Skip to main content
IT BIBLESign in

Forum / General / Questions & Answers

Run passes tests but only links at 100 Mbps

RackGhostDwayne OP
#1

The question: how do you tell a bad punch-down at the patch panel from a bad cable when a run beats a cheap continuity tester but the link only negotiates 100 Mbps? I have a method, but I would like to compare it with yours.

The background is simple. A basement run, about 40 feet, Cat5e, terminated at a keystone in a panel and at a wall jack. The tester lights all eight pins in order. Both ends still come up at 100 Mbps full duplex, never gigabit.

A continuity tester only proves that each pin reaches the far end. It says nothing about pair twist, a split pair, or a conductor that is touching the contact but not seated well. Gigabit uses all four pairs at once. 100 Mbps uses only two, pins 1, 2, 3 and 6. So a link that settles at 100 is often telling you that the orange and green pairs work and the blue or brown pair does not.

My order of work:

  1. Check the physical layer first. Swap the patch cords at both ends for known good ones. Cords fail more than people admit.
  2. Look at which pairs are in trouble. Pins 4, 5, 7 and 8 are the usual suspects.
  3. Re-punch the blue and brown pairs at the panel. Keep the untwist under half an inch, and use a fresh punch-down blade, since a worn one leaves a conductor half seated.
  4. If it is still at 100, put a short known good cable in place of the whole run, and see if the port negotiates 1000. That separates the switch and NIC from the wiring.

A tester that reports pair length or cable faults would settle this faster. I have not used the mid-priced ones enough to recommend a model, and they cost real money for something you may use twice a year.

What do you reach for first, the punch tool or the cable tester? And has anyone had a run that looked clean and still failed because of a bad keystone jack and not the punch-down?

Sign in to reply.