1

The extension to that question is does TP improve DSL over long runs? The reason I ask is because when I plug my DSL modem inside my home, I get a 1/3 speed reduction over when I plug it directly into the demarc. It's been recommended that I run Cat-5 from my demarc into my house, into which I'll plug my modem. I thought that this would be because TP ensures that the impedance is consistent, which would allow the signal to travel over large distances, but according to the Wikipedia page TP only protects against RFI. There isn't a lot of RF where I live, and the telephone cable carrying the DSL doesn't have a lot of neighbours, so I'm not sure how this is going to help.

The bottom-line for me is that my demarc is miles from the exchange, so how does the last 40' of my cheap telephone cable degrade the signal so much and how would a replacement cable make those last 40' easier?

I should also point out that I'm running a dry loop so the telephones and their network are completely isolated from the DSL and the demarc. There is absolutely nothing else connected to the DSL line.

PenguinLust
  • 125
  • 1
  • 9
  • It looks like [crosstalk](https://en.wikipedia.org/wiki/Crosstalk_%28electronics%29) is the issue. – bwDraco Feb 10 '15 at 20:29
  • Might also be that TP cable is intended for other frequencies than the frequencies used for DSL. See http://en.wikipedia.org/wiki/Asymmetric_digital_subscriber_line - with a quick search I have not found a similar source of frequencies for TP (the above wikipedia page lists BANDWIDTH not frequency). – Hannu Feb 10 '15 at 20:36

4 Answers4

1

There are two separate issues here:

  1. Twisted pair cabling, along with differential signaling, allows RF interference to be cancelled out.

  2. Running a dedicated line of new high-quality copper (even if it weren't twisted pair) from your demarc to your DSL modem, and separating it from your legacy in-home voice telephone wiring by putting a filter right in your demarc, ensures the shortest wiring run and isolates your DSL signal path from all the noise and reflections of your legacy voice wiring.

I doubt that any of this has anything to do with impedance.

Spiff
  • 101,729
  • 17
  • 175
  • 229
  • So, you're saying w/#2 that this is a sheer resistance issue? It's not a matter of AC problems? Also, a quick Google of "twisted-pair impedance" seems to indicate that your postscript is wrong, unless you are saying that for the purposes of this circuit, impedance mismatching isn't a problem. – PenguinLust Feb 11 '15 at 01:00
  • @PenguinLust With #2 I'm not so much talking about resistance, I'm mainly talking about how all the legacy telephone wiring in your house can act as a massive antenna (gathering RF noise) and every stub of that cabling causes signal reflections of the end of the cable, and any legacy telephone equipment including security equipment could be introducing noise and even stray voltage. And yes, I'm saying impedance matching is usually not a concern in home telephone (including DSL) wiring. – Spiff Feb 11 '15 at 01:15
  • So it sounds like #1 and #2 are basically the same thing: RFI. I forgot to mention that I added another paragraph explaining why typical parasitic losses don't apply to my situation. Having said that, I have a hard time imagining how RFI is causing problems on my line. Still, the fact remains that if I plug the modem directly into the demarc, I get a 50% speed boost – PenguinLust Feb 11 '15 at 01:33
0

A change in impedance between different media (conductors or even optical media, such as glass and air) causes reflections at the interface. The best signal transmission is through matched impedance in the line and at connectors. See http://www.highfrequencyelectronics.com/Nov02/HFE1102_Lao.pdf for telephone TP, and https://electronics.stackexchange.com/questions/33455/is-cat5-cable-good-enough-for-rs-485-vs-true-rs-485-cable for an explanation of impedance mismatch.

Another issue may be that there are other parasitic losses within the house, such as other devices connected to the cable and surge protectors (which exhibit capacitance to ground) etc.

Community
  • 1
DrMoishe Pippik
  • 25,661
  • 4
  • 36
  • 54
0

The issue could be the connection of the phone lines to the demarc. I checked the connection in my own home and redid it to resolve the speed loss.

Nelson
  • 1,363
  • 10
  • 13
  • Thanks. I've added another paragraph to explain why it can't be this. – PenguinLust Feb 11 '15 at 00:52
  • I'm not thinking of the interference though. I'm thinking the telephone panel may not be connected properly, or the contacts may have oxidized. When I was fixing mine, I manually stripped the wire and that fixed my connection speed. The "punch down terminals" may have oxidized between the telephone panel and the wiring. Although normally you don't need to strip the wire because the connection is suppose to do it, the contact area can still oxidize or the insulation may not be stripped off completely. – Nelson Feb 11 '15 at 13:06
0

The only after I can think of, after reviewing your comments on the other answers, is that your "cheap phone cable" attenuate the high frequency components of the ADSL signal.

Cat-5 UTP cables can reliably transfer up to 20 MHz without perceptible attenuation, and Cat-5e up to 200 MHz.

You need to test your cable against a sweep oscillator, and see where the 'cut-off' is.

pepoluan
  • 1,271
  • 2
  • 12
  • 27