Electronics Guide

Articles in This Category

Why Harnesses Are Harder Than Boards

On a printed circuit board, trace width, dielectric thickness, and the distance to the reference plane are fixed at lamination and never change. A harness offers none of those guarantees. Conductor spacing shifts when a bundle is tied, pulled around a bracket, or flexed in service; the return path may be a shield, a dedicated return wire, or the chassis, and it may change along the run; and the assembly is built largely by hand, so workmanship becomes a genuine electrical variable. Failures in cabled systems typically concentrate at terminations and at points of mechanical stress rather than in the body of the cable.

Three requirement sets therefore run in parallel and frequently conflict. Electrical requirements set conductor size, impedance, and loss. Electromagnetic requirements set shielding construction and, above all, how the shield is terminated. Mechanical and environmental requirements set stranding, bend radius, clamping, and jacket material. A harness that meets its loss budget but chafes through against structure has failed, and so has one that survives every vibration test but radiates above the applicable emissions limit. Cost enters as assembly labor rather than as component price, which is why designs that reduce the number of distinct terminations often win even when their bill of materials looks more expensive.

Conductors, Insulation, and Wire Selection

Conductor size follows the American Wire Gauge scale, in which the number rises as the wire gets smaller and the cross-sectional area roughly doubles for every three-gauge decrease. Sizing is driven by three constraints at once: current-carrying capacity, allowable voltage drop over the full length of the run, and mechanical robustness. The last constraint often dominates, because a conductor thin enough to carry a milliampere signal is too fragile to crimp and handle reliably, which is why aerospace and military wiring practice imposes a minimum gauge for general airframe wiring regardless of the current involved. Current ratings must also be derated when many current-carrying conductors share a bundle, since wires in the interior cannot shed heat, and derated again for altitude in airborne installations.

Stranding trades manufacturability against flex life. Solid conductors are inexpensive and hold their geometry well, but any assembly that moves requires stranded construction, and continuous-flex applications require high strand counts of fine wire. Finer stranding costs diameter and makes termination more delicate. Conductor plating protects the copper and sets the practical temperature ceiling: tin, silver, and nickel plating are the common choices in ascending order of temperature capability, and the aerospace cable specification MIL-DTL-27500 covers constructions with usable temperatures from roughly 150 °C to 260 °C depending on the wire and jacket selected.

Insulation choice is simultaneously a mechanical, thermal, and high-frequency decision. Polyvinyl chloride is cheap and flexible but limited in temperature and lossy at high frequency. Polyethylene and the fluoropolymers, particularly PTFE and FEP, offer low dielectric loss and are standard in coaxial and high-speed cable. Foamed dielectrics dilute the plastic with gas, lowering the effective dielectric constant toward that of air, which raises propagation velocity and reduces loss at the cost of crush resistance. In aerospace, thin-wall fluoropolymer wire to MIL-DTL-22759, also issued as SAE AS22759, is the workhorse because it combines high temperature capability, flame resistance, fluid resistance, and low weight.

Controlled-Impedance Cable Types

Any cable carrying a signal whose edge rate is fast relative to its propagation delay must be treated as a transmission line, and the industry has converged on a small set of characteristic impedances. Coaxial cable is predominantly 50 ohms for general radio-frequency and instrumentation use and 75 ohms for video and broadcast distribution. Balanced pairs cluster near 100 ohms differential for Ethernet and most high-speed serial links and 120 ohms for the CAN vehicle bus. Some standards fix the cable explicitly: MIL-STD-1553 requires a bus cable with a characteristic impedance of 70 to 85 ohms measured at 1 MHz, typically 78 ohms, terminated in a matched resistance at both ends.

Loss in copper cable comes from two mechanisms with different frequency signatures. Conductor loss grows roughly with the square root of frequency as skin effect confines current to an ever-thinner surface layer, while dielectric loss grows roughly in proportion to frequency. At low frequencies the conductor term dominates and a larger conductor is the remedy; above a few gigahertz the dielectric term takes over and only a lower-loss insulation helps. This is why high-speed cable assemblies grow thicker and stiffer as their required reach increases, in direct conflict with the bend radius, weight, and connector density the mechanical designer wants. Passive copper assemblies therefore become reach-limited at multi-gigabit rates, and the response is either to place active circuitry in the connector shell or to convert to optics inside the cable.

Skew is the timing error that cable construction introduces. Intra-pair skew, the delay difference between the two conductors of a differential pair, converts part of the differential signal into a common-mode signal that radiates and erodes margin. It arises from asymmetry in the dielectric and from unequal effective path lengths in the twist. Inter-pair skew, the delay difference between lanes, matters for parallel interfaces where lanes must be deskewed at the receiver. Both are controlled by tight, uniform, matched construction and are verified by measurement rather than assumed.

Shielding and Shield Termination

Cable shields come in two basic forms with complementary weaknesses. Foil shields provide essentially complete optical coverage, add little weight, and cost little, but aluminum foil has far higher direct-current resistance than braid and tears under repeated flexing. Braided shields typically provide somewhere between 60 and 95 percent optical coverage, and although the gaps in the weave leak at high frequency, the braid offers low direct-current resistance and survives flexing. Combination constructions wrap a foil under a braid to obtain broadband coverage and a low-resistance path together, and demanding electromagnetic environments use double or triple braids.

The figure of merit for a shield is transfer impedance, expressed in milliohms per meter: the voltage developed on the inside of the shield per unit of interference current flowing on the outside. Lower is better. At low frequency the transfer impedance of a braid equals its direct-current resistance; it then falls as skin effect confines interference current to the outer surface, and rises again at higher frequencies as leakage through the apertures of the weave takes over. That rising aperture term is what limits braided cable at high frequency and what additional braid layers are meant to suppress.

Termination usually decides whether a shield performs at all. A shield brought to ground through a short twisted wire, the so-called pigtail, behaves as a series inductor and loses most of its high-frequency effectiveness; a 360-degree circumferential bond to a conductive backshell or connector shell preserves it. Pigtail termination is among the most common and most consequential harness electromagnetic compatibility defects, and it is expensive to discover late because it is found in the emissions chamber rather than at the bench. The choice of how many ends to bond is a separate decision: bonding both ends is required for high-frequency shielding and for magnetic-field attenuation, while bonding one end only avoids a power-frequency ground loop. A hybrid termination, direct at one end and through a capacitor at the other, resolves the conflict by bonding at high frequency while blocking low-frequency circulating current.

Connectors, Terminations, and Routing

The connector is normally the weakest link electrically and mechanically. It introduces an impedance discontinuity where the cable geometry opens out into contacts, it concentrates crosstalk where pairs are forced close together, and it is the site where corrosion, fretting, and contact wear appear. High-speed connectors answer with shielded, impedance-controlled contact geometry and careful attention to the launch from cable into connector, which time-domain reflectometry exposes as a signature of discontinuities.

For wire termination, a properly executed crimp forms a gas-tight, cold-welded joint that outperforms solder under vibration, because solder wicks into the strands and creates a stiff stress riser exactly where the wire needs to flex. Crimp quality is controlled by measuring crimp height, by pull testing to verify tensile strength, and by cross-sectioning sample crimps to confirm that the strands have deformed into a void-free mass. Contact plating follows the circuit: gold is used for low-level signal contacts where a stable, low contact resistance matters, and tin is used where cost dominates, with the caveat that tin contacts are vulnerable to fretting corrosion under vibration and micro-motion, and mixed plating between mating halves is avoided.

Routing then determines much of the harness's electromagnetic and mechanical fate. Noisy and sensitive circuits are segregated into separate bundles with physical separation between them, and are crossed at right angles where crossing cannot be avoided. Minimum bend radius is specified as a multiple of the cable's outside diameter, with continuously flexing runs requiring substantially larger radii than static installations; violating it deforms the dielectric and permanently shifts impedance and return loss. Strain relief must ensure that the conductor is never the load path, with clamps at defined intervals, backshell clamping at connectors, and chafe protection wherever the bundle passes an edge or bulkhead. Weight is a first-order concern in vehicles and aircraft, where the harness is among the heaviest non-structural assemblies; it is attacked by replacing point-to-point wiring with multiplexed data buses and, in road vehicles, by raising the distribution voltage, since a 48-volt system carries one-quarter the current of a 12-volt system for the same power and permits correspondingly smaller conductors.

Standards, Testing, and Qualification

Workmanship for cable and harness assemblies is governed principally by IPC/WHMA-A-620, Requirements and Acceptance for Cable and Wire Harness Assemblies, whose Revision F was released in 2025. It defines materials, methods, and acceptance criteria for crimped, soldered, ultrasonically welded, and mechanically secured terminations, along with splices, shielding, marking, protective coverings, and testing, and it sorts products into three classes: Class 1 for general electronic products, Class 2 for dedicated service electronic products, and Class 3 for high-performance and harsh-environment products where continued operation is required.

Aerospace practice adds installation and selection requirements. SAE AS50881, Wiring, Aerospace Vehicle, which succeeded the military specification MIL-W-5088, covers the electrical wiring interconnection system from wire selection through installation. Detail specifications such as MIL-DTL-22759 for fluoropolymer-insulated airframe wire and MIL-DTL-27500 for shielded and jacketed multiconductor cable define the constructions themselves. Protocol standards can constrain the harness just as tightly: beyond fixing cable impedance, MIL-STD-1553 limits stub length to roughly 6 meters (20 feet) for transformer-coupled stubs and roughly 0.3 meter (1 foot) for direct-coupled stubs, so that stub loading does not disturb the main bus.

Verification proceeds in layers. Every assembly is tested for continuity and correct point-to-point wiring against the wire list, then for insulation resistance and dielectric withstanding voltage to confirm that insulation and clearances are intact. Terminations are sampled for contact retention and crimp pull strength. High-speed assemblies add time-domain reflectometry to profile impedance along the assembly and vector network analyzer measurement of insertion loss, return loss, crosstalk, and mode conversion. Qualification then subjects representative assemblies to the environment they will see: temperature cycling, humidity, vibration, flex endurance, salt fog, and exposure to the fuels, hydraulic fluids, and cleaning agents present in service.

Conclusion

Cable and harness design succeeds when the electrical, electromagnetic, and mechanical requirements are resolved together rather than in sequence. The cable chosen for its loss budget must survive its bend radius; the shield chosen for its coverage must be terminated circumferentially to deliver that coverage; the connector chosen for its density must still be assembled reliably by hand. Because the resulting assembly is built and installed by people and then lives in a moving, heated, contaminated environment, disciplined workmanship standards and layered testing are as much a part of the design as the impedance calculation. Treating the harness as a designed system, rather than as wiring added after the boxes are defined, is what keeps signals recoverable and installations reliable over a product's service life.

Related Topics