The Engineer's Guide to EMI/RFI Shield DFM for Progressive Stamping

An IoT gateway passed pre-compliance with 52 dB of measured shielding effectiveness. Three weeks later, the same product failed final radiated emission testing at 2.4 GHz. The only change was a die revision that enlarged the shield vent holes from 1.0 mm to 1.6 mm. Each hole now admitted 2.6 times more open area. The certification retest cost USD 18,000 (typical for a small product), and the launch slipped four weeks.

Key Takeaways

Orient all flat blank bend lines perpendicular to the material grain to prevent micro-cracking.
Design bend reliefs and maintain proper radii to prevent flange tearing and tooling breakage.
Add stiffening ribs to flat surfaces to defeat thermal warping and guarantee SMT coplanarity.

The shield looked identical. It measured identically on a caliper. Its perforation grid just stopped behaving like a solid plane at 2.4 GHz. That is the gap between a stamped EMI shield and a DFM-correct one. The difference hides in holes, bends, contacts and tolerances that a drawing review will not catch.

This guide is the design for manufacturability (DFM) playbook for stamped EMI shields. It is written for hardware engineers, PCB designers and procurement teams. It covers materials, skin depth, aperture rules, grounding, tolerances, plating, tooling and cost. Every production number is labeled typical.

The Snapshot

Stamped shields deliver 40-80 dB of shielding effectiveness in the 30-100 MHz band from 0.08-0.5 mm stock (typical)
Skin depth at 100 MHz is about 6.6 µm in copper and 65 µm in magnetic steel (typical). Above 300 MHz, thickness is a stiffness decision, not a shielding one
Keep every aperture below one twentieth of the wavelength at your highest frequency, about 6 mm at 2.4 GHz (typical guidance)
Progressive dies hold critical dimensions to ±0.01-0.03 mm (typical) at up to 800 SPM on 25-300 ton presses
SECC, SUS301 and AL5052 cover most shield programs; tin or nickel plating handles solderability and salt spray per ASTM B117 for 48-72 hours (typical)

Stamped EMI shields, ground clips and cover frames come off one progressive die line.

Why EMI Shields Fail: The Test That Exposes Bad DFM

Radiated emission limits are the real customer for every shield. Commercial electronics typically must hold emissions below FCC Part 15 or CISPR limits. Receivers must also survive nearby transmitters. A shield that passes in the lab and fails in the module is almost always a DFM problem. Apertures, grounding or material changed during tooling.

Shielding effectiveness is measured in decibels. Every 20 dB is a tenfold reduction in field strength. A 40 dB shield stops 99 percent of the field. A 60 dB shield stops 99.9 percent. Stamped enclosures with good grounding typically land between 40 and 80 dB in the 30-100 MHz range. That band covers most immunity and emission work on real products.

The math behind the test is worth internalizing. A 3 dB shift doubles the field strength. A 6 dB shift quadruples it. Small mechanical changes, like a slightly larger hole grid, move the number far more than most teams expect.

Shielding effectiveness	Field reduction	Typical stamped shield result
20 dB	90%	poor grounding, large apertures
40 dB	99%	solid panel with modest apertures
60 dB	99.9%	small holes, frequent ground contacts
80 dB	99.99%	solid walls, sealed seams, best practice

Three factors set the number: reflection at the surface, absorption inside the material, and leakage through apertures and seams. DFM controls the third one. It is also the factor that drifts in production.

How a Stamped EMI Shield Works

A metal shield attenuates a field by reflection and by absorption. Reflection dominates in high-conductivity, non-magnetic metals like copper and aluminum. Absorption dominates in magnetic steel, which converts field energy into heat. Both mechanisms follow skin depth, the depth at which the field falls to about 37 percent of its surface value.

The skin depth formula is δ = 1 / √(π f μ σ). Here f is frequency, μ is permeability and σ is conductivity. Three practical lessons follow. First, skin depth shrinks as frequency rises, so thin material works at high frequency. Second, steel’s high permeability shrinks skin depth despite lower conductivity, which is why steel absorbs well. Third, material thicker than three skin depths behaves like a solid wall (typical rule of thumb).

Total shielding effectiveness is the sum of reflection loss, absorption loss and a re-reflection correction. Reflection depends on surface conductivity. Absorption depends on thickness and permeability. For a stamped shield in the 30-100 MHz band, both mechanisms work because skin depth is a fraction of the sheet thickness.

A 0.2 mm SECC can works below 100 MHz mainly through absorption. An aluminum can of the same geometry works through reflection. Choose the mechanism first, then the metal.

Below about 1 MHz, the picture changes. Magnetic fields at power-line frequencies pass through thin copper almost untouched. Reflection loss collapses, and skin depth grows beyond the sheet. Shielding low-frequency magnetic fields is a steel thickness game, not a stamping game. If your module faces magnetic interference below 1 MHz, say so at the RFI review before the die is cut.

Materials for EMI Shield Stamping

Four materials carry nearly all stamped shield programs. The right choice balances shielding mechanism, cost, spring behavior and plating compatibility. The table lists verified properties from our production data.

Material	Verified property	Shielding mechanism	Typical shield use
SECC electro-galvanized steel	tensile 270-380 MPa	magnetic absorption, stiff, low cost	can covers, board shields, 0.2-0.4 mm
SUS301 stainless steel	tensile 520-1200 MPa, temper dependent	spring temper, corrosion resistance	ground clips, spring fingers, covers
AL5052 aluminum	thermal conductivity 138 W/m·K	reflection, light weight, heat path	RF cans, heat-shield combos, 0.3-0.5 mm
C1100 copper	conductivity 101% IACS	best reflection, solderable	finger stock, high-frequency grounding

SECC is the default for cost-driven programs. Its zinc layer protects the surface, and tin plating makes it solderable. SUS301 adds spring behavior for clips and fingers that must press against a mating surface. AL5052 suits modules that need light weight or a heat path. C1100 appears where conductivity must be the best in the system, usually as finger stock or ground clips.

Material cost follows the same order. SECC is the cheapest. SUS301 runs two to three times higher, and C1100 is the premium (typical). Plating adds five to fifteen percent to the stamped part cost (typical).

Stamped C1100 copper clips and finger stock carry the grounding job on high-frequency shields.

Material Thickness and Skin Depth

Engineers often assume thicker metal shields better. Above a few hundred MHz, that assumption is wrong. The table lists typical skin depths at four frequencies. Values assume bulk conductivity and, for steel, a relative permeability near 100 (typical for low-carbon magnetic steel).

Frequency	Copper (µm)	Aluminum (µm)	Magnetic steel (µm)
30 MHz	12	16	120
100 MHz	6.6	8.5	65
300 MHz	3.8	4.9	38
1 GHz	2.1	2.7	21

Three skin depths is the typical rule of thumb for an effective wall. At 1 GHz, copper needs only about 6 µm, far below the thinnest practical stamping stock. The conclusion is practical. Above 300 MHz, choose thickness for stiffness, flatness and handleability, not for shielding.

In the 30-100 MHz band, magnetic steel needs roughly 0.2-0.4 mm to reach three skin depths (typical). That is why thin steel cans still carry an absorption story at these frequencies. Pure copper foil starts to need help below 100 MHz.

The standard 0.08-0.5 mm range for stamped shields exists for mechanical reasons. Stock below 0.15 mm is hard to feed flat in a progressive die and distorts during plating. Stock above 0.5 mm raises bending force and tooling cost without adding shielding value at GHz frequencies.

Aperture Design: Holes, Slots and Gaps

Every opening in a shield is a leak. A slot radiates like an antenna when its long dimension approaches half a wavelength. The practical rule keeps the longest aperture below one twentieth of the wavelength at the highest frequency of concern (typical guidance).

Frequency	Wavelength	λ/20 aperture limit (typical)
100 MHz	3000 mm	150 mm
300 MHz	1000 mm	50 mm
1 GHz	300 mm	15 mm
2.4 GHz	125 mm	6 mm
5 GHz	60 mm	3 mm

For a product with a 2.4 GHz radio, keep perforations at 2 mm or smaller and slots shorter than 6 mm (typical). The perforation grid must also be dense enough to behave like a solid plane. Typical practice holds the open area below five percent of the surface. Keep hole spacing at least equal to the hole diameter so the metal bridges stay strong.

Prefer round holes over slots for vents. A row of round holes leaks far less than one long slot of the same open area, because the slot’s long dimension radiates. If a slot is unavoidable, break it with a metal bridge every few millimeters (typical). The same logic applies to seams between shield halves and to the gap under a snap-on cover.

Grounding: The Contact That Makes or Breaks the Shield

A floating shield is a capacitor, not a shield. Seam leakage joins aperture leakage, and both grow when grounding is sparse. This is where stamped parts earn their keep. Die-formed fingers, clips and pads make dozens of contacts in one press stroke.

Typical design targets: a ground contact every 10-15 mm along the shield edge for GHz applications, and at least one contact in every corner. Contact resistance should sit below 10 mΩ at the PCB pad. Stamped spring fingers deliver 0.2-0.6 N of contact force per finger (typical). That force keeps the joint stable through shock and vibration.

Solder-down shields need ground pads wide enough for a reliable fillet. Clip-on shields need finger geometry that the die can form without work-hardening the metal past its elastic limit. SUS301 in a spring temper is the standard for fingers because it keeps its force for the life of the product.

Soldered ground pins are the most predictable joint on a shield. A pin that goes through a plated through-hole and wets on both sides gives a repeatable connection that clip contact can only approach. Use pins at the corners and clips along the edges, and the shield stays grounded everywhere it matters.

Let the shield vendor review the PCB footprint before the layout is frozen. A ground pad under a bend radius, or a finger that springs against silkscreen, will fail the first drop test.

Shield Architecture: Cans, Frames, Clips and Fingers

Stamped shields come in a few proven architectures. Each has a cost, a shielding ceiling and a DFM personality.

Architecture	Typical construction	Best for
One-piece can	drawn or formed from one blank	simple modules, lowest cost
Two-piece can and frame	frame soldered to PCB, can snaps on	rework access, higher part count
Finger and clip sets	SUS301 spring parts	grounding and retention
Gasket carrier	stamped channel holds conductive gasket	high SE, seams that must seal
Snap-on cover with tabs	formed tabs lock into frame	serviceable designs

One-piece cans are the cheapest and the most common. They shield well when apertures stay small and grounding is frequent. Two-piece designs cost more but let you open the module for rework. Finger sets are the quiet workhorse of EMI stamping; one press stroke becomes ten ground contacts.

The architecture decision belongs early in the project. It sets the die count, the plating scope and the assembly process. Changing architecture after tooling starts is the most expensive DFM mistake in this category.

DFM Rules for Stamped EMI Shields

These rules come from the die side of the business. They keep the part cheap, the die stable and the shield consistent across millions of pieces.

Feature	Rule (typical)	Why it matters
Material thickness	0.08-0.5 mm; prefer 0.2-0.3 mm	flat feeding, stable forming, plating without distortion
Bend radius	at least 1× stock; more for SUS301	prevents cracks and springback surprises
Hole to edge	at least 1.5× stock	prevents edge distortion and weak bridges
Aperture diameter	at or below λ/20	holds shielding effectiveness
Large flat spans	add ribs or emboss	controls flatness on thin stock
Bend dimensions	±0.05-0.10 mm	springback varies with material lot

Springback is the biggest DFM risk on SUS301 and high-temper steel. The released bend angle differs from the die angle, and the amount varies with temper, rolling direction and lot. Design the shield so its function does not depend on a tight angular tolerance. If it must, plan a coining step in the die (typical practice).

Flatness matters more than people expect. A shield that bows 0.2 mm after plating lifts its ground fingers off the PCB and can drop shielding effectiveness by 10-20 dB (typical). Ribs, embossed patterns and symmetric forming keep thin stock flat.

Tolerances and Measurement

Stamped shields are not precision machine parts. They need a few tight dimensions for fit and function, and loose ones everywhere else. Spending tolerance budget on non-critical features raises cost with no benefit.

Dimension type	Typical capability
Blanked outline and hole diameter	±0.03-0.05 mm
Critical hole position (pitch)	±0.01-0.03 mm
Bend and form dimensions	±0.05-0.10 mm
Flatness over a 50 mm span	0.10-0.20 mm
Production capability on a sharp die	Cpk ≥ 1.33 (typical target)

Hold the critical set to ±0.01-0.03 mm (typical) and confirm capability with a short run before production. Measure shields on a CMM or optical comparator. Check flatness with a surface plate and feeler gauge. Do not measure formed parts the same way you measure flat blanks; springback makes the datum ambiguous.

Flatness measurement deserves a fixture. A shield that springs in your hand will measure differently on the bench. Use a dedicated checking fixture with the same datum scheme as the assembly, and the measurement will match the module.

Plating and Corrosion Protection

Plating on a stamped shield is rarely cosmetic. It decides solderability, contact resistance and salt spray survival.

Finish	Thickness (typical)	Purpose
Tin	2-5 µm	solderability, corrosion control on SECC
Nickel	2-5 µm	wear, corrosion, hard contact surface
Zinc, SECC factory coating	base coat	cost baseline, light corrosion
Clear passivate	thin conversion coat	aluminum corrosion control

The common acceptance test is salt spray per ASTM B117 for 48-72 hours (typical for consumer and industrial electronics). Sheared edges on SECC expose bare steel. Edge corrosion shows up exactly where the test watches. Tin plating after forming covers the sheared edge and closes the failure mode.

Plating thickness tolerance matters on contact surfaces. Too little tin oxidizes and raises contact resistance. Too much tin builds up on sharp bends and flakes. Specify the finish on the drawing, not in a side conversation.

The Progressive Die Process at Scale

Shields are stamped on progressive dies. A strip of coil feeds through stations, and a finished part exits with every press stroke. A typical shield die runs 8-20 stations: pierce, blank, draw or emboss, form and trim.

High-speed presses reach up to 800 SPM for small parts. Formed shields usually run 150-400 SPM because the metal needs time to move (typical). Press tonnage for shields is modest, 25-100 tons typical, well inside the 25-300 ton range of a full stamping shop. Shield dies typically cost USD 10,000-35,000, depending on stations and draw depth.

The die itself is a precision stack: die plates, punches, strippers and pilots. Pilots locate the strip at every station, and their wear shows up first as hole position drift. A shield die with good piloting holds ±0.01-0.03 mm pitch for millions of strokes (typical).

Die clearance is the quality lever. For thin stock, typical clearance is 8-10 percent of material thickness per side. Too much clearance leaves burrs. Too little wears the die and leaves a rough edge. Burrs on a shield are not cosmetic. A burr that touches a PCB trace or a neighbor component can arc or short.

Process control on the strip matters as much as the die. Lubricant type and flow, coil flatness and strip width all change forming results. A shield program that skips first-article inspection after die maintenance will rediscover this the expensive way.

Telecom module shields: tolerance, plating and flatness decide pass or fail.

Cost Drivers: What Determines Shield Price

Shield stamping cost follows a predictable ladder. Knowing the ladder keeps the quote honest.

Driver	Typical impact
Tooling	USD 10,000-35,000 for a shield progressive die
Material	SECC cheapest; SUS301 2-3×; C1100 premium
Plating	adds 5-15% to part cost
Tolerances	tighter than ±0.03 mm adds tryout and slows the die
Volume	stamping wins above 20,000-50,000 pieces per year

Below roughly 20,000 pieces a year, a laser-cut and hand-folded shield can beat stamping on total cost (typical crossover). Above it, the progressive die pays back quickly because each press stroke costs pennies. The crossover moves up if the design needs tight tolerances or plating, because those cost the same per piece either way.

Die tryout is the hidden line item. A shield with deep draws or spring-temper fingers can need three to six tryout iterations before dimensions hold (typical). Each iteration costs toolroom time and delays the first article. Asking the vendor how many iterations the geometry implies is a legitimate sourcing question.

Ask the supplier for the tooling amortization plan in writing. The cheapest quote is often the one that amortizes tooling over a volume the program will never reach.

Common Failure Modes and Fixes

The failure modes below account for most field returns and test failures on stamped shields (typical experience). Each has a DFM fix.

Failure mode	Root cause	DFM fix
SE drop of 15-20 dB	aperture too large or grounding sparse	shrink apertures, ground every 10-15 mm
Rust at salt spray	bare sheared edges on SECC	tin plating after forming
Shield pops off in drop test	low finger force, over-stressed spring	SUS301 spring temper, relaxed bend radius
Burr shorting to PCB	die clearance worn or too wide	8-10% clearance per side, die maintenance
Distortion after plating	thin stock, unbalanced forming	ribs, symmetric form, 0.2 mm+ stock
Ground pad under a bend	footprint not reviewed with vendor	align pads to flat zones

The DFM Checklist Before You Send the Drawing

Material thickness in the 0.08-0.5 mm range, preferred 0.2-0.3 mm
Highest frequency identified, apertures at or below λ/20
Open area below 5 percent, round holes over slots
Ground contact every 10-15 mm, one in each corner
Contact resistance target below 10 mΩ at the pad
Critical dimensions marked at ±0.01-0.03 mm, the rest loose
Finish specified on the drawing, salt spray 48-72 hours per ASTM B117
Bend radius at least 1× stock, more for SUS301
Ribs or emboss on large flat spans
Vendor footprint review before PCB layout is frozen

The Bottom Line

A stamped EMI shield is a mechanical part with an electromagnetic job. The electromagnetic job fails when the mechanical details drift: apertures, grounding, flatness, plating. Good DFM keeps the shield a cheap, repeatable stamped part instead of a custom metal box.

Design with the vendor’s die in mind from the first layout. Hold critical dimensions to ±0.01-0.03 mm. Keep apertures below λ/20 and ground every 10-15 mm. Pick the material for the frequency band, and put the finish on the drawing. That is the whole game.

For the production side of shield stamping, see our metal stamping services overview. If material selection is still open, the materials page compares SECC, SUS301 and AL5052 in detail, and related articles live in the blog.