The Complete Guide to EMI Shielding for Electronics

Electromagnetic interference (EMI) shielding is the practice of blocking electromagnetic fields from interfering with electronic circuits. At the board level, this typically involves metal enclosures—often called shield cans—that surround sensitive components.

EMI shielding serves two purposes:

1. Containment — Preventing your circuit from emitting interference that affects nearby devices
2. Immunity — Protecting your circuit from external electromagnetic interference

Modern electronics require effective EMI management to meet regulatory standards (FCC, CE, CISPR) and ensure reliable operation in the field.

EMI problems manifest in multiple ways:

• Regulatory failure — Products that exceed emission limits cannot be sold in most markets
• Performance degradation — Unshielded RF circuits suffer reduced sensitivity and range
• System interference — High-speed digital circuits can disrupt analog sections on the same board
• Safety concerns — In automotive and medical applications, EMI can cause safety-critical failures

The cost of addressing EMI late in development far exceeds the cost of proper shielding design upfront. Redesigns, re-certification, and delayed launches can cost 10-100x more than early EMI planning.

Board-level EMI shields come in several configurations. The right one is set by one question: will the assembly need rework or test access after the shield is on?

One-Piece Shield Cans

• Single stamped enclosure soldered directly to the PCB ground perimeter
• Most cost-effective for production; lowest part count
• No rework access—the can must be desoldered (and usually scrapped) to reach components

Two-Piece Shields (Frame + Lid)

• A solderable frame stays on the board; a stamped lid snaps on and off
• Allows probing, rework, and component replacement without desoldering
• Higher piece price, but lower total cost where rework or post-assembly tuning is expected
• Lid retention and grounding come from spring fingers on the lid or a separate shield-can clip

Surface-Mount Fence + Cover

• A picked-and-placed "fence" (open frame) plus a separate snap-on cover
• The fence is reflow-soldered like any SMT part—no through-hole, no hand soldering
• Cover removal force and grounding continuity are set by the clip/finger pitch around the fence

Multi-Cavity Shields

• One frame with internal walls that isolate multiple circuit sections (e.g. RF front-end from digital)
• Reduces part count and board area versus separate cans, and kills cavity-to-cavity coupling

Board-Level Shielding (BLS)

• General term for the perimeter shield-can approach above, as distinct from enclosure- or chassis-level shielding

Most cans are stamped and formed from thin sheet (drawn corners or tabbed-and-folded), which is why footprint, wall thickness, and aperture pattern are all tooling decisions made early in the design.

Base Materials

The counter-intuitive part: at RF, almost any metal reflects electric and plane-wave fields extremely well—reflection loss for a good conductor is typically well over 100 dB, far more than a board-level can ever actually delivers in practice. So above a few MHz, material choice is rarely about raw shielding capability. It is driven by formability, solderability, corrosion resistance, and cost. Nickel silver is the workhorse because it stamps and solders cleanly and resists corrosion.

Material *does* become decisive for low-frequency magnetic fields (below ~100 kHz—motor, transformer, and current-sense noise). Here reflection is weak and you need high permeability to absorb the field, which is why mu-metal—not a better conductor—is the answer in that band.

Plating Options

• Tin — Standard, solderable, cost-effective; the default for reflow assembly
• Nickel — Corrosion resistance and a harder surface; common under other finishes
• Gold — Best surface conductivity and contact reliability; highest cost, used at mating/clip surfaces
• Custom — Application-specific stack-ups on request

Plating mainly affects the surface (where RF currents actually flow, due to skin effect) and the solder/contact interface—not the bulk shielding of the wall.

Why one shield works at 2 GHz but does little at 2 kHz comes down to two mechanisms—reflection and absorption—and how they shift with frequency.

Shielding effectiveness has two main parts: reflection loss (the impedance mismatch a wave sees at the metal surface) and absorption loss (attenuation as it travels through the wall). Absorption depends on how many skin depths thick the wall is. Skin depth is the depth at which the field falls to about 37% (1/e):

δ = 1 / √(π · f · μ · σ)

where f is frequency, μ is permeability, and σ is conductivity. As frequency rises, δ collapses:

What this means for a shield-can design:

• At RF (MHz–GHz), wall thickness is almost irrelevant. A 0.15 mm wall is thousands of skin depths at 1 GHz, so absorption is enormous and reflection already dominates. You cannot meaningfully improve a working RF can by making the metal thicker—you improve it by closing apertures and seams.
• At low frequency, thickness and permeability are everything. Near DC the skin depth is millimetres, so a thin non-magnetic can barely touches a magnetic field. You need a high-permeability material (mu-metal) and real thickness.
• Currents ride on the surface. Because the field lives in the first few skin depths, the *continuity* of the conductive surface across seams and lids matters far more than the bulk metal behind it.

Grounding — a shield is only as good as its ground.

• Run a continuous ground ring on the PCB directly under the shield footprint
• Stitch ground vias close together along that ring (a useful rule: via spacing under λ/20 at your top frequency, the same limit that governs apertures below)
• Treat every break in the ground ring as a slot antenna, not a cosmetic gap

Apertures and Seams — usually the real limit on a shield's performance.

• Keep the longest dimension of any opening under λ/20 at the highest frequency of concern (see the worked numbers in Apertures, Seams & Leakage below)
• Overlapping or flanged seams beat butt joints; the goal is continuous metal-to-metal contact
• On two-piece designs, lid spring fingers and clip pitch set the effective seam length—closer pitch, higher frequency coverage

Thermal Management

• Solid shields trap heat; vent with many small holes (perforation or honeycomb) rather than a few large ones—total open area can be high while each hole stays electrically small
• Use thermal interface materials or a thermal pad to the lid for hot components instead of an oversized vent
• Two-piece designs allow airflow paths a sealed can cannot

Manufacturing (DFM)

• Design the footprint for pick-and-place: solderable tabs or a continuous wall, clear fiducials, defined keep-outs
• Reuse standard footprints where possible—custom outer dimensions are fine, but a sane footprint reduces tooling and assembly risk
• Decide rework strategy (one-piece vs. frame and lid) before tooling, because it drives the whole part

There is no single "standard" shield-can size, but board-level cans cluster in a well-understood design space. Knowing the ranges helps you decide between an existing footprint and a custom tool.

Typical dimensional envelope

• Footprint: from a few mm per side up to ~50 mm and beyond; larger areas usually move to multi-cavity or fence-and-cover construction
• Height: commonly ~1–6 mm, set by the tallest enclosed component plus clearance
• Wall thickness: typically ~0.10–0.20 mm for nickel silver—thin enough to stamp and solder, and (per skin depth) far more metal than RF shielding needs

Footprint is a tooling decision. The outer dimensions, wall, and any aperture pattern are all defined by the stamping tool, so the practical choice is:

• Match an existing footprint — fastest and cheapest, no new tooling
• Tool a custom can to your board — when no standard footprint fits your keep-out and height

POCONS is a direct stamper: parts are precision-stamped at an IATF 16949 facility in Korea, with sales, stock, and design-in support based in San Diego. If you have a competitor part number or a footprint, the fastest path is to check it against real dimensioned parts—browse the product catalog and run the cross-reference finder (both linked under *Continue Learning* below) to find an off-the-shelf match where one fits, or a custom-tooled drop-in where it doesn't.

A note on honesty: we publish real part dimensions, but we do not publish shielding-effectiveness numbers we have not measured for a given part and configuration. Ask us for the dB-versus-frequency data that applies to your design.

Shielding effectiveness (SE) measures how well a shield attenuates a field, expressed in decibels (dB). Formally it is the sum of three terms (the Schelkunoff model): reflection loss + absorption loss + a multiple-reflection correction. Every 20 dB is another 10× reduction in field strength:

Two practical points engineers miss:

• The metal almost never sets the limit. As shown above, a solid metal wall offers far more than 100 dB at RF. Real board-level shields land around 40–80 dB because of *apertures and seams*, not the material—the next section shows the math.
• SE is measured, not assumed. Effectiveness is specific to geometry, frequency, and field type; lab methods such as IEEE Std 299 (enclosures) and the coaxial/transfer-impedance methods exist precisely because you cannot read SE off a datasheet line. Ask for SE data tied to your configuration rather than a single headline number.

Because a solid metal wall is effectively perfect at RF, the openings decide your shielding effectiveness. An aperture behaves like a slot antenna: it leaks based on its longest linear dimension, not its total area.

For a slot of longest dimension L, a useful first-order estimate is:

SE (dB) = 20 · log10( λ / 2L ) — valid while L is below λ/2

When L reaches λ/2 the slot resonates and SE falls to about 0 dB: it radiates freely. The familiar λ/20 rule of thumb gives ~20 dB. Rearranged, the largest opening you can allow for a target SE is L = λ / (2 · 10^(SE/20)):

This one table explains most shielding surprises:

• Frequency is brutal. An opening that's fine at 1 GHz leaks badly at 28 GHz. A shield that passed at an older band can fail after a frequency bump with no other change.
• Many small holes beat one big slot. Fifty 1 mm vent holes and one 50 mm slot have similar open area, but the slot leaks like a 50 mm antenna while each small hole stays electrically tiny. That is why honeycomb and perforated vents work.
• A seam is just a long, thin slot. Two surfaces touching only at the corners leave a slot as long as the gap between contacts. Tighter spring-finger or clip pitch shortens the effective slot and pushes usable shielding to higher frequency.
• Cables and connector gaps count too. An unshielded cable or an ungrounded connector opening is an aperture—often the largest one on the whole product.

The design rule that falls out: size every opening *and* every contact gap under the λ/20 limit for the highest frequency you must pass, and prefer an array of small holes over any single large cut.

For New Designs

1. Start with your shielding effectiveness requirements
2. Define the frequency range of concern
3. Consider rework and testing access needs
4. Evaluate thermal constraints
5. Request samples for prototype validation

For Cost Reduction

1. Review current shield specifications—often over-engineered
2. Consider one-piece shields if rework access isn't critical
3. Evaluate standard dimensions before custom tooling
4. Compare material options for equivalent performance

POCONS Advantage

As a direct manufacturer, we can help optimize your design for performance and cost. Our sales team works with our Korea R&D team to review applications and recommend solutions based on actual requirements—not catalog assumptions.

EMI shielding components for different industries require specific certifications:

Automotive

• IATF 16949 — Quality management system
• AEC-Q200 — Passive component stress test qualification
• OEM-specific requirements (PPAP documentation)

Medical

• ISO 13485 — Medical device quality management
• IEC 60601 — EMC requirements for medical electrical equipment

General

• ISO 9001 — Quality management system
• RoHS/REACH — Environmental compliance

POCONS holds IATF 16949, ISO 9001, and ISO 14001 certifications. We provide complete documentation packages for your qualification process.

Verify Your Supply Chain

• Work with manufacturers, not brokers, for traceability
• Request certificates of conformance and lot documentation
• Audit manufacturing facilities when possible

Plan for Lead Times

• Sample lead time: Typically 2-4 weeks depending on specification
• Production lead time: 8-16 weeks depending on complexity
• Build buffer for qualification testing

Manage Costs Effectively

• Custom tooling amortizes over volume—factor into piece price
• Standard dimensions reduce or eliminate tooling charges
• Direct manufacturers typically offer 20-30% savings vs. distribution

Ready to Get Started?

Request samples to evaluate POCONS components for your application. Our sales team coordinates specification review and design-in support with our Korea R&D team.