Cable Construction: Gauge, Shielding and Length
Every page in this section ends in the same three questions: what is on each end, how long is it, and how is it built. This page is the third one.
Connector choice decides what an assembly can mate with. Construction decides whether it works at the length you need, in the place you are putting it.
On this page
- What does conductor gauge change in a cable assembly?
- What does shielding actually do, and when do you need it?
- Why does impedance control matter on a differential pair?
- How is cable length specified, and why is it usually wrong?
- What environmental requirements change the build?
- What construction information does EPS need to quote?
- Frequently asked questions
What does conductor gauge change in a cable assembly?
Conductor gauge is stated in AWG, and the scale runs backwards: a lower AWG number is a thicker conductor.
Thicker conductors carry less loss over distance, which is what buys length on a high-rate link, and carry more current, which is what matters on a power assembly. They also produce a stiffer, bulkier cable with a larger bend radius — so gauge is a trade against the routing space the chassis actually has.
On a data assembly, gauge and length are a single decision. On a power assembly, gauge follows the current the load draws. Neither is chosen in isolation.
What does shielding actually do, and when do you need it?
Shielding keeps external noise out of the pair and keeps the pair from radiating into everything around it. It matters most where the rate is high, where the cable runs near switching power supplies or motors, and where the assembly leaves a chassis.
Two constructions dominate. Foil gives complete coverage and is effective at high frequency, but it is fragile and adds little mechanical strength. Braid gives lower coverage but survives flexing and handling far better. Many assemblies use both, and per-pair shielding inside an overall shield is common where crosstalk between pairs is the concern.
A shield only works if it is terminated. An unterminated shield is a length of metal in the assembly doing nothing useful. How and where the shield lands is part of the specification, not an assembly detail.
Why does impedance control matter on a differential pair?
High-speed interfaces run as differential pairs against a controlled impedance. Where that impedance changes along the path — at a termination, at a connector, at a point where the geometry of the pair shifts — part of the signal reflects rather than continuing.
Reflections consume the same budget that distance consumes. This is why a shorter assembly built loosely can perform worse than a longer one built to a controlled geometry, and why construction is not a detail that can be left to the builder on a high-rate link.
How is cable length specified, and why is it usually wrong?
Length is measured along the route, not across the gap. It is the single most commonly mis-supplied input on a custom assembly.
A cable leaves a connector, turns, follows a tray or a channel, clears an obstruction, and arrives at the other end with enough slack for service access and enough bend radius that the pair geometry survives the turn. The straight-line distance between two connectors is rarely any of that.
Over-specifying is not free either: on a high-rate link, surplus length is spent budget. Measure the path you intend to use.
What environmental requirements change the build?
- Temperature range, which decides the jacket and insulation material.
- Flex duty. A cable that moves in service is a different construction from one that is installed once.
- Vibration, which drives latching, strain relief and overmold decisions.
- Proximity to heat or to switching noise, which affects both routing and shielding.
- Service access, which decides orientation, pull tabs and whether a connector can be reached at all once the system is closed.
What construction information does EPS need to quote?
- The interface and generation the assembly has to carry
- Routed length, measured along the intended path, with the slack you need
- Any gauge requirement, or the current draw if this is a power assembly
- Shielding requirement, and how the shield is to be terminated
- Temperature range, flex duty and vibration exposure
- Bend radius available at the tightest point on the route
- Quantity, and whether this is a prototype or a production run
Where you do not know a figure, say so rather than estimating it. Engineering review exists to close those gaps before anything is built.
Frequently asked questions
Does a lower AWG number mean a thicker or thinner cable?
Thicker. The AWG scale runs backwards, so a lower number is a larger conductor — less loss over distance and more current capacity, at the cost of a stiffer cable and a larger bend radius.
What is the difference between foil and braid shielding?
Foil gives complete coverage and works well at high frequency but is fragile. Braid gives lower coverage but survives flexing and handling far better. Many assemblies use both, and per-pair shields inside an overall shield are common where crosstalk is the concern.
How should I measure the length of a custom cable?
Along the route the cable will actually take, including turns, obstructions, bend radius and the slack needed for service access — not the straight-line distance between the two connectors.
Is a longer cable always worse on a high-speed link?
Length consumes budget, but so do reflections, connectors and mated interfaces. A shorter assembly built without controlled impedance can perform worse than a longer one built to a controlled geometry.
Does a cable that moves in service need a different construction?
Yes. Flex duty changes conductor stranding, shield choice and strain relief. A cable installed once and a cable that flexes continuously are different builds even at the same connectors and length.
Looking for a stocked assembly rather than a custom build? See the full ranges at MicroSATACables.
Have a custom interconnect challenge that off-the-shelf parts won’t solve?
EPS engineers and manufactures precision cable assemblies, signal-conditioning AICs, and custom enclosures for OEM partners, in Acton, Massachusetts. Lead time is confirmed with your quote against the actual specification and volume.