How to read a cable datasheet: the five numbers that decide a job

A datasheet is a promise you can enforce

Most cable disputes are not about copper quality. They are about a number nobody wrote down. The datasheet exists so that the purchase order can say "meets the stated limits" and a test report can be compared against a figure instead of an opinion.

Here is what each of the five numbers on a structured cabling datasheet actually constrains, using the limits we publish for our own range.

1. Characteristic impedance: 100 ± 15 ohm

Impedance is the ratio of voltage to current for a wave travelling down the cable. It is set by the geometry — conductor diameter, insulation wall, pair spacing — not by the copper purity. The tolerance band matters more than the centre value: 100 ± 15 ohm accepts 85 to 115 ohm.

Why it is a purchasing criterion: a channel built from cable at one end of the band and connectors at the other produces reflections at every transition. The link may pass continuity and still fail a certifier sweep, and the installer will blame the reel.

In this range the 100 ± 15 ohm limit is stated for the network, multi-pair and patch cord families. The coaxial range is a different animal — 75 ohm, matched to video equipment rather than to Ethernet.

2. Conductor resistance: ≤ 9.38 ohm per 100 m at 20 °C

This is DC resistance, and it is the number that decides whether the cable will run Power over Ethernet at distance. Resistance rises with temperature, so a limit quoted at 20 °C is a colder-day best case; in a hot ceiling void the same conductor is measurably worse.

9.38 ohm per 100 m converts to roughly 23 AWG-equivalent behaviour. If a customer is pushing PoE++ over 90 m, ask for the conductor size before quoting the reel — a 0.50 mm conductor and a 0.57 mm conductor are both "CAT6" and behave differently under load.

3. Mutual capacitance: ≤ 5.6 nF per 100 m

Capacitance between the two conductors of a pair, measured per unit length. It is the second half of the impedance story: too much capacitance and the pair behaves like a low-pass filter, which shows up as return loss and eventually as a link that trains down to a lower speed.

Our network cable limit is 5.6 nF/100 m, and the multi-pair telephone range is looser at 6.6 nF/100 m because it is not carrying Ethernet.

4. Insulation resistance: ≥ 5000 MΩ·km

A leakage figure. It is quoted in megohm-kilometres because insulation resistance scales with length — the longer the cable, the lower the measured value for the same compound. A factory measuring a 500 m reel and one measuring 100 m are not comparing the same number unless they normalise.

A failing insulation resistance reading usually means moisture or a damaged jacket, not a bad compound. It is the first test to run when a reel has been stored in a humid yard.

5. The test standard, not the marketing category

A category on the box is a claim; the standard is the method. Our range lists UL758, UL1581, UL444, TIA/EIA-568, CSA C22.2, ISO/IEC 11801, YD/T1019-2001, GY/T135-1998 and SJ/T11138-1997 across the range — different markets, different acceptance tests, and a customer in one market may not accept evidence produced under another.

When you quote, ask which standard the acceptance test will follow. That single question prevents the common end-of-job argument about whether a 100 MHz sweep or a 500 MHz sweep was the agreed proof.

What a datasheet cannot tell you

Conductor size in AWG is not published for this part number; it is stated as a diameter in millimetres. If the customer's specification is written in AWG, convert before quoting rather than assuming the nearest size matches.

And no datasheet tells you the installation quality. A cable inside its limits, pulled round a tight corner, is out of limits at the far end. The numbers are a floor, not a guarantee.