Today’s buzzword is bandwidth, and coaxing more from the local loop is a service provider’s preoccupation.
For years, telephone technicians have confidently approached phone-system troubleshooting with their trusty analog test sets. Also known as buttinsky or butt sets because they allow a repair person to butt into a phone call, these test sets have many features: touchtone and rotary dialing; last-number redial; a ringer with on/off; bent-nose bed-of-nails clips with a spike; durable butt-set cords; a data-safe monitor; a big, easy-to-push dial; moisture resistance; a large butt-set clip; a loud receiver; dual LED polarity test; and a mic mute button.
Some of the functions performed are common to all telephones, such as dialing and talking/listening. Other items address the lineman’s specific physical-layer test problems.
For example, bed-of-nail clips consist of alligator-like test clips but with many fine, sharp spikes. A service technician doesn’t have the time, nor does he want to strip insulation from wires so that he can connect to a loop pair. This type of clip pierces the insulation to make contact. The elasticity of the insulation causes the very small holes to close up after the clips are removed.
Similarly, a loud receiver means that you can hear the tone or conversation while the test set still is connected to your belt. Alternatively, some butt sets have a shoulder-cradle accessory that provides hands-free operation with the test set held near the user’s head. And big buttons are a must if you’re wearing gloves.
The tip and ring sides of the line are intended to be electrically positive and negative, respectively, but, inversions do occur. The actual polarity can be displayed by LEDs on the test set.
Of course, all phone lines aren’t carrying analog conversations these days. Some data-safe butt sets automatically determine that data is flowing on the line. The test-set talk function may be inhibited or, alternatively, the talk bandwidth limited to 3 kHz. The instrument impedance is high enough that simply connecting it won’t disrupt data on the line.
The Data Conundrum
High-speed data transmission implies a large bandwidth requirement. But the predominate access medium for exchange of data among the world’s telephone customers is the existing restricted-bandwidth, twisted-pair copper, local-loop wiring. That is the conundrum—how to squeeze a quart into a pint pot.
As a minimum, the FCC requires telephone companies to provide customers 2,700 Hz of bandwidth between 300 Hz and 3 kHz at a level from +4 to -9 dBm. The actual received level at the customer’s phone will be within a narrower range because the central office (CO) clips anything above 0 dBm.
Modems use a variety of encoding schemes to provide high data rates within the available voice bandwidth. Two-tone frequency shift keying (FSK) has a top data rate of about 1,300 baud. Faster modems, such as ITU-T types V.22 and V.27, use four- or eight-phase encoding of a 1,800-Hz carrier. Still higher rates are achieved by combining phase and amplitude modulation.
The fastest modem, type V.92, provides asymmetrical speeds of up to 44-kb/s upload from your computer and up to 56-kb/s download to your computer. A direct digital link from the CO to an Internet service-provider’s computer is required to achieve the 56-kb/s rate. However, fast modems may not deliver what you think you have bought.
A 56-kb/s modem needs a 3,600-Hz bandwidth, 33% more than the phone company is required to provide.1 Also, best performance comes with phone lines that have low noise, no bridge taps, no load coils, no wire gauge changes, and low attenuation. Customers close to a CO may achieve full-speed operation most of the time. Other users’ modems will revert to a lower speed if that is the maximum that can be sustained on their particular line.
Streaking Past Voiceband
The long time required to download a large file even at a 56-kb/s rate is a problem, especially now that very fast and relatively inexpensive desktop PCs are available. To remove the acoustic modem bottleneck, various digital subscriber line (xDSL) technologies have been developed. They don’t use the voiceband at all, instead transferring data from 100 to 1,000 times faster than V.92 modems by using the spectrum above 4 kHz.
Local-loop wiring never was intended to provide more than voiceband communications, yet it does so very well over reasonable distances. Similar restrictions apply to xDSL wiring as to wiring used with 56-kb/s modems: no bridge taps, load coils, or wire gauge changes are allowed. But, in this case, these considerations avoid impairments to the high-frequency characteristics of the line.
Additional restrictions result from the dual function being performed by the local loop. For example, xDSL cannot be combined with plain old telephone service (POTS) on loops longer than 18,000 ft because POTS requires a load coil on these loops.2
One popular type of xDSL is asymmetrical DSL (ADSL), which provides up to 6,144 kb/s downstream and 786 kb/s upstream. It operates in two distinct bands from just above the voiceband to 125.0625 kHz upstream and from 163.875 kHz to 1,104 kHz downstream.
Discrete multitone (DMT) modulation subdivides the total bandwidth into 256 separate, 4.3-kHz-wide tones. Each tone theoretically is capable of supporting up to 15 bits of information. The ADSL frequencies are separated from the POTS voiceband by a filter called a splitter.
There are many other initial letters to choose, such as H for high-speed DSL, S for symmetrical DSL, and V for very high-speed DSL. These and other types of xDSL services all have advantages in particular circumstances, but none of them can be tested with a butt set.
A data-safe butt set still is required to verify POTS operation. In addition, the field technician may need the following:
- A hand-held ADSL modem.
- A hand-held splitter to substitute for a suspected faulty unit.
- A test meter to troubleshoot inside wiring.
- A load-coil detector to aid in their location and removal.
- A short-range time-domain reflectometer (TDR) to locate loop impairments.
Telephone companies realized long ago that only automated testing could maintain a high quality of voice service for the huge number of installed loop pairs. Routine POTS testing is done at the CO by a mechanized loop test (MLT) system that continuously performs measurements on a succession of loop pairs temporarily taken out of service. Ideally, xDSL loop testing should have a similar automated capability.
xDSL Service Prequalification
Because the copper access medium is being used far beyond its POTS design specifications, impairments that cause noise can be very important. These include split pairs, untwisted drop cables, RFI from external sources, and crosstalk from other lines in the same 25-pair binder group or from other binders in a large cable. Due to these impairments, 10% to 20% of xDSL modems initially cannot link up with the corresponding DSL access multiplexer (DSLAM) termination in the CO.3
To avoid the cost of blindly configuring loop pairs for xDSL service only to find that many don’t work, loop prequalification is needed. The vision shared by communications companies is of a low-cost xDSL service that is easily, quickly, and inexpensively deployed. If prequalification could be automatically implemented entirely from the CO end of the loop, truck rolls and technician labor would be saved.
TDR Measurements
The June 2002 edition of the IEEE Journal on Selected Areas in Communications presented 21 papers related to aspects of twisted-pair transmission. In their article, Galli and Waring explored time-domain reflectometry as a means of determining the precise makeup of a loop. Reflectograms were presented as examples of the distinctive signatures caused by a gauge change, a bridged tap, or simply the end of the line.
If only a single impairment is affecting a loop and close to the CO, a good case can be made for automated analysis of the reflectogram. However, “…conventional metallic TDRs are not capable of detecting all echoes. In fact, conventional metallic TDRs cannot detect gauge changes and, moreover, have a serious range limitation that prevents them from detecting echoes farther than some number of kilofeet from the CO. This limitation is due to the presence of a slowly decaying signal, caused by the distributed RLC nature of the loop, that overlaps with and masks the echoes generated by impedance changes.”4
The authors proposed subtracting the overall slow decay signal from the TDR reflectogram to expose the underlying impairment echoes. A theoretically derived slowly decaying wave shape compared well to measurements made on an actual 30,000-ft loop. Using the subtraction approach, signatures from different types of single discontinuities 9,000 ft from the CO had distinctive signatures, a prerequisite for automated impairment recognition.
Transfer Function Estimation
In another article, Bostoen et al proposed using a network analyzer to determine the S11 parameter of a loop pair, the ratio of the reflected to the incident wave power. From knowledge of S11 and by using a five-parameter model of a twisted-pair telephone line, the transfer function of the actual line could be estimated. Theoretical and actual results were shown to correlate well for the cases of a single line, a homogeneous network with a bridged tap, and a cascade of sections with differing wire gauges.
A network analyzer applied a single-frequency sine wave to the loop and measured the S11 ratio as the signal frequency was swept over the range of interest. In their work, the authors assumed an additive white Gaussian noise model with a power spectral density (PSD) of N(f) = -130 dBm/Hz. The upstream and downstream PSD masks were as defined in the standards, -38 dBm/Hz and -40 dBm/Hz, respectively.
The focus of the paper was to determine channel capacity rather than impairment location. “It has been shown by simulation that the relative error on the ADSL capacity of a 26 AWG line is below 1% up to 4,000 m and below 3% up to 5,000 m if one assumes a measurement noise level of -90 dB and a simple noise model….”5
Actual Loop Prequalification
Several approaches to loop prequalification are being pursued with a few already in place. However, the situation has been complicated by the FCC decision in June 2000 that requires incumbent local exchange carriers (ILECs) to unbundle the xDSL-related spectrum of a local loop. In simple English, this means that a competitive local exchange carrier (CLEC) now is allowed to provide xDSL services on the same line used by the ILEC to provide POTS voice service. Providing xDSL and POTS on the same line is called DSL over POTS regardless of which carrier supplies which service.
Only by close cooperation can an ILEC and collocated CLECs arrange their equipment to support loop sharing and retain full test access. A splitter is used on each xDSL line to separate the low-frequency POTS signals from the high-frequency xDSL signals: POTS signals going to the usual Class 5 Switch and xDSL signals going to the DSLAM. If the splitter is not optimally placed in the signal path, high-frequency loop signals can’t reach the test equipment, nor can the equipment apply them to the loop.
One solution is the Smart Splitter developed by Turnstone Systems that combines traditional splitter functionality with metallic test access. “A service provider can…see through the splitter for DSL-specific testing to the subscriber. The provider also can see back from the splitter to determine whether CO wiring is connected correctly and verify that both voice and DSL services operate correctly.”6
In another example, Tollgrade Communication has developed the DigiTest xDSL Line Qualification System to interoperate with Lucent’s MLT equipment. Claimed to have a large installed base of more than 140 million POTS lines, the Lucent MLT system provides the CO metallic-loop access upon which both DigiTest and the Lucent-developed LoopCare™ xDSL Test Software are based.
The Pragmatic Approach: Self-Installation
Prequalification and up-to-date line-test database information will help service providers satisfy their customers’ need for wideband xDSL. Customer frustration, the cost of truck rolls, plus the amount of technician time used during xDSL installation should be reduced.
Given that many loops can carry xDSL services without modification, self-installation is progressing in parallel with prequalification testing. In the past, CLECs were required to provide xDSL over a separate line, often of unknown capacity. But with xDSL over POTS, self-installation is worth a try on a known-good POTS loop.
“A self-install kit typically consists of a CD-ROM that automatically configures the customer’s PC for service and a modem that plugs into a universal serial bus (USB) or Ethernet port. With a well-designed kit, most customers can set up their own service in a few minutes.
“The ILECs have been deploying DSL over POTS since DSL’s beginning. Today, ILECs report that as many as 95% of new sales opt for self-installation and that approximately 80% to 90% of those customers succeed.”7
The role of the technician and hand-held test equipment won’t be eliminated. There will remain marginal service problems, and if demand for xDSL should grow very large, currently unusable loops will have to be conditioned and brought into service. New local loops will be required with greater demand.
As CO prequalification equipment is installed, compatibility between remote testers and the centralized test facilities will become more important. For example, a technician in the field should be able to query the wiring plant database and select and run tests from the CO using a hand-held terminal as a controller. It also should be possible to verify correct reception at the subscriber’s premises by using a terminal in a DSL-emulation mode.
The clear trend, however, is to drastically reduce the need to send technicians into the field for all but maintenance and repair work. CO-based loop testing is being extended to include xDSL parameters as well as POTS, with the test results used to update the wiring plant database. Self-installation helps to reduce the cost of on-site turn-up and generates revenue quickly from those loops already capable of xDSL service.
References
1. www.tschmidt.com/writings/POTS_Modem_impairments.htm
2. Galli, S. and Waring, D., “Loop Makup Identification Via Single Ended Testing: Beyond Mere Loop Qualification,” IEEE Journal on Selected Areas in Communications, Vol. 20, No. 5, June 2002, p. 933.
3. Bostoen, T. et al, “Estimation of the Transfer Function of a Subscriber Loop by Means of a One-Port Scattering Parameter Measurement at the Central Office,” IEEE Journal on Selected Areas in Communications, Vol. 20, No. 5, June 2002, p. 936.
4. Galli, S. and Waring, D., op cit., p. 930.
5. Bostoen, T. et al, op cit., p. 946.
6. “Effective Loop Management and Testing for Residential DSL,” TeleChoice and Turnstone Systems, 2000, p. 12.
7. Ibid., p. 5.
Return to EE Home Page
Published by EE-Evaluation Engineering
All contents © 2002 Nelson Publishing Inc.
No reprint, distribution, or reuse in any medium is permitted
without the express written consent of the publisher.
November 2002