Generally the plugs and sockets make no difference.
What counts is the cable. I would highly suggest that you just run CAT5 cable.
In fact though, any quality UTP (unshielded twisted pair) telephone or data cable will be fine. Just make sure that it is
*twisted**pair*! Any of the "flat satin" type cables, and some that has four wires (red, green, yellow and black) and looks just like other cable with the same colors that is twisted pair, are *not* suitable for anything longer than maybe 10-12 feet.
I used a splitter and then a filter on the cord going to the phone. This phone line only had one connection - at the marina. Nice and clean.
The other part of the split line went to the DSL modem. Yes my initial cable was "dime store" flat, not twisted phone cable, 50 ft as I recall. Also as I recall the shorter cable, which Verizon supplied was about 7-8 ft of round, I assume, twisted phone cable.
Also, below is the url to a dsl-reports thread about this topic
I have a Belkin wireless adsl router mounted on a roof approx 70meters from my incoming phone point. Well, I have a lot of trouble getting the router to login to BT, so just to test it I connected it directly to my Phone socket and it works perfectly. So my question is :- As I am using normal phone extension plug and sockets from my local hardware shop, are these the correct cables to use?
Not so. The DSL splitter and microfilters have a direct through connection between the MPOE (minimum point of entry) at the NID (network interface device) also known as the protector. It's exactly like an extension cord directly to the phone line. With a splitter, cable length is not an issue because there is only one pair of wires running at DSL frequencies (about 100Khz -> 1.2MHz). However, if you use microfilters, the cable between the MPOE and the splitter can act as a "bridged tap", stub, unterminated line, or whatever. If long enough, these might affect the ADSL frequency response and reduce performance. If really long, the can put a notch in the frequency response. This is the real advantage of using a splitter over microfilters in that it isolates the ratty house wiring from the ADSL part of the circuit. Usually, there's little effect on loops of less than 2 miles at 1500/256 kbit/sec. However at the loop length approaches 3 miles, the line characteristics become somewhat critical. The extra garbage wiring between the MPOE and the microfilters can (and do) make a difference. If you're at the bitter edge of the 3 mile limit, use a splitter.
Anyway, the length of wire between the MPOE and the DSL modem makes little difference. However, the length wire between the MPOE and the numerous microfilters might be a problem. Use CAT5 or twisted pair so that it doesn't pickup garbage from AM radio stations, motors, lights, etc. The flat telco cable will work, but I wouldn't use more than about 50ft of it due to losses (it's made of tinsel) and lack of twisting (it will pickup some EMI).
Amazing. Are you using a splitter at the MPOE (or inside the NID) or microfilters? Cable length might make a small difference with microfilters, but it should have no affect with a splitter.
My guess(tm) is that it may have been a type of cable problem. If the original 50ft cable were flat tinsel cable, it might have been slightly inductive from the way the cable is made. It might also be somewhat high resistance. What type of cable was replaced?
Incidentally, I've installed quite a few DSL modems and troubleshot quite a few "slow" installations. In most cases, I can fix it with a splitter. I have run long 100ft or more cables, but I always use CAT5 for the long runs, never flat cable or round untwisted station wire. There may be a short flat cable between the wall plug and the ADSL modem.
I just did a quick test. I added 3ea 50ft sections of flat 4 conductor stranded (not tinsel) flat telco wires to our ancient Alcatel 1000 ADSL office modem and ran the SBC Speed Test. Same speed with and without the added 150ft of cable. Incidentally, we use a splitter, not microfilters. It's not the length, it's the type of cable.
Incidentally, if your ADSL modem has internal diagnostics, you can tell if you have a screwed up system by looking at the S/N ratio and xmit levels. If your modem doesn't, then my standard test for internal wiring goofs is to unplug the entire house at the NID, and install the ADSL modem directly to the incoming phone line with a short cable. If that works, look for wiring or microfilter problems.
Hmmm... Just a double check. Are you *SURE* that you have a microfilter on every telco device plugged into the line? The usual omissions are kitchen wall phones, CATV set-top boxes, Satellite receivers, burglar alarms, credit card machines, secret phones in the kids room, and garage phones. A missing microfilter will cause all kinds of horrible problems (another reason why I like splitters).
Any cords I've seen supplied by telcos have been 6-12 feet. So I presume your DSL modem was not 50 feet from the jack if you just swapped over to the telco cable. I also presume you just had 50 feet of cable spooled up on the floor (either in a nice looped pile, or a spagetti mess).
You may have induced crosstalk into the cord by having 50 feet bundled up.
Also, the cables provided by the telco (that I've seen), are all cat3 UTP... if you used a 50 foot flat cord, you will most definitly get crosstalk which would likely degrade the DSL signal. If it was round "station" cable (the kind with 4 solid wires colored red, green, yellow, black), then it too will develop crosstalk.
Unless you happened to have a cat3 or better UTP cord of 50 feet, and wound it with alternating loops (loop one way, then loop the other to break the inductance pattern, much the way you loop live mic cables), or had it stretched out and not near any sources of inductance (flourecent lights), then I can't say I'm surprised in the least that your cable failed.
Oh, and depending on who you talked to at Verizon, don't believe everything they tell you. Often they just know you need the cable supplied by them, and make up their own reasons why (or similar 'I know what is needed, I just don't know why, and no one here trains us in those details any more' type of issues. It has been a LONG time since Verizon has had high standards on who they higher (and anyone with talent has moved up the line to where they don't see the customer very often any more).
I no longer let Verizon Com-Techs touch ANY of my wiring without me present, and if I hear one was in messing with the UG box, I verify all my lines to see what they knocked out this time while working on someone else's lines in the building.
Bingo. You don't have to go to a dime store to buy junk. You can also get it at Radio Shack.
I just happen to have a Verizon DSL kit. It's a bit old. Westell B99-211015-00 modem setup for 0/35 VPI/VCI. The included phone cord is two wires, apparently twisted, "Line Tech AWM 20251" #26AWG stranded (not tinsel). Measured length is 15ft.
It pretty much agrees with my previous mumbling. It's the wire type, not the length.
Not spooled at least would lessen the chance of it creating its own RFI. But since it was a flat cable, it will still suffer from crosstalk. Not enough to get noticed on a voice call, but could be on something like DSL (and possibly since you saw that it failed).
Of course, it also could have just been a fluke with a slightly defective cable. I just did a DSL installation the other day and had one of those flukes. I always split the line at the NID, install a filter to isolate the house, and run a line dedicated for the DSL from the NID. I did this as always, and the DSL wouldn't sync. Voice service was great, DSL just wouldn't work.
After some head bashing trying to find the problem, it turned out to be a bad jack. Brand new jack, Cat 3 keystone jack (cat 5 from NID to jack). I verified and repunched it TWICE. Finally replaced the jack with another brand new one, everything worked fine.
So sometimes, you have all the right parts, one of them is just defective.
Just for fun, I happen to have several rolls of random cable at home. Using clip leads (yeah, I know... they're not kosher) I tried adding to the phone line between the wall jack and my Efficient 5260 ADSL modem: 800ft of CAT5 1000ft of 18AWG shielded pair (mic wire) 1000ft of 14AWG zip cord (AC power cord) 400ft of #26 round station wire on rolls and all together at the same time. No problem, and no observed speed change. I only have a few short pieces of flat telco cable handy so I didn't try that. I would have added several lengths of AC extension cords if I had more clip leads. Anyway, like I said, it's not the length of the cable, but the type of cable.
I then decided to see if I could make it fail. I tried a 1000ft roll of RG-6/u coax cable. Guaranteed to be unbalanced and pickup all manner of garbage. Works fine with no obvious problems. I plopped a power strip full of wall warts and my flourescent desk lamp on top of the roll to see if I can create some interference. That made it slow down slightly (1160kbits/sec down to 1000kbits/sec) but it was still useable. So much for the alleged need for perfect balanced pair, twisted pair, and proper impedance matching. My guess(tm), based up the above anecdotal evidence implies that you can use just about anything and it will work.
That leaves the question of why your flat cable didn't work. My best guess(tm) is that it was either defective, highly inductive due to the twisted nature of tinsel wrapped around silk, high resistance, or some combination of these. It might also be possible that your ADSL modem is excessively sensitive to line matching and levels. Lastly, methinks you might be at the bitter end of the DSL cable limit (about
3 miles) and that almost anything will perturb the connection. Note that I previously stuffed 150ft of flat cable in my office DSL without a problem.
Incidentally, is your DSL service RADSL? (Rate Adaptive DSL). This is where the DSLAM measures the error rate and adjusts the maximum speed accordingly. If it sees lots of errors, you go down to about
300Kbit/sec. This is good because you can use this feature to tweak your phone line. In the SF Bay area, we're stuck with SBC, who does such a test during the first few days after activation, and then "rate caps" the line to whatever speed they consider useable. If someone forgets to install a microfilter, they get PERMANENTLY stuck with
384/128 Kbits/sec, even though there's nothing wrong with the line. Yelling at SBC to remove the rate cap is a major problem.
But, on the other hand I've spent not hours but *years* chasing all manner of strange circuit problems, and on occasion that has included bad cable one or two times... ;-)
The only reason you didn't see much effect is that you left all that cable rolled up! A five foot stretch of cable with a 1000 foot roll in the middle is still a five foot stretch of cable, though it does have more DC resistance than a short length of the same cable (okay, at lot more distributed reactance too...).
Take that coax cable and unreal enough to circle your house, and give it another try! Or, just try enough to circle a fair sized room, and include one loop around that fluorescent lamp, and see if that does anything fun!
Of those cables, the coax will be the worst. The zip cord won't be too bad for short enough runs, and will be just about like flat satin cables. And while everyone recommends against them, it is also true that many telco's (GTE for one) used to use untwisted pair as the standard drop wire from a pole to a house, and only replaced it with twisted pair if it didn't check out okay.
The whole point with an unbalanced cable is that it becomes an antenna. Of course the longer an antenna is, the better it is. And generally the term "long" is relative to wavelength, which in this case is hundreds or thousands of feet. Hence a coiled up cable with five foot stubs on each end is essentially a point source, and makes a really really poor antenna.
And yes, up to about a dozen feet or so just about anything including barbed wire will work. At the 70 yards the original article mentioned, it had better be well balanced (which requires twisting) or there had better not be significant external electric fields. That last is hard to find inside a house, given all of the electricity we use.
Because it was *long*. Or, long enough to be an antenna. (At
2000 feet it would pick up some really good signals though! Maybe even some of those VLF transmissions used to talk with submarines, eh?)
None of the above.
But adding 200 feet, either to a 1/2 mile of cable or to 2 1/2 miles of cable, doesn't do much. So that isn't likely either.
However, that does suggest one other possible cause that is closely related. Because while the unbalance of a non-twisted pair doesn't make the 3 miles of cable back to the telco office an antenna, other kinds of imbalance can do exactly that. Hence if even a little leakage to ground occurs (from a damp junction box, or wet cable with cracked insulation), the results on even a short section of cable will be dramatic if that is connected to a longer cable. The entire length of the cable becomes unbalanced, and the antenna will be so good that you can listen on the phone and hear other people talking on other cable pairs! Clear enough to be understood too!
Sounds interesting! How do you tweak the line using that feature?
Sounds bad. (As far as I know, *all* telephone companies are easy to yell at. What's hard is getting one to *listen* when you want to talk, which makes you want holler and shout, at which point they lock access to the brain activation switch, which of course is by default set to off, and never do turn it on.)
Go look up the Dilbert cartoon archives from the middle 90's, when Scott Adams was still an ISDN applications engineer (which Dilbert is modeled after) with Bell South or whoever it was. That was about the time that I really began to appreciate the character that Wally has been developed into since then... (I used to have wonderful coffee cup at work, just for management conferences, it said "I LOVE my job! I LOVE my job! I need the money..."
(But, all that said... what I specialized in for the last 20 years or so was dealing with local telco's to "encourage" them to trouble shoot customer complaints, mostly about data communications. So I do have a perspective on yelling at telco's that is vastly different than most people. I had the almost unique experience of being able to pull rank on virtually anyone in a local telco and cause the sky to begin falling if they didn't listen when I asked them to. You have no idea the effect it has on a telco if you threaten to cut them off from all long distance, and calmly provide the phone number of a lawyer in Carrier Relations who will back up the threat, and then give them a day to verify if it is an idle threat or not. I've actually done that, and it makes for some *very* cooperative miscreants.)
Sigh. First I drag every roll of cable I could find up the stairs and now you want me to make a huge mess by unrolling the cable? No thanks. However, you are correct. The ability of a roll of cable to play LF/BC antenna is rather limited. If there was going to be any EMI pickup, it would certainly be worse if I unrolled the cable. I'll be installing a "Free to Air" 30" dish on a nearby tree with about
200ft of coax when it stops raining. I'll try the coax cable test again when I unroll (and untangle) the cable.
Yep. I was trying to determine if it was the type of cable or the length of cable. I hadn't considered EMI pickup in the test.
I have a metal box with a flourescent tube, coil, and broadband amp that I use for a broadband noise source. It's flat to about 300MHz and slowly rolls off to about 3Ghz. I use it for RF and IF alignment instead of a speed generator. I've never bothered to see what it does at low frequencies. Yeah, they're noisy.
Now, if you really want the ultimate nightmare RF noise source (other than BPL or Broadband Power Line), it's the little 117VAC to 12VDC track lighting power supplies. Dirty comb line every 40KHz from DC to light.
I charcter I know (name withheld) runs DSL over the pair of barbed wired wire fence on his farm. I think it was 192/192 kbit/sec SDSL for something like 10 miles. SDSL (CAP) requires adaptive equalizers at each end to deal with goofly looking frequency response and group delay. However ADSL (DMT) doesn't because the frequency response is going to be flat over a 4KHz channel. It just turns on and off channels depending on line quality.
Those are around 15-20Khz. That's below the ADSL working frequencies of 120KHz to about 1.5MHz. The Navy beacons and Broadcast Band are the most likely culprits.
Yep. A lousy termination or capacitive unbalance to ground will also turn the cable bundle into an antenna. However, my experience is that RF ingress isn't that fatal with ADSL. If there's a carrier in the middle of the working bandwidth, only the single channel occupied by that frequency is trashed. Unless the source of RF is broadband, there won't be that much of an effect. I could easily run a test to see for myself, but I'm lazy.
Incidentally, the Efficient 5260 is a great ADSL modem for such exercises because it has LOTS of diagnostic output accessible via telnet. |
Agreed. However, if that were the cause of Bob's problem, then the replacement cable should have had the same problem. While the unbalanced line problem is real, I don't think it was the cause of his slow performance.
I don't have any RADSL customers, so I don't know. You don't actually tweak the line. The ISP purchases bandwidth from SBC. The ASI (Advanced Services Inc) division of SBC runs the DSLAM. The ISP pays ASI for bandwidth. ASI sets the rate cap at the DSLAM. The speed can be less, but not more than the rate cap.
The big advantage of RADSL is NOT that it will compensate for crappy lines by lowering the speed. It's that it will work furthur than ADSL but at slower speeds. It also means that the ISP does not need to sell tiered service and pre-qualify the lines. The customer will get whatever speed the service can handle.
He was with Pacific Bell in San Ramon. He was gone before SBC bought Pacific Bell.
Where were you when I needed you? I have customers paying for "business grade" lease line service (T1 or better) that used to get one hour service. Now days, it's considered a miracle if I can get anyone's attention in less than a day or two. Multi-day outages are common. Fixing one problem and breaking two others is also common. Borrowing line cards from one customer to fix another or leaving lines in maintenance mode seems to be standard proceedure. Provisioning rarely agrees with billing. Complaining to SBC in Texas is a waste of time. The local PUC (Public Utilities Commission) directs all datacomm to the FCC, which favors playing monopoly and has no effective method of dealing with end user complaints (unless you're a VIP or Congress Critter).
Wheeeee. Nice if you can do that. I don't have any clout. I post my SBC horror stories to ba.internet. Back in the daze of PBI (PacBell Internet), I used to get email or phone calls from concerned PBI managers wondering if they could do something to help. As soon as SBC arrived and outsource the everything including the test board, that stopped. Now, it's NBC (NoBody Cares).
Ok, compromise. I found a roll of about 300ft of quad shielded RG-6/u cable that I rolled half around the house and half outside in a big messy loop. Attach with clip leads to my DSL modem phone line input. Nothing, dead. Check dial tone. Nothing, dead. Dig out the TDR and find where I had stapled the coax in two places (which was why it was buried in back with unreadable scribbling on a big red tag). Peel jacket around damaged areas and remove short. Dialtone and DSL are now back.
DSL works without any problems. Piled the wall warts and flourescent lights on top of the cable. No effect. Some audible 60Hz hummmmm on the phone handset, but nothing serious. Theory says that such an unbalanced line should pickup substantial EMI causing the DSL to drop channels and slow down. It shouldn't work. But it does. Weird.
[Incidentally, I found a small pile of 446A lighthouse tubes, which are similar to the 416A. Also, 2K25 klystrons, and 955 acorn and 7586 Nuvistor tubes. Why I saved these, I dunno. Maybe tubes will make a comeback.]