Shielding is how a wire harness keeps electrical noise from getting into — or out of — its signal wires. A shield is a conductive layer, usually foil, braid, or both, wrapped around a conductor or a bundle and grounded so it drains interference away instead of letting it couple onto the signals. Whether you need shielding, and which type, depends on the signals you carry, the electrical environment around the harness, and the standard you build to. This guide covers foil versus braid versus combination shields, shield coverage and transfer impedance, how to ground and terminate a shield correctly — the part most builds get wrong — when a harness actually needs shielding, and how to specify it in an RFQ.
It is written for engineers and buyers specifying a shielded harness for the first time: enough to choose a shield, brief a supplier, and read a quote without over- or under-specifying. Shielding is cheap insurance when a signal needs it and wasted cost and weight when it does not, so the goal is to match the shield to the job.
What a shield does
Electromagnetic interference (EMI) is unwanted electrical energy that couples from one circuit onto another — capacitively through electric fields, inductively through magnetic fields, or by radiation at higher frequencies. A shield intercepts that energy and gives it a low-impedance path to ground before it reaches the signal conductors. It works in both directions: it keeps outside noise out of a sensitive signal, and it keeps a noisy signal from radiating out and disturbing its neighbours. The single most important thing to understand is that a shield only works when it is grounded. An ungrounded shield does not just do nothing — it can act as an antenna and make the problem worse.
Types of shield: foil, braid, and combination
Three constructions cover almost every harness. They trade coverage, flexibility, frequency performance, and cost against each other.
- Foil shield — a thin aluminized polyester tape wrapped around the bundle. It gives essentially 100% coverage, is light and inexpensive, and performs well at high frequencies. Because the foil itself is thin and has relatively high resistance, it is always paired with a drain wire — a bare or tinned conductor run in contact with the foil that carries the shield current to ground. Foil is not meant for continuous flexing; it can crack and lose coverage if it is bent back and forth.
- Braid shield — interwoven strands of tinned copper. It is flexible and durable, has low resistance for a strong low-frequency and ground path, and adds mechanical strength. Its optical coverage is typically in the 70–95% range because the weave leaves small diamond gaps, so a braid alone can be less effective at very high frequencies. Braid is heavier and more expensive than foil.
- Combination (foil + braid) — a foil layer for full coverage and high-frequency performance, wrapped by a braid for a low-resistance ground path and mechanical strength. This pairing gives the broadest frequency coverage and is the usual choice for demanding EMC environments.
- Spiral (served) shield — strands wound helically. It is very flexible but performs poorly at high frequency because the spiral behaves like an inductor, so it is used only where flexibility outweighs high-frequency shielding.
Coverage and transfer impedance
Shield coverage — the percentage of the surface a braid covers — is a useful first cut, but it is not the whole story. The real figure of merit is transfer impedance, usually given in milliohms per metre: it measures how much voltage the shield lets through onto the inner conductor for a given shield current, and lower is better. A foil-plus-braid construction has a lower transfer impedance than braid alone, which is why it holds up better as frequency rises. When you compare two shielded cables, compare transfer impedance where you can, not just the headline coverage number, and remember that a shield terminated badly at the ends can undo an excellent cable.
Grounding and termination: the part most builds get wrong
A shield is only as good as its connection to ground, and this is where most real-world shielding fails. Two rules matter most:
- Terminate the shield 360°, not with a pigtail. The best practice is a full circumferential (360°) bond of the shield to a connector backshell or metal shell, so the shield current returns evenly around the whole circumference. Twisting the shield into a short wire — a “pigtail” — and landing it on a pin adds inductance, and that inductance destroys high-frequency shielding no matter how good the cable is. Where a drain wire is used with foil, keep it as short as possible.
- Ground to chassis, and decide one end or both. The shield connects to chassis (earth) ground, not signal ground. Grounding at both ends gives the best high-frequency shielding but can create a ground loop if the two ends sit at different potentials, which shows up as hum or noise. Grounding at one end only avoids the loop and suits low-frequency and analog signals. The right choice depends on the signal and the system, so it is a decision to make deliberately and state in the drawing, not leave to the build floor.
When a harness needs shielding — and when it does not
Shield the circuits that need it, and leave the rest unshielded to save cost and weight. A harness generally needs shielding when it carries low-level analog or sensor signals, high-speed data such as encoder, USB, Ethernet, or some CAN runs, or any signal that runs near motors, variable-frequency drives, switching power supplies, or long parallel power runs. It also needs shielding when the product has to pass EMC compliance for CE or FCC. Conversely, robust digital lines, short runs, and power-only conductors often do not need a shield, and shielding everything “to be safe” simply adds cost, weight, and termination labour. Robots and motion systems are a common case where a few shielded signal circuits share a bundle with unshielded power — see our specialty cable capability for those constructions.
Common shielding mistakes
- Pigtail grounding. The most common one: a long twisted drain instead of a 360° bond, which ruins high-frequency performance.
- A floating shield. Leaving a shield ungrounded at both ends can make it act as an antenna and couple more noise than no shield at all.
- Grounding both ends without thinking. On a low-frequency or analog circuit this can create a ground loop and inject the very hum the shield was meant to stop.
- Forgetting the drain wire. A foil shield with no drain wire, or a drain wire left unconnected, gives a shield that is not actually grounded.
- Mixing shielded and unshielded wires without separation. A noisy unshielded power wire run tight against a sensitive shielded signal can still couple through the ends and gaps; keep them separated in the bundle.
How to specify shielding in an RFQ
Give a supplier enough to build and test the shield the same way you designed it. In the harness drawing and RFQ, state the shield type (foil, braid, or foil-plus-braid) and, where it matters, a coverage or transfer-impedance target; the termination method (360° to a backshell or connector shell versus a drain wire); the grounding scheme (one end or both, and to which ground); and the workmanship standard the assembly is built and inspected to. Naming the connector and backshell that make the 360° bond possible is part of the same decision — the connector system, not just the cable, determines whether a clean shield termination is even achievable. A harness built to a recognised workmanship standard such as IPC/WHMA-A-620 gives both sides a shared definition of an acceptable shield termination. For related background, see automotive wire harness basics, and browse the full wire harness blog for more on construction and sourcing.
FAQ
What is cable shielding in a wire harness?
Cable shielding is a conductive layer — foil, braid, or both — placed around a conductor or bundle and connected to ground, so that electromagnetic interference drains to ground instead of coupling onto the signal wires. It protects sensitive signals from outside noise and stops noisy signals from radiating out to their neighbours.
Foil or braid shield — which is better?
Neither is universally better; they trade off. Foil gives full coverage, is light and cheap, and performs well at high frequency but needs a drain wire and does not survive flexing. Braid is flexible and durable with a strong low-resistance ground path but has small gaps and is heavier and costlier. For demanding EMC, a combination of foil plus braid gives the broadest performance, which is why it is common on critical signal circuits.
Should a shield be grounded at one end or both?
It depends on the signal. Grounding both ends gives the best high-frequency shielding but risks a ground loop if the ends sit at different potentials. Grounding one end only avoids the loop and suits low-frequency and analog signals. Decide deliberately based on the signal and the system, and state the choice on the drawing rather than leaving it to the build floor.
What is a pigtail, and why avoid it?
A pigtail is a short wire made by twisting a shield or drain into a lead and landing it on a single pin instead of bonding the shield 360° to a backshell or shell. The pigtail adds inductance, and that inductance destroys high-frequency shielding no matter how good the cable is. A full circumferential (360°) termination is the correct practice.
Do I need shielded cable for my harness?
Shield the circuits that carry low-level analog or sensor signals, high-speed data, or any signal running near motors, drives, or switching supplies, and shield where the product must pass EMC compliance. Robust digital lines, short runs, and power-only conductors usually do not need a shield — shielding everything just adds cost, weight, and termination labour.
What shield coverage percentage do I need?
Coverage is a first cut, not a specification on its own. Higher braid coverage helps, but the more meaningful figure is transfer impedance (lower is better), and correct 360° termination matters more than a few percent of coverage. For critical high-frequency signals, a foil-plus-braid construction terminated properly outperforms a high-coverage braid that is pigtailed at the ends.
