Eight continuity lights show that conductors connect from one end to the other. They do not prove correct pairing or gigabit performance. First confirm both ports support gigabit and negotiate automatically. Then substitute a known-good short lead, isolate the installed run, and check for split pairs, poor terminations and intermittent faults.
Link rate or internet speed: which should you check first?
Read the Ethernet link rate in the computer’s network settings or the switch’s port status. A broadband speed test also includes the router, internet connection and remote server. It cannot isolate the installed run or tell you which part of a cable is faulty.
A 100 Mb/s connection may be entirely normal: many cameras and other low-bandwidth devices have 100 Mb/s interfaces. Confirm the capabilities of both ends before changing the wiring. Record the switch port, device, link rate and whether the link drops.
How do you isolate the faulty part of the link?
Isolate the fault by comparing the same devices through a known-good short lead and then through the installed run. Change one component at a time, including patch leads and ports. Disconnect active equipment before passive wiremap tests, inspect the pairing, and repeat the original checks after any repair.
- Check the port settings. Confirm the device and switch support 1000BASE-T and use suitable auto-negotiation settings. Do not force a speed to hide an unresolved fault.
- Connect the same devices with a known-good short patch lead. If the link is still 100 Mb/s, investigate the ports, settings or device before blaming the installed cable.
- Put the installed run back into the path. Substitute its patch leads separately. This isolates a bad lead from a wall socket, patch panel or fixed cable.
- Disconnect active equipment before passive wiremap tests. Use the tester’s prescribed remote and follow its live-port restrictions. Photograph the result.
- Test pairing and inspect terminations. A simple LED sequencer may miss a split pair. Use an instrument explicitly capable of finding that fault.
- Retest after repair. Record negotiated rate, port error counters and a local transfer test under the same conditions.
Why can a split pair pass continuity?
Ethernet uses particular twisted pairs: pins 1–2, 3–6, 4–5 and 7–8. If the same wrong conductor arrangement is repeated at both plugs, all eight pins can still connect one-to-one. The required signal conductors are nevertheless no longer twisted together, which can increase interference and crosstalk.
A split pair is different from swapping the two wires within one pair, and different from crossing complete pairs. Compare the actual conductors with a correct T568A or T568B diagram, rather than judging only the LED order.
What if the wiremap really is correct?
1000BASE-T uses all four pairs, while 100BASE-TX uses two. An open pair can therefore prevent gigabit while still allowing some equipment to establish a 100 Mb/s link. Actual fallback behaviour depends on the devices and fault; it is not a universal diagnosis.
Also consider high-resistance contacts, excessive untwisting, a damaged conductor and an intermittent termination. A tester can miss a fault that is absent during its brief measurement. Inspect without pulling or disturbing in-service connections, and arrange a maintenance window for replacement tests.
Gough’s published case describes a link that dropped to 100 Mb/s even though a switch cable diagnostic passed. Further resistance measurements identified abnormal conductors. This is an external case, not a WireTrace UK measurement.
When should you move beyond a basic tester?
Use qualification when you need evidence that a link supports a stated network application. Use correctly configured certification when your project requires compliance with a cabling category or Class. A successful file transfer, a green link light and a wiremap PASS are useful observations, but none is a substitute for a certification report.
What else should you check before relying on the result?
A test result is useful only within the instrument’s documented capability and the setup used for that measurement. Check the exact model, permitted connection state and acceptance requirement before acting on a PASS. The questions below clarify common assumptions that can otherwise turn a useful observation into an unsupported conclusion.
Does a cheap tester always miss split pairs?
No. Some inexpensive verification testers explicitly detect split pairs. Check the exact model, minimum cable length and manual. Price alone does not define the capability.
Does 1 Gb/s negotiation certify the cable?
No. Negotiation is a link observation. It does not measure the full set of parameters required for cabling certification.
Which sources support this guide, and what are their limits?
These references support the distinctions and troubleshooting methods described above. Manufacturer capabilities apply to the named model and operating mode; external cases describe the author’s reported observations. Neither is presented as a WireTrace UK measurement. References were checked on 22 September 2026, and project requirements still govern any acceptance test.
- Fluke Networks — split-pair identification ↗
- Fluke Networks — cable verification, qualification and certification ↗
- Gough’s Tech Zone — a gigabit link falling back to 100 Mb/s ↗
KEEP A RECORD OF THE FIX
What should you record for the next visit?
Record the original symptom, equipment models, cable details, connection state and measurement method alongside the result. After a repair, repeat the conditions that produced the problem and save the retest. The blank field record helps keep expected behaviour separate from observed readings; it contains no example values that could be mistaken for measurements.
Download field record ↓