Folded phone line can mess up DSL.

The networking ng is dormant so I thought I'd report this here.

Folded phone line can mess up DSL.

I think a couple years ago when I had 100 feet of phone line but was only using 40 feet, but wanted to save the whole 100' in one piec, I had

40 feet played out and 60 feet rolled on the spool, and I asked here or somewhere else if that could cause a slow DSL connection. And iirc the answer was no, but I'm not sure of that.

So a couple months ago I finally finish installing my home burglar alarm and I want to connect it to the monitoring company via the phone line. Since my main phone connection goes in via the second floor, to the computer on the 2nd floor, I go to the NIC and put a Y connector in it, adding back the original wire to the basement. Somewhere I have a 6" piece of phone wire with modular plugs on the end, but I can't find it, so I put in one of the cords that comes with most phone devices, maybe 8 feet folded up, maybe 6 circle's-worth, squeezed flat.

This is some time between Thursday and Saturday, and Sunday morning I leave for two months.

I got back two weeks ago today and the house and car were just the way I left them, and the computer worked fine. For 3 or 4 days. Then it stopped loading webpages, except very intermittently (so little as to be unusable), loaded email and newsgrooups only 20% or 30% of the time. People told me to call Verizon, as if I hadn't thought of that. But they charge about $60 if it's not their fault, and anyhow, the point is to figure it out on my own.

Check the NIC. It's popped open. It's under a roof but maybe it got wet. Doesn't look wet. After a couple days with no better ideas, I decide to unplug things in the NIC to give them a chance to dry, and only then I notice the folded phone cord (with the Y-connector). Take it out, and everything works fine again.

But how come it worked fine for the first 3 or 4 days I was back? And at least 12 hours, maybe 2 days, before I left?

Reply to
micky
Loading thread data ...

Are there DSL isolators in line where they should be? I couldn't tell from your post.

Reply to
Jason

Micky, I am not going to answer your question as to why. Don't know. However, with you alarm system connection you need to place the alarm dialer as the first thing on your phone line from the NIC. This is done with a RJ31X block properly installed. Also, depending on your specific use of DSL Filters the line to the alarm dialer needs to be a DSL Filtered line. Otherwise the communication can be compromised. Twisted pair cable is most desirable.

Good luck.

Les

Reply to
ABLE1

On similar lines I was using a network cable that was way longer than needed and the excess was coiled up. At 1 gigabit packet loss was around 50%. I replaced it with a much shorter cable and no packet loss.

Reply to
Lucifer Morningstar

formatting link

Reply to
Lucifer Morningstar

Lucifer Morningstar wrote in news: snipped-for-privacy@4ax.com:

That one is easy. Gigabit eathernet is very sensitive to interference. Having the wires coiled like that very likely degraded the line electrical characteristics to the point that you were getting a lot of lost/error packets. CAT5/CAT6 is very sensitive to proper conditions for transmission and reception. It doesn't take much to degrade a packet to the point that it is not decipherable at the other end.

Reply to
Tim

Where I live, they have two practices for ADSL.

The ADSL original setup, they handed out filters per phone jack in the house. So they would put four filters in their "kit", as back then the modem was a rental. This was back in the days of "no truck roll" installs.

When ADSL2 rolled out, they started using a tiny whole-house filter, installed at the demarc. What's supposed to happen in that case, is the "existing" house phone wiring, is put on the filtered port. (And the alarm dialer could go on that network of wires.) Whereas the customer can run a cable to the ADSL (unfiltered) port on the filter, to make the ADSL2 modem work properly. The filter box was pretty small, and only

2x3 inches or so.

Paul

Reply to
Paul

There is a minimum bend radius of at least 1" for CAT5. See

*
formatting link
  • Silver Satin cables are bad news. I don't know if that's the type of cable your using. The wires are all parallel and not twisted.

My house has many phone locations, more than 8, and not home run. Nearly everything is wired with quad wire. This won;t support DSL filters,

So, the best way to handle it was to install a DSL splitter at the NID.

When I upgraded from WIFI based dial-up to DSL, I did things right. The NID to DSL port is only about 4 foot long. Then i found a 5' length of CAT4/CAT5 6P4S straight-thru cables that run to teh DSL modem. Now that distance is short.

Then I ran the LAN port via power line Ethernet to the center of the basement where it should be. The AP (Wireless) should really be centrally located which it is.

Later, that power line Ethernet connection was upgraded to wired. It also broke.

The telco guy said that I had pristine DSL. During one of the service calls, the guy brought back a 12" piece of cable where the insulation was eaten. He also said he removed 1000' of length of cable.

There is no way I could have this pristine DSL if it wasn't done right from the beginning.

I do have some upgrade plans, but they have not been implemented yet..

One thing I must have is the ability to easily switch out a stand-alone modem to wired `Ethernet for the normal one in Bridge mode. I will eventually be able to do this with a push of a few buttons. Anytime trouble shooting needs to be done, I have to eliminate nearly everything at my end. The older modems have better trouble-shooting information.

Then I started the "structured wiring" panels. They are mounted, but not connected. Some locations have low voltage plates, but unconnected too. Everything will be CAT6 RJ45's. If the jacks are telco, they will be reduced to 6 pin jacks using sleeves.

I want to install a 24 port POE switch as well which I also have.

Reply to
Ron D.

Why? Twisted pairs are resistant to interference.

Reply to
Jason

Absolutely, the twisted repairs counteract interference, and each pair is twisted at a slightly different rate to reduce crosstalk between the pairs. Here's a simple summary.

formatting link

Reply to
Mike S

Yes, but coiling the cable means that pairs (with twist pitch X) interfere with other turns of themselves (still with pitch X), so coiling the cable can increase scattering.

Reply to
Clifford Heath

Jason wrote in news: snipped-for-privacy@news.eternal-september.org:

Yes, twisted pairs are more resistant to interference, but the conductor diameters, wire spacing, and insulation for CAT5 and CAT6 cables are very carefully selected for among other things interconductor capacitance. For the same reason that puting a CAT3 connector on a CAT5 cable limits the cable to CAT3 specifications, having the cable being pinched at any spot will change the electrical characteristics of the cable in a negative manner, with a corresponding drop in performance.

Reply to
Tim

Agreed, I used to install DSL and I never did that.

Reply to
Mike S

Mike S wrote in news:pd915k$26e$ snipped-for-privacy@dont-email.me:

Have you never heard of insulation creep? Under pressure, the plastic used for modern insulation will slowly flow in such a way as to relieve that pressure. This can result in direct shorts, changes in spatial relationships between wires, and other problems. Any of the above will change the capacitance and inductance of the conductors involved, in a negative way. Changes in capacitance can cause the digital signals to round off too rapidly, to the point thay are no longer detectable as signal pulses at the far end, while changes in inductance can cause mismatches in impedence that can cause unwanted reflections of the digital signal, which can also degrade the digital pulses. That is why on of the first things I was taught about dressing cables was to never pull cable ties too taut, as that is a prime way of causing both problems, either in tying up loops or in dressing cable runs.

Reply to
Tim

That all makes sense. We used Cat5e. On a few occasions the customer supplied Cat6 for us to run, it had substantial nylon wire guides... it's clear that a lot of science has gone into ethernet cables.

Reply to
Mike S

You may have the clue in your response, "excess was coiled up" A coil of wire makes an inductor. An inductor is an impedance to high frequencies. Years ago, a mobile welder couldn't get his tig welder to work right, after much consternation, he uncoiled the 75ft of welding cable that was left on his trailer. The inductance of that coil blocked the high frequency and it was not helping sustaining the arc. I don't have an answer why it worked 3 or 4 days. Mikek

Reply to
amdx

Coiling the cord makes something of an inductor. ADSL2+ goes up to

2MHz. If your coil of wire has enough inductance to act as a choke at 2MHz or less, the speed will slow down, but will not quit as it did in your situation. That's because ADSL has 256 "tones" or channels, while ADSL2+ has 512. The added inductance might cause the higher frequency channels to drop, but the lower frequency channels and the uplink channels will continue to work. The ADSL modem also contains an adaptive equalizer that will partially compensate for the loss in high frequency channels. You would need to add quite a bit of inductance for the signal to completely disappear.

From my experience with DSL, most of the problems come from rotten wire connections and noise coupled from nearby sources. Bad terminals, mangled connectors, weird adapters, crude microfilters, and such are all too common. Since the ADSL apparently quit completely, I would tend to favor a bad connection as the most likely culprit.

It could also be your friendly local telco playing with the DSLAM, router, or provisioning, which would also make the internet disappear. I had this happen to me recently which caused me to go through internet withdrawal for about a week.

There's also the possibility of local interference. Running the phone line parallel to power lines or other services that use the same frequency range, such as T1, can cause problems. I had someone in another newsgroup ask why her internet would disappear at night. I turned out that when she turned on the lights at night, the associated noisy light dimmer would couple enough junk into the phone line to kill the DSL signal. Rewiring the DSL to a "home run" configuration solve the problem. If the coil of wire were sitting on top of a length of lamp cord attached to a light dimmer, or switching power supply low voltage light, or LED light bulb, you're likely to see some noise coupled to the coil. Even though the balanced pair offers some noise immunity, the lack of a twisted pair (as in CAT5 cable) in most flat phone cords, can make the common mode rejection less effective.

I would look for a bad, loose, intermittent, or corroded connection somewhere in the phone line.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 
 Click to see the full signature
Reply to
Jeff Liebermann

Would that apply so much to a two-core cable, where at any instant, a current is flowing one way in one conductor (eg on the way to the appliance) and the opposite way (on the return) in the other? Would the fact that there are two conductors tend to reinforce or cancel the effect?

If you have a cable that is too long and don't want to cut it to shorten it, what is the best way of arranging it so as to avoid/minimise inductance? Is it better to arrange it in a zig-zag rather than loops?

Reply to
NY

"NY" wrote in news:v5OdnTChrerT3WDHnZ2dnUU78f snipped-for-privacy@brightview.co.uk:

At the frequencies DSL works at, inductance is not an issue. More of an issue is the fact that the wire was kinked. OP doesn't state how severely, but any kink is going to disrupt the twist pattern, and thus has the potential to interfere with transmission. The other factor is that if the kink is severe enough, it can cause partial reflectance of the signal in the reverse direction, potentially degradeing the signal to some extent. Since the DSL signal can approach 4mhz at the high end, skin effect is starting to enter the mix (barely, but there). This is going to be more noticable in solid core wire, since the single conductor is larger. In a stranded cable the multiple strands of smaller gauge will tend to lessen any impact from skin effect. I am going to assume that the wire is question is stranded as opposed to solid core conductors. In that case it is entirely possible that one or more of the multiple strands could actually have been broken by the kink, thus pretty much guaranteeing some reflection of the signal, with attending degradation.

Granted, the effect of any of the conditions mentioned above will probably be slight, but they will be additive. The ideal solution would be to rewire the link to remove any excess wire. The only reason not to would be that the connection is considered to be only temporary, and even then I can make a case for not having any excess wire involved.

Reply to
Tim

I agree with previous posters that a bad connector, or a severe kink, sound like the most likely cause of the problem, in that order.

Reply to
Mike S

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.