1
Background
I recently had an electrician run a cat6 ethernet cable through the walls of my house. It is terminated with standard cat6 faceplates, the cable is shielded and the shield is earthed. It passes a continuity test.
Symptoms
Further testing has revealed intermittent a drops in performance (topping out at speeds of 300 Mbits/s) and in the worse case auto-negotiating at 100BASE-T (topping out at 90 Mbit/s, clearly not negotiating to gigabit) These tests were performed over several different days using iperf with two machines connected via the LAN. Tests on most days are typically as expected (speeds of around 800-900 Mbits/s)
Cable condition
A fluke microscanner does not detect any issues with the cable, but having inspected the terminations I can see that roughly 1-2inches of twisted pairs have been untwisted up to the contacts and no attempt has been made to isolate each wire up to the contacts (i.e. some wires cross past others at some points)
Questions
- Given the intermittent nature of the fault, is it reasonable to assume that the symptoms are triggered by an external factor (mains interference or possibly EM interference?)
- Could a poor termination (as described above) lead to a cable that is more highly susceptible to external interference, despite it's shielding throughout the rest of the length of the cable?
- What is the best way to create conditions under which the cable would consistently perform poorly so that I can confirm that a re-termination has resolved the issue?
Note, the cable is plastered in to the wall, so there are naturally a limited number of re-terminations that are now possible!
Red Taz
Posted 2019-07-18T17:47:08.233
Reputation: 143
"the cable is plastered in to the wall" Oh dear. You should have run cables through conduit or drop spaces. – DavidPostill – 2019-07-18T20:32:01.530
1Fluke makes a lot of different products with a lot of different capabilities. Exactly which Fluke product did you use, and which tests did you run? Did you visually verify that both ends are terminated according to TIA-568b (or both 568a)? I ask because a split pair could definitely cause this. Is the cable less than 100m in length? Untwisting of twisted pairs at the ends can decrease the cable's resistance to interference and crosstalk, but it's not the first thing I'd look at. – Spiff – 2019-07-18T21:08:58.617
If it's not terminated properly this can happen. The pairs need to be terminated with less than an inch of the pairs exposed from the shielding. You may also want to check to make sure this electrician did infact ground the shielding properly. – Tim_Stewart – 2019-07-19T16:41:04.237