DSL loop test

Digital subscriber line offers many benefits to carriers. It uses existing wiring to deliver high-speed data and voice services. Yet it also presents challenges-the largest of which is outside plant prequalification.

Before DSL can be deployed, local loops must be tested to see whether they can support it. This requires detecting load coils and measuring loop length as well as bridge taps and wideband noise in some cases.

The significant number of DSL services the industry projects means that automated testing will be crucial to handling the growing line volume. Today's manual methods for testing loops simply will not keep pace with doubling and quadrupling line volumes.

Testing methodologies Prequalification testing can be done in two ways-single-ended and double-ended.

Single-ended testing requires test equipment at only the central office or digital loop carrier, whereas double-ended testing requires equipment at both ends of the loop.

For single-ended testing, a carrier can test a copper pair from the CO or DLC without involving on-site technicians (Figure 1). These tests also enable pre-service loop testing from a central location and provide information on loop repairs such as removing load coils.

Double-ended testing involves dispatching a technician to the customer's location to install a modem or test equipment that communicates with the reference modem in the CO. If the service doesn't work, a work order is issued to clean up the pair.

Double-ended testing works well, but the extra expense of dispatching the technician is difficult to justify, particularly for a local exchange carrier marketing a data service that's competing with cable modems.

Prequalification tests should be executed either in a batch mode or on-demand.

The batch mode is used to profile an office or area to determine which subscribers can be offered DSL service and estimate the amount of work required to condition the routes to support DSL. The on-demand capability is used to determine if a customer's line can support DSL transmission for data or voice/data services. The predominant POTS-testing operation support systems can provide either batch or on-demand testing.

Useful testing parameters in single-ended DSL loop prequalification are detecting load coils, measuring line length, detecting bridge taps, ensuring spectral compatibility, conducting metallic tests and determining longitudinal balance.

Each area has its own complications. Load coils are added to subscriber loops longer than 18,000 feet to improve voice quality. But when they are added, the attenuation in the voice band is reduced while that of the higher frequencies is dramatically increased. Because DSL transmission techniques rely on frequencies above the voice band, the DSL cannot be transmitted on a loop with load coils. Load coils impair 33.6 and 56 kb/s modem operations for similar reasons.

The accepted practice for engineering local loops is revised resistance design. These rules specify that all loops longer than 18,000 feet must be loaded.

A 1983 Bellcore subscriber loop survey estimated that 24% of loops are loaded but only 12% are 18,000 feet long-implying that only 12% of loops should be loaded. This discrepancy can be attributed to the installation of new COs or DLCs, which increased the number of loop reconfigurations. LECs' outside plant records are usually inadequate in finding load coils because they don't reflect the age of the plant and the significant number of plant changes made in recent years.

DSLs are sensitive to the line length and the gauge of wire between the CO or DLC and the customer. DSL transmission rate is inversely proportional to loop distance. The legacy POTS testing systems measure loop length using simple capacitance measurements.

The recent introduction of rate-adaptive asymmetrical DSL and the trend toward offering lower bit-rate services have helped increase the acceptable line lengths. The wire gauge is also important. Typically, the distance supported for 26 American wire gauge is two-thirds the distance supported for 24 AWG.

Although the specifications for DSL transmission are typically given for 24 or 26 AWG cable, it is difficult to determine what percentage of the loop plant is of uniform gauge, assuming that the smallest gauge for a given length is consistent with revised resistance design and yields the slowest transmission speed (lower bound) that distance can be supported.

Bridge tap detection A bridge tap is any unterminated portion of a loop not in the direct talking path between the CO or DLC and the customer premises equipment. As the outside plant has evolved, cable splices and cable-pair swaps, combined with poor documentation, have made locating and removing bridge taps time-consuming and costly.

Bellcore reported last year that 56% of the loop population has bridge taps. This percentage alone is not reason for concern because a bridge tap's length and location-not its mere presence-are the primary factors that determine whether DSL transmission will be impaired. In only 5% of loops with bridge taps would the tap adversely affect DSL transmission, according to a report in the December 1995 IEEE Journal. And some of the newer DSLs and 2B1Q-based technologies claim to be unaffected by bridge taps.

Spectral interference or crosstalk can affect DSL throughput. Crosstalk can be caused by insufficient cable shielding, excessively large disparity between signal levels in adjacent circuits, unbalanced lines and overloaded carrier systems.

Crosstalk can be categorized two ways: near-end crosstalk and far-end crosstalk (Figure 2). Near-end crosstalk occurs when a transmitting signal affects a receiving signal on the same end of the cable. Far-end crosstalk occurs when a far-end signal affects the near end.

Far-end crosstalk is usually less damaging because the signal is attenuated as it traverses the loop. Repeatered T-1 circuits and ADSL transmission usually cannot be mixed in the same binder group, but ADSL, high bit-rate DSL/multirate rate DSL and ISDN can be freely mixed. Some of the newer DSL technologies claim to be non-interfering in binder groups and are unaffected by spectral disturbers.

Wideband noise measurement can help identify spectral interference and isolate DSL transmission problems. The ability to recognize signature frequencies will equip the LEC with more information to aid in loop conditioning. T-1 circuits or other DSLs can be identified by looking for their characteristic frequencies (Table 1).

Prequalification testing should include standard metallic fault testing. The quality of the OSP can be determined with the standard tip-to-ring, tip-to-ground and ring-to-ground parameters, including DC voltage and resistance, along with AC voltage, resistance and capacity.

The loop's longitudinal balance measurements are important to ensure that DSL services are efficiently delivered. An unbalanced circuit caused by unbalanced longitudinal currents or power line harmonics can cause crosstalk noise, causing bit errors that slow the DSL throughput.

Two aspects of access raise questions: * Where can a test device establish metallic test access?

* Which OSS has the information to identify the access point?

Because the majority of copper pairs in a residential area carry switched voice traffic, the metallic test access points for pre-service qualification are located at the switch or DLC. To prequalify existing residential loops, the most prolific access point is the switch.

The number test trunk access at the switch places the highest restrictions on the testing that can be performed. The bypass-pair access on DLCs should not restrict tests, but it adds considerable complications for the testing operating system controlling the metallic test access.

The access methods listed above assume that the pair being tested is already providing service-in other words, spare pairs cannot be tested without manual access at the main distribution frame. These tests, which provide pre-service qualification, require metallic test access with full splitting capability on the customer's line. After a DSL service has been provisioned, the types of test access expand to include number test trunk, DLC metallic test access unit (MTAU), DSL access multiplexer (DSLAM) and external MTAUs.

Access at this point limits the tests outlined above. The limitation arises because the switch-test access would most likely be on the CO side of the POTS splitter, and the frequency range for the test access is typically limited to the voice band, less than 4 kHz (Figure 3). This will not support detection of bridge taps and wideband noise measurements.

DLCs The bypass-pair test access for switched voice services would be used for pre-service qualification, which is controlled by the pair-gain test controller under the direction of the test operating system.

A different test access is necessary for in-service testing because most legacy test operating systems do not support ADSL testing. Some DLCs provide a metallic test access on the customer's side of the POTS splitter. The access could be controlled by the DLC's internal MTAU using Transaction Language 1 (TL1) commands with the access point identified.

This type of access would be useful for in-service testing. Some DSLAM manufacturers provide a metallic test access capability. There are no standards for controlling this test access.

This type of access would be useful for in-service testing. The external MTAU could gain access using TL1 commands and be wired into DSL circuits at the main distribution frame. These devices must support the entire frequency range required in DSL in-service testing.

>From the perspective of testing and monitoring, operations support systems can be broadly classified as supporting two types of circuits-designed (special services, T-1 circuits) and switched.

For services targeted at residential users, the predominance of pairs in service-such as pairs the LEC would like to prequalify for DSL-will be supporting switched services.

In today's environment, switched services are tested with POTS test operating systems, using number test trunk access at the switch or the bypass pair of a DLC. The number test trunk access is "addressed" using the telephone number, while the bypass pair is controlled by the pair gain test controller.

For designed circuits, a work order record, design document is created in TIRKS that will list all the equipment in the circuit. These documents should contain the necessary test access information before the circuit is provisioned.

If LECs were treating DSL services as special services, the test access information would be required for all devices that address an access point, such as next generation DLCs, MTAUs and DSLAMs.

LECs require single-ended test equipment that can be controlled by the legacy testing operation system, providing the automated ability to prequalify copper loops to support DSL transmissions. The test solution should cost-effectively identify obstacles to deploying DSL-based services over traditional copper telephone lines.

The solution should provide single-ended prequalification for both the CO and DLC environments without involving on-site technicians to perform tests. To meet DSL deployment goals, the LECs must have an automated solution to pre-qualify loops quickly and cost effectively.