What Your Cable Is Actually Rated For

Cable is the component people specify last and think about least, and it fails more often than the equipment it connects. Not dramatically — a cable rarely dies in a way that announces itself. It develops an intermittent fault, and the intermittent fault gets blamed on the sensor, then on the controller, then on software, and eventually somebody wiggles the cable and the fault moves.

The conductor size is the easy part and usually the only part that gets calculated. Everything else about a cable is a set of ratings that describe the conditions it was designed for, and installations routinely violate two or three of them at once.

Bend radius is the one violated most often and noticed least. Every cable has a minimum, usually quoted as a multiple of the outside diameter — often five to ten times for a fixed installation, more for anything that moves. Pull a shielded cable round a corner tighter than that and the screen deforms, the coverage opens up, and the cable's noise immunity quietly drops. It still works. It works slightly worse than it should, and nobody can tell you when that changed, because the installation photograph nobody took would have shown a neat right angle behind a panel.

Flex life is the rating that separates a cable that costs three euros a metre from one that costs eleven. A standard cable is designed to be installed once and then sit still. A drag-chain cable is built with stranded conductors laid in bundles around a central element, a specific lay length, and a jacket that slides rather than grips, and it is tested for millions of cycles at a stated radius. Putting fixed-installation cable in a moving application does not fail immediately. It fails after four months, usually on a Sunday, and the conductor break is inside the jacket where nothing is visible.

Then temperature, which has two separate numbers that get conflated. There is the installation temperature, below which the jacket becomes stiff enough to crack when bent, and the operating temperature, above which the insulation degrades over time. A PVC cable installed outdoors in a cold climate can be perfectly happy at minus twenty once it is in place and split if it is uncoiled and routed at that temperature. This is worth knowing before a winter installation rather than after.

Screening deserves more attention than it gets because the type matters, not just its presence. A foil screen gives excellent high-frequency coverage and is mechanically fragile; a braid screen is robust and slightly less effective at high frequency; a combination is what most decent industrial cable uses. What actually decides performance in practice is the termination. A screen connected at one end by a pigtail — a thin twisted wire to a terminal — is barely a screen at all above a few megahertz, because the pigtail is an inductor. A 360-degree clamp at a gland is the difference between a screen that works and a screen that is a comfort blanket.

Segregation is where the installation defeats the specification. Running a screened analogue signal cable in the same tray as the output of a variable-speed drive puts a fast-switching, high-voltage source a few centimetres away, and no realistic amount of screening compensates for that. The remedies are distance, separation by a metal divider, and crossing at right angles rather than running parallel. These cost almost nothing at design time and are effectively impossible to retrofit once the tray is full.

One practical habit closes most of this. When a cable schedule is written, put the actual conditions next to each run — moving or static, indoor or outdoor, temperature range, what else shares the route, total length — and make the supplier recommend against that description instead of against a part number someone used last time. It takes an afternoon, and it converts the cable from the thing that gets blamed into the thing that was specified.