Shielded vs unshielded connectors

Choosing between a shielded and an unshielded connector is not a choice between better and worse. A shield adds cost, diameter, stiffness and assembly steps, and it only delivers when the whole shielding path — cable shield, termination, connector housing, mating interface and panel — is continuous and low in impedance at the frequencies that matter. A shielded cable terminated through a pigtail into an unshielded connector can perform little better than no shield at high frequency. This article explains what a shield does, how cable and connector shields are built and measured, how to terminate and ground them, and when an unshielded connection is the sound choice.

What a shield does, and what it cannot do

Interference couples into and out of cables in three main ways. Capacitive (electric-field) coupling dominates near high-voltage, fast-switching conductors. Inductive (magnetic-field) coupling dominates near conductors carrying large or fast-changing currents, such as motor and welding cables. Radiated coupling becomes significant when cable lengths are comparable to the wavelength — at tens of megahertz and above for typical machine cabling. Coupling also works in reverse: a cable that picks up interference will radiate it, so shielding serves both immunity and emission requirements.

A conductive shield connected to a reference intercepts electric fields well, even at low frequencies. Against magnetic fields it works by carrying an induced current that opposes the external field, which requires the shield to be connected at both ends and to be effective at the frequency concerned. At power frequencies, a thin shield gives little magnetic protection; there, twisted pairs and balanced circuits do most of the work, because the induced voltages in the two conductors largely cancel. Shielding and balanced transmission are therefore complementary measures, not alternatives.

Shield (screen)

A conductive layer around one or more conductors — foil, braid, spiral wires or a combination — intended to reduce electromagnetic coupling to and from the conductors inside.

Transfer impedance (Z~T~)

The ratio of the longitudinal voltage induced on the inside of a shield to the current flowing on its outside, usually per unit length (Ω/m or mΩ/m). Lower is better. It describes the shield itself, independent of the installation.

Screening attenuation

The ratio, in dB, between the power fed into the shielded system and the maximum power radiated from it, used at higher frequencies where transfer impedance alone does not describe performance.

360° termination

A connection that contacts the shield around its full circumference and transfers it to the connector housing without an intermediate wire.

Cable shield constructions

The connector can only continue the shield the cable provides, so the cable construction comes first. Most shielded cables use foil, braid or both.

Spiral (served) shields of helically wound wires are flexible and easy to terminate, but behave inductively at high frequency and are mostly used for audio and low-frequency signals. For balanced data cables, the naming convention used with ISO/IEC 11801 describes the construction compactly: the part before the slash is the overall shield, the part after it the shielding of the pairs. U/UTP is fully unshielded, F/UTP has an overall foil, SF/UTP an overall braid and foil, and S/FTP an overall braid with individually foiled pairs.

Transfer impedance and the weakest link

Transfer impedance is the most useful single quantity for comparing shields. At low frequency it approaches the DC resistance of the shield. As frequency rises, a solid tube improves because of the skin effect, whereas a braid eventually gets worse, because magnetic fields leak through its apertures and the woven strands add an inductive component. Datasheets of high-quality cables and connectors may state transfer impedance or screening attenuation as a function of frequency; where they do, compare them at the frequencies of the actual interference, not at a single headline point.

For cables, surface transfer impedance is commonly measured with the triaxial method of IEC 62153-4-3. For mated connectors — including the cable-to-connector transition — and for cable assemblies, IEC 62153-4-7 defines the triaxial tube-in-tube method for transfer impedance, screening attenuation and coupling attenuation. The existence of a dedicated connector method reflects a practical reality: in a shielded link, the connector and its termination are usually the weakest part. A cable with excellent shielding over its length can be undermined by a few centimetres of poorly terminated shield at each end.

Shield termination: 360° versus pigtail

A shielded connector has to make three low-impedance transitions: from cable shield to connector housing, from housing to mating housing, and from the fixed connector to the panel or enclosure. Each transition should be circumferential.

1. 01Cable to housing: the braid is folded back over a cone or ferrule and clamped, crimped or pressed by spring elements so that it contacts the metal housing all the way round. Foil shields are normally combined with a braid or a drain wire; where only a drain wire exists, the foil must still be brought into the clamping zone.
2. 02Housing to housing: metal or metallised housings, shield springs or contact fingers ensure continuity across the mating interface. The coupling nut alone, if it is free to rotate and not in defined contact, is not a reliable shield path.
3. 03Housing to panel: a conductive flange, EMC gasket or serrated contact on an unpainted, conductive surface. Paint, anodising and powder coating are insulators at the joint.

A pigtail — the shield twisted into a wire and connected to a pin or a screw — replaces a circumferential connection with a short inductance. At low frequency this hardly matters. At high frequency the pigtail's reactance can dominate the whole shield path, and shield current concentrated in the pigtail produces a local magnetic field that couples directly onto the inner conductors. The gap where the conductors leave the shield acts as an unshielded section.

Where a connector series offers only a shield pin, that pin is still better than no connection, but the wire to it should be as short as possible and the shield should be opened over the shortest possible length. For demanding EMC requirements, select a series that offers a 360° shield clamp as a design feature rather than adapting an unshielded design.

Grounding the shield: one end or both?

Few connectivity questions produce more debate. The answer depends on the frequency of the interference and on the potential differences in the installation.

In machinery and automation, connecting shields at both ends with 360° terminations, combined with a well-meshed equipotential bonding system, is the prevailing practice, because most interference sources — drives, switched-mode supplies, fast logic — are high-frequency. One-end grounding remains a valid tool for specific low-frequency analogue problems. In either case, the shield is not a substitute for the protective earth conductor and should not be used as a signal return unless the interface is designed that way, as in coaxial connections.

When to choose shielded or unshielded

Unshielded connections have genuine advantages: smaller and more flexible cables, simpler and cheaper termination, no ground-loop paths, and fewer failure points on moving machinery. Where balanced transmission, segregation from noisy wiring and filtering at the equipment boundary already achieve the required margin, adding a shield brings cost without benefit. Conversely, where a shield is needed, a half-implemented one gives a false sense of security.

Many common interfaces exist in both versions — shielded and unshielded M12 variants, shielded RJ45 and D-Sub connectors with metal hoods — so the decision is usually made per link rather than per product family. Shielded connector types can be explored in the catalogue via shielded connectors; for complete links, shield termination is one of the details best settled in a cable assembly design. Related topics are covered in how to select an industrial connector and RF connector selection and signal integrity.