How do the different types of materials used in Ethernet cables affect reliability?

Question

There are many metals used in different types of network cables, such as aluminum and copper. Some are expensive and some are cheap. What scenarios constitute the use of these types of different wire materials?

Answer 1

metal

Aluminum is less expensive, but will develop cracks if it is flexed too many times. Copper is more expensive, but is more ductile; So it is much more forgiving of being repeatedly flexed.

specifications

The Cat-N specifications determine various limits; Parameters such as:

• how much a signal may be attenuated per unit of distance
• how much cross-talk (signal leaking from one pair to another) is allowed between the pairs
• how much radiation ("leakage" of the signals) is allowed from the cable
• the range of frequencies which the cable's pairs are designed to carry

(and much, much more...)

use cases

There are differences in the insulation which jackets the conductors and the entire cable. For example, "plenum" cable -- meaning it is designed to be used in a plenum, a space within a building which is part of the air handling path for climate control -- is designed to limit how much, and at what temperature, it will emit toxic fumes.

There are options such as full length shielding (grounding) which prevents interference from other sources. (This is not common with Ethernet cables.)

The overall physical diameter of the cable may be a use case limiting factor. If you are install a very large number of cables, that might matter.

Some cables contain a "strain relief" feature; A small nylon string (looks like dental floss) can be added within the cable to help prevent stretching the conductors when cables are pulled into place.

Answer 2

This is more of a physics question than a networking question actually. Different kinds of metals have different physical characteristics, such as durability (can it rust/degrade?), conductivity (how much does the signal degrade per meter?), flexibility (does the inner cable easily crack?) and sensitivity to interference/crosstalk.

Network cables all carry a Category specification, which roughly groups them in how fast they can reliably connect two endpoints over a maximum distance. Cat5e for example is specified to be able to carry a 1000BASE-T (Gigabit) signal up to 100 metres per segment, with tighter crosstalk specifications than its predecessor Cat5.

While different metals might change a cable's category, the category only specifies rough minimum characteristics, such as impedance, resistance, propagation speed and even operating temperature range. Within a category you could still have notable differences in cable pricing for improvements on specific characteristics that do not make it fall in another category yet.

Answer 3

Excluding plenum-rating & speed, if any of this matters you probably should already be using fibre.

For EMI, grounding, etc. fibre is cheap insurance even when it's not strictly necessary.

For distance, if it's even possible you're running up to distance limits you should switch to fibre (obviously this doesn't really work with access runs). For access I'd try and stay on the conservative side of distance limits and verify runs with a certification tester. As long as you're on cat-5e or better you're fine for speed, although the better ratings do increase the distance limits on gig, but again you should be running cable certification tests if you're likely coming up on those limits.

Answer 4

The reliability of copper Ethernet cables is defined by Category -- such as Cat3, Cat5, Cat5e, Cat6, etc. The higher the category, the higher the reliability (and therefore the higher allowed speed standard). Higher categories represent more twists in the cables per foot, which alleviates crosstalk and EMI interference. For example, Cat5e cable has more twists per foot than Cat3, which is what allows gigabit Ethernet to work with Cat5e but not Cat3.

It should be noted that Ethernet does not always mean copper Unshielded Twist Pair cables. Ethernet is an OSI Layer-2 concept with different acceptable Layer-1 standards (such as fiber optic cables).

Answer 5

Aluminium is cheaper than copper but has higher resistivity and is more prone to unreliable terminations. Copper clad aluminium is an attempt to get some of the advantages of copper while keeping the price down. It's a clever idea in theory (in a high frequency system most current flows in the surface of the conductor anyway) but it's nonstandard and i've seen reports that it is unreliable in practice.

The categories of cable are mostly about the high frequency performance. Higher categories have better high frequency performance. This is acheived through higher twist rates (which better reject high frequency capactive and inductive coupling), higher quality materials, more careful control of the geometry and in some cases shielding. The category you need will depend on the speeds you are operating at and the distance you are operating over.

I would generally suggest going for cat6 solid copper cable for installation wiring to end devices, it's not terriblly expensive will support 1000BASE-T with ease ans on short runs should support 10GBASE-T . If you want to run 10GBASE-T on long links you might want to splash out on cat6A but last I checked it was substantially more expensive than cat6 and it's not clear if 10GBASE-T will ever become more than a niche product.

For backbones I would seriously consider fiber at this point.