A shielding effectiveness rating quoted as a single decibel figure tells you almost nothing. The same material can offer eighty decibels at one gigahertz and a handful at power frequency, and a well-chosen alloy loses most of its benefit the moment you cut a slot in it. The number is a curve, and the enclosure geometry matters more than the metal.
Three separate mechanisms make up one number
Shielding works by reflection, absorption, and multiple internal reflection. Reflection dominates when there is a large impedance mismatch between the incoming wave and the metal, which is why thin, highly conductive foils work well against electric fields and against plane waves at radio frequencies. Absorption dominates when the material is thick relative to skin depth, so it grows with frequency, conductivity and permeability. Multiple internal reflection is a correction term that matters mainly in thin shields at low frequency, where it reduces the total. Because these three depend on frequency in different directions, the summed curve has shape rather than a plateau. Quoting one value is like quoting a car’s speed without saying whether it is uphill. Ask which frequency, which field type, and which test method produced the figure.
Low-frequency magnetic fields are the hard case
Most shielding intuition comes from radio work, where a thin conductive box does very well. Magnetic fields at power frequency behave differently: skin depth in copper at sixty hertz is several millimeters, so a foil is close to transparent. What works instead is a high-permeability alloy that offers the flux an easier path than the air around the protected volume, effectively diverting rather than blocking it. Those alloys are expensive, must be annealed after forming, and lose permeability if they are bent, dropped or saturated by a strong field. That is why a mumetal can around a sensitive sensor is a precision component and why a steel box often outperforms an aluminum one for hum, even though aluminum is the better conductor. Choose the material for the field you actually have.
Apertures dominate real enclosures
In practice the metal is rarely the limit. Ventilation slots, display cutouts, cable entries, and the seams between panels all leak. A slot behaves like a slot antenna, and its leakage rises sharply as its longest dimension approaches a half wavelength, which is why a long thin gap can be far worse than a round hole of the same area. This drives several standard design moves: many small holes rather than one large one, honeycomb vents for airflow, conductive gaskets along every mating surface, and screw spacing chosen so the panel cannot bow open between fasteners. Paint, anodizing and corrosion all raise contact resistance across a seam and quietly erode the shield over years. An enclosure that tested well when new can degrade in service without anything visible changing.
Cables carry energy straight through a perfect shield
The most common failure is not a leak in the metal at all. A cable entering the box acts as a conductor that carries common-mode current from outside to inside, bypassing the shield entirely. The cure is to terminate the cable shield to the enclosure wall in a full three hundred and sixty degree bond at the point of entry, not with a pigtail wire to an internal ground stud. A pigtail adds inductance, and at a few hundred megahertz even a short one presents enough impedance to make the termination nearly useless. Filtered connectors and feedthrough capacitors serve the same purpose for unshielded conductors. Treating the enclosure boundary as the place where every conductor is controlled is the single most productive habit in shield design.
How test numbers are produced and why they travel badly
Published figures come from standardized methods: a dual-chamber test fixture for flat material samples, a reverberation chamber for enclosures, or a nested-loop arrangement for magnetic shielding. Each imposes its own geometry, source impedance and frequency range. A flat-sample figure describes the material, not your box, because it deliberately eliminates the seams and apertures that will limit your result. An enclosure figure describes that enclosure with that specific set of gaskets, cables and covers. Comparing a material number from one supplier to an enclosure number from another is not a comparison at all. When you need a defensible answer, test the assembly you intend to ship, in the configuration you intend to ship it, and record the setup alongside the number.
What to check before you act
- Ask for the frequency and field type behind any decibel figure.
- Inspect seams, vents and cutouts before blaming the material.
- Bond cable shields fully around the entry point, not with pigtails.
- Re-check enclosure joints for corrosion and paint after service work.
Common questions
Does a thicker shield always work better?
Thickness helps absorption at higher frequencies but does little for low-frequency magnetic fields, where permeability and flux path matter far more than mass.
Will conductive spray paint shield an enclosure?
Conductive coatings can provide useful attenuation at radio frequencies, but they are thin, easily scratched, and depend on good electrical contact at every seam to work as claimed.
The practical takeaway
Treat shielding effectiveness as a frequency-dependent property of a complete assembly, not a grade of metal. Fix the apertures and the cable entries first, choose the material for the field type you face, and insist that any quoted number comes with the frequency and test method attached.
References and further reading
- FCC: electromagnetic compatibility background and radio frequency questions
- FCC: how radio frequency energy interacts with equipment and people
- NFPA: electrical installation and equipment safety resources
Numerical scenarios are illustrative unless identified otherwise. Follow the exact product instructions; component ratings and local installation requirements can differ.
