Unwanted Noise on Cable Pairs

A field technician called me to help him with a noise complaint that affected the customer’s POTS circuit and the Internet. The customer and the technician could hear the noise.

When testing the vacant cable pair, the technician found the following:
• AC voltage tip and ring to ground measured 8VAC: acceptable
• Longitudinal balance measured 58dB: marginal
• DC volts tip and ring to ground 0VDC: acceptable
• Insulation resistance tip to ring, tip to ground, and ring to ground 999 megohms: very acceptable
• Opens tip to ground 5,400 feet, ring to ground 5,350 feet. From that information, a capacitive balance calculation shows 99%: acceptable

The technician then ran a transmission test and found:
• The circuit loop current measured 37 Ma: acceptable
• Station ground measured 4 ohms: acceptable
• Circuit loss measured 5.2 dBm: acceptable
• Circuit noise measured 30dBrnC: unacceptable
• Power influence measured 90dBrnC: unacceptable

The technician then measured power influence with the C filter removed and measured 106dBrn, an increase of less than 20dB indicating a bonding and grounding issue. Not only is that customer affected by the high-power influence, so are all other circuits in the cable.

Just listening for noise on a circuit can be an almost futile task. For example, here’s an intermittent noise complaint that I’ve used as an example in the past but it’s very relevant: the customer complains of high noise in the evening. A technician is dispatched in the morning and checks the circuit at 10:00 A.M., when everyone is at work and the power influence in the area is down. The circuit sounds good. The customer comes home, and like everyone else in the neighborhood, turns on the TV, the stove, and the washer and dryer. The power load increases, power influence goes up, and the line noise goes up.

This noise heard by the customer is often caused by such physical factors as pair trouble, an open lateral, water, or bad splicing. If all physical factors test OK, the noise is most probably caused by induced AC current from adjacent power lines. This is often called line born noise and is caused by an imbalance in the capacitance or resistance of the pair.

If the pair is balanced, the AC current flow is equal and opposite, thus canceling, and no current flows to ground. Any unbalance, either resistive or capacitive, will cause current — equal to the difference in the balance between the two conductors — to flow to ground and "noiseup" the circuit.

Circuit noise (sometimes referred to as Noise Metallic) and Power Influence are interrelated, each being dependent upon the other to form acceptable balance.

When Noise Metallic is marginal or unacceptable, and Power Influence is acceptable, suspect a pair problem and go after it. The pair is unbalanced either resistively (going open) or capacitively (one side open on a lateral, beyond the workout terminal, or crossed with a nonworking pair). In most instances, a capacitive imbalance can be identified and isolated with an open meter or a TDR. If not, use the location techniques for finding a resistive unbalance. This type of problem pertains only to this pair and this customer, and can be isolated and repaired by the first man out. If you transfer the circuit to another cable pair rather than fixing the pair, you can bet the next tech, looking for a good cable pair, will transfer another customer back to it.

When both Noise Metallic and Power Influence are marginal or unacceptable, suspect a grounding, bonding, or associated power company problem (open capacitor bank, bad transformer, open neutral, etc.). These are problems for a transmission team. But record the test results. The records of all field tests allow computer analysis of the approximate area of the problem, and the trouble can quickly be pinned down by the transmission team. They wouldn’t have the slightest clue as to where to start without the technician’s original input.

One more thing: while at the customer’s, check the network interface ground. Use the transmission test set to measure the resistance between the CO ground and the protector ground. The station ground should test 25 Ohms or less to assure the carbons at the protector will fire when power or lightning is present on the circuit, and protect the customer and station equipment from damage. If the test shows resistance greater than 25 Ohms, remake the ground.

Circuit quality is an individual responsibility. No one but the field technician, standing at the network interface, can guarantee that the service is acceptable. If this is understood in all levels of outside plant responsibilities, long-term and quality customer service can be guaranteed.

As Yogi Berra said, "It ain’t over until it’s over." The same applies to getting a customer’s service restored. A quick fix or a "found OK" will suffice for a few hours, days, maybe weeks, but it’s likely going to fail. A few extra minutes spent scoping transmission characteristics of the circuit will end repeat dispatches on the same trouble.

Signing off
Noise is one of the bugaboos that keep us all up at night. It’s like trying to kill that gopher that is destroying your yard but you can never quite find him. I’d like to hear about your noise challenges: [email protected] or 831.818.3930.