Frequency-based vs. application-based testers

Project requirements and capabilities of testers determine the right tool for the job.

Th 315049

Project requirements and capabilities of testers determine the right tool for the job.

Modern LAN cable testers can perform an astounding range of network tests, and produce pages and pages of data. But in a world where “time is money,” more is not always better. The industry has seen the development of simpler, cheaper, and more streamlined testers that focus on Ethernet testing—the most common commercial and residential network transmission technology. The rise of these new “application-based” testers has raised important questions: Do they satisfy industry standards? How do they differ from traditional LAN testers?

Though the technology is simple, the answers to these questions are not.

Th 315049
Click here to enlarge image

Field-testing telecommunications cables in the LAN environment developed partially as a result of the need to use existing cable infrastructure to support newly installed network servers and terminals. The most common tools in the arsenal were a break-out-box (BoB), volt-ohm meter (VOM) and, budget-permitting, a time-domain reflectometer (TDR). The BoB allowed technicians to re-map connectors with jumper wire and route the signals to get the system running. They could then make a permanent connector that duplicated the crossover pattern they had created with the BoB. The VOM measured voltages and checked for shorts or opens on their custom wired connectors. The TDR allowed the technician to determine the length of a cable, if it was terminated into an open, short, or load, and if midspan connectors were causing significant impedance mismatches. As helpful as these tools of the trade were for providing information about connec- tivity, none provided information about the ability of the cabling to support data transmission without error.

568-A brings costs, benefits

Industry efforts led to the adoption of the American National Standards Institute/Telecommunications Industry Association/Electronic Industries Alliance (ANSI/TIA/EIA)-568-A structured wiring standard for commercial buildings, which defined common or generic practices in the design, installation and testing of communications cabling.

The field-testing requirements revolutionized the standard. By having a set of performance requirements that could be tested in the field, installers could test the cabling they were installing to ensure it provided the performance the networking hardware required.

ANSI/TIA/EIA-568-A required technicians to measure near-end crosstalk (NEXT) and attenuation. Both para- meters were frequency-specific and required a field tester that could sweep a cable across a given frequency range, and provide test results for all pairs and their unique combinations. Existing testers evolved, and other field cable testers came to market from many now-defunct manufacturers, but the field-certification arms race had begun.

The cost of the arms race

Standards are living documents. Because technology never stops advancing, industry must also continue to update its guidelines to take advantage of new technologies. As a result, cabling grades quickly advanced from Category 3 to 4, 5, 5e, 6, and now 6A (16, 20, 100, 100, 250 and 500 MHz respectively). To remain competitive, tester manufacturers began introducing testers that swept to higher frequencies than the standards required (hence the arms race).

Today, in the parts of the world that follow TIA, Category 6A is superior. It requires the following tests to be run from 1 to 500 MHz: Wire map, length, delay and skew, attenuation (now called insertion loss), near-end crosstalk (NEXT), power sum NEXT, equal level far-end crosstalk (ELFEXT), power sum ELFEXT, return loss, attenuation-to-crosstalk ratio (ACR), and power sum ACR. Unfortunately, the circuitry required to perform these tests is not simple, nor can ready-made circuits be found readily obtainable. This means that engineers are tasked with devising very sophisticated test equipment that is affordable, durable enough to be used in the field, and can perform all the above tests quickly. By comparison, the reference equipment used in the lab to test network cables cost as much as $100,000, weighs dozens of pounds, is as big as a dorm-fridge, and can take up to an hour to test one cable.

Counting the costs

In addition to testing higher frequencies, LAN cable certifiers must also test a defined number of frequency points or steps. It is easy to forget that the original purpose of the ANSI/TIA/EIA-568-A standard was to provide a generic cabling infrastructure that was application independent.

Today, we in the LAN industry are only concerned with Ethernet in all of its various forms; however, the ANSI/TIA/EIA-568-B still requires the cable to be tested across a broad range of frequencies to support current and legacy voice and data signaling. This broad requirement of the standard keeps the development and manufacturing cost of LAN cable certifiers high. A LAN certifier limited to testing 1 Gigabit Ethernet (GbE) would be much simpler to design and build because it would need to test only a very small range of frequencies.

Finally, the accuracy requirements of the LAN cable certifiers drive the costs up. As a result of the thoroughness and accuracy of LAN cable certifiers, cabling system manufacturers are willing to extend warranties to installations that are certified to comply with the appropriate standard. While all LAN cable certifier manufacturers use third-party testing agencies, such as Intertek/ETL and Underwriters Laboratories, to verify their accuracy claims, most cabling manufacturers still have internal programs to audit the accuracy and approve a particular field tester for use in certification of its cabling systems. In many cases, it has become the job of the field tester to verify the headroom and performance claims made by the cabling manufacturers.

New products keep costs down

A question often arises: Is there a cheaper way to test the performance of LAN cabling? The answer, in many cases, is yes. It can be done by streamlining testers to test only the technology most often deployed by businesses—Ethernet. Virtually all of the critical voice, data, and video systems in use today are Internet Protocol (IP) or are rapidly becoming IP. Fast Ethernet and GbE are the standard deployment for data networks. IP phone systems, favored for voice and closed-circuit television (CCTV) camera systems, are quickly transitioning to IP.

One reason for the rapid deployment of IP systems is the low development cost of the devices. Unlike with custom-designed circuits, IP device engineers can simply choose from a huge number of off-the-shelf components to inspire their designs and create the software to make it all work. The use of off-the-shelf components and circuits allow manufacturers to take advantage of economies of scale.

Using this same logic, test equipment manufacturers are now developing application-based testers specifically designed to certify only Ethernet. Because these streamlined testers no longer have to verify that any application in existence can run over a given cable, many of the design limitations are removed and designers can employ off-the-shelf components—much like the designers of IP telephones, CCTV cameras, and network adapters. In short, by limiting the scope of field-testing to just one application, much of the cost and complexity is removed.

Today, the issue of alternative certification has not been entirely embraced by the industry. The greatest fear is that the term “certifier” has definite meaning and that the availability of two completely different types of field testers, both called certifiers, will cause confusion. Indeed, there is much debate about what a “real” certifier is. Convention in this industry is that a “certifier” refers to the type of equipment that performs a number of frequency-sweep tests on a cabling link to test its compliance with the complete ANSI/TIA/EIA-568 field testing standards, not just the interconnect portion.

The testers

There are currently four true ANSI/TIA/EIA-568 field certifiers in production, which provide complete radio frequency (RF) tests of the cabling. All of these certifiers sell for $5,000 or more, depending on the capabilities of the particular model in question. A number of manufacturers also sell alternative, cost- effective, application-based testers that may claim to certify a cabling link, but universally test the performance characteristics of LAN cabling. These certifiers cost between $1,000 to $1,500 for a copper-only version and $2,000 to $3,000 for a model that can test both copper and fiber.

Most of the world uses standards written by the TIA, ISO, and International Electrotechnical Commission (IEC), which all describe testing methods that span the cabling across a range of frequencies and aim to determine if the installed cabling actually meets the desired category rating.

Application-based testers do not comply with these standards and may or may not comply with other international standards when determining the pass/fail status of a cabling link. Application-based testers are useful for certain jobs. Some manufacturers choose to test to the Institute of Electrical and Electronics Engineers (IEEE) 802.3x standard. This set of standards provides the signaling requirements for Ethernet in all of its various flavors, and is a favorite for alternative certification since it does determine if a cabling link has “passed” or “failed.” This test is adequate for smaller commercial and residential installations because many building owners and tenants do not particularly care about the cable; they just want to know that data can be passed throughout the building without problems.

This type of practical application testing is less sophisti- cated than the analytical method used by traditional LAN cable certifiers, but it is useful in answering the most critical question: Can my network cabling pass data?

This form of application certification only applies to the data that is sent across the cabling during the test—in this case, Ethernet at 10-, 100-, or 1,000-Megabits per second (Mbits/sec) data rates.

It should be clear now that there are two basic methods of LAN cable certification available today. What is not clear is an absolute way to identify what type of tester you are considering purchasing or may have already purchased. The test equipment industry uses several different naming conventions, so a user must look at the features to determine what type of tester he or she has.

Which is best for you? First, determine the needs of your customer. The most important task for the cabling installer is to meet the obligations spelled out in the contract:

  • Does the contract make reference to accuracy requirements of the testers used to certify the cabling?
  • Does the installation need to carry a warranty from a system vendor?
  • Does the contract require data reports to comply with the requirements of ANSI/TIA/EIA-568-B, or that frequency and dB data for NEXT, return loss, ACR, etc., be reported?

If the answer to any of these questions is yes, then a traditional frequency-based certifier is required. Application-based testers do not provide this information or capability and cannot be used to satisfy the requirements of most commercial cabling contracts.

If there is no need to provide a vendor warranty, and the obligations of the cabling installer are simply to prove performance of the cabling and provide documentation of such performance, an application-based tester is perfectly acceptable.

An additional benefit that application-based testers provide, which traditional certifiers do not, is they can be affected by electromagnetic interference/radio frequency interference (EMI/RFI). A traditional frequency-based LAN cable certifier is designed to have very high common mode rejection (CMR) performance. This means the certifier does not “see” the effects of external noise on the cabling. Its receiver circuitry measures only the signals that its own transmitters apply to the cable under test. The effect is that even when cabling is installed in an electrically noisy environment, the certifier is not going to measure those links any differently than it would if those same links were installed in perfect laboratory conditions.

At first, this sounds like a flaw, but one must keep in mind the purpose of the test equipment. It is specifically designed to measure the internal effects of NEXT, insertion loss, return loss, and the litany of other parameters required by the TIA/ISO. It goes back to answering one question: Does this cabling link meet the requirements of the performance standard, regardless of its environment?

Application-based testers do not ignore the effects of external noise on the cabling. Because they monitor the flow of data packets across a link, any EMI/RFI influence that is bad enough to cause data corruption and retransmission of packets will be detected by the tester and reported to the user.

Going beyond the basics

Other advantages of application-based certifiers go beyond the initial testing of new cabling. For example, end users are always performing changes to their networks. An application-based certifier is the perfect low-cost tool to ensure that cabling can indeed perform as expected when changes are made.

One example is the rollout of GbE in existing networks. Over the years, many thousands of buildings were wired with Category 5 and 5e cabling, with the expectation that someday GbE would be introduced to all users in the network. GbE is designed to run over Category 5 and does not need 5e cable. Since the Category 5 standard is officially obsolete according to TIA and no longer acceptable for new installation certifications, many traditional LAN certifiers do not even have it as an option in their firmware anymore. Therefore, testing the Category 5 cable to the current Category 5e test limits is bound to cause problems.

While the intent in this case is to verify the ability of the cabling to support GbE, testing it with an application-based certifier will basically stress-test the cabling to confirm its ability to support GbE without excessive errors. If the old Category 5 cabling can handle the task, the certifier will pass it with flying colors. On the other hand, if for example, some of the links exhibit excessive error rates because they are too long and have insufficient ACR margins, the IT manager can flag those drops as being able to support only 10/100 Ethernet.

Making an informed, educated decision when choosing which field tester to use on a job will prevent problems and may even improve profitability by allowing lower-cost test equipment to be used on jobs that do not require traditional certification.

DAN PAYERLE is business unit manager for data communications test products at IDEAL Industries, Inc. (www.idealindustries.com).

More in Home