Avalon Networks Avalon Networks
Copper Wire Tests

Copper Wire Tests

Copper Wire from the Telco
Our testing was done on copper facilities obtained from our ILEC (USWest) under the LADS (Local Area Data Service) tariff. Iowa tariff #3 part 105.2.5 specifies: "These circuits are furnished on either two-wire or four-wire non-loaded facilities" This tariff however does not address bridge taps.

Location of Facilities
Avalon's POP is located across the street from the USWest central office. Despite this, there is still about 1800 feet of copper between our machine room and the MDF (main distribution frame) at the CO. This test was done using newly installed facilities at our office. For reference, we paid $11,000 to have 200 new pairs installed at our office.

Test Wire Plant Description
This trial was done on 24 different LADS loops.
Endpoint Locations Average Wire Length Check Wire Continuity
Continuity of all wire pairs was confirmed using a Progressive 200EP tone probe and tone source. It may seem like an unnecessary step, but in our experience, the telco had a difficult time successfully installing LADS circuits. 20% of our lines lacked continuity on installation and several more lines were not tagged at one end or the other. Additionally, after having uswest fix the original nonfunctional lines, we found several of the previously working lines were no longer end to end. This experience suggest we may have some non-trivial headaches during deployment.

Check for Load Coils
Load coils are devices that are placed in-line with copper telco lines to improve reach for voice circuits. Typically they're not found on lines less than 18,000 feet; however, given that telcos often splice into lines, it's hard to know what you might get by luck. In our case, the tariff under which we purchase our lines specifically states that the lines are free of load coils.

To verify that the lines were indeed free of load coils, we used a Progresive Model 88 Load Coil Detector. This is a simple device that will let you detect up to four load coils, although all you need is one load coil to render the line useless for DSL.

No load coils were found on our lines.

Length by Capacitance
A very useful feature of telco wire is that the capacitance is fixed at 83 nF per mile, irregardless of the wire type (PIC or Pulp), wire gauge, or temperature. (ANSI standard T1.601-1992 Annex G). This feature can be used as a method of measuring wire length. For this test, we used a Fluke Model 87 multimeter which has capacitance measurement mode. Lines were measured from the central office end with the far end open. The formula for distance then is: 
                                   capacitance (nF)     
		Length (feet) =   ------------------ * 5280
                                        83


Testing wire using Time Domain Reflectometers
In this trial, we used two different TDRs: Riser-Bond 1205T and Tektronix TS100.

For those who are not familiar with the theory and operation of TDRs, we suggest that you read Riser-Bond's online training materials or ask them for a copy of their "Product Catalong and Applications Guide".

Measureing Length with a TDR
When measuring any line with a TDR, it should be done with the far end open. TDRs in our experience only are useful out to about 8,000 feet - therefore to test longer lines, you will have to test from each end, and on some lines, it may not be possible to see the entire line. In our case, we were primarily interested in looking for bridge taps, and almost always the bridge taps occur close to the customer's demarc, so testing was mostly done from the customer demarcation point. Testing length with a wire that is less than 8K feet is easily done with a TDR. The end of the wire will be an upgoing peak as illustrated in the example below:



This example wire is about 1800 feet long.

Comparing Methods of Measurement
To establish that we had consistency in measurement methods, we compared three methods of line length measurement on several lines. We found that in all cases there was only a small percentage of error between the methods. As an example:
Line #10


Bridge Taps
Bridge taps refer to a situation where a lateral wire is connected to an existing cable forming a sort of "Y" in the cable. Typically the "bridge tap" goes to the customer while the original wire continues down the street and eventually terminates in an open. This causes a problem with ADSL modems by causing a null in the spectrum if the wire is short or a uniform 3.5 dB or so loss if the wire is long.

Bridge taps can be identified by using a TDR. An example of a line with a bridge tap is as follows:

In this example, the TDR was connected to the customer's demarcation point. There is a 459 foot segment of cable to the Y. The part labeled "Bridge Tap" is actually the lateral that runs down the street and terminates in an open. The start of the bridge tap is a sharp downgoing curve on the TDR display, and the end of the bridge tap is a matching rising spike.

Effects of Bridge Taps
The typical effect of a bridge tap is to cause increased signal loss in a particular frequency band. A formula for calculating the center of the first null is (provided by an engineer at Westell): 

                                        164
            First Null at (MHz) =   ------------
                                     Length (ft)
As an example of this using the Riser-Bond TDR, Line 22 looks like:



In this line we might expect to see signal loss at 164/759 = 0.21 MHz. In fact, we see a null at 0.19 MHz, but that may be because we chose the wrong velocity of propagation for the type of wire in this particular case.



This image deserves a bit of explanation. This is a screen capture from a HP spectrum analyzer. The top trace represents the training tones from a DMT ADSL modem. Each of the 250 or so spikes represents a channel. The lower trace is the same training tones across line 22, which is a 15,000 foot line. The first thing that you notice about this trace is that the signal becomes lost in the noise at about 400 KHz - this accounts for the slower speed at this distance - there is far less usable spectrum for the ADSL modem. The second thing we're showing here is a small dip at around 200 KHz (red arrow) that is caused by the bridge tap.

Inteferers
There were no other hi-cap services in the same binder groups as the LADS lines. However, we do have services in adjacent binder groups as follows: The following image is a noise curve from the CO end of a LADS circuit with no other ADSL units connected at our facility. The image was obtained using a HP spectrum analyzer looking from 10Khz 1 MHz. Primarily we see noise from the HDSL lines, but also interestingly, we see a spike at 780 KHz, which is a local AM radio station. This AM signal was even stronger at several customer sites.



Copyright (c)1998 Avalon Networks

Dave Lacey and Mike Lutz
To contact us check out the contacts page.