A metal enclosure shields RF until you cut it. The leaks are the openings: door seams, screw-to-screw gaps, cutouts and gland holes. The single rule of thumb that orders all of them is the half-wavelength cutoff: an opening whose largest dimension approaches half the wavelength of the frequency in question passes RF through with almost no loss, while an opening well under that dimension attenuates it. Work out that number first and you know which openings matter.
Our stainless enclosures are rated NEMA 4X / IP66 by the manufacturer; not UL listed. Shielding, like the water rating, is a property of the finished assembly, so it lives or dies on the apertures you specify.
The half-wavelength rule
Wavelength shrinks as frequency rises: roughly 12 in (300 mm) at 1 GHz, 4.8 in (120 mm) at 2.4 GHz, and about 0.7 in (17 mm) at 17 GHz. Half of that number is the cutoff length for any opening.
The behavior splits three ways:
- Opening near or above half a wavelength: the aperture is an efficient slot antenna. Expect little to no shielding at that frequency, no matter how good the metal is.
- Opening well under half a wavelength, shallow wall: attenuation grows as the opening shrinks. A rough estimate for a single aperture is on the order of 20 dB once the opening is a tenth of a wavelength, improving as it gets smaller, capped by what the solid metal and the seams would give you anyway.
- Opening well under half a wavelength, with depth: the hole behaves like a waveguide below cutoff. A hole of diameter
athrough a wall of depthdgives extra attenuation roughly proportional to the depth-to-diameter ratio. Doubling the wall thickness at a gland hole roughly doubles the below-cutoff loss in decibels.
The number to remember: a 1 in (25 mm) hole goes transparent around 6 GHz, because half a wavelength at 6 GHz is about an inch. Below that, a 1 in hole in a thick wall is still doing useful work.
What this means for common frequencies
| Frequency band | Half wavelength | Openings that leak |
|---|---|---|
| 433 MHz ISM | 13.7 in (348 mm) | Door seams, long gasket gaps |
| 900 MHz ISM | 6.4 in (164 mm) | Long seams, large display cutouts |
| 1 GHz (test limit region) | 5.9 in (150 mm) | Seams between widely spaced screws |
| 2.4 GHz Wi-Fi / Bluetooth | 2.4 in (61 mm) | Screw gaps, hinge-side gaps, vents |
| 5 GHz Wi-Fi | 1.2 in (30 mm) | Small cutouts, unlatched corners |
| 10+ GHz, harmonics | < 0.6 in (15 mm) | Gasket voids, punch-hole clearances |
Two conclusions follow. First, low frequencies forgive a lot: a continuous welded body with a decent door seam is quiet at 433 MHz without heroics. Second, above about 2 GHz the enemy is not the 1 in gland hole but the fraction-of-an-inch gap where a gasket does not contact, or the slot between two screws on a cover. A 0.020 in (0.5 mm) crack several inches long can radiate at 2.4 GHz far better than its width suggests, because the length is what sets the cutoff.
This is the same physics behind Wi-Fi signal through a stainless steel enclosure: if you want signal out, you engineer the opening; if you do not, you control the openings the same way.
Seams and screws: the longest gap wins
A seam without continuous metal contact is a slot, and a slot is judged by its longest continuous run. The shielding of a screwed cover is set by the screw pitch, not by the number of screws.
Practical rules for a screwed or hinged door:
- Keep the longest unsupported gap under a tenth of a wavelength at your highest frequency of concern. At 2.4 GHz that is about 0.24 in (6 mm), which is why screwed seams alone rarely get you there.
- Above roughly 1 GHz, plan a conductive or wire-reinforced gasket in the door seam so contact is continuous, not point-to-point. The gasket discussion in sheet-metal enclosure gasket design for IP64 and IP65 covers water sealing; for RF the added requirement is conductivity, so specify a gasket with a conductive layer or a knitted wire mesh core.
- Hinges and latches are classic dead spots. A latch that pulls the gasket tight at one corner often leaves the opposite corner open; that corner gap is your antenna.
- Paint and anodizing are insulators. Contact surfaces for conductive gaskets must be bare metal, the same principle NEC 250.12 applies to grounding contact surfaces. Our enclosures are stainless, so no coating has to be stripped, but any painted or plastic-coated box loses seam conductivity where the coating sits.
For the water-sealing side of door and gasket choices, the cutouts, glands and HMI windows guide collects the fundamentals.
Cutouts and cable entries
Cable and vent openings are unavoidable. Treat each one as a shield decision, not just a hole:
Cable glands. A gland hole in a stainless wall, say 0.875 in (22 mm) for a common metric gland, is far under half a wavelength at panel-relevant frequencies, and the wall depth gives a little waveguide-below-cutoff bonus. The aperture is rarely the leak. The cable is: an unshielded conductor is a direct path through the wall. Where shielding matters, use shielded cable terminated in a metal gland that bonds the shield to the enclosure wall, or pass the line through a bulkhead filter. Sizing and sealing details are in getting a connector through a cable gland without leaks and two cables through one cable gland.
Vents and breathers. A breather vent is a hole by design. Metallic breather vents solve the conflict: many small holes, each far below cutoff, with the depth of the vent body adding waveguide attenuation. See do breather vents keep a NEMA 4X enclosure rating for the water side; for RF, specify a metal-bodied vent, not a plastic one, if the enclosure is expected to shield. Many holes close together leak a little more than one hole, as an array, so keep the total vent area modest and the individual holes small.
Display and HMI windows. A viewing window is an aperture with plastic in it, which RF ignores. Above a few hundred MHz, a window of a few inches is a wide-open door. If you need both shielding and a display, specify a window with a conductive mesh or a conductive-coated glass, and bond its edge to the door. The trade-offs in a lockable HMI cover on a NEMA 4X door apply to shielding as much as to water.
Antennas. If a radio lives inside and must reach outside, do not fight the physics. Feed the antenna out through a dedicated feedthrough, as covered in sealing outdoor cable entries with an antenna feedthrough, and let the rest of the box stay quiet.
What to specify
Put the shielding decisions on the drawing, not in the field:
- List every opening with its largest dimension, and note the highest frequency the enclosure must contain or keep out. Anything within a factor of three of half a wavelength at that frequency gets redesigned.
- Prefer many small, deep holes to one large shallow opening where airflow allows; each hole must sit well under a tenth of the wavelength at your critical frequency.
- Specify conductive gaskets in door seams when the design works above roughly 1 GHz, with bare-metal contact surfaces.
- Keep interior noise sources at least a few times the largest aperture dimension away from that aperture. A switching supply pressed against a vented panel radiates through it efficiently.
- Use shielded cable and metal, wall-bonding glands for every line crossing the wall, or filter at the bulkhead.
- Lay out all openings in the cutout designer so spacing and edge distances are checked before metal is cut; the fundamentals are in the cutouts, glands and HMI windows guide.
If the openings you need exceed what shielding-by-geometry can deliver, the honest fix is a filtered or shielded enclosure product designed for the job, or a deliberate antenna strategy. What you should not do is cut first and measure later: a slot that is transparent at your frequency costs nothing to shrink on the drawing and a rework cycle to fix in the field. Any custom stainless enclosure order lets you set every opening before anything is cut.
Enclosures for this job
Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 12 × 10 × 6 in | 304 stainless | 12 x 10 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 16 × 14 × 8 in | 304 stainless | 16 x 14 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 20 × 16 × 8 in | 304 stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
What is the half-wavelength rule for enclosure shielding?
An opening passes RF freely when its largest dimension approaches half the wavelength of the frequency of interest, and attenuates RF when the dimension is a small fraction of that. At 1 GHz the wavelength is about 11.8 in (300 mm), so half a wavelength is about 5.9 in (150 mm). Any slot, seam gap or cutout approaching that length is an antenna at 1 GHz. Keep every opening well under a tenth of the wavelength you care about and the metal does most of the shielding for you.
Do stainless steel enclosures shield EMI?
Yes. Stainless steel is conductive, so a welded or continuously contacted stainless body behaves close to a Faraday cage for the frequencies panel builders usually worry about. The leaks come from the apertures, not the metal: door seams, screw spacing, cutouts and gland holes set the shielding performance. Fix those and the box itself is rarely the limit. Our enclosures are rated NEMA 4X / IP66 by the manufacturer; not UL listed.
How many screws do I need on a door for RF shielding?
Screw spacing sets the longest seam gap between screws, and the longest gap sets the cutoff. Space screws so the gap between them is well under half a wavelength at your highest frequency of concern, and under a tenth of it if you need serious attenuation. For a 2.4 GHz radio, half a wavelength is about 2.4 in (61 mm), so screw pitches under roughly 0.5 in (13 mm) or a conductive gasket is the practical answer.
Does a cable gland hole leak RF?
The hole itself is small and, in a metal wall of typical thickness, acts like a short waveguide below cutoff, so the aperture adds little leakage. The cable through it is the real path: an unfiltered, unshielded conductor carries RF straight through the wall. Use shielded cable with a conductive gland, or filter the line at the bulkhead, and the opening stops being the problem.
