
When a structured cabling installation is complete, the cable runs may be terminated, labeled and connected to the patch panels and outlets. The next question is whether the installed system meets the project’s performance requirements.
This is where field testing becomes important. A cable tester can report a pass, but the result only has meaning when the correct test configuration and limits have been selected. Testing the fixed cabling is not the same as testing the complete connection that will carry network traffic.
Understanding the difference between a permanent link and a channel helps contractors, system integrators and project teams interpret test reports correctly and avoid disputes during handover.
What a test report actually certifies
A field test report is a statement about a defined boundary, not a statement about “the cable”. Two boundaries dominate copper acceptance work, and the result is only meaningful once the boundary is named.
The permanent link is the horizontal cable from the patch panel port to the work area outlet, including the connection at each end and, where one is installed, a consolidation point. Patch cords and equipment cords sit outside it. This is the part built into walls and ceilings, and the part a tenant inherits and cannot cheaply replace.
The channel is the complete passive path: the permanent link plus the patch cord at the panel and the cord at the device end. It is what actually carries traffic.
The distinction is not bookkeeping. Cords are the section of a link that gets unplugged, moved, replaced and bought from a different supplier. Certifying them at handover certifies something the project does not control; excluding them certifies only the portion the contractor installed. Neither choice is wrong, but a report that does not say which boundary it covers cannot settle an argument about either. The two boundaries come with two length figures, and the arithmetic between them is where a lot of specifications go wrong.
| Boundary | What it includes | Typical maximum |
|---|---|---|
| Permanent link | Horizontal cable and one connection at each end | 90 m |
| Channel | Permanent link plus patch cords and equipment cords | 100 m |
The 10 m difference is the total allowance for cords at both ends, not 10 m per end. A 90 m permanent link leaves 10 m for everything plugged in at the two ends combined. A 95 m permanent link is already outside the model, whatever the cords measure.
That model is a design allowance for horizontal twisted-pair cabling, not a rule for every topology. It assumes cords of a construction comparable to the horizontal cable; an ultra-thin 32 AWG cord changes the loss budget and is not covered by the same arithmetic. Length also cuts both ways as an argument: 90 m is a limit, not a target, and a 70 m link can still fail insertion loss, because the category sets what has to be met and the installed condition decides whether it is.
Two tests, two setups, two limit sets
The test configuration is part of the test. A permanent link is measured with permanent link adapters and reference cords at both ends, and the instrument is set to the permanent link limit. A channel is measured with channel adapters and the patch cords that will actually be used, against the channel limit.
The two limit sets are not interchangeable. Permanent link limits are tighter because they exclude the loss the cords would contribute; channel limits are written to accommodate it. Run a permanent link measurement against channel limits and marginal links pass that should not. Run a channel measurement against permanent link limits and sound links fail. Either mistake produces a report full of numbers that mean nothing.
Each limit set is also tied to a category. A Cat6A permanent link is measured against Cat6A limits across the full frequency range that category defines, not against Cat6 limits with the upper range ignored. The category sets the envelope; the boundary sets where the measurement starts and stops. A report that names one category while applying another is worth less than no report at all.
A PASS is a comparison between a measured value and a limit at a frequency, under a specific configuration. It is not a verdict on the network, and it is not evidence about any part of the link that the configuration excluded. The instrument itself also has to be within calibration and meet the accuracy level the applicable field test standard requires for the category being certified; a tester that is out of date on either point produces results nobody can defend.
What to require, and where handover goes wrong
There is no universal answer, and pretending otherwise causes more disputes than it prevents. Three situations cover most projects. Permanent link certification is the usual basis for handing over installed cabling. It records what was built into the building, independent of cords that will be swapped within a year, and it gives the owner a baseline that can be re-tested later against the same boundary and the same limits. Every link can be re-measured years later, by a different contractor, with a comparable result. When a project is buying infrastructure rather than a working connection, this is the test that matches the purchase.
A channel test answers a different question: does the assembled path, with the cords in place, meet the limits? It fits projects where the cords are part of the supplied package, where the acceptance requirement is a working connection rather than a cable run, or where the owner wants evidence of what the tenant will actually run on. The result is bound to the cords present on the day. Replace one patch cord and the certificate describes a link that no longer exists.
Acceptance criteria come from the project specification, the applicable standard and the contract, not from a general preference for one test. Some projects require permanent link certification on every link plus channel tests on a sample; some require channels throughout; some accept either, provided the report states the configuration and the limit set used. The failure mode is a specification that never names the boundary and a contractor who picks the cheaper one.
Most test disputes at handover trace back to a handful of practical errors rather than to cable defects. They tend to surface at the point where retention or final payment depends on the report, which is the worst possible moment to discover that the report answers a different question from the one the specification asked.
- The wrong limit set: Cat6 limits applied to a Cat6A job, or permanent link limits left in place for a channel test.
- Adapter mismatch: channel adapters used for a permanent link measurement, or reference cords that were never re-referenced after a change.
- A channel report that outlived its cords: the link was re-patched and the certificate was not reissued.
- A PASS screenshot instead of a report: no margins, no parameter list, no tester model and no calibration date.
- Results with no link identity: numbers that cannot be tied to the outlet labels on the drawing are not evidence about any specific location.
A usable report carries the link identifier, the test configuration and limit set, the tester model with its calibration status, the adapter type, and every measured parameter with its margin, including the failures. The margin is the part a project team actually uses, because it says how much room the link has before the next change. A pass tells you that a link met a limit. Only the report tells you which link, under which configuration, and by how much.
