PCB-Level EMI Shielding: Shield Can Design, Via Stitching, and Grounding Strategies for RF Compliance
Engineering guide covering shield can selection, via stitching pitch rules, ground plane continuity, and reflow soldering profiles for CISPR 25 and IEC 61000-4-3 compliance.
Executive Summary
Radiated emissions failures at the PCB level remain the most expensive compliance failure mode in consumer electronics and automotive ECU development — a single re-spin costs $50K–$250K and delays certification by 8–16 weeks. The root cause in the majority of cases is not the IC itself but the shielding implementation: inadequate ground return paths beneath shield cans, via stitching pitched too wide to contain harmonics above 2 GHz, and shield-to-board contact impedance degraded by poor solder joint design. This application note addresses these failure modes with quantified design rules mapped to CISPR 25 Class 5, IEC 61000-4-3 (10 V/m, 80 MHz–6 GHz), and MIL-STD-461G RE102. POCONS USA's two-piece shield can assemblies and BeCu spring contacts are engineered to close the mechanical and electrical gaps that defeat PCB-level shielding.
Technical Specifications & Attenuation Data
Shield effectiveness (SE) is the sum of reflection loss, absorption loss, and correction factors for apertures and seams. A well-implemented single-wall shield can on a four-layer PCB with continuous ground plane achieves the following measured performance:
| Parameter | Specification | Reference Standard | |---|---|---| | Shielding effectiveness, 200 MHz–1 GHz | ≥60 dB | IEEE 299 / CISPR 25 Class 5 | | Shielding effectiveness, 1 GHz–3 GHz | ≥50 dB | IEEE 299 | | Shielding effectiveness, 3 GHz–6 GHz | ≥40 dB | IEEE 299 | | Wall material: C5210 phosphor bronze | 0.20 mm nominal thickness | — | | Wall material: tin-plated cold-rolled steel | 0.30 mm nominal thickness | — | | Sheet resistance (tin-plated steel, 0.30 mm) | ≤0.5 mΩ/sq | ASTM B568 | | Relative permeability (steel variant) | μr ≈ 200 at 1 MHz | — | | Contact resistance per spring contact | 20–30 mΩ at 100 gf | EIA-364-06 | | Spring contact cycle life | ≥10,000 cycles at ≤50 mΩ | EIA-364-09 | | Via stitching pitch (target ≤λ/20 at 6 GHz) | ≤2.5 mm center-to-center | IPC-2221B §6.4 | | Solder pad width (shield can perimeter) | 1.0–1.5 mm | IPC-7351B | | Maximum aperture longest dimension | ≤λ/20 at highest frequency | — |
Material selection trade-offs. Tin-plated cold-rolled steel delivers superior low-frequency magnetic shielding (μr ≈ 200) and is the correct choice for applications with switching power supply noise below 30 MHz co-located with RF circuitry. Phosphor bronze and nickel silver offer better corrosion resistance and solderability at a lower profile (0.15–0.20 mm wall) but provide negligible magnetic permeability — they attenuate through reflection and are preferred for pure RF environments above 100 MHz. For dual-threat environments (conducted EMI below 30 MHz plus radiated above 500 MHz), POCONS offers hybrid-plated steel cans with tin over nickel base coat, achieving both magnetic absorption and consistent solder wetting.
Skin depth and absorption. At 1 GHz, skin depth in steel is approximately 0.5 μm. A 0.20 mm wall presents roughly 400 skin depths of absorption — absorption loss alone exceeds 100 dB at this frequency. Practical SE is therefore limited entirely by aperture leakage and contact impedance, not by wall penetration. This is why via stitching and spring contact quality dominate real-world shield performance.
Common Design Pitfalls
1. Via stitching pitch too wide for target frequency. Root cause: Designers apply a blanket 3–5 mm via pitch rule derived from legacy sub-1 GHz requirements without recalculating for modern wireless bands (Wi-Fi 6E at 5.9–7.1 GHz, 5G NR FR1 at 3.3–4.2 GHz). At 6 GHz, λ = 50 mm, and a via pitch of 5 mm corresponds to λ/10 — this creates slot antenna behavior along the shield perimeter, re-radiating energy that the shield can walls successfully contained. Observable consequence: radiated emissions spikes at frequencies where the via fence gap approaches λ/10 or wider, typically 3–6 dB above the limit line in CISPR 25 scans. Mitigation: enforce ≤λ/20 pitch at the highest harmonic of concern. For 6 GHz compliance, use 2.5 mm maximum pitch. Place vias on both sides of every shield can solder pad.
2. Broken ground plane beneath the shield can footprint. Root cause: Signal traces or power copper routed through the L2 ground layer directly below the shield perimeter. Even a single 0.15 mm trace crossing beneath a solder pad creates a slot in the return current path. Observable consequence: localized SE degradation of 15–25 dB at frequencies where the slot length resonates (typically 1–4 GHz for 20–50 mm slots). This manifests as a narrow-band emissions spike that is extremely difficult to diagnose without near-field probing. Mitigation: establish a routing keep-out zone on all inner layers extending 0.5 mm beyond the shield can footprint on all sides. Route all signals entering or exiting the shielded zone through designated feedthrough points with matched-impedance transitions.
3. Insufficient solder paste volume on shield can pads. Root cause: Using the same stencil aperture ratio for shield can perimeter pads as for standard SMD components. Shield can feet have larger thermal mass and wetting area, demanding more paste volume. Observable consequence: intermittent or cold solder joints along the shield perimeter, creating resistive contact points (>100 mΩ) that degrade SE by 10–20 dB and introduce reliability failures under thermal cycling. Mitigation: use a paste aperture ratio of 80–90% for shield can pads versus the standard 70–75% for passive components. If the shield can pad width is 1.0 mm, the stencil aperture should be 0.80–0.90 mm wide. Consider a step-up stencil with 0.15 mm thickness in the shield region versus 0.12 mm elsewhere.
4. Cavity resonance from untreated internal geometry. Root cause: The shield can interior dimensions create a resonant cavity. The dominant TE₁₀ mode resonant frequency is f = c/(2L) where L is the longest internal dimension. A 30 mm shield can resonates at approximately 5 GHz. Observable consequence: amplification of emissions at the resonant frequency and its harmonics — the shield can actually worsens emissions at these specific frequencies, sometimes by 6–10 dB above the unshielded baseline. Mitigation: partition large shield cans using internal fences (available on POCONS two-piece assemblies as welded or stamped dividers). Reduce the longest unsupported internal dimension to ensure the first cavity resonance falls above your highest frequency of concern, or apply RF-absorbing material to the can interior for suppression above 3 GHz.
5. Shield can lid contact degradation over product lifecycle. Root cause: Two-piece shield assemblies require reliable removable contact between base frame and lid for rework and debug access. Flat lid-to-frame contact relies on solder or friction, which degrades under thermal cycling and mechanical shock. Observable consequence: SE degrades by 10–30 dB after 500–1,000 thermal cycles (–40 °C to +85 °C per IEC 60068-2-14) as contact resistance rises above 100 mΩ per point. Products that pass initial certification fail in the field. Mitigation: specify spring contacts (BeCu finger stock or discrete pogo-style contacts) at the lid-to-frame interface. POCONS BeCu spring contacts maintain ≤50 mΩ through 10,000+ mating cycles and provide 50–100 gf contact force independent of solder joint integrity.
PCB Footprint & Soldering Profile Guidelines
Pad Geometry
The shield can perimeter pad must satisfy three constraints simultaneously: mechanical retention during reflow, low-impedance ground connection, and rework accessibility.
- Pad width: 1.0–1.5 mm, depending on shield can foot width. The pad should extend 0.3 mm beyond the shield foot on each side to allow visual solder joint inspection per IPC-A-610H Class 2.
- Pad length (along perimeter): Match the shield can foot segment length. For continuous perimeter feet, use a single continuous pad — do not segment the pad arbitrarily, as gaps create impedance discontinuities.
- Courtyard clearance: 0.50 mm minimum from the outer edge of the shield can body to any adjacent component courtyard. This allows rework nozzle access and prevents thermal shadowing during reflow.
- Via-in-pad: Place ground vias directly in the shield can solder pad. Use via diameter 0.25–0.30 mm with 0.50 mm pad annular ring. Fill and cap vias per IPC-4761 Type VII to prevent solder wicking. This approach eliminates the inductance of a via-to-pad trace and is critical for performance above 3 GHz.
- Thermal relief: Do not use thermal relief on shield can ground vias. Full connection to all ground plane layers is mandatory. The additional thermal demand is manageable with appropriate reflow profile adjustment.
Stencil Design
- Stencil thickness: 0.15 mm in the shield can region. If using a global 0.12 mm stencil, specify a step-up to 0.15 mm on shield can pads.
- Aperture ratio: 80–90% (aperture area / pad area). For a 1.2 mm wide pad, use a 1.0 mm wide aperture.
- Aperture shape: Rectangular with 0.05 mm corner radius to improve paste release.
- Orientation: Align aperture long axis parallel to squeegee travel direction.
Reflow Soldering Profile (SAC305, per J-STD-020E)
| Phase | Parameter | Value | |---|---|---| | Preheat | Ramp rate | 1.0–2.0 °C/s | | Soak | Temperature range | 150–200 °C | | Soak | Duration | 60–120 s | | Reflow | Peak temperature | 245 ± 5 °C | | Reflow | Time above liquidus (TAL, >217 °C) | 60–90 s | | Cooling | Ramp rate | ≤3.0 °C/s (–6 °C/s max) |
Shield cans present a significant thermal load. Profile the board with a thermocouple placed directly on a shield can solder pad to ensure the pad reaches ≥230 °C for at least 30 seconds. Insufficient TAL on shield pads is the most common root cause of intermittent contact failures in volume production.
For hand rework of individual shield cans, use a hot-air rework station at 350–380 °C with a nozzle sized to the shield can footprint. Apply flux per J-STD-004B, Type ROL0 or ROL1. Reference IPC-7711/7721 Section 8.3 for BGA-adjacent rework procedures, which apply to shield can removal near sensitive components.
Recommended POCONS Components
Custom Two-Piece Shield Cans
POCONS two-piece shield assemblies consist of a solder-mounted base frame and a removable snap-fit or spring-loaded lid. Available in tin-plated steel (0.20–0.30 mm) and phosphor bronze (0.15–0.20 mm). Custom dimensions from 5 × 5 mm to 80 × 80 mm with internal partitioning available. The base frame includes pre-formed solder feet at pitch intervals matched to your via stitching requirements — specify your target frequency and POCONS engineering calculates the foot pitch to ensure ≤λ/20 compliance.
Two-piece construction allows post-reflow debug access without desoldering, eliminating the board damage risk inherent in single-piece shield removal. This directly addresses the rework cost that drives many teams to skip shielding during prototype builds — a decision that frequently results in late-stage compliance failures.
Spring Contacts and Pogo Pins
POCONS BeCu spring contacts provide the reliable lid-to-frame electrical connection that maintains shielding effectiveness over the product lifecycle. Available in surface-mount and through-hole configurations with 50–150 gf contact force range. Gold-over-nickel plating ensures ≤30 mΩ contact resistance at initial mate and ≤50 mΩ at 10,000 cycles.
For applications requiring board-to-board grounding or shield-to-chassis bonding, POCONS pogo pin assemblies provide compliant, low-impedance connections that absorb tolerance stack-up between PCB and enclosure — eliminating the rigid gasket compression problems that cause field SE degradation.
SMD Pan Nuts
For shield cans secured with machine screws to achieve higher contact pressure than snap-fit retention, POCONS SMD pan nuts provide a surface-mount threaded fastening point. These reflow-soldered nuts create a direct, low-impedance ground bond at the screw location while allowing controlled compression force on the shield can lid. Available in M2, M2.5, and M3 thread sizes with soldering temperature compatibility up to 260 °C peak.
SMD pan nuts are the preferred solution when vibration requirements (per IEC 60068-2-6) exceed the retention capability of spring clips, or when shield cans exceed 40 × 40 mm and require distributed mechanical fastening to prevent lid bowing that creates aperture gaps along the perimeter.
Application note produced by POCONS USA engineering team. Contact applications@poconsusa.com for design review.
Frequently Asked Questions
What via stitching pitch is required to maintain shield effectiveness above 3 GHz?
Via pitch must be ≤λ/20 at the highest frequency of concern. At 6 GHz (λ = 50 mm), this means ≤2.5 mm center-to-center pitch around the full shield can perimeter. POCONS two-piece shield cans are designed with solder pad spacing that enforces this rule.
What contact resistance should spring contacts achieve for effective EMI grounding?
Spring contacts used for shield can lids should achieve ≤50 mΩ initial contact resistance per point, measured per EIA-364-06. POCONS BeCu spring contacts typically measure 20–30 mΩ per contact at 100 gf, maintaining ≤50 mΩ through 10,000+ mating cycles.
Can POCONS provide custom shield can dimensions matched to our PCB footprint?
Yes. POCONS manufactures custom two-piece shield cans with lead times starting at 2–3 weeks for prototypes. Minimum order quantities begin at 500 pieces for custom tooling. Submit your PCB outline and keep-out zone dimensions to applications@poconsusa.com for a DFM review and quotation.