EMI shield aperture size
Photochemical machining for EMI shield is often sold as a way to make holes and outlines, which is true and beside the point. A shield works or fails on four things — the size of its apertures relative to the wavelength, the continuity of its seams, the resistance of its contact points, and the thickness of its walls against skin depth — and etching touches three of them directly. This page is about the electromagnetic side of the decision: what the physics asks for, what the process can deliver, and where a supplier should refuse to promise a number.
1. What Actually Decides Shielding
A shield is a conductive enclosure, and its performance is limited by the largest opening in it rather than by its area.
Four openings matter. Vent and access apertures are the obvious ones. Seams — the joints between a can and its lid, or between the shield and the board — are the second. Contact points, where fingers or gaskets touch, are the third. And the wall itself is the fourth, which only becomes relevant at high frequency where skin depth is small.
That ranking is the useful one, because it tells a designer where to spend effort: a shield with perfect walls and a bad seam is a bad shield, and no amount of etching precision will fix a gap that runs the length of the lid.
Etching contributes to the first, second and third; it does nothing for the fourth except keep the wall uniform. The fourth is also the one that decides whether a painted shield is a shield at all. The ranking is also why a shield is designed from the seams inward rather than from the can outward, and why a wall thick enough for the lowest frequency is usually thicker than the skin depth needs at the highest.
2. Apertures and Frequency
An aperture stops behaving as a hole when its largest dimension approaches a fraction of the wavelength.
The rule of thumb is expressed as a fraction of wavelength at the frequency of interest, and the practical consequence is that the largest dimension matters, not the area. A long narrow slot leaks more than a round hole of the same open area, because the slot is a better antenna.
That is why vent patterns are made of small round or square holes rather than long slots, A round hole is a worse antenna than a slot of the same area, and that asymmetry is the whole design rule. That is why vent patterns are made of small round or square holes rather than long slots, and why a decorative slot pattern is a shielding liability. The fraction itself is a rule of thumb that differs between standards and consultants, and a hole array behaves differently again when the holes are close enough to couple.
The peak frequency of an emission also matters: an aperture that is acceptable at one frequency can be the worst one at another, and a shield is usually designed against a specific band rather than against a single number.
3. The Seam Is the Weak Link
Most shielding failures are seam failures, and they are mechanical rather than chemical.
A lid that does not seat, a can that is distorted by reflow, a fence that lifts at one corner: each of those opens a slot along its length, and a slot is the most efficient leak an enclosure can have.
That is why contact fingers, gaskets and screw pitch exist, A seam that is mechanically tight is usually electrically tight, though the reverse is not always true. That is why contact fingers, gaskets and screw pitch exist, and why the mechanical design of the seam matters more than the wall thickness of the can. Screw pitch is a shielding specification, not a mechanical convenience, and a gasket that compresses over time turns a tight seam into a loose one at year three.
For an etched part, the relevant contribution is flatness and edge quality: a can that is flat seats better, and a finger that is burr-free makes contact where it is drawn rather than where the burr happens to touch.
4. Contact Resistance and Plating
Every contact point is a resistor, and the shield’s performance depends on how many of them there are and how low each one is.
Tin plating is common because it is cheap, solderable and reasonably conductive; nickel is harder and more corrosion resistant but more resistive; gold is used where the contact is critical and the volume justifies it.
A plated contact that oxidises becomes a worse contact, A gold contact that is a tenth of a micron thin is a gold-coloured contact, not a gold contact. A plated contact that oxidises becomes a worse contact, which is why the plating is a shielding decision and not only a finishing one. Contact resistance is measured in milliohms, and a few milliohms at every finger is a shunt path for the current the shield is trying to divert; a thin tin over a rough etch is a short-lived contact.
Etching contributes here through consistency: fingers cut from one sheet have the same thickness and the same temper, so their contact force is the same, and a row of equal fingers is a better shield than a row of unequal ones.
5. What Etching Improves
Three things, and they are worth stating precisely.
The first is aperture geometry. Etching holds a defined aperture across a field, so a vent pattern that is designed for a frequency is the pattern that gets built, rather than a pattern that drifts with tool wear.
The second is integral features. Etching also lets a shield be made in one piece, which removes a seam altogether. The second is integral features. A shield can carry its fingers, its locating tabs, its gasket groove and its part number in one etch, which removes assembly steps and the tolerances they bring.
The third is edge quality. A burr-free edge seats better, contacts better and does not scrape the plating off the mating part.
None of those is a shielding number; they are the mechanical conditions that let the shield reach the number the design predicted. A vent pattern etched to a drawing is also a pattern that can be simulated before it is made, and integral fingers mean the finger and the can share a temper, which is not true when they are riveted on.
6. What Etching Cannot Promise
A shield supplier cannot guarantee shielding effectiveness in decibels, and a supplier who offers to should be treated with suspicion.
Shielding effectiveness is a system property: it depends on the enclosure, the seams, the gaskets, the cable penetrations, the board layout and the test method. A part can be perfect and the system can still fail, because the leak is at the connector.
What a supplier can do is hold the dimensions that the design assumed, and say so on a first-article report. That is a defensible position, and it is the one this page argues for. The honest deliverable is a part to drawing plus a report plus a conversation about what the drawing assumes, and a supplier who has seen many shields can say which features usually cause trouble — which is worth more than a fabricated number.
7. Design Rules for an Etched Shield
Four rules cover most of the work.
Keep apertures small and round or square where possible, and avoid long slots unless they are needed for a specific reason.
Space the contact fingers evenly and make them all the same width, so the contact force is uniform around the perimeter.
Put the seam where the lid is stiff, not where it is flexible; a seam along a floppy edge will gap when the assembly is handled.
Leave the etch allowance for the plating, because a plated aperture is smaller than an etched one and a plated finger is thicker than a drawn one. A slot that is needed for a connector should be as short as the connector allows and placed away from a current-carrying trace, and finger length matters too: a long finger is a softer spring but a worse antenna.
8. Specifying and Testing
An order for an etched shield should state the alloy and thickness, the aperture pattern with its largest dimension, the finger pattern with its pitch and width, the plating, the flatness of the seating face, and the packing.
It should also state what will be tested: A drawing that names the test and the face is a drawing that can be accepted. It should also state what will be tested: a dimensional first article, a plating thickness check, a flatness check. A first article should be measured on the seating face and on the finger row, because those are what the assembly feels, and plating thickness should be measured on a finger rather than on a flat area, because the finger is where the current crosses. Shielding effectiveness is tested on the assembly, by the customer, and the drawing should say so rather than imply otherwise.
The recurring mistakes are a slot pattern designed for appearance, a finger row with uneven widths, a plating allowance forgotten, and a purchase order that asks the etcher to guarantee a dB figure.
