Featured Articles
Announcing Pi
Technical Specifications
Raspberry Pi FAQs
Win a Pi
Raspberry Pi Wishlist
Actions
Forum Thread Details
  • State Not Answered
  • Replies 8 replies
  • Subscribers 678 subscribers
  • Views 2006 views
  • Users 0 members are here
Related

Using 4-pin Ethernet on the Pi 3?

Former Member

Hey everyone. Just bought a Pi 3 recently. I am trying to use the 10/100 Ethernet on the Pi, but the cable I am using is a four wire cable. 10/100 should only use 4 pins, however, so this shouldn't be a problem. My MacBook can use the internet fine with this cable, but the Pi fights with me. When plugged in, the green light on the port flashes quickly, continuously, and my router never seems to assign the Pi an IP. The Pi shows no signs of an outgoing connection, either, even though my terminal says that it recognizes the 10/100 link has been connected.

 

Any reason the 4 pin ethernet isn't working? And possible solution? The cable actually has six wires, but I have only connected the four used in 10/100 ethernet wiring.

Parents
  • 0

    Hi Preston,

     

    How long is the cable? It may be a very marginal connection that just-about works with the Mac. 4 wires is okay, but the cable needs to be twisted in specific pairs. Without the twists, you may only be able to go a metre or less (I've never measured it, but 1 metre sounds about realistic). You could twist yourself, but the amount of twists is important too (as is wire thickness and insulation material). All these factors determine how effective (in terms of error) and how far it can go. For these reasons, it is strongly advised to use the normal network cable i.e. Cat5.

Reply
  • 0

    Hi Preston,

     

    How long is the cable? It may be a very marginal connection that just-about works with the Mac. 4 wires is okay, but the cable needs to be twisted in specific pairs. Without the twists, you may only be able to go a metre or less (I've never measured it, but 1 metre sounds about realistic). You could twist yourself, but the amount of twists is important too (as is wire thickness and insulation material). All these factors determine how effective (in terms of error) and how far it can go. For these reasons, it is strongly advised to use the normal network cable i.e. Cat5.

Children
  • 0 in reply to shabaz

    Just FYI: I don't think the number of twists is that important. What IS important is that the four pairs in a cat5 cable each have a DIFFERENT twist frequency. If the twists-per-meter are the same, the pairs can influence each other. By having the twists-per-meter different this effect quickly cancels itself out: If there is a positive effect during a few turns that will be compensated by a negative effect a few turns later.

     

    @preston: four wire cable is often "cat3". And that is not twisted. That's bad at 100mbps. A client of mine wired a whole building with CAT5, but used the wrong pinout. So effectively the pairs weren't pairs. That worked at 10mpbs but not at 100. (Actually it worked at 100, but lowsy, worse than at 10).

  • 0 in reply to rew

    Hi Roger,

     

    Exactly as you say, cable length (and hence overall number of twists) is unimportant, I'd meant amount of twists for the cable length that he had. With no twists* at all for his length, he'd be lucky to achieve a metre perhaps (I once deliberately untwisted and separated a network cable to see what impact there was - I did need to separate perhaps 50cm to a metre of the twists before the eye-pattern completely degraded into complete uselessness, and probably high bit error prior to that).

     

    However since he mentions 200ft further below, he's probably got some older standard as you say or generic 'phone extension' grade twisted wire which is still not as controlled as Cat5.

    Preston, you really do need Cat5 over such a distance, otherwise you're at risk of either flakiness (or like complete unreachability) or high bit error which will give you network connectivity but will result in poor throughput.

     

    *well, at a first level explanation to help the OP's practical situation; no twists could be fine providing they were well defined in their parallel-ness which is not the case here.