123 Main Street, New York, NY 10001

← Back to: eFuse / Hot-Swap / OR-ing Protection

Definition & Scope

Port-level protection for Ethernet focuses on differential ESD/EFT at the pair and common-mode lightning surge that must return to chassis. Physical stack: RJ-45 → Magnetics (Transformer + CMC) → CT → Chassis → PHY. Residual voltage/energy defines the upper boundary for downstream eFuse/Hot-Swap/OR-ing.

Boundaries (Anti-overlap)

  • No PHY registers/protocol flows here.
  • No PoE classification/power budget math; we only state constraints so protection does not break signature/classification.
  • No energy metering or long logs (those live in the Telemetry sibling page).

Threat Mapping

  • Differential path: IEC 61000-4-2 ESD / 4-4 EFT couple at the pair → impacts eye/BER/link stability.
  • Common-mode path: external lightning/GR-1089 couples to cable shield/pairs → returns to chassis through CT clamp and surge stages.
  • PoE constraint: leakage in the ~57 V window must not spoof signature/classification.

Interface to eFuse/Hot-Swap/OR-ing

This page outputs residual limits (VRESIDUAL, energy, duration) and a fault semantic (active-low FAULT / active-high PG). Downstream power path shall survive the residual and latch/report per the unified polarity.

Key Parameters

  • Differential: Cdiff (array effective), VBR, TLP peak, Rdyn, package L/C.
  • Common-mode: VRESIDUAL, ISURGE(8/20), Lloop (return inductance), leakage at ~57 V (PoE safe).
  • Placement: array-to-RJ-45 distance ≤ X mm; array→chassis return loop ≤ Ltarget.

Validation Hooks

  • Pass/fail wording for IEC 61000-4-2/4-4/4-5 and GR-1089 (link recovery time, residual peak, no PHY latch-up).
  • Unified PG/FAULT polarity and debounce window for event logging.
Ethernet port stack and threat paths RJ-45 to Magnetics+CMC to CT to Chassis to PHY; red = differential ESD/EFT near pair; blue = common-mode surge to chassis; residual defines downstream limits. RJ-45 MagneticsTransformer + CMC CT PHY Differential ESD/EFT Lightning Chassis return (minimize Lloop) Residual to chassis defines limits for downstream power path
F1 — Overview: differential threats at the pair; common-mode surge flows to chassis via CT. Residual sets the downstream limit.

Threat Model: ESD / EFT / Surge / GR-1089

Event Sources & Waveforms

  • IEC 61000-4-2 ESD: ±8 kV contact / ±15 kV air (typical system levels); fast edge, low energy but high dv/dt.
  • IEC 61000-4-4 EFT: burst trains with ns rise; stresses routing discontinuities and return paths.
  • IEC 61000-4-5 Surge: 1.2/50-8/20; for Ethernet, common-mode to chassis is typical (e.g., 1 kV CM).
  • GR-1089 external line lightning tests (telecom/operator grade) — higher energy CM events.

Differential vs Common-Mode Propagation

  • Differential (near-end injection): treat with TLP/TDR equivalence; watch eye/BER, Sdd21, bias stability.
  • Common-mode (far-end coupling): energy couples to cable; correct path is CT→chassis with minimal Lloop.

PoE-Specific Constraint

Protection must not break IEEE detection/classification. Control leakage near ~57 V and keep signature Rsig intact. CT clamp path must bypass the signature network.

Metrics & Logging

  • Record LINK_RECOVERY_MS, frame error rate delta, PHY auto-recovery (no manual reset needed).
  • Thermal hotspot map during surge to guide via fences and spring placement.
Differential vs common-mode threat paths Red = differential ESD/EFT at the pair; Blue = common-mode surge flowing to chassis with minimized loop inductance. RJ-45 PHY ESD/EFT (diff) Chassis return — keep Lloop low Measure residual & recovery; keep PoE signature intact
F2 — Red path affects eye/BER; blue path must dump to chassis with minimal loop. Use these paths to plan tests and limits.

Differential Pair ESD Topologies

Deliver a minimal, repeatable recipe: select ultra-low-capacitance arrays, place them within X mm of the RJ-45 pads, and provide a low-inductance return to chassis. Reduce residual without degrading the eye; for substitutions, satisfy Cdiff and VBR before Ipp.

Selection

  • Cdiff target: ≤ 0.2–0.5 pF (speed-tiered; use the lower bound for 5G/10G).
  • Secondary metrics: VBR (breakdown/clamp), TLP peak (pulse capability), Rdyn (lower is better).
  • Unidirectional vs bidirectional: default to bidirectional; only use unidirectional when DC bias and polarity are guaranteed safe.
  • Package parasitics: prefer short-lead, symmetric arrays; account for Lpkg/Cpkg in bandwidth budgeting.

Placement & Return

  • Distance: array to RJ-45 pad centerline ≤ X mm; array→chassis loop inductance ≤ Ltarget.
  • Return path: nearby chassis vias/spring fingers; avoid digital-ground return to prevent ground bounce.
  • Via strategy: 2–3 parallel vias from array to chassis; bridge any splits with springs/straps to reduce Lloop.

π vs T Topology

  • T (default): better bandwidth/eye preservation; use when margin is tight.
  • π: lowest residual but adds more effective C; use only when high-speed margin allows.
  • Compare by eye/BER and TLP clamp-slope; do not decide by Ipp alone.

Failure Modes & Replacement Rules

  • Clamp miss (VBR drift), leakage rise (PoE signature impact), package open/detach.
  • A→A / A→B: keep Cdiff and VBR first, then Rdyn, then Ipp. Re-validate eye + IEC 61000-4-2.

Key Parameters

Cdiff, VBR, TLP peak, Rdyn, Lpkg/Cpkg, layout thresholds (≤ X mm, loop ≤ Ltarget).

Validation Hooks

  • IEC 61000-4-2 injection + eye/BER comparison; record LINK_RECOVERY_MS, BER_delta, residual.
  • S-parameters: Sdd21 (bandwidth/insertion loss) and Sdc (bias integrity).
Low-cap ESD arrays near RJ-45 with short chassis return Arrays within X mm; short return to chassis; arrows show low-inductance path; tiny inset compares T vs π trade-off. RJ-45 PHY X mm Chassis
F3 — Place arrays ≤ X mm from RJ-45; provide low-L chassis return; pick T or π per bandwidth vs residual.

Common-Mode Lightning Suppression

Use a GDT/TSS → low-Rdyn TVS cascade to shed energy then limit residual, and a short CT→chassis clamp that bypasses PoE signature. Targets: low VRESIDUAL, limited PCB energy, PoE-safe leakage.

Cascade & CT-Clamp

  • GDT/TSS → TVS: front stage carries I8/20, TVS sets residual and slope (low Rdyn).
  • CT clamp: dedicated short path to chassis away from the IEEE signature/classification network; control leakage around 57 V.

Loop Inductance & CMC Order

  • Minimize Lloop: straight return to chassis, via fences, springs/straps across slots.
  • CMC helps CM→DM rejection, but surge energy must exit via GDT/TSS + TVS + CT-clamp first.

Targets & Metrics

  • VRESIDUAL below magnetics/PHY withstand with ≥30% margin.
  • Design for I8/20, Ppp, low Rdyn; verify leakage near PoE 57 V window.

Validation Hooks

  • IEC 61000-4-5 / GR-1089: measure residual waveform/peak, thermal hotspots, link recovery time.
  • Chassis connection impedance scan (near-field + inductance estimate); event telemetry flag with unified polarity.
GDT/TSS + TVS cascade with CT clamp to chassis Common-mode surge path to chassis; CT clamp bypasses PoE signature; labels indicate loop and residual goals. RJ-45 PHY Chassis PoE signature Vresidual Minimize Lloop
F4 — GDT/TSS sheds energy → TVS limits residual; CT→chassis path bypasses PoE signature; aim for low Vresidual and low Lloop.

Layout Rules & Return Paths

Turn “short • straight • wide • decoupled” into executable rules for Ethernet port protection. Give numeric thresholds and probe standards so the layout can be reviewed and reproduced.

Differential ESD Arrays

  • Placement: array center to RJ-45 pin centerline ≤ X = 2–4 mm (1G ≤3 mm; 10G ≤2 mm).
  • Symmetry: matched pads/vias; avoid stubs ≥ 0.5 mm; length-compensate paired vias.
  • Return: array → chassis direct; do not use digital GND flood; keep path straight.
  • Loop inductance target: Ltarget1.5 nH (10G ≤ 1.0 nH).
  • Via bundle: parallel chassis vias N ≥ 3 (Ø 0.30–0.35 mm; pitch 0.6–1.0 mm) adjacent to array.
  • Copper width: return copper / neck width ≥ 0.8–1.2 mm.

Common-Mode Surge Path (GDT/TVS → Chassis)

  • Cascade: GDT/TSS (energy) → low-Rdyn TVS (residual) → chassis.
  • Via fence: along GDT/TVS-to-chassis copper with fence pitch ≤ 2–3 mm.
  • Across splits: use chassis springs/straps. If a capacitor is required, single-point Climit ≤ 100 pF (SRF ≥ 300 MHz), far from the pair.
  • Residual target: internal node VRESIDUAL80–120 V with ≥30% margin to magnetics/PHY withstand.

CT-Clamp & PoE Constraints

  • CT clamp route to chassis is isolated from IEEE detection/classification network.
  • Leakage near the 57 V PoE window: Ileak100 µA (prefer ≤ 50 µA).

CMC Position vs ESD/TVS

Prefer RJ-45 → Magnetics/CMC → ESD (diff) / CM suppression → CT → chassis. Let surge energy exit to chassis before crossing the CMC.

Probe & Measurement Points

  • Provide SMA/probe pads near arrays and near PHY (both sides of the channel).
  • Leave a 10–15 mm clamp region on the GDT/TVS-to-chassis copper for current probe.
  • Place 0603 thermistor/TC pads next to TVS/GDT for ΔT logging.

Quick math · Residual ≈ Lloop·(di/dt) + Rdyn·I; use IEC-4-2 edge di/dt ≈ 20–40 A/ns. Added capacitance per line ΔCtotal ≈ Cpkg + Cstub; target ≤ 0.2 pF (1G) with stricter budget for 10G.

Layout Checklist (copy to review)

  • X (array→RJ-45) ≤ 3 mm (10G ≤ 2 mm)
  • N (parallel chassis vias) ≥ 3; fence pitch ≤ 3 mm
  • Ltarget ≤ 1.5 nH (10G ≤ 1.0 nH)
  • Return copper width ≥ 0.8–1.2 mm; route straight
  • Climit (across split) ≤ 100 pF, SRF ≥ 300 MHz
  • CT clamp isolated from PoE signature path; Ileak ≤ 100 µA@57 V
  • Residual goal VRESIDUAL ≤ 80–120 V
  • SMA/probe pads + current-clamp window + ΔT pads present
Placement & return rules with via fences and chassis bridges Top-view PCB sketch showing ESD arrays near RJ-45, direct low-L return to chassis with via fences, chassis bridge across split, and standardized probe points. RJ-45 ESD X mm Direct to chassis Chassis bridge SMA Clamp ΔT
F5 — Arrays close to RJ-45, straight low-inductance return with via fences, chassis bridge across splits, and standardized probe points.

Compliance & Validation Playbook

Reusable test steps for IEC 61000-4-2/-4/-5 and GR-1089 with PoE compatibility. Each block defines inputs, injection points, records, and pass/fail criteria plus unified log keys.

Log schema: PORT_ESD_EVENT, PORT_EFT_EVENT, PORT_SURGE_EVENT, PORT_POE_EVENT, LINK_RECOVERY_MS, V_RESIDUAL_VPK, HOTSPOT_DT_MAX, FAULT_POLARITY_OK.

ESD — IEC 61000-4-2

  • Levels: ±8 kV contact / ±15 kV air (option ±10/±20).
  • Matrix: RJ-45 shell, grouped pins (to chassis), near-array pads; 10 shots/polarity/point.
  • Record: LINK_RECOVERY_MS, BER/eye delta, FAULT polarity.
  • Pass: no PHY latch; LINK_RECOVERY_MS ≤ 1 s; leakage stable.

EFT — IEC 61000-4-4

  • Setup: coupling clamp 1.0 kV; 5 kHz/100 kHz, 60 s each; cable & enclosure paths.
  • Record: FRAME_ERR_RATE; target Perr ≤ 10−6 (project-specific).
  • Pass: no link drop; PoE not falsely interrupted; FAULT_POLARITY_OK = true.

Surge — IEC 61000-4-5 (Common-Mode to Chassis)

  • Wave: 1.2/50-8/20; typical 1 kV CM (extendable to 2 kV); 5 shots/polarity.
  • Return: coupled to cable, return via chassis using the Ch.5 low-L path.
  • Record: V_RESIDUAL_VPK, HOTSPOT_DT_MAX at TVS/GDT, LINK_RECOVERY_MS.
  • Pass: V_RESIDUAL ≤ 80–120 V; ΔT ≤ 60 °C (or within SOA); recovery ≤ 2 s.

GR-1089 (If Applicable)

  • External line lightning CM levels per carrier spec; wiring between outside plant and chassis as specified.
  • Pass includes no permanent FER increase plus Surge criteria above.

PoE Compatibility

  • IEEE detection 2.7–10 V and classification 15.5–20.5 V succeed continuously ≥ 20 cycles.
  • Leakage near 57 V within Ileak ≤ 100 µA; CT clamp path bypasses signature network.
ESD/EFT/Surge/GR-1089/PoE test setups Four mini-scenes: ESD gun to RJ-45, EFT clamp around cable, Surge CM injection/return to chassis with probes, and PoE signature test. ESD (IEC 61000-4-2) RJ-45 EFT (IEC 61000-4-4) Coupling clamp Surge (IEC 61000-4-5) Chassis return TVS GDT/TSS Clamp PoE Signature/Class Signature net CT clamp (isolated)
F6 — Standard setups: ESD gun points, EFT clamp on cable, Surge CM injection with chassis return & probes, and PoE signature/class test.

Telemetry & PG/FAULT Semantics

Map port-level events to a unified PG/FAULT bus with consistent polarity, deglitch, timestamps, and cloud fields. Prevent semantic drift when parts are substituted across brands, and define how port events drive downstream eFuse/Hot-Swap/OR-ing behaviors.

Event Set & Severity

  • Events: PORT_ESD_EVENT, PORT_EFT_EVENT, PORT_SURGE_EVENT, PORT_LINK_DOWN, PORT_LINK_RECOVERY.
  • Severity tiers: INFO (transient), WARN (limited protection engaged), SEVERE (needs lock/inspection).

Polarity Unification

  • FAULT = active-low, PG = active-high.
  • If a device uses opposite polarity: (1) board-level inverter or MCU/FPGA logic flip; (2) update polarity_map in cloud; (3) verify once in NVM self-test.

Timestamps & Deglitching

  • Deglitch tdg: 2–5 ms (ESD/EFT), 5–20 ms (Surge/Link).
  • Minimum event width tmin_event: 1–2 ms. Report window report_window_ms: 1000 ms aggregation.
  • Log keys: ts_ms, port_id, event, severity, pg, fault, optional v_residual_vpk, link_recovery_ms.

Interface to eFuse / Hot-Swap / OR-ing

  • PORT_SURGE_EVENT(severe) → limit current / switch to preferred source / brief disconnect (holdoff_ms=200–500).
  • PORT_LINK_DOWN(persist > t_hold) → reduce port power, or switch power source.
  • EVENT_BURST_RATE > R_max → raise protection tier; assert PG=0, FAULT=1 and log.

Key Parameters (defaults)

  • Input thresholds (3V3): V_il ≤ 0.8 V, V_ih ≥ 2.0 V; pull-ups 10–47 kΩ.
  • Deglitch: 2–5 ms (ESD/EFT), 5–20 ms (Surge/Link). Report window 1000 ms.
  • Error rate gate: FER_thresh ≤ 1e-6. Auto-rearm: 1000–3000 ms (≤ 5 s max).

Example log payload

{
  "port_id": "ETH0",
  "event": "PORT_SURGE_EVENT",
  "severity": "WARN",
  "ts_ms": 1730712345678,
  "v_residual_vpk": 96.4,
  "link_recovery_ms": 420,
  "pg": 1, "fault": 0,
  "polarity_map": {"pg":"active_high","fault":"active_low"},
  "deglitch_ms": 8, "burst_rate": 3
}
      
Port → PG/FAULT bus → MCU/FPGA → Cloud Mapping ESD/Surge/Link events to unified PG↑/FAULT↓ with deglitch and polarity rules, then to cloud analytics. Ethernet Port PG/FAULT Bus PG↑ FAULT↓ MCU / FPGA t_dg, polarity map Cloud PORT_* schema ESD · Surge · Link PG↑ FAULT↓ t_dg report_window 1000 ms
F7 — Port events are normalized to PG/FAULT (PG↑, FAULT↓) with deglitch and polarity mapping, then logged to cloud fields.

BOM Remarks & Procurement Hooks

Lock guardrails into the BOM so small-batch substitutions won’t break Ethernet port protection. Four windows to control: Cdiff, VRESIDUAL, Lloop, and PoE leakage. Define A→A / A→B / A→C paths and the telemetry semantics that must remain stable.

Copy-Paste BOM Notes (fill placeholders)

Differential ESD

“Use ultra-low-C arrays: Cdiff ≤ {X_pF}, VBR ≥ {Y_V}; place ≤ {X_mm} mm from RJ-45; direct chassis return via via-fence (pitch ≤ 3 mm).”

Common-Mode Surge

“Cascade GDT/TSS + low-Rdyn TVS; design for Lloop ≤ {L_max_nH}; ensure internal node VRESIDUAL ≤ {V_target_V}.”

CT-Clamp / PoE

“CT clamp path is isolated from PoE signature/classification; leakage at 57 V ≤ {I_leak_max_uA} μA; must not alter Rsig.”

Telemetry

“Expose PG (active-high) / FAULT (active-low); deglitch {t_dg_ms} ms; report window {report_ms} ms; cloud fields PORT_* present.”

Default Windows (recommended)

  • Cdiff ≤ 0.2–0.5 pF (1G ≤ 0.3 pF; 10G ≤ 0.2 pF).
  • VRESIDUAL ≤ 80–120 V (internal nodes).
  • Lloop ≤ 1.5 nH (10G ≤ 1.0 nH).
  • Ileak@57 V ≤ 100 μA (prefer ≤ 50 μA).
  • X_mm ≤ 3 mm (10G ≤ 2 mm) from RJ-45 to array center.

Receiving / Replacement Checklist

  • Datasheet revision locked; package and pad pitch compatible.
  • Cdiff ≤ X, VBR ≥ Y; TVS Rdyn and Ppp adequate.
  • Leakage @57 V ≤ Ileak_max; CT clamp not tied to signature net.
  • Via-fence geometry (pitch/diameter/distance to device) implemented.
  • PG/FAULT polarity and tdg match template; cloud mapping verified.
  • Bring-up proves V_residual_vpk, LINK_RECOVERY_MS, and ΔT within limits.

Substitution Paths

  • A→A (same brand/series): pin-compatible; first lock Cdiff / VBR, then verify TVS/GDT Ipp/Ppp; keep X_mm target.
  • A→B (cross-brand): match Cdiff / leakage / VRESIDUAL window; re-run IEC 61000-4-5 CM; update polarity_map if needed.
  • A→C (de-feature, only when no PoE): allow slightly larger Cdiff or omit CT-Clamp; mark “limited scope” and attach a re-test job.
BOM guardrails: windows & substitution paths Four parameter windows (Cdiff, Vresidual, Lloop, PoE leakage) and three substitution paths A→A, A→B, A→C. Cdiff ≤ 0.2–0.5 pF Vresidual ≤ 80–120 V Lloop ≤ 1.5 nH PoE leakage ≤ 100 μA @57 V A → A A → B A → C (no PoE)
F8 — Guardrails: four numeric windows + three substitution paths. Keep these constraints when buying or replacing parts.

Brand Mapping (TI / ST / NXP / Renesas / onsemi / Microchip / Melexis)

Capability-level mapping for Ethernet port protection. Focus on differential ultra-low-C ESD/TVS families and PHY-side constraints. Common-mode surge parts are usually third-party (use numeric windows). Melexis: no direct Ethernet port protection/PHY line.

Differential ESD/TVS Families — Representative Part Numbers

  • TI (TPD family): TPD1E05U06, TPD2E2U06, TPD4E05U06, TPD1E10B06.
  • ST (ESDA/ESDALC/ECMF): ESDALC6V1-1, ESDALC6V1-2, ESDALC6V1-5M2, ECMF02-2AMX6 (verify bandwidth).
  • NXP (Nexperia) (PESD/TrEOS): PESD5V0S1UL, PESD5V0R1BSF, PESD5V0R2B, PESD2ETH1.
  • onsemi (ESD/low-C): ESD8002, ESD7002, ESD9B5.0ST5G, ESD11B5.0ST5G.
  • Renesas: typically pair with third-party ultra-low-C arrays; follow windows below.
  • Microchip: focus on PHY/Sw; use above low-C arrays with windows below.
  • Melexis: — (no direct Ethernet port ESD/TVS line).

PHY-Side Constraints — Families & Examples

  • Microchip KSZ* (e.g., KSZ9031RNX – RGMII GbE; KSZ9131MNX – GbE; KSZ8081MNX – 10/100): keep added C per line ≤ 0.2–0.3 pF (GbE); arrays ≤ 2–3 mm from RJ-45; direct low-L chassis return.
  • TI DP83* (DP83867IR – GbE; DP83822I – 10/100; DP83848I – 10/100): symmetric placement; ensure PoE leakage window when applicable.
  • NXP TJA110x (100BASE-T1, automotive single pair, not RJ-45): different magnetics/CT topology; follow auto-Ethernet rules (this page maps capabilities only).
  • Renesas RZ/N (e.g., RZ/N2L, RZ/N1D) with external PHY: apply chosen PHY’s low-C/placement constraints.
  • onsemi / ST: when using their PHY/SerDes, apply the same low-C and low-L rules.
  • Melexis: — (no direct Ethernet PHY line).

Numeric windows to enforce across brands: Cdiff ≤ 0.2–0.5 pF (GbE ≤ 0.3 pF; 10G ≤ 0.2 pF), VRESIDUAL ≤ 80–120 V (internal nodes), Lloop ≤ 1.5 nH (10G ≤ 1.0 nH), PoE leakage @57 V ≤ 100 μA (prefer ≤ 50 μA).

Brand × Capability Matrix Seven-brand capability mapping for Ethernet port protection: Diff ESD/TVS, PHY-side constraints, CM windows. Brand Diff ESD/TVS PHY constraints CM windows TI ST NXP onsemi Renesas Microchip Melexis TPD (✓) ESDA/ESDALC (✓) PESD / TrEOS (✓) ESD8xx / ESD7xx (✓) 3rd-party low-C (→) 3rd-party low-C (→) DP83* families — / follow PHY used TJA110x: 100BASE-T1 — / follow PHY used RZ/N + ext PHY KSZ* families use windows use windows use windows use windows use windows use windows Note: TJA110x = 100BASE-T1 (not RJ-45). Melexis — no direct Ethernet port/PHY line.
F9 — Capability mapping across the seven brands for Ethernet port protection (families & constraints).

A→A/A→B: lock Cdiff and VBR first, then compare Rdyn/Ipp. Maintain ≤ 2–3 mm to RJ-45 and a low-inductance chassis return.

Worked Examples

Two copy-ready port-level baselines: 1G non-PoE and 1G PoE-PD. Include numeric windows and validation cues. Use Ch.9 families/PNs for A→A/A→B selection.

Example A — 1G non-PoE (RJ-45)

  • Diff-ESD: Cdiff ≤ 0.3 pF; π-network or single array (e.g., TI TPD2E2U06; NXP PESD2ETH1; onsemi ESD8002; ST ESDALC6V1-2).
  • CM surge: GDT/TSS + low-Rdyn TVS to chassis; VRESIDUAL ≤ 600 V (node example; finalize per product spec).
  • Layout: ≤ 2–3 mm from RJ-45; via-fence to chassis; use chassis tabs across splits.
  • Validation: IEC 61000-4-2 (±8 kV contact), -4-4 (burst Perr), -4-5 (CM surge, residual & ΔT), link auto-recovery ≤ Trec.

Example B — 1G PoE-PD (RJ-45)

  • Signature/classification: CT-clamp must not interfere with the PoE signature path; leakage @57 V ≤ 100 μA (prefer ≤ 50 μA).
  • Diff-ESD: tighter Cdiff (≤ 0.25–0.3 pF) with symmetric placement; same low-L chassis return.
  • CM surge: same cascade; ensure PD front-end SOA/energy limiting.
  • Validation (extra): PoE signature/classification/power-up robustness; GR-1089 external line (if applicable).
Worked Examples: 1G non-PoE and 1G PoE-PD Two baseline port setups with numeric windows and validation cues. 1G non-PoE Cdiff ≤ 0.3 pF (π or single array) RJ-45 distance ≤ 2–3 mm Vresidual ≤ 600 V (node) Via-fence → chassis, low-L Validation: 4-2 / 4-4 / 4-5 1G PoE-PD Cdiff ≤ 0.25–0.3 pF CT-clamp not in signature path Leakage @57 V ≤ 100 μA Surge cascade + PD SOA Validation: add PoE + GR-1089 IEC 61000-4-2 IEC 61000-4-4 IEC 61000-4-5 PoE Sign/Class GR-1089
F10 — Two baseline Ethernet port setups: non-PoE and PoE-PD with key windows and validation cues.

Request a Quote

Accepted Formats

pdf, csv, xls, xlsx, zip

Attachment

Drag & drop files here or use the button below.

FAQ

Why do designs pass IEC 61000-4-2 yet fail 61000-4-5 on Ethernet?

IEC 61000-4-2 is a short, low-energy ESD strike; IEC 61000-4-5 injects far more energy as a surge. Designs without a GDT/TSS plus low-Rdyn TVS cascade and a low-inductance chassis path often meet 4-2 but miss 4-5. Reduce loop inductance (≤ 1.5 nH), use via-fences to chassis, and target acceptable residual voltage.

How close must the differential ESD array be to the RJ-45 pins?

For GbE, place ultra-low-capacitance arrays within 2–3 mm of the RJ-45 pins, symmetrically per pair. Return to chassis through a direct, low-L path rather than a generic ground flood. Keep stubs short, lengths matched, and do not route protection currents under the PHY. Validate with eye-diagram and BER under ESD injection.

Can a single TVS handle common-mode lightning events by itself?

Not reliably. Common-mode lightning surges require a two-stage approach: a GDT or TSS to divert bulk energy to chassis, followed by a low-Rdyn TVS to clamp residuals. A lone TVS often overheats or leaves excessive residual voltage. Use short, wide returns, and verify with IEC 61000-4-5 and operator GR-1089 where applicable.

What does a GDT+TVS cascade reduce in practice—energy or residual voltage?

Both, in sequence. The GDT/TSS conducts first and shunts the majority of surge energy into the chassis, while the TVS—with low dynamic resistance—controls the remaining residual voltage and dV/dt. This combination lowers component stress on magnetics and PHY. Confirm with node probes and thermal imaging during surge testing.

How do I minimize residuals—shorter path or bigger TVS?

Start with geometry: shorten and widen the path to chassis, add via-fences, and avoid slot crossings to reduce inductance. Then size the TVS for adequate Ipp and low Rdyn. Increasing TVS size alone cannot compensate for a high-inductance return. Measure residual voltage at defined nodes during 4-5 testing to confirm improvements.

Will CT clamping break IEEE 802.3 signature/classification for PoE?

It can, if the clamp path intercepts the PoE signature branch. Route CT-to-chassis so it bypasses the signature and classification network. Specify acceptable leakage around 57 V (≤ 100 μA, preferably ≤ 50 μA) in the BOM. During validation, re-run signature, classification, and power-up sequences with surge applied.

How do I pick low-cap ESD arrays without killing the eye-diagram?

Choose arrays with Cdiff in the 0.25–0.3 pF range for GbE (tighter for 10G), adequate VBR, and low Rdyn. Place them within 2–3 mm of the connector with symmetrical routing and minimal stubs. Validate with S-parameters (Sdd21) and eye-diagram under injected ESD to ensure margin remains acceptable.

Should the CMC be before or after the clamp network for best immunity?

A common, robust order is RJ-45 → ESD/clamp network → CMC → PHY. Clamps work closest to the connector for current shunting; the CMC then attenuates remaining common-mode noise. Keep returns low-L and avoid routing surge currents through the CMC. Verify by comparing link stability during 4-5 with both orders.

Why do links drop only during nearby lightning but not in bench tests?

Nearby lightning couples large common-mode energy into long cables and chassis loops not replicated on benches. Improve chassis bonding, add via-fences, and refine clamp placement. Reproduce with external line coupling and long-cable setups. Log link-down duration and recovery time to confirm improvements match field behavior.

How do I log “PORT_SURGE_EVENT” on the PG/FAULT bus with correct polarity?

Standardize polarity as PG active-high and FAULT active-low. Debounce with t_dg ≥ 5–20 ms so brief spikes do not raise events. Buffer timestamps and report windows (e.g., 1000 ms) to the host. Map fields one-to-one in firmware and cloud analytics so cross-brand replacements preserve semantics without code changes.

What leakage is acceptable around 57 V for PoE during surge tests?

Specify leakage ≤ 100 μA at 57 V, with a preferred target ≤ 50 μA to avoid disturbing signature/classification. Add this window to the BOM remarks and vendor checks. During IEC 61000-4-5 runs, confirm signature, classification, and power-up remain stable with the clamp network installed and stressed.

How do I A→B replace an ESD array across brands without re-certifying everything?

Match the windows first: Cdiff, VBR, Rdyn, Ipp, and package parasitics. Keep placement and return geometry unchanged. Perform a minimal regression: eye/BER under ESD, IEC 61000-4-4 bursts, and key-node IEC 61000-4-5 surges. If residual voltage and link recovery match prior results, broader re-certification is often unnecessary.