With encircled flux, things are finally getting real

From the December, 2013 Issue of Cabling Installation & Maintenance Magazine

Commercial availability of EF devices brings the topic from the theoretical to the practical, and highlights the need for associated fiber cleaning and inspection.

by Patrick McLaughlin

It has been more than three years since the Telecommunications Industry Association (TIA; www.tiaonline.org) adopted the TIA-526-14-B standard. Technically, the standard is titled Optical Power Loss Measurements of Installed Multimode Fiber Cable. But for all the other information it contains, TIA-526-14-B has become known as the "encircled flux" standard. And if the mythological Helen of Troy was the face that launched a thousand ships, then the very real TIA-526-14-B has been the standard that launched a thousand discussions, debates, comments, articles, blog posts and the like throughout the cabling industry. That has been the case because turning the theory of the encircled flux (EF) launch condition into the practical application of EF in field-test instruments has been "easier said than done," as the cliché goes. And it has taken most of the three-plus years since the standard's adoption to reach the point of EF's field application. With EF-compliance devices now commercially available, the questions are likely to shift from "What?" to "How?"

But the "what" is worth revisiting to trace the steps that have gotten the industry to this point. What is EF, and why should anyone care about it? It begins with the "why." The intention of the EF specification is to ensure that regardless of the light source used in a light-source/power-meter test setup, the test results on any given multimode link will be consistent. Inconsistencies result from the optics used in the light source. Many light sources using light-emitting diodes (LED) emit what is called an overfilled launch--so named because the LED-emitted light more than fills the multimode fiber under test. It emits more light than the fiber core captures. When certain VCSELs are used in the light source, what is called a slightly underfilled launch is created. The VCSEL source almost but not completely fills that multimode core with light. An underfilled launch is created by some VCSELs or by an edge-emitting LED.

These light sources, of course, are used in combination with power meters to test the insertion loss of fiber circuits. Because the launch conditions of the light sources can vary like this, the resulting insertion loss measurements are very likely to vary as well. In other words, if you tested one multimode circuit three times, using the same power meter on one end all three times but switching out the light source on the other end each time, there is a decent likelihood that you'd get three different insertion loss results. Nothing changed about the circuit itself between these tests. Its real performance didn't suddenly get better or worse. But the test results would indicate it did just that--got better or worse--because of the different launch conditions within the different light sources.

This situation with light sources is nothing new. It has been known for some time, and the idea that insertion loss results can vary by as much as 1 full dB, depending on the light source used, was simply tolerated. Generally speaking, it was tolerated because for many users a 1-dB margin of error in insertion-loss performance was not going to make or break the network's performance. That's changing. It already has changed for many. When speaking particularly about Ethernet, as transmission speeds have gone up, allowable loss has gone down. For 1-Gigabit transmission in the short-wave, 850-nm operating window, the loss limit is 4.5 dB over Om³ and 4.8 dB over Om⁴. When you move up to 10-Gigabit, it's 2.6 dB over Om³ and 3.1 dB over Om⁴. And the early adopters of 40/100-Gigabit must meet loss budgets of 1.9 dB over Om³ or 1.5 dB over Om⁴.

Standard and TSB

So the long-existing situation of varying test results coming from different light sources could no longer go unaddressed. And at the heart of these variances is the light source's launch condition. So in 2010 the TIA took action to address the situation by publishing TIA-526-14-B, titled Optical Power Loss Measurements of Installed Multimode Fiber Cable. The standard was back-adopted from IEC 61280-4-1. In practical terms that means the standard was adopted word-for-word. In IEC 61280-4-1, and now also in TIA-526-14-B, the EF launch condition is defined and referenced. The TIA standard does include a foreword, which provides additional information to what is in the IEC spec.

When the TIA adopted that standard, it solved everything, right? Not exactly. That adoption took place approximately three years ago and EF remains a much-discussed but not-very-well-understood phenomenon. For one thing, EF itself is a measurement of the light source's launch condition. Beyond that, when initiated, EF was aimed at laboratory test equipment as opposed to equipment used in the field. Its transition from lab testing to being a condition of field testing has proven to be an interesting ride so far.

This year, 2013, was the year in which EF went from the lab to the field, with the commercial availability of devices that, when used in conjunction with light sources, bring them into EF compliance. Because the mere appearance of these devices did not answer every single question or clarify every detail, the TIA issued Telecommunications Systems Bulletin (TSB) 4979 in August 2013. The TSB is titled Practical Considerations for Implementation of Encircled Flux Launch Conditions in the Field.

Like many standards and TSBs from the TIA, TSB-4979 contains a foreword. The final paragraph of that foreword provides a clear delineation of responsibility for installers and technicians who conduct multimode-fiber field testing. It reads: "It is not the intent of this TSB to provide users a prescriptive method for implementing EF compliance, nor is it required that they construct equipment to produce or verify the launch condition. Rather, it is the responsibility of manufacturers of test equipment to offer devices that have been built and tested for EF compliance."

With that understanding of what TSB-4979 is not, let's turn attention now to what it is. Its main thrust is the description of two separate types of controllers that, when used with light sources, bring those sources into EF compliance.

Universal and matched

TSB-4979 describes a "universal" controller as intended for use with what is called a "legacy" light source, for which the launch type is not known. As the name "universal" indicates, this type of controller can be used with any light source. As for what a universal controller is as a physical device, the TSB uses the term "black box" for a device to which input and output connectorized cords are attached. The black box serves to redistribute then filter the modes that are emitted by the legacy light source. It is through that redistribution and filtering that EF compliance is achieved. So the universal controller is a single item that comprises this black box with connectorized input and output cords.

A matched controller, on the other hand, is used with a light source in which the source characteristics are well-controlled. And the name has meaning here too. The "matched" part means that a certain light source model can be used with a certain launch-cord model to achieve EF compliance. This opens the door for users of this test equipment to be able to choose from multiple vendors, provided of course the light source of choice and the launch cord of choice are "matched" appropriately.

Because the light source in this scenario is matched up with a certain type of cord, the light source itself is part of the controller. With the universal controller, what is defined as the controller excludes the light source. With the matched controller, the definition includes the light source.

The TSB includes practical considerations for using each of these controller types. The size and weight of the universal controller are noted. It is heavier than the typical launch cord and light plastic mandrel that have been commonly used. Also--not unique to the universal controller, but noted in the TSB-- the controller is rated for a certain number of matings. What that really means is, if you follow that to the letter, it's a consumable. The TSB both acknowledges and cautions against using a controller past its useful life, for the very purpose of ensuring test-result accuracy.

Specifically, the TSB says, "It may not be practical to verify EF compliance at the output of a launch cord after repeated use. Such validation may include testing by near-field scanning measurements or using an insertion loss artifact as suggested in TIA-526-14-B. However, at the time of this writing, no manufacturer has commercialized an artifact to check EF compliance in the field. In general, users of test equipment should follow the recommendations of the original supplier of the universal controller with regards to validation of the launch condition at the output of the launch cord after repeated use." In other words, it is not safe to trust a controller for more than the specified number of matings--typically 500.

But the TSB doesn't end with that statement. It adds, "While users may ignore that a launch cord has a limited life, it is still considered an expendable item. The universal controller, because it is more complex, is not as likely to be readily discarded."

TIA cabling standard documents do not consider or discuss pricing of systems or components. The reference to a universal controller being a "more complex" device equates to a more expensive device. Hence the acknowledgement that users will want to maximize the return on that investment, raising the likelihood that these devices will be used for more than the manufacturer-recommended number of insertions.

The TSB continues: "These controllers may be non-adjustable so rework is unlikely. However, several connector reterminations may be possible. If a connector becomes damaged and requires retermination, the universal controller needs to be recertified by the original supplier or by any supplier capable of measuring and certifying EF over a wide range of source mode power distributions."

Standards in motion

To use a bad pun, the state of multimode fiber-optic testing can be described as being in flux. At the same time that commercially available EF-compliance devices are coming onto the market, the TIA is in the early stages of revising the standard that contains the EF requirement. In July the TIA issued a call for interest for the revision of TIA-562-14 from its current "B" version to a "C" version. When issuing the call for interest, the TIA said the revision process will include an effort "to modify the current foreword in the standard from adoption to adaptation of IEC 61280-4-1-ed2 for regional variances. These variances would change presently normative aspects of the standard to become informative."

At least on its surface, the change from normative to informative appears to be noteworthy. A user must comply with all normative portions of a standard in order to claim adherence to it. The user does not have to comply with informative portions to claim adherence. As is the case with many standards-related developments, the practical impact of these potential changes will be determined over time.

Broader perspective

What already has been a long journey to this point of EF compliance in the field is going to get even longer, as standards-development processes move along and as more methods come onto the market to achieve compliance. Along with the challenging technical questions for everyone involved, the issue raises difficult questions for the "commercial food chain" of cabling-system manufacturers, installation contractors and end users. Will contractors purchase and use EF-compliant test sets before their end-user customers demand they do so? And when will user organizations begin demanding such test-launch conditions?

Take a step back and consider why cabling-system testing is carried out to begin with, whether the system is optical or copper based. Of course it ensures the installed system achieves the electrical-performance levels necessary to successfully support the user's applications of choice. As mentioned earlier, in the realm of optically based Ethernet, as speeds increase, cabling-system performance must increase also in the form of lower insertion loss levels.

But as a practical business matter, installed systems are tested because the manufacturers of those systems require test reports in order to issue a warranty. That requirement means end users include testing in project contracts. To this point, that lever has not been pulled with EF-compliant testing. Generally speaking, manufacturers of fiber-optic cabling systems are not requiring their systems be tested using EF-compliant launch conditions. And how could they, with the field application of EF being scarcely defined? Now that the proverbial dust is settling on that front, it may be sooner rather than later when manufacturers begin requiring EF-compliant test results before issuing a warranty.

Remember, the reason EF came into the picture to begin with was the inconsistency of test results from different light sources. Imagine a scenario in which a user organization or an independent third-party test technician (or both) gets different test results than those generated by the installation contractor. The contractor's test results are called into question, yet are repeated when the contractor is called back on-site to spot test some circuits. It is well within reason to believe that in this scenario, at some point the cabling manufacturer will be brought into the process to determine the cabling's role in these inconsistencies. This very realistic scenario--caused by light-source inconsistencies but requiring cabling-system manufacturer resources to resolve--no doubt has played out numerous times. It is in the best interests of cabling manufacturers for test results to be consistent, and as of today that consistency best can be achieved through EF.

Cleaning and inspection

The dual topics of cabling-system manufacturer test requirements and the shrinking loss budgets necessary to achieve high-speed transmission like 10-, 40- and 100-Gbits/sec bring to light the importance of the inspection and cleaning of fiber-optic connections. For example, when preterminated fiber-optic cabling systems were being introduced, they were positioned as plug-and-play systems. Having been terminated and tested at the manufacturer facility, these preterminated systems--most or all of which included multimode fibers terminated to multi-fiber, MPO-style connectors--held the promise of quick turnup. The user simply plugged the system in and it was ready to operate without the need to field test the links. Manufacturers guaranteed their performance.

As these systems have been gaining increasing deployment, instances have occurred in which some did not in fact perform as expected when plugged in. In most of these cases, upon receipt of the returned systems, manufacturers determined the culprit behind the inadequate performance was contamination of connector endfaces. Lost in the fact that insertion-loss testing could be skipped was the reality that inspection, and cleaning when necessary, should still take place.

In August 2012 the TIA published TIA-568-C.0-2. The 2 at the end of the document name indicates it is the second addendum to the 568-C.⁰ standard. The document is titled General Updates, and covers a number of items. Of specific interest to this discussion is Annex E of TIA-568-C.0-2, which covers field testing of optical fiber cabling and addresses connector endface quality. That annex references the standard IEC 61300-3-35, Examinations and Measurements--Fibre optic Connector Endface Visual and Automated Inspection.

The section of the annex titled "General" explains that it, the annex, "describes field-testing of length, optical attenuation and polarity in optical fiber cabling using an inspection microscope, OLTS, OTDR, a visible light source such as a VFL, and cleaning supplies." It also states that its purpose is to clarify rather than replace standards TIA-562-7 and TIA-526-14-B. TIA-526-7 addresses the testing of installed singlemode fiber-optic cabling.

Elsewhere, the annex makes several direct references to the aforementioned IEC specification. For example, it says that test jumpers, and ports under test, should be clean and free from damage in accordance with IEC-61300-3-35. One precaution in the annex is, "Ensuring that all connectors, mating adapters and test cords are clean in accordance with IEC-61300-3-35 prior to and during the test measurement." Another precaution specified in the annex is, "Systematically using an inspection microscope to verify connector quality per IEC-61300-3-35."

The section of the annex called "Optical Link Attenuation" includes the statement, "To verify that cords are in acceptable condition, first inspect the connector endface per IEC-61300-3-35. … each connector should be inspected per IEC-61300-3-35 prior to being connected, each time they are mated."

The IEC standard specifies pass/fail requirements for endface quality and how to verify that the requirements are met. It includes different criteria for different connection types, such as APC and UPC, as well as multifiber connectors like the MPO. In keeping with the notion that documented evaluation of performance is desirable for the end user as well as other involved parties, using automated inspection equipment enables a technician to produce such documentation. Several inspection sets are programmed to provide pass/ fail results.

In a sense, both the fiber-test and fiber-inspection technologies discussed here have evolved from the manual to the automated. For years fiber technicians have been accustomed to using a mandrel to strip out higher-order modes when testing installed multimode fiber. The need to address inconsistent light sources brought about the now-available EF compliance devices, which obviate the need for the mandrel. Likewise, the growing industrywide awareness of the need to inspect and potentially clean fiber connections before every mating also has raised consciousness about the value of documented inspection. When TIA took the opportunity to reference IEC 61300-3-5 in Addendum 2 to its 568-C.⁰ standard, it literally put in writing what already was a prudent practice.

One reason the inspection of fiber connections is a prudent practice is that a single contaminant, even one not visible to the naked eye, can significantly adversely affect one or more fiber circuits. For example, when the 12 fibers in an MPO connection make physical contact, the impact can break apart a contaminant and spread it across several of the fibers in the connector, bringing down or at least slowing down multiple connections. In extreme cases, even an unseen contaminant can be large enough to cause air gaps between two fibers that should be mated. These compromised connections are practically impossible to detect by any means other than a highly magnified visual inspection.

For both issues--EF and fiber cleanliness--the final chapter remains to be written. But for installers, technicians, network owners and consultants who specify fiber-installation projects, decisions must be made and actions must be taken every day to ensure the best possible results of a fiber-system deployment. Today, EF compliance and documented inspection results are marketplace realities.

Editor's note: The text of this article largely comes from a web seminar titled "Fiber-Optic Cable Cleaning and Testing Practicalities," which was delivered on November 6. AFL (www.aflglobal.com), EXFO (www.exfo.com) and JDSU (www.jdsu.com) sponsored that seminar and provided some of the technical information that was delivered in it--and included in this article. The author thanks those organizations for their contributions. Any errors or omissions are the work of the author, not those organizations.

Patrick McLaughlin is our chief editor.

TIA standards documents

This article, and the web seminar upon which it is based, heavily reference standards documents produced by the Telecommunications Industry Association (www.tiaonline.org). Although the article quotes directly from TSB-4979 and Annex E of TIA-568-C.0-2, those and other TIA standards documents are copyrighted.

The author wishes to take this opportunity to thank the TIA for the association's willingness to allow quoting from their documents, and for the cooperation that has enabled Cabling Installation & Maintenance to provide news and articles on completed and in-progress TIA standards documents.

TSB-4979, TIA-568-C.0-2, and all other standards documents produced by the TIA are available from IHS Global. That organization has established a specific Web page called the "TIA Standards Store," where users can search for and purchase TIA standards documents. The TIA Store is located at this URL: http://global.ihs.com/?rid=TIA