Connections +

Cutting down on noise

While determining separation distances between power and telco cables can be problematic, help is on the way from a TIA task group

May 1, 2004  

Print this page

It’s definitely a challenge to determine the separation distances between power and telecommunications cables. Part of the problem is that it is not obvious to gauge what the minimum distance between each should be. The codes quote one number, standards bodies quote another and manufacturers still another.

I was involved with the TIA task group that did a lot of work to determine the effect of power line noise transients on the operation of Gigabit Ethernet (1000BASE-T) over Category 5e and Category 6 cabling.

The results of this work are summarized in an Annex of the soon to be published TIA 569-B standard, which is currently awaiting ANSI approval.

The chart on the right shows the relative effect of coupled noise between power and telecommunications cables measured under different conditions.

The reference condition corresponds to the case where the noise Reduction Factor is equal to 1.0, i.e. Category 5e cabling with a balance (LCL/TCL) of 40 dB from 1 to 30 MHz, good NIC cards, maximum coupled length and horizontal link length of 90 meters, unbundled power conductors and zero separation distance (~ 6 mm).

How do we interpret and make use of this table? The tests performed by TIA confirmed error free operation for 1000BASE -T for the reference condition (Reduction Factor = 1) with 500V peak to peak high frequency (EFT) noise signals on power cable.

It is also noted in informative Annex C of TIA 569-B that EFT testing using repetitive 500 V pulses represents a worst case condition that is unlikely to be encountered in commercial office environments.

For those engineers that want to design with an additional margin of safety to accommodate poor NIC cards or for noisy environments, the table can be used to calculate the expected improvement in noise immunity.

For example, using Category 6 cabling and two-inch separation would reduce the coupled noise by a factor of (0.5 x 0.6) = 0.3, or a 70% improvement compared to the reference condition, which is quite significant.

What do other codes and standards have to say about separation of power and telecommunications circuits? The ANSI/NFPA 70 National Electrical Code 1999 specifies that communications wires and cables shall be separated at least 2 in. (50.8 mm) from conductors of any electric light, power, Class 1, non-power-limited fire alarm, or medium power network-powered broadband communications circuits with two exceptions.

The exceptions involve various means of isolating the power and communications cables, such as separate raceway or non-conductor tubing.

The BICSI NECA 568 2001 document provides a table of separation distances that are more conservative than TIA. Therefore, the minimum separation distance requires a judgment call, depending on how conservative you want to be.

If it is any consolation, the TIA notes that excessive noise on power branch circuits is an abnormal condition and advises that potential noise sources need to be isolated or eliminated and in situations where these sources cannot be removed, they should be suppressed using line conditioners or surge protection.

Noise reduction factors under different test variables

Test VariableReduction

Effect of cable categoryFactor

Category 5e 1

Category 6 0.5

Effect of cable balance

LCL/TCL > 40 dB from 1 to 30 MHz 1

LCL/TCL > 50 dB from 1 to 30 MHz 0.5

NIC performance

Good 1

Poor 2

Effect of coupled length near equipment

Length = 90m 1

Length = 10m 1

Length < 2m 0.7

Effect of horizontal link length

Length = 90 m 1

Length = 60 m 0.7

Length = 40 m or less 0.6

Effect of Power Cabling

Unbundled power conductors 1

Bundled power conductors 0.7

Effect of separation distance

Zero (~ 6 mm) 1

1 inch (25 mm) 0.8

2 inch (50 mm) 0.6

3 inch (75 mm) 0.5

6 inch (100 mm) 0.3

Paul Kish is Director, IBDN Systems & Standards at NORDX/CDT. He is also vice chair of the TR-42 engineering committee.

Disclaimer: The information presented is the author’s view and is not official TIA correspondence.

Print this page