38-EFT.WP.EDX.EMI v1.0 | Chapter 11 — Engineering Rules: Shielding / Grounding / Return Paths (Design Rules.EMI)

Home ／ Docs-Technical WhitePaper ／ 38-EFT.WP.EDX.EMI v1.0

Chapter 11 — Engineering Rules: Shielding / Grounding / Return Paths (Design Rules.EMI)

I. Chapter Objectives & Structure

1. Objective: Convert shielding, grounding, and return-path geometry/process into executable engineering rules and acceptance gates, using the monotonic consistency of Re{ΔZ_rad(omega)}, I_CM(omega), and field envelopes E_rad/H_rad as core controls, so port–path–field are quantifiable, regressable, and falsifiable in emission/immunity contexts.
2. Structure: Symbols & domain → Golden rules → Rule subsets (shielding/returns/connectors/cables/PI) → KPIs & acceptance → Records & QA → Falsifiability → Compliance templates → Cross-chapter closure.
3. Shared time-of-arrival dialect (equivalent; explicit gamma(ell) & d ell; record delta_form):
   • Constant-factored: T_arr = ( 1 / c_ref ) * ( ∫ n_eff d ell )
   • General: T_arr = ( ∫ ( n_eff / c_ref ) d ell )

II. Symbols & Domain

• Ports/normalization: S(omega), Znorm(omega), (if used) T_mm/Z0_mm (DM/CM), Z_eft(omega), Z_c(omega).
• Radiation & fields: ΔZ_rad(omega) (require Re{ΔZ_rad} ≥ 0), P_rad(omega), E_rad(omega,r), H_rad(omega,r), I_CM(omega).
• Shielding geometry: sigma_seam, L_seam, pitch_seam, mesh_ratio.
• Paths/weights: gamma(ell), d ell, {γ_main, γ_side}, w_{p,m}(omega), ΔW.
• Hard QA: check_dim = pass, passivity (Re{Z_eft} ≥ 0), KK_consistency = pass, two-dialect T_arr agreement, and (if enabled) Re{ΔZ_rad} ≥ 0.

III. Golden Rules (EMI)

• Seal-first: prioritize high sigma_seam, long seams/apertures, open returns, and long stubs; reduce Re{ΔZ_rad} and ΔW first.
• Mixed-mode first: normalize DM/CM before back-projecting to SE; must pass passivity/KK after S↔Z mapping.
• Shortest-return closure: at cross-layer/plane transitions, use via bridges/rings to close returns locally and suppress γ_side and I_CM.
• In-window evaluation: fit phase/arrival and evaluate power–field consistency in-window; use energy composition out-of-window—no extrapolations.
• Unified records: update binding_ref/arrival and shielding geometry for all design changes to preserve regression comparability.

IV. Engineering Rules (verifiable)
A. Shielding continuity & seam mitigation

• Rule A1: reduce sigma_seam, shorten L_seam, decrease pitch_seam, increase mesh_ratio; add bridge bars/via fences where needed.
• Acceptance: after mitigation, Re{ΔZ_rad}, |E_rad|/|H_rad|, and I_CM drop together; ΔW decreases.

B. Return paths & bridges

• Rule B1: place ring/grid stitching vias near critical traces and at plane transitions; via pitch ≤ 20× trace height; add via bridges at cuts.
• Acceptance: denser stitching lowers ΔW; I_CM and Re{ΔZ_rad} decrease; phase KPIs (E_phase/GDR) do not worsen.

C. Stubs/vias & back-drilling

• Rule C1: limit stub_len; prefer back-drilling or absorption; label long stubs as γ_side and link to ΔZ_rad.
• Acceptance: after back-drill, w_{side}, Re{ΔZ_rad}, and |E_rad|/|H_rad| decrease together.

D. Connectors/launches & port DM/CM

• Rule D1: use tapered transitions + optimized ground fingers; normalize DM/CM at ports before back-projection; constrain |C_{DM→CM}|.
• Acceptance: I_CM and Re{ΔZ_rad} do not worsen due to the transition; S↔Z loop preserves power/reciprocity and passes passivity/KK.

E. Cables & common-mode returns

• Rule E1: route bundles with paired returns close to critical signals; if needed, add absorbers/ferrites in the ΔZ_rad channel (while keeping Re{Z_eft} ≥ 0).
• Acceptance: improvements show co-decreases in I_CM and Re{ΔZ_rad} and a drop in ΔW.

F. PI/grounding grids

• Rule F1: place near-end decoupling and low-impedance returns close to high-speed loops; avoid large plane voids and neck-downs.
• Acceptance: Re{ΔZ_rad}, I_CM, and ΔW meet gates; no new leakage paths appear.

V. KPIs & Acceptance Gates (aligned with M20-EMI)

• Engineering: mean(Re{ΔZ_rad}) decreases after sealing; corr(P_rad, E/H) ≥ ρ_gate (suggest ≥ 0.8); I_CM and Re{ΔZ_rad} monotonic; ΔW ≤ ΔW_gate.
• Release: engineering-level plus passivity/KK = pass, two-dialect T_arr agreement; (if applicable) |C_{DM→CM}| under gate; dataset card complete.

VI. Records & QA (aligned with data/pipeline cards)

• Required: binding_ref, arrival{form,gamma,measure,c_ref,Tarr,u_Tarr,delta_form}, shielding geometry {sigma_seam,L_seam,pitch_seam,mesh_ratio}, Znorm(ω), (if used) T_mm/Z0_mm, Z_eft/argZ, ΔZ_rad, I_port/I_CM, field calibration & distance E_rad/H_rad, ΔW, qa_gates{check_dim,passivity,KK}.
• Phase correction: arg Z_corr(ω) = arg Z_raw(ω) − ( ω · Δt_sync ) , then unwrap before fitting.

VII. Falsifiability (for Design Rules.EMI)

• J-DR-EMI-1 (Sealing verification): if Re{ΔZ_rad} and |E|/|H|/I_CM do not drop together post-sealing/meshing, reject seam equivalence or geometry records.
• J-DR-EMI-2 (Return bridging): if ΔW does not drop or KPIs worsen after stitching/bridging, reject return strategy or path annotations.
• J-DR-EMI-3 (Back-drill verification): if w_{side} and Re{ΔZ_rad} do not decrease post back-drill, reject stub parameters or execution.
• J-DR-EMI-4 (Mixed-mode normalization): if post-normalization/back-projection fails passivity/K–K, reject port basis or normalization config.
• J-DR-EMI-5 (Two-dialect agreement): if |T_arr^{(1)}−T_arr^{(2)}| > u(T_arr), reject release.

VIII. Compliance Templates (copy-ready)

• Design acceptance card (YAML)
• design_rules_emi:
• binding_ref: "LAY2PATH-EMI-0001"
• arrival:
• form: "n_over_c" # or "one_over_c_times_n"
• gamma: "explicit"
• measure: "d_ell"
• c_ref: 299792458.0
• Tarr_s: 1.234e-09
• u_Tarr_s: 6.0e-12
• delta_form: "n_over_c"
• shielding:
• seams: {sigma_seam_S: 0.12, L_seam_mm: 18.0, pitch_seam_mm: 3.0, mesh_ratio: 0.6}
• bridges:{via_fence: true, pitch_mm: 2.0}
• returns:
• stitch_vias: {ring: true, pitch_mm: 2.0}
• backdrill: {enabled: true, max_stub_mm: 0.5}
• connectors:
• launch: {tapered: true, ground_fingers: "optimized"}
• modal_coupling_limit: {abs_C_DM2CM_max: 0.2}
• cables:
• pairing: {cm_return_paired: true}
• absorbers: {enabled: false}
• targets:
• Re_Zrad_drop: {required: true}
• rho_PRad_E: {gate: 0.8}
• rho_PRad_H: {gate: 0.8}
• ΔW_gate: 0.20
• qa_gates: ["check_dim","passivity(Re{Z}≥0)","KK_consistency"]
• Acceptance script (pseudocode)
• # Positive-real / causality
• assert min(Re(ΔZ_rad)) >= 0.0 and KK_consistency(Z_eft)
• # Pre-vs-post sealing/bridging/back-drill
• assert mean(Re(ΔZ_rad_after)) < mean(Re(ΔZ_rad_before))
• assert mean(E_after) < mean(E_before) and mean(I_CM_after) < mean(I_CM_before)
• # Mixed-mode loop
• assert passivity(Z_eft) and reciprocity_preserved(S_se, T_mm, Z0_mm)
• # ΔW
• ΔW = sum_abs(w[p,m](ω2) - w[p,m](ω1))
• assert ΔW <= ΔW_gate and monotone(ΔW, Re(ΔZ_rad))
• # Two-dialect arrival
• assert abs(Tarr_n_over_c - Tarr_one_over_c_times_n) <= u_Tarr

IX. Cross-Chapter Links & Closure

• Dependencies: Chapter 4 (Min. Equations & Equivalents), Chapter 5 (Seams/Apertures/Stubs), Chapter 6 (Cables/Connectors & Mixed-Mode), Chapter 8 (S↔Z & Field Mapping), Chapter 9 (Metrology Chain), Chapter 10 (Experimental Design & Falsification).
• Downstream: Chapter 14 (SimStack & Benchmark Cases — regression & layout/process review), Chapter 16 (Design Protocol & Checklist — integrate these gates into sign-off).