NEXT failing at the panel end

Likely causes

• Untwist and bad dress at the termination

  ~40%

  Deficit sits right at the connector on the failing end, and the failing pair combination tells you which two pairs were fanned out and dressed side by side at the punch.

• Patch cord (channel) or worn permanent link adapter tip (PL)

  ~20%

  The fault follows the cord or the adapter when you move it to a known-good link. Costs nothing to prove, so prove it first.

• Component category mismatch

  ~15%

  Fails across the band with a wide deficit and looks similar from both ends. Read the part number on the jack and the panel: the jacket print on the cable tells you nothing about the connector.

• Jack not fully seated, or punch cap not fully closed

  ~10%

  Margin moves when you press on the jack, and repeat tests give different numbers. A solid fault does not wander.

• Cable crushed or over-dressed at the termination

  ~10%

  Long stripped sheath, kinked cable behind the panel or a tie wrap you can't spin the bundle inside. Return loss is usually marginal on the same end.

• Split pair

  ~5%

  Deficit is double-digit dB across the whole sweep, ACR-F collapses with it, and the wiremap calls the split. Different animal. Go to Split pair, continuity fine, NEXT gone.

What to bring

• Certifier with both permanent link and channel adapters
• Spare adapter tip
• Two certified reference cords
• Punch tool with correct blade
• Jack instruction sheet (phone)
• Flush cutters
• Inspection light

Steps

1. 1

   Step 1: Swap both patch cords, or move the adapter tip

   Certified cords, as short as practical. On a permanent link test, fit a fresh adapter tip. Adapter plugs are rated for a limited number of insertions (brand-specific, so get the figure from your tester's documentation) and they drift long before they fail outright.

   If that doesn’t do it

   Margin unchanged: the fault is in the fixed link.

2. 2

   Step 2: Check the reference and the selected limit before anything else

   Reference set with the exact adapters in use and inside your tester's interval, correct limit, correct configuration. One wrong setting fails a whole building the same way.

   If that doesn’t do it

   Setup is right and only some links fail: keep going. If every link fails identically, go to Whole job failing, or passing too well.

3. 3

   Step 3: Read the crosstalk-vs-distance trace if your tester has one

   This is a vendor diagnostic, not a standards-defined result: names differ (crosstalk analyser, HDTDX-class trace) and lower tiers don't have it. Spike at 0 m is the end you're plugged into, spike at link length is the far end, elevated the whole way is cable or bundle.

   If that doesn’t do it

   No such feature on your unit: test from both ends and see which end holds the deficit, substitute a known-good jack at one end then the other, and use the TDR length and any return loss anomaly distance to place a mid-run fault.

4. 4

   Step 4: Test from the other end

   NEXT is measured at both ends. Whichever end holds the deficit is the end you open. Saves opening the wrong one.

   If that doesn’t do it

   Both ends show a similar deficit: suspect components, split pair or cable grade rather than dress.

5. 5

   Step 5: Open the failing end and inspect the dress

   What good looks like: sheath stripped only as far as the jack requires, pairs twisted right up to the punch, no combing or fanning, cap fully seated. Cat5e tolerates roughly 13 mm of untwist; Cat6 and 6A go by the jack's sheet, usually far tighter.

   If that doesn’t do it

   Dress is clean and it still fails: verify the components.

6. 6

   Step 6: Re-terminate to the manufacturer's instruction sheet and retest

   Unplug the switch port first. One re-termination, done properly, resolves most of these. If margin improves but still fails, you are chasing a second cause on the same link.

   If that doesn’t do it

   No improvement: check the jack and panel part numbers against the class you are certifying to.

7. 7

   Step 7: Verify every component in the link is rated for the class

   A Cat 6 connector is only specified to the Cat 6 band; it cannot meet Class EA / Cat 6A limits. The deficit shows above that band in whichever parameter that connector's compensation is weakest, often return loss or ACR-F as much as NEXT. Re-punching cannot fix it, only the right class part can, and the same link may pass cleanly against the Cat 6 limit, which is why the report header matters.

   If that doesn’t do it

   Components correct, both ends clean, still failing across the band: suspect cable grade or a pinch and re-pull.

More Copper & Certification guides

• Wiremap fail, crossed or reversed pairs
• Split pair, continuity fine, NEXT gone
• Whole job failing, or passing too well
• Comes up at 100 Mb, not 1 Gb
• PoE device won't power, or reboots under load

All 54 guides