Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

Would someone explain whether using two pairs (four cables) stripped out of CAT 5, as opposed to one pair (2 single cables) for DSL connection, (from the box on the house to home, essentially replacing old CAT 3) would be better? In other words, if not better, what are the pottential problems? I started doing some research and ran into a lot of different terminologies such as:

resistance impedance cable type cable diameter voltage current skin effect coax copper

and few others, and at the end I've read that at high frequencies cables behave totally differently.. So, I'm looking for a electrical engineer or perhaps physicist, or some other amateur cable expert who could lend me a hand and explain in some detail.

EDIT: Little more explanation as I understand it's confusing: Connecting end to end with four cables, to draw an analogy: let's say you're connecting a voltage socket (120V with just two wires - live and neutral) and instead of using one for live and one for neutral, you use two for live wire and two for neutral, essentially doubling each. Now I think what happened, someone (not me) just twisted them at each end. So instead of taking two single wires out of CAT 5 cable, he took four (as there's eight total inside) and used it. It works, but is it better, worse?

From what I've read thicker copper cable has less resistance so it would be better, but if he twisted only the ends, not sure if it's better at this point..

share|improve this question
add comment

4 Answers

up vote 0 down vote accepted

Worse. This will create a higher impedance and more resistance on the line. This will show up on your ISPs end as a lower DB margin because there will be more noise on the line and this will affect your DSL. It might not immediately, but as the copper begins to degrade, your connection will become intermittent much faster than if you were just using a single pair of wires. Anytime you are doing IW (inside wiring) for any type of connection (EoC, T1, DSL) , you always want to use a single pair in a cat5 for each pair from the telco. Generally speaking, it is really best just to use a single cat5 for each pair from the telco, unless you have a 25 pair run as your IW, then using multiple pairs on the same cable with have little negative effect (as 25 pairs cables are designed for handling different signals through the same cable while cat5 is not).

share|improve this answer
    
Thicker wire creates lower, not higher resistance. –  CarlF Oct 21 '10 at 15:53
    
@CarlF - right, that's what I thought. I can see that it's bad overall and higher impedance, but bigger copper has less resistance..Although, he might be right that once you twist two copper wires just at the ends (and they're covered with jacket in the middle) then the resistance might be higher. –  ra170 Oct 21 '10 at 16:08
    
when it is over twisted pair and you combine them, it doesn't make for a thicker wire, it makes for two wires that a signal is traveling down that are split inside the cable, it is not the same effect. As I said, it decreases your Db margins on the line (signal to noise ratio and the noise goes up) and makes the line less usable for DSL. You have to keep in mind that DSL is a very high frequency signal and is very very easily disrupted. Modifying the cable path in such a way disrupts that signal. –  MaQleod Oct 21 '10 at 16:50
    
I agree that it might cause trouble, particularly if the wires are of significantly different lengths (which would interfere with phase matching of the two signals) but the resistance would be lower. –  CarlF Oct 21 '10 at 17:49
add comment

Signals on the line have to switch on and off. That takes some finite time, and a more time is needed to assure that the measured value (high or low) is real and not just noise.

How much time is a complicated business and investigating it closely will lead you into that thicket of vocabulary. But the long and the short of it is: for a particular cable geometry and switching electronics there is an upper limit to the (bandwidth * signal-to-noise ratio) that you can push down a pair. More pairs means more through-put.

This is, of course vastly simplified but it is a place to start.

share|improve this answer
    
Thanks, it already led me down the rabbit hole..hence the post. –  ra170 Oct 21 '10 at 16:09
add comment

You can't use more pairs than the port on your PC will read. You'd have to design and build your own adapter.

EDIT after the OP edited the question.

Resistance in the wires is rarely a factor in networking. The limiting factor is the ability of the medium to handle rapid voltage changes (frequency of oscillation).

Doubling the wire thickness should have very little effect on bandwidth, since signal frequencies are set by the adapter, not the wires per se. You might get a little better fault tolerance, since damage to one of the wired-together conductors wouldn't interrupt the signal. In practice, it probably has no effect at all.

share|improve this answer
    
I guess, I didn't explain it enough, there wasn't enough space here so I added an EDIT. Please take a look. –  ra170 Oct 21 '10 at 15:12
add comment

Doubling up the wiring will do the following:

1) Decrease overall line resistance.

2) Reduce the line impedance (resistance with frequency factored in) due to the cable pairings looking like two inductors in parallel.

3) Affect crosstalk 'in some way' (due to the magnetic coupling between the extra wiring) according to how the pairs are joined - it may help, it may have no overall effect, or it may make it worse!

4) Possibly increase noise due to signal loops and signal reflection - each signal run from end to end is now a loop.

5) Possibly result in slightly worse signal transitions (voltage jumps from 0->1 and 1->0) as the signal is taking two paths and the arrival of each signal may be really really small fractions of a second apart due to slight differences in the length and diameter of the copper wires in each leg. This may be seen as 'noise' but be within allowable tolerances.

6) Change the skin effect characteristics of the link (the tendency for high frequency signals to travel along the outer edge of a conductor) - reducing impedance/resistance (let's not split hairs over the difference)

Executive summary:

There will be measurable effects but whether they improve, degrade or have no effect on data will only be determined by try-and-see.

It's not the way it's supposed to be done - but hey, it's working!

Edit: Oops, I missed capacitance, but I think you get the picture!

share|improve this answer
    
ok, cool, thanks..so what if one used a stripped coax cable (two in that case) ? –  ra170 Oct 21 '10 at 17:58
    
Not sure what you mean by "stripped", but if the effective impedance of the coax and its capacitance aren't in line with the requirements of the signal then it's try-it-and-see time again. –  Linker3000 Oct 21 '10 at 18:52
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.