I made a DIY Cat 6 ethernet cable using Cat 6 23 AWG Solid Core cable & an RJ45 pass through connector that is known to work on a 23 AWG Solid Core cable.

I noticed today that my DIY cable gives me the expected 1Gbps speed when I connect it to a device using an Ethernet to USB C dongle (I used it to test all the Ethernet ports around the house), but whenever I connect it directly to my Desktop's motherboard's (ASRock X570 Taichi) ethernet port, the link speed would drop to 100 Mbps.

From what I know, this typically means the DIY cable was made incorrectly and I confirmed that a store-bought cable does work fine on the ethernet port on the Desktop.

I've also confirmed that the DIY cable passes the continuity test via a tester.

I've installed different types of LAN drivers for my mother board which didn't solve the issue.

So, to summarize:

  • I have a DIY Cable paired with an Ethernet - USB C dongle that gives me 1Gbps on any device.
  • I have a Desktop that gets 100Mbps speeds on my DIY cable when plugged directly into the motherboard's ethernet port but gets 1Gbps speeds with the same DIY cable when plugged into the USB C port via a dongle.

Honestly, the easiest solution is to probably just buy a pre-made ethernet cable.

Question: Is it possible for a DIY ethernet cable to deliver different speeds depending on the device it's connecting to? Both devices support Gigabit speed.

  • 1
    Is it possible for a DIY ethernet cable to deliver different speeds depending on the device it's connecting to? .... If a connection inside the plug is flaky, then yes.
    – John
    Sep 7, 2022 at 20:05
  • @John Just to clarify, do you mean the connection between the rj45 connector & the ethernet port from the device? If that's the case, maybe the rj45 connector I bought might be the problem (or just that specific connector I used.)
    – D_Pain
    Sep 7, 2022 at 20:34
  • 1
    That or the wire into the RJ45 connector. That was DIY and the wiring inside the connector could be flaky.
    – John
    Sep 7, 2022 at 20:36

1 Answer 1


Yes, it's possible for a flaky cable to work at gigabit speeds with one device and only 100Mbps speeds with another device. Different Ethernet NIC/dongle implementations may have differing levels of noise tolerance, so some may be able to handle the decreased signal or increased noise of an out-of-spec cable, whereas another device might not.

In my experience, the most common ways DIYers make out-of-spec cables includes:

  1. Split pairs. A typical continuity/pinout tester (the kind that lights up LEDs 1-8) cannot detect a split pair. Which pairs of pins are connected the the same twisted pair matters. Follow T568b on both ends. For example, note that pin 3 is paired with pin 6, not pin 4.
  2. Bad crimp job. If all 8 conductors didn't get all the way into the plug, or the teeth didn't get crimped down all the way, or if the person untwisted the pairs too far back from the connector (as much as possible, you want the twists to continue right up to the shortest possible distance from where they need to go parallel to fit into the slots below the teeth).
  3. Cable/connector mismatch. It sounds like you were mostly on top of this one. This is about making sure the grade and gauge and solid vs. stranded core design of the plugs matches the cable being used. Using the wrong connector design for the cable type often results in flaky connections.
  4. Using solid-core cable in living spaces (i.e. not buried inside the walls or ceiling or cable trays or conduits), where they get handled too much and break. Solid-core cable is great for long runs in the walls where it's not being touched, and it only gets connected to other cables. Stranded cable is strongly preferred for cables in human-accessible spaces, like patch panel cables and the cables you plug into your devices. This is because stranded cable is more flexible and less likely to break when getting flexed in various ways.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .