2 Ethernet cables combined

[kennyprost]

So I ran cat5e to different rooms of my house so that they can all connect to my router from a central point. Pretty simple there. For the sake of not having to do it again if a cable breaks, I ran 2 individual wires to each room.

Here is the question:

If I connect both cables to a single wall outlet jack, cables in parallel essentially connected to each other, would that set up the redundancy or would I see loss in performance?

I could always leave one wire disconnected at both ends, but I figured if both ends were connected in parallel, if even 1 wire in one of the cables broke, the other would still be pushing info.

Thoughts?

 

[Someone Somewhere]

That would cause massive packet loss and maybe total link failure, and likely still wouldn't work even if one of the cables was chopped, before you even consider things like shorts.

Just have two jacks on each plate.

 

[bill001g]

I agree just install 2 jacks to start. It is so trivial to plug into the other jack if you would get a failure. It is extremely rare to get any kind of wire failure after cable is installed. If anything has issues it is almost always the connection to the jack at the end of the cable not the wires in the wall.

Ethernet actually uses tiny voltage differences between the pair of wires to do the signalling. The pair twisted together allows this to work, it has a lot of issue if you use wires from other pairs that are not twisted together. I am pretty sure it would have issues with a second pair of wires connected because the 2 pair would be just slightly differnet.

 

[Mousemonkey]

Or just have a single socket and coil the spare in the back box.

 

[Someone Somewhere]

PM I received - please post to the thread:

Well that wouldn't be a short since they are the same wire sending the same information. Really no different than twisting two copper wires to run power through them. A short would be two different wires sending different signals. At Least from an electrical standpoint. However, I'm still not sure if the two wires would still ha e the packet loss even though nothing is different between the two wires. I could see how two different lengths may result in a confusion due to travel time by the two wires.

I was talking more about what happens if someone drives a nail through one of the cables and shorts two of the conductors together - that absorbs (and reflects) the signal and will stop it from reaching the other wire.

Look up characteristic impedance. Two pairs in parallel will have a different impedance to one pair, and this will cause reflections both injecting signal in, and extracting it out. Plus, some signal will come down one cable and, instead of entering the device, will travel back up the other cable.

Gigabit has a wavelength of about 2m. A length difference of a tenth of that is likely to cause major issues reading the arrived signal. That's pretty hard to avoid.

AC reflects off short and open circuits not terminated with the right resistance - and by paralleling them, that resistance changes. For mains, this isn't a problem - the wavelength is thousands of kilometers long, so every part of the system is pretty much at the same point in the cycle. Not so for high-frequency; it works with a completely different set of rules.