Clamp RC RANGE attenuate • dual-path • hold ATTEN DUAL HOLD SHAPE filter • envelope LPF ENV ADC + Windowed Capture bind samples to the pulse window CLIP_FLAG RANGE_ID NOISE_FLOOR Feature Extract + Classifier Inputs stable features outperform raw waveform storage PEAK DECAY WINDOW_SUM PATTERN CONF ZONE_ID AFE validity policy Reject classification when CLIP_FLAG is set or WINDOW_ID is misaligned; require multi-pulse confirmation valid sample window no clipping repeatability score
Figure F6. Pulse detection AFE chain. The design must survive the interface (PROTECT), keep measurements within headroom (RANGE + CLIP_FLAG), and output stable, window-aligned features for logging and classification.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F6 (Pulse Detection AFE Chain). Link

H2-7. Vibration Cable Front-End: AFE, Blanking, and Signature Features

A vibration cable channel must detect small, repeatable motion signatures while coexisting with strong pulse-related interference. This chapter defines an evidence-first pipeline: protected sensing → pulse-synchronized blanking → stable features (VIB_ENV, VIB_BAND_EN, VIB_DURATION) → confidence-backed events bound to WINDOW_ID and ZONE_ID.

7.1 What the vibration cable “looks like” electrically (design constraints)
  • High-impedance + low-level variation is common; the front end must protect inputs without masking small deltas.
  • Wide spectral content is expected (slow sway, rhythmic tapping, sharp impacts); features must preserve frequency separation.
  • Strong common-mode and impulse interference can couple from pulse activity and long outdoor wiring; measurement validity must be explicit.
Evidence fields (minimum set)
  • VIB_RAW (raw samples) + VIB_ENV (envelope / energy proxy)
  • VIB_BAND_EN[1..N] (multi-band energy) + VIB_DURATION (event length)
  • VIB_REPEAT_SCORE (pattern repeatability) + CONF (confidence)
  • VIB_RANGE_ID + VIB_SAT_FLAG (dynamic range validity)
  • PULSE_BLANKING_WIN + WINDOW_ID (pulse-synchronized gating)
7.2 Front-end chain: protect → bias/excite → filter → capture
  • Input protection: include a survivable impedance/clamp concept before any high-gain node; log saturation via VIB_SAT_FLAG.
  • Bias/excitation (optional): provide a controlled operating point so the AFE sees consistent impedance behavior across environment.
  • Filtering: prefer feature-oriented filtering (multi-band or band-limited energy) rather than “pretty waveforms”.
  • Capture policy: tie sampling windows to a timebase and expose validity flags (range/saturation/window alignment).
7.3 Pulse coexistence: blanking is mandatory, not optional
Design rule: vibration features are valid only outside the pulse-coupled disturbance window. Use a synchronized gate PULSE_BLANKING_WIN and compute features only from “valid segments”. VIB_VALID = (NOT PULSE_BLANKING_WIN) AND (VIB_SAT_FLAG == 0) FEATURES computed over valid windows → EVENT built with WINDOW_ID + ZONE_ID
7.4 Signature features that remain stable outdoors
Feature bundle (recommended)
  • Multi-band energy: VIB_BAND_EN separates slow sway vs sharp impacts vs repetitive tapping.
  • Duration + burst structure: VIB_DURATION and short/long burst ratios improve discrimination without heavy compute.
  • Repeatability: VIB_REPEAT_SCORE reduces one-shot false alarms from random environmental noise.
  • Zone baseline: VIB_BASELINE per ZONE_ID prevents “one threshold fits all” failures.
Vibration Cable Channel — AFE + Blanking + Features Extract stable signatures while rejecting pulse-coupled disturbance Vibration Cable VIB_IN PROTECT limit • clamp • ESD Rlim Clamp BIAS / EXCITE optional Bias Test AFE gain • filter Gain BPF ADC capture Pulse coexistence gate Compute features only outside the disturbance window PULSE_BLANKING_WIN VIB_SAT_FLAG VIB_RANGE_ID WINDOW_ID Feature vector (logged) VIB_ENV VIB_BAND_EN[ ] VIB_DURATION VIB_REPEAT_SCORE CONF ZONE_ID Validity rule Only compute / classify using non-blanked, non-saturated windows VALID = !BLANK && !SAT
Figure F7. Vibration cable sensing chain with pulse-synchronized blanking. The front end remains deterministic by computing features only from valid windows and logging range/saturation state alongside the feature vector.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F7 (Vibration Cable AFE + Blanking + Features). Link

H2-8. Timebase & Event Timestamping: EVENT_TS, SEQ_ID, and Zone Correlation

Timestamping must be designed as an end-to-end chain: time source → window alignment → event creation → log commit → transmit. This chapter defines a minimal, audit-friendly set of fields so events remain explainable across resets, outages, and link delays.

8.1 Separate “event time” from “write time”
  • EVENT_TS: when the event is determined to have occurred (bound to a capture window).
  • LOG_COMMIT_TS: when the record is safely stored (may lag due to buffering, retries, or brownouts).
  • Why it matters: without both, field investigations cannot distinguish real timing from backlog effects.
Minimum timestamp & ordering fields
  • TIME_SRC (time source mode) + RTC_STATUS (valid/invalid)
  • WINDOW_ID (capture window) + SEQ_ID (monotonic event sequence)
  • BOOT_ID (power-cycle epoch) + RESET_CAUSE (POR/WDT/UVLO)
  • EVENT_TS + LOG_COMMIT_TS (two clocks, two meanings)
  • ZONE_ID (baseline + correlation scope)
8.2 Window alignment: bind all channels to the same reference
  • V/I pulse features and vibration features must share a common WINDOW_ID to avoid “phantom correlation”.
  • When a blanking window is active (PULSE_BLANKING_WIN), mark vibration-derived features as invalid for that window.
  • Attach validity bits (range/saturation/blanking) to every event record so post-analysis does not guess.
8.3 Reset & outage integrity: prevent duplicates and gaps
Policy: the tuple (BOOT_ID, SEQ_ID) must uniquely identify an event, even across retransmissions. EVENT_ID := (BOOT_ID, SEQ_ID) On reset: increment BOOT_ID; preserve next SEQ_ID monotonicity if storage allows Always log RESET_CAUSE + RTC_STATUS to explain time validity
8.4 What “good” logs enable in field debug
  • Correlate: V_SENSE + I_SENSE + VIB_ENV within one WINDOW_ID.
  • Explain anomalies: a spike in alarms aligned with RESET_CAUSE or invalid RTC_STATUS points to power integrity, not “intrusions”.
  • Prove ordering: SEQ_ID and LOG_COMMIT_TS expose backlog and retransmit behavior without platform assumptions.
Timebase & Timestamp Pipeline EVENT_TS (event time) ≠ LOG_COMMIT_TS (write time) TIME_SRC timer • crystal • RTC hold-up MCU timer RTC RTC_STATUS + RESET_CAUSE Window alignment bind channels to WINDOW_ID WINDOW_ID ZONE_ID Validity bits (blank/sat/range) Event build event time + ordering identity EVENT_TS SEQ_ID BOOT_ID EVENT_ID = (BOOT_ID, SEQ_ID) Store → Commit → Transmit (independent clocks) Commit time may lag event time; ordering must stay deterministic Ring buffer EVENT_REC Commit to storage LOG_COMMIT_TS Tx queue retry / ack Remote comms PACKET_SEQ + ACK Integrity rules Use EVENT_ID=(BOOT_ID,SEQ_ID); keep EVENT_TS separate from LOG_COMMIT_TS EVENT_ID = (BOOT_ID, SEQ_ID)
Figure F8. Timestamp pipeline from time source to transmit. Reliable field investigations require deterministic ordering (BOOT_ID/SEQ_ID) and explicit separation of EVENT_TS vs LOG_COMMIT_TS.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F8 (Timebase & Timestamp Pipeline). Link

H2-9. Remote Communications & Alarm Reporting: RS-485/Ethernet/Cellular with Security Hooks

Remote reporting is the final link in the evidence chain: alarms must be replayable, de-duplicable, and explainable across outages and resets. This chapter focuses on edge-device behavior only—event records, transmit queues, acknowledgements, offline buffering, and minimal identity hooks—without relying on any platform details.

9.1 Decouple detection from reporting (never let comms stall sensing)
  • Rule: detection/classification runs on a fixed schedule; reporting runs on a separate queue with backpressure.
  • Implication: link loss increases OFFLINE_Q_DEPTH but does not change capture windows or feature extraction.
  • Explainability: log backlog and retries so a “storm” can be attributed to comms, not intrusions.
Core reliability fields (log these)
  • EVENT_ID (recommended = BOOT_ID + SEQ_ID)
  • PACKET_SEQ (per-packet ordering; separate from SEQ_ID)
  • TX_RETRY_CNT + ACK_LAT_MS (retries + ack timing)
  • DUP_DROP_CNT (de-dup drops) + OFFLINE_Q_DEPTH (queue depth)
  • LAST_TX_FAIL_CAUSE (timeout / link down / auth fail)
9.2 Reliable alarm reporting: replay, de-dup, and commit semantics
  • Replayable: a report is derived from an immutable event record; retransmission must not mutate the event.
  • De-duplicable: receiver-side and sender-side de-dup both key off EVENT_ID.
  • Commit semantics: differentiate EVENT_TS (event time) vs LOG_COMMIT_TS (stored time).
Minimal policy (platform-agnostic) EVENT_ID := (BOOT_ID, SEQ_ID) Report carries: EVENT_ID + ZONE_ID + EVENT_TS + CONF + key features Sender keeps: PACKET_SEQ + TX_RETRY_CNT + OFFLINE_Q_DEPTH Receiver can ACK: EVENT_ID (idempotent)
9.3 Remote health & diagnostics: counters that explain false alarms
Recommended remote-readable counters
  • Link: LINK_UP, LINK_DOWN_CNT, PHY_ERR_CNT (Ethernet) or RSSI/RSRP (cellular)
  • Resets/power: RESET_CAUSE, BROWNOUT_CNT, UVLO_FLAG
  • Time validity: RTC_STATUS, TIME_SRC
  • Outdoor stress: SURGE_HIT_CNT (if available), GND_HEALTH (score / trend)
9.4 Security hooks (minimal, non-instructional)
  • Device identity: include DEV_ID and firmware/config versions (FW_VER, CFG_VER) in diagnostics.
  • Key storage abstraction: log only a handle such as KEY_SLOT_ID (no key material, no provisioning steps).
  • State counters: SEC_STATE + AUTH_FAIL_CNT to explain “silent” dropouts caused by authentication or integrity checks.
Remote Communications & Alarm Reporting (Edge-only) Replayable events • De-dup • Offline queue • Minimal identity hooks Edge Device Event Record (immutable) EVENT_ID • ZONE_ID • EVENT_TS • CONF • key features EVENT_ID Tx Queue (decoupled) PACKET_SEQ • ACK • TX_RETRY_CNT • OFFLINE_Q_DEPTH PACKET_SEQ TX_RETRY_CNT OFFLINE_Q Link Monitor LINK_UP • ERR_CNT ACK_LAT_MS Security Hooks DEV_ID • KEY_SLOT_ID AUTH_FAIL_CNT Remote diagnostics (read-only) RESET_CAUSE • BROWNOUT_CNT • RTC_STATUS • GND_HEALTH Uplink options RS-485 isolated PHY Ethernet link + counters Cellular RSSI/RSRP Controller / Panel (abstract) Alarm ingest idempotent by EVENT_ID ACK(EVENT_ID) Diagnostics read Read counters / health RESET_CAUSE GND_HEALTH Interface only No platform specifics Key rule: EVENT_ID drives de-dup; PACKET_SEQ drives transport ordering
Figure F9. Edge-only alarm reporting. An immutable event record feeds a decoupled transmit queue with transport-level sequencing. Receiver acknowledgements are idempotent by EVENT_ID, enabling safe retries without duplicates.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F9 (Remote Comms & Alarm Reporting Interfaces). Link

H2-10. Outdoor Robustness: Ground Health, Lightning Coupling, and Cable-Fault Isolation

Outdoor performance is dominated by long-cable physics: ground conditions drift, common-mode impulses couple into wiring, and a single degraded segment can destabilize an entire zone. This chapter stays specific to fence-loop robustness: quantify ground health as a trend, segment the system for isolation, and prioritize safety and survivability before measurement fidelity.

10.1 Ground health is a trend, not a one-time check
  • Goal: estimate loop health from observable behavior (V/I features + drift), then watch it over time.
  • Why trend matters: corrosion, moisture, and seasonal changes create slow drifts that look like “random false alarms”.
  • Field value: trend-based health prevents misclassification by flagging suspect conditions before alarms spike.
Ground-health evidence fields (log + report)
  • R_LOOP_EST (equivalent loop impedance estimate) + LEAK_INDEX (leakage signature score)
  • GND_DRIFT_24H (trend metric) + WET_FLAG (environmental suspicion flag)
  • GND_HEALTH_SCORE (compressed score) + ALARM_SUSPECT_PWR (alarm likely power/ground related)
10.2 Lightning / impulse coupling: handle common-mode stress without turning this into a “surge textbook”
  • Common-mode first: long cables act as antennas; impulsive coupling can cause resets, saturation, or momentary validity loss.
  • Engineering requirement: every protective action must be observable in logs (validity bits, reset causes, saturation flags).
  • Outcome: operators can separate “stress event + reset” from “true intrusion signature”.
Observability rule If protection / coupling causes saturation or invalid windows: set CLIP_FLAG / RANGE_ID / VALID_WIN_RATIO If stress causes reboot: log RESET_CAUSE + BOOT_ID If time becomes untrusted: log RTC_STATUS + TIME_SRC
10.3 Cable-fault isolation: segment first, then localize (concept-level)
  • Zone segmentation: isolate baseline and health per ZONE_ID to prevent one fault from poisoning all zones.
  • Segment scanning (concept): optionally rotate a measurement focus across segments to compare deltas.
  • Response-shape clues (concept): coarse localization can use response delay/shape bins without exposing implementation details.
Isolation & localization evidence fields (concept-friendly)
  • ZONE_HEALTH + ZONE_BASELINE (per-zone stability)
  • SEG_ID + SEG_SCAN_TS (segment scan bookkeeping)
  • SEG_DELTA_SCORE (segment deviation score) + RESP_DELAY_BIN (coarse delay bin)
10.4 Protection priority: safety → survivability → measurement fidelity
  • Layer 0 (safety): log interlock/isolation state (INTERLOCK_OK, ISO_STATUS).
  • Layer 1 (survivability): prove why the system reset or browned out (UVLO_FLAG, RESET_CAUSE, WDT_CNT).
  • Layer 2 (fidelity): keep measurement validity explicit (CLIP_FLAG, RANGE_ID, VALID_WIN_RATIO).
Outdoor Robustness — Ground Health + Zoning + Isolation Quantify drift • Segment the loop • Make protection observable Fence loop segmentation (concept) Zone A ZONE_ID=A • ZONE_HEALTH LEAK_INDEX DRIFT_24H Zone B ZONE_ID=B • ZONE_BASELINE R_LOOP_EST WET_FLAG Zone C ZONE_ID=C • SEG_DELTA_SCORE RESP_DELAY_BIN SEG_ID Ground health node (observable) GND_HEALTH_NODE trend + score R_LOOP_EST DRIFT_24H Common-mode stress (concept) coupling → validity loss / reset CLIP_FLAG RESET_CAUSE Isolation paths (concept) segment boundaries + bypass/isolation markers ISO_STATUS INTERLOCK Protection priority (observable) Layer 0 — Safety INTERLOCK_OK • ISO_STATUS Layer 1 — Survivability UVLO_FLAG • RESET_CAUSE • WDT_CNT Layer 2 — Measurement fidelity RANGE_ID • VALID_WIN_RATIO • CLIP_FLAG Rule: make robustness measurable (health + drift + validity), not anecdotal
Figure F10. Outdoor robustness is best treated as a measurable, zone-scoped health problem: quantify loop/ground drift, segment the system for isolation, and log protection/validity states so stress events remain explainable.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F10 (Ground Health, Zoning & Cable-Fault Isolation). Link

H2-11. Validation & Field Debug Playbook (Symptom → Evidence → Isolate → Fix)

This chapter is the repeatable “moat”: every symptom is handled with the same template and the same small set of measurable evidence. The goal is fast root-cause separation across environment/ground, measurement chain, and logging/uplink—without relying on any platform-side features.

Template used for every symptom First 2 measurements: (2 waveforms/log fields with exact names) Discriminator: (1–2 rules that route to Environment/Ground vs Measurement vs Reporting) First fix: (one “first-cut” change that improves correctness fast)
Common evidence fields (keep consistent naming)
  • Fence sensing: V_FENCE, I_RETURN, R_LOOP_EST, LEAK_INDEX, GND_DRIFT_24H
  • Validity & range: VALID_WIN_RATIO, CLIP_FLAG, RANGE_ID
  • Event integrity: EVENT_ID (= BOOT_ID + SEQ_ID), EVENT_TS, LOG_COMMIT_TS
  • Uplink: PACKET_SEQ, TX_RETRY_CNT, ACK_LAT_MS, OFFLINE_Q_DEPTH, LAST_TX_FAIL_CAUSE
  • Power/stress: RESET_CAUSE, BROWNOUT_CNT, UVLO_FLAG

Symptom 1 — Frequent false alarms in rain/fog

  • First 2 measurements: LEAK_INDEX trend + GND_DRIFT_24H (or R_LOOP_EST)
  • Discriminator: multi-zone simultaneous drift → environment/ground; single-zone step change → local segment degradation
  • First fix: enable “wet-state gating”: when wet/drift flags rise, freeze baseline updates and raise intrusion confidence threshold while tagging ALARM_SUSPECT_PWR
MPN examples (for implementable evidence & gating)
  • Humidity/temperature sensor: Sensirion SHT31-D or SHTC3 (wet-state hint)
  • Precision temperature sensor for drift compensation: TI TMP117
  • Isolated measurement bitstream/ADC interface (conceptual safety boundary): ADI ADuM7701 (isolated sigma-delta) or TI AMC1311 (isolated amplifier)

Symptom 2 — One zone always reports “low confidence”

  • First 2 measurements: VALID_WIN_RATIO + per-zone baseline stability (ZONE_BASELINE / ZONE_HEALTH)
  • Discriminator: low validity ratio → clipping/range/protection; valid windows but unstable baseline → environment/ground or segment condition
  • First fix: correct dynamic range first (range step/peak-hold policy), then re-learn baseline only with valid windows
MPN examples (range & sampling robustness)
  • Low-noise op-amp for front-end filtering stages: TI OPA320 / OPA197
  • Analog switch for range selection (auto-ranging concept): ADI ADG704
  • Delta-sigma ADC for slow/robust feature capture: TI ADS1220 (common in rugged sensing)

Symptom 3 — Night-time false alarms rise (suspected EMI or supply noise)

  • First 2 measurements: RESET_CAUSE/BROWNOUT_CNT + CLIP_FLAG/RANGE_ID
  • Discriminator: resets correlate with alarms → supply/common-mode stress; no resets but clipping rises → measurement chain susceptibility
  • First fix: make validity observable and conservative at night: log VALID_WIN_RATIO, gate classification when invalid/clipped, then harden supply sequencing
MPN examples (survivability & observability)
  • eFuse / inrush & fault protection: TI TPS25982
  • Supervisor / reset IC: TI TPS3890 (reset/threshold monitoring)
  • EMI common-mode choke (for comm/supply lines where applicable): Würth 744231091

Symptom 4 — Events are missing after outages (dropouts / gaps)

  • First 2 measurements: OFFLINE_Q_DEPTH peak + EVENT_ID continuity (watch BOOT_ID changes)
  • Discriminator: queue overflow with stable boot → buffering/retry policy; frequent boot changes → power integrity or brownout commits
  • First fix: prioritize “event commit before transmit”: write event record first (LOG_COMMIT_TS), then allow retries/acks to be transport-only
MPN examples (durable logging & power-fail tolerance)
  • Non-volatile FRAM (fast, power-fail friendly): Infineon/Cypress FM25V10
  • SPI NOR flash (common log store): Winbond W25Q64JV
  • Power-path / hold-up controller (system-level concept): ADI LTC4040 (backup manager) or simpler “supervisor + energy storage” approach with TI TPS3890

Symptom 5 — Fence output looks “normal”, but cut events are not detected

  • First 2 measurements: V_FENCE (or V-sense) + I_RETURN (or I-sense)
  • Discriminator: V/I show no meaningful delta during known cut → sensing point/clamp/auto-range issue; V/I show delta but not classified → feature/threshold logic issue
  • First fix: fix measurement chain before classification: verify sensing is taken before heavy clamping and that auto-ranging/peak-hold does not compress deltas
MPN examples (sensing chain building blocks)
  • Current sense amplifier with strong PWM/edge rejection: TI INA240
  • Precision op-amp for conditioning/rectification stages: TI OPA197 / OPA320
  • Isolated measurement interface (conceptual boundary): ADI ADuM7701 or TI AMC1311

Symptom 6 — Vibration cable triggers on vehicles (unwanted sensitivity)

  • First 2 measurements: band-energy proxy (BAND_ENERGY_LOW/HIGH) + duration/pattern score (DURATION_MS/PATTERN_SCORE)
  • Discriminator: low-frequency dominant + long duration → ground/vehicle coupling; high-frequency impulsive + consistent patterns → likely intrusion contact signatures
  • First fix: implement band-gated 2-stage trigger: low-frequency long events require higher threshold or second confirmation window; keep false-alarm tagging explicit
MPN examples (AFE + compute for band features)
  • Dual op-amp for active filtering: TI TLV9062 (rail-to-rail, general AFE)
  • Low-power MCU for feature extraction & logging: STM32G0 series (e.g., STM32G071)
  • Delta-sigma ADC for robust envelope/band metrics: TI ADS1220

Symptom 7 — Tamper (enclosure open) does not trigger

  • First 2 measurements: TAMPER_STATE raw input + TAMPER_CNT (debounce / interrupt count)
  • Discriminator: state never changes → physical path/pull-up/isolation issue; state changes but not recorded → debounce/interrupt masking or logging priority issue
  • First fix: treat tamper as a high-priority event: always log a tamper edge with EVENT_ID even under comms backlog
MPN examples (tamper sensing options)
  • Micro switch (simple tamper): Omron D2F series
  • Hall switch (sealed tamper option): TI DRV5032
  • Optocoupler for isolated input paths (when needed): Vishay VO615A

Symptom 8 — Remote comms intermittently drops (after stress or during load)

  • First 2 measurements: LINK_DOWN_CNT/PHY_ERR_CNT + RESET_CAUSE/UVLO_FLAG
  • Discriminator: dropouts coincide with resets → supply/stress; no resets but error counters rise → physical-layer/common-mode margin issue
  • First fix: improve explainability first (log LAST_TX_FAIL_CAUSE, ACK_LAT_MS), then harden interface protection and isolation where needed
MPN examples (comms PHY + protection)
  • Isolated RS-485 transceiver: ADI ADM2587E (integrated isolation) or TI ISO1410
  • RS-485 TVS (line protection): Littelfuse SM712
  • Ethernet PHY: TI DP83826 (industrial 10/100) or Microchip KSZ8081

Symptom 9 — Ground health score “jitters” (unstable day-to-day)

  • First 2 measurements: R_LOOP_EST short-term variance + VALID_WIN_RATIO (are updates based on valid windows?)
  • Discriminator: jitter occurs when validity is low → measurement window problem; jitter persists with valid windows → real environment variability or segment intermittency
  • First fix: freeze/trust gating: update ground-health trend only when validity is high; tag invalid updates explicitly for audit
MPN examples (time + sensing stability)
  • RTC for reliable timestamps: Micro Crystal RV-3028-C7 (ultra-low power)
  • Precision temperature sensor (drift correlation): TI TMP117
  • Current sense amplifier (stable I features): TI INA240
MPN quick list (commonly used building blocks in this playbook)
  • Logging memory: FM25V10 (FRAM), W25Q64JV (SPI NOR)
  • Supervision & protection: TPS3890 (reset), TPS25982 (eFuse)
  • Comms: ADM2587E / ISO1410 (isolated RS-485), SM712 (RS-485 TVS), DP83826 (Ethernet PHY)
  • Sensing chain: INA240 (current sense), OPA197/OPA320 (op-amps), ADS1220 (ADC), ADG704 (range switch)
  • Environment/time: SHT31-D (humidity), TMP117 (temperature), RV-3028-C7 (RTC)
  • Tamper: Omron D2F (switch), DRV5032 (Hall), VO615A (opto)
Field Debug Decision Tree (Symptom → Evidence → Fix) Use 2 measurements per step • Route to the right root-cause bucket START: Symptom observed Reset / brownout? RESET_CAUSE + BROWNOUT Bucket A: Power / Stress First fix: log validity + harden supply MPN: TPS25982 • TPS3890 YES Valid windows OK? VALID_WIN_RATIO + CLIP_FLAG NO Bucket B: Measurement chain First fix: range/clip/peak-hold before thresholds MPN: INA240 • OPA197 • ADG704 ADC: ADS1220 NO Multi-zone drift? GND_DRIFT_24H + LEAK_INDEX YES Bucket C: Ground / Environment First fix: freeze trend on invalid windows MPN: SHT31-D • TMP117 • RV-3028 YES Event integrity OK? EVENT_ID + OFFLINE_Q_DEPTH NO Bucket D: Reporting / Queue First fix: commit-before-transmit + idempotent ACK MPN: FM25V10 • W25Q64JV PHY: ADM2587E • SM712 NO Bucket E: Segment / Signature First fix: band-gated 2-stage trigger MPN: TLV9062 • ADS1220 YES
Figure F11. A minimal decision tree that routes symptoms using two measurements per step: power/reset evidence → window validity/clipping → multi-zone drift vs single-zone → event integrity/uplink queue.
Cite this figure: ICNavigator — Electronic Fence & Vibration Cable, Fig. F11 (Field Debug Decision Tree). Link

Request a Quote

Accepted Formats

pdf, csv, xls, xlsx, zip

Attachment

Drag & drop files here or use the button below.

H2-12. FAQs ×12

Each answer is intentionally short but evidence-based: Measure (2) → Rule → First fix, plus 2–3 example MPNs. (Edge device only; no platform/VMS scope.)

FAQ 01Fence voltage looks normal, but cut events are missed—sensor range or sampling window?

Answer: Most misses come from clipped sensing or an invalid sampling window, not the fence itself. Measure: V_FENCE + I_RETURN, plus VALID_WIN_RATIO/CLIP_FLAG. Rule: if CLIP_FLAG rises or RANGE_ID never changes, it’s range/clamp; if VALID_WIN_RATIO collapses, it’s windowing. Fix: sense pre-clamp, enable auto-ranging, align window. MPN: INA240, ADS1220, OPA197.

See: H2-5 / H2-6
FAQ 02False alarms only on rainy mornings—leakage signature or threshold drift?

Answer: Rain-morning spikes usually look like leakage drift, unless thresholds move with temperature/firmware. Measure: LEAK_INDEX + GND_DRIFT_24H (or R_LOOP_EST). Rule: multi-zone correlated rise → wet leakage/ground; single-zone or post-update shift → threshold drift. Fix: wet-state gating + freeze baseline updates; log humidity/temperature. MPN: SHT31-D, TMP117, FM25V10.

See: H2-3 / H2-6 / H2-10
FAQ 03One zone always reports low confidence—ground health or cable coupling?

Answer: Persistent low confidence is either invalid windows (measurement chain) or real zone coupling/ground issues. Measure: VALID_WIN_RATIO + ZONE_BASELINE variance (or R_LOOP_EST). Rule: low validity → clipping/range; good validity but drifting baseline → coupling/ground. Fix: restore valid windows first, then re-learn baseline for that zone only. MPN: ADG704, ADS1220, TMP117.

See: H2-2 / H2-10
FAQ 04Arc/noise triggers intrusion—how to separate EMI from real contact?

Answer: Arcs/impulse noise becomes false intrusion when the AFE clips or the classifier trusts invalid windows. Measure: CLIP_FLAG + band energy (BAND_ENERGY_HIGH/LOW). Rule: broad clipping + erratic spectra → EMI; repeatable high-band bursts with stable windows → real contact. Fix: validity gating + band-gated 2-stage trigger. MPN: TLV9062, INA240, SM712.

See: H2-6 / H2-10
FAQ 05Vibration cable reacts to nearby traffic—filter band or installation resonance?

Answer: Traffic sensitivity is usually low-frequency resonance and long-duration energy, not “too much gain.” Measure: BAND_ENERGY_LOW ratio + DURATION_MS (or PATTERN_SCORE). Rule: low-band dominant and long duration → installation resonance; high-band impulsive with consistent pattern → intrusion. Fix: retune bandpass, add duration hysteresis/confirmation. MPN: TLV9062, STM32G071, ADS1220.

See: H2-7 / H2-10
FAQ 06Tamper switch works on bench but not in field—wiring or debounce policy?

Answer: Bench OK but field fail is almost always wiring/common-mode or debounce/priority. Measure: raw TAMPER_STATE edges + TAMPER_CNT (and check RESET_CAUSE). Rule: no edges → wiring/pullup path; edges present but no event → debounce/interrupt masking. Fix: log tamper edge as highest-priority EVENT_ID. MPN: Omron D2F, DRV5032, VO615A.

See: H2-3 / H2-11
FAQ 07Events appear out of order after power loss—timestamp source or log commit?

Answer: Out-of-order after power loss is a time source jump or non-atomic log commit. Measure: EVENT_TS vs LOG_COMMIT_TS + BOOT_ID continuity. Rule: BOOT_ID changes around gaps → power-fail commit; stable BOOT_ID but TS jumps → RTC/timebase issue. Fix: commit-before-transmit with monotonic SEQ_ID. MPN: RV-3028-C7, FM25V10, TPS3890.

See: H2-8 / H2-11
FAQ 08Remote alarms arrive duplicated—sequence ID or retry logic?

Answer: Duplicates come from retry semantics unless EVENT_ID isn’t idempotent. Measure: PACKET_SEQ + TX_RETRY_CNT (or DUP_DROP_CNT) tied to EVENT_ID. Rule: same EVENT_ID repeats with rising retries → transport retry; different EVENT_ID for one physical event → generation bug. Fix: de-dup by EVENT_ID, ack idempotently. MPN: FM25V10, W25Q64JV, ADM2587E.

See: H2-9
FAQ 09Communication drops when pulse fires—ground bounce or isolation issue?

Answer: Drops during pulse firing are either UVLO/ground bounce or common-mode margin on the interface. Measure: LINK_DOWN_CNT + UVLO_FLAG/RESET_CAUSE. Rule: drops coincide with UVLO/reset → bounce/supply; no reset but PHY/CRC errors → isolation/common-mode. Fix: decouple comm rail, add isolation + line TVS. MPN: TPS25982, ADM2587E, SM712.

See: H2-4 / H2-9 / H2-10
FAQ 10Good detection in dry weather, poor in humid nights—insulation aging or AFE saturation?

Answer: Dry-good/humid-bad is either insulation aging (slow leakage trend) or AFE saturation (sudden clipping). Measure: LEAK_INDEX trend + CLIP_FLAG/VALID_WIN_RATIO. Rule: gradual LEAK_INDEX rise → aging/leakage; sudden clip spikes at night → saturation/window issue. Fix: wet baseline + range step; freeze updates when invalid. MPN: SHT31-D, TMP117, ADS1220.

See: H2-6 / H2-10
FAQ 11Firmware update changes false-alarm rate—feature extraction or threshold table?

Answer: A firmware update changes false alarms via feature extraction or threshold table. Measure: FEATURE_VER + THRESH_TABLE_VER alongside FALSE_ALARM_CNT. Rule: FEATURE_VER change shifts signatures → extraction; table-only change shifts decision boundary → thresholds. Fix: version-pin features, store signed table, keep rollback counter. MPN: STM32G0, W25Q64JV, ATECC608B.

See: H2-7 / H2-8
FAQ 12How to validate zones quickly on site with minimal tools?

Answer: Fast on-site validation needs two checks: loop health and event integrity. Measure: R_LOOP_EST + EVENT_ID continuity (BOOT_ID+SEQ_ID) while applying a controlled zone stimulus. Rule: stable R_LOOP_EST but missing increments → logging/uplink; drifting R_LOOP_EST → ground/cable. Fix: run zone self-test, commit events locally before transmit. MPN: RV-3028-C7, FM25V10, TPS3890.

See: H2-11
icnavigator

ICNavigator helps engineers find verified ICs, alternatives, and fast quotes.

Explore
Categories
  • Power Management
  • MCU
  • Automotive ICs
  • Sensors
  • Interface & Transceivers
  • Analog Front-End
Get in Touch
  • Email: hello@icnavigator.com
  • WeChat/WhatsApp: +86-xxxx

© 2025 ICNavigator. All rights reserved.