The handover binder had 384 clean test reports in it, every link a pass. Eleven months later a conference room went down, and the investigation ended with a permanent link that had never met specification — it had been tested as a channel, with high-quality patch cords at both ends, and the report said pass because the cords carried it.
Nothing was falsified. The wrong test was run, and the paperwork looked perfect.
Three Activities, Three Different Questions
The word "test" covers three distinct activities, not tiers of the same thing. Each answers a different question.
Activity | Question it answers | What it measures |
|---|
Verification | Is this cable connected correctly? | Basic continuity: wiremap, toning, and sometimes length by TDR |
|---|
Qualification | Can this cable support my desired network technology? | Whether the link supports 100BASE-TX, VoIP, Gigabit Ethernet and so on |
|---|
Certification | Does this cable comply with cabling standards? | Full parameter sweep compared against TIA or ISO limits |
|---|
The distinction is easy to demonstrate. Two cables can both pass a verification wiremap test, and qualification testing can then show that one supports only 10BASE-T while the other handles Gigabit Ethernet. Same wiremap result, entirely different cable.
Only One of Them Produces a Pass or Fail
Certification tools are the only tools that provide pass or fail information in accordance with TIA or ISO standards. Qualification tools, however sophisticated, do not perform certification, and cable manufacturers do not accept them for warranty purposes.
The commercial consequence is easy to miss at quotation stage. If you want a warranty backed by the cable manufacturer, the certification tool has to be one the manufacturer approves. Anything else leaves the installer holding the liability alone.
Fluke Networks puts the exposure in numbers: final cost per link averages about $100, so a 1,000-link job is roughly a $100,000 project, and a Category 6 installation commands about a 20 percent premium on cable and terminations, taking it to around $120,000. Being the only party standing behind that result is an expensive position.
Verification and qualification tools typically test the channel; certification tools can also test the permanent link — the part a contractor installs and cannot change later.
Certification Itself Has Accuracy Grades
Certification is not a single standard of accuracy, and the required accuracy scales with the cable category.
Under ANSI/TIA-1152, Level IIe is required for up to and including Category 5e, Level III for Category 6, and Level IIIe for Category 6A. The A revision added Level 2G for Category 8. IEC 61935-1 shares those values but adds Level IV for Class F to 600 MHz, with later editions extending to Level V for Class I and Class II.
Category 8 pushes further still, needing testers capable of 2000 MHz for 25 and 40 Gbps transmission. So a tester can be perfectly functional and still be the wrong instrument: a Level III unit testing a Cat6A link is not producing a certification.
The mandatory report set from Category 5e through 6A covers wiremap including shield connection where present, insertion loss, length, NEXT at both ends, PSNEXT, ACRF, PSACRF, return loss at both ends, propagation delay and delay skew.
Five Ways a Report Full of Passes Means Nothing
This is the part worth taking to a project meeting, because every item below produces green results on non-compliant work.
Testing the channel instead of the permanent link. A channel runs from active device to active device and is limited to 100 m — up to 90 m of cabling and no more than 10 m of patch cords. The permanent link is the fixed portion, up to 90 m. Compliant patch cords on a compliant permanent link always produce a passing channel, so testing only the channel with high-quality cords lets permanent link problems go undetected. Patch cords are moved, replaced and handled more than any other component, which is why they are the weakest link. Test the permanent link.
Testing a modular plug terminated link with the wrong far-end adapter. An MPTL terminates the horizontal cable in a plug that goes directly into a device — common for PoE lights and cameras. Putting a channel adapter at the far end excludes the mated connection from the test, producing optimistic results and hiding problems with the field-terminated plug. The correct setup is a permanent link adapter at the near end and a patch cord adapter at the far end.
Never agreeing on which parameters are being tested. Standards specify the basic set, but the parameters actually tested are set by the customer’s project specification — and a cable plant warranty brings the manufacturer’s list into play. One major tester’s default limit is essentially the PoE set including DC resistance unbalance; the extended limit adds TCL and ELTCTL. Agree the list before the tester is configured, not after 384 links have been run.
Not enabling plot data. Plot data is the graphical record of every measured parameter and is required for certification testing. It is also the only practical way to diagnose a failure, because time domain data shows *where* the crosstalk or return loss is excessive. Reports without it look empty to a customer, and a support engineer asked to analyze a failure will send the job back for retesting.
Accepting a marginal pass without agreeing what marginal means. A marginal pass sits closer to the limit than the tester’s own published accuracy. TIA and ISO standards treat any marginal pass as a pass and compliant, and require it to be marked with an asterisk — the honest signal that measurement accuracy is now part of the answer. Any tester that lets you disable the asterisk is non-compliant. Unless the contract states in writing that marginal results are unacceptable, a marginal pass still passes. Settle that in writing rather than in an argument at handover.
Alien Crosstalk: Required, Optional, or Already Covered
Two reputable sources describe the same standard differently, and both are right because they answer different questions.
Belden’s reading of ANSI/TIA-1152 is precise: alien crosstalk is not on the list of standard parameters to be reported. The standard provides annex methods that *can* be used in the field, requiring two sets of testers with a communications link between them. Fluke’s reading is equally precise from the compliance side: for Category 6A, alien crosstalk — power-sum ANEXT and power-sum AACR-F — is the only way to demonstrate compliance for 10GBASE-T, and most vendors will not grant a system warranty without it.
The reconciliation: the standard’s required *report* set and the obligation to *demonstrate application compliance* are different things. Where alien crosstalk data is required, sampling is allowed: an equal number of short, medium and long disturbed links, stopping once three of each exceed a margin of 5 dB. Two selection rules decide whether the sample means anything — the disturbed link should be surrounded by connectors above and below rather than sitting at the end of a row, and the disturbers must be in the same bundle, because alien crosstalk across bundles is not considered significant.
One trap is specific to shielded cable. An open, unconnected shield can fail an alien crosstalk test, and the usual continuity check will not catch it, because a signal will find any path to the remote unit — including through the building’s common ground to which racks and patch panels are bonded. The tester reports a connected shield even when it is open.
Fiber Has Its Own Version of Every Trap
Reference methods, not parameters, decide fiber results. Standards recommend the one-jumper reference for Tier 1 testing because it includes the connections at both ends. The two-jumper alternative includes only one end connection and can produce overly optimistic or even negative loss results — and many vendors reject two-cord reference results, which can cost you the warranty.
Reference cords are consumables with a specification: reference-grade, better than 0.1 dB multimode and 0.2 dB single-mode, verified every 288 tests so a rising failure rate can be attributed to the link rather than the cords. Multimode also requires encircled flux compliance for vendor warranty. And on end faces, contamination remains the leading cause of fiber failures — IEC 61300-3-35 supplies the grading criteria so inspection is not a matter of how tired the technician is.
How to Specify
Name the activity, not just the outcome. Write "certification to ANSI/TIA-1152 Level IIIe, permanent link" rather than "test all links." That single change prevents the channel-versus-permanent-link substitution.
Agree the parameter list, the test limit and the marginal-pass rule before testing starts. All three are contract terms rather than field decisions.
Require plot data in the deliverable. It costs nothing at test time and it separates a diagnosable report from a number.
Match the tester’s accuracy level to the category. Level IIIe for Cat6A is a floor, not a target.
Agree the alien crosstalk position explicitly — in scope, out of scope, or covered by design.
Our test and inspection process covers assemblies and patch cords against the parameters the project specifies, with results available on request, and the certifications we hold are listed by standard. The Cat6 premises cable and Cat8 ranges are built to the category they are sold as, since no test report can certify a cable that was not built to the limit. How Ethernet cables work explains the construction those tests are measuring. If a report is not telling you what you need, send us the summary page — the test limit, adapter type and configuration usually say what was actually run.
Frequently Asked Questions
Does a qualification test count as certification?
No. Qualification tells you whether existing cabling supports a network technology. It produces no standards-based pass or fail, and manufacturers do not accept it for warranty.
Why test the permanent link instead of the whole channel?
The permanent link is the fixed portion you install and cannot easily change. A passing permanent link plus compliant patch cords produces a passing channel; the reverse is not guaranteed.
What does the asterisk next to a pass mean?
That the result is marginal — closer to the limit than the tester’s published accuracy. Standards treat it as a pass, require it to be marked, and prohibit hiding it. If you need marginal results excluded, that belongs in the contract.
The Short Version
Verification asks whether the cable is connected correctly, qualification asks whether it supports your network technology, and certification asks whether it complies with standards. Only certification produces a standards-based pass or fail, and only certification from a manufacturer-approved tool protects the installer rather than leaving them liable.
Certification has accuracy grades — Level IIe for Cat5e, Level III for Cat6, Level IIIe for Cat6A, Level 2G for Cat8 — and the wrong grade is not a certification at all.
Then configuration decides whether any of it means anything. Test the permanent link, not the channel. Use the right far-end adapter on a modular plug terminated link. Agree the parameters, the test limit and the marginal-pass rule before you start. Enable plot data. Treat the asterisk as information rather than an inconvenience. Settle the alien crosstalk position in writing, and if you test it, remember that an open shield can pass a continuity check through the building ground.
And on fiber, use the one-jumper reference, verify reference cords every 288 tests, and inspect end faces against IEC 61300-3-35.
Sources
- Fluke Networks — Verification Qualification and Certification, covering the three hierarchical groups of test tools and the distinct question each answers, verification tools as basic continuity checking including wiremap and toning with optional TDR for length and switch detection, the visual fault locator as a fiber verification tool, qualification tools as determining whether cabling supports specific network technologies such as 100BASE-TX, VoIP and Gigabit Ethernet, the worked example of two cables both passing wiremap where qualification shows one supports only 10BASE-T and the other Gigabit Ethernet, the statement that qualification tools do not perform the certification required by cable manufacturers, certification tools as the only tools providing pass or fail information in accordance with TIA or ISO standards, the requirement for a manufacturer-approved certification tool to obtain a manufacturer-backed warranty and the consequent placing of sole warranty responsibility on the installer otherwise, the ability of certification tools to test the permanent link where verification and qualification tools typically test the channel, the average final cost per link of about $100 and the $100,000 figure for a 1,000-link installation, the roughly 20 percent premium on higher quality cable and terminations for a Category 6 installation and the resulting $120,000 project value, and the conclusion that a technique short of standards-based certification makes the installer liable for the performance of that installation — https://www.flukenetworks.com/knowledge-base/cableiq/verification-qualification-and-certification
- Fluke Networks — Top 10 Cable Testing Mistakes to Avoid, covering mistake one on agreeing test parameters including insertion loss, return loss, near end and far end crosstalk and the alien crosstalk requirement for Category 6A, the optional balance parameters of DC resistance unbalance, TCL and ELTCTL, the default PoE test limit including DC resistance unbalance and the extended limit adding TCL and ELTCTL, the statement that the parameters tested are ultimately set by the customer’s project specification and that a cable plant warranty brings the manufacturer’s requirements into play, Tier 1 optical loss testing with an OLTS as required by standards and providing the most accurate insertion loss measurement over the link with Tier 2 OTDR testing as extended testing that still requires an OLTS for application compliance, mistake two on marginal test results including the marginal pass region defined as a result closer to the test limit than the accuracy published by the field tester manufacturer, the statement that any marginal pass is still a pass and compliant under TIA and ISO/IEC standards and must be marked with an asterisk, the statement that a tester allowing the asterisk to be disabled to hide marginal results is non-compliant, and the guidance to agree upon marginal test results up front including keeping firmware current, servicing and calibrating the tester and ensuring permanent link adapters are not worn out, mistake three on alien crosstalk including power-sum ANEXT and power-sum AACR-F as the only way to demonstrate compliance for 10GBASE-T and the withholding of system warranties without it, the permitted sampling approach recommending an equal number of short, medium and long disturbed links and the acceptance of a margin exceeding 5 dB on three of each as grounds to stop testing, the selection rules that disturbed links should be surrounded by connectors above and below rather than at the end of a row and that disturbers must be in the same bundle because alien crosstalk across bundles is not considered significant, and the shielded cable trap in which an open shield can fail an alien crosstalk test while the continuity check still reports a connected shield because a signal will find any path including through the common building ground, mistake four on testing the channel instead of the permanent link including the 100 meter channel limit comprising up to 90 meters of cabling and no more than 10 meters of patch cords, the permanent link as the fixed portion up to 90 meters, the statement that compliant patch cords on a compliant permanent link always produce a passing channel and that testing only the channel with very high quality patch cords allows permanent link problems to go undetected, the patch cord as the weakest link because it is handled and replaced most often, and the guidance to test the permanent link as the true foundation of the network, mistake five on modular plug terminated links including the definition as a horizontal cable terminated at one end to a plug and plugged directly into a device as an option where an outlet and equipment cord are impractical or unsafe, the common applications of PoE lights and surveillance cameras, the error of using a channel adapter at the far end which excludes the mated connection from the test and produces optimistic results while overlooking problems with the field-terminated plug, and the standards requirement for a permanent link adapter at the near end and a patch cord adapter at the far end, mistake six on enabling plot data including plot data as the graphical depiction of measured parameters required for certification testing, the graph axes of decibels against frequency with the frequency range depending on category covering up to 100 MHz for Category 5e, 250 MHz for Category 6 and 500 MHz for Category 6A, the representation of test limits and per-pair results, the diagnostic value of time domain crosstalk and reflectometry data in locating where crosstalk or return loss is excessive, and the observation that reports without plot data will look empty to the customer, mistake seven on fiber inspection and cleaning including contaminated connections as the number one cause of fiber network problems and failures, the risk of particles on the fiber core causing loss and reflections, the unreliability of manual inspection by microscope given experience, ambient lighting, eyesight and fatigue, and the IEC 61300-3-35 standard providing cleanliness grading criteria based on the number and size of scratches and defects, mistake eight on the one-jumper reference method including connector loss as the loss of a mated pair of connectors, the practical impossibility of measuring a single connector, the recommendation of the one-jumper method to include the loss of connections at both ends and deliver the highest accuracy, the two-jumper alternative referencing out the connection between the two jumpers and including only one end connection, and the risk of overly optimistic or negative loss results and vendor rejection of two-cord reference results preventing warranty acquisition, mistake nine on reference cords including the requirement to use manufacturer-recommended test reference cords with loss better than 0.1 dB for multimode and 0.2 dB for single-mode, the recommendation to verify reference cord performance every 288 tests and document the results, and the encircled flux requirement for multimode testing to reduce measurement uncertainty and satisfy vendor warranty requirements, and mistake ten on relying on a duplex tester for MPO certification including the about 15 steps required for a three-jumper method on 12-fiber MPO connectors, and the complexity, time and inconsistency removed by testers with an on-board MPO connector — https://www.flukenetworks.com/blog/cabling-chronicles/top-10-cable-testing-mistakes
- Fluke Networks — ANSI/TIA-1152-A: Requirements for Field Test Instruments for Twisted-Pair Cabling, covering the 2007 decision of the TIA TR-42 Engineering Committee to begin the ANSI/TIA-568-C series, the 2008 agreement to remove field test instrument specifications from that series into the standalone ANSI/TIA-1152 document, publication of the A revision in November 2016, the standard defining what tests should be run and how accurate they should be with different accuracy levels called out for different cabling categories, the addition of Level 2G accuracy for Category 8 installations in the A revision, the DSX-5000 meeting the Level IIIe accuracy level used for Category 6A cabling and the DSX-8000 additionally meeting Level 2G for Category 8, and the international equivalent IEC 61935-1 sharing the same values for Level IIe, III and IIIe while adding Level IV for Class F testing to 600 MHz and, in later editions, Class Fa to 1000 MHz and Class I and Class II to 2 GHz with a Level V accuracy specification — https://www.flukenetworks.com/knowledge-base/applicationstandards-articles-copper/ansitia-1152-twisted-pair-field-testing-standard
- Belden — What ANSI/TIA-1152 Says About Category 6A Testing, covering the role of testing in verifying overall cable performance, the ANSI/TIA-1152 standard specifying the reporting and accuracy performance requirements for field testers of permanent links and channels for Category 5e through Category 6A, Level IIe performance required for up to and including Category 5e, Level III for up to and including Category 6 and Level IIIe for up to and including Category 6A, the required permanent link and channel tests and reported parameters comprising wiremap including shield connection where present, insertion loss, length, near-end crosstalk at local and far end, power-sum near-end crosstalk, attenuation to crosstalk ratio at the far end, power-sum ACRF, return loss at local and far end, propagation delay and delay skew, the definition of alien crosstalk as an unwanted disturbing signal coupled from one Category 6A component, channel or permanent link to another, the note that alien crosstalk is not on the list of ANSI/TIA-1152 standard parameters to be reported, the provision of methods in an annex that can be used to measure alien crosstalk in installed Category 6A cabling requiring two sets of testers with a communications link between them, and the statement that Belden Category 6A connectivity meets alien performance specifications by design so that alien field testing is not required to obtain the 25-year product warranty when the system is properly installed by a certified partner — https://www.belden.com/blog/ansi-tia-1152-and-cat-6a-testing
