On Sunday, December 26, 2010 11:22:47 am James McKenzie did opine: > On 12/25/10 9:23 PM, Tim wrote: > > On Sat, 2010-12-25 at 12:51 -0800, Joe Zeff wrote: > >> Part of the trouble-shooting was making sure there weren't any > >> cabling issues, so the client brought out an electrician. Not only > >> weren't the colors on the cable standard, they were different at > >> each end! Our only guess was that there was a splice down in the > >> pipe because whoever'd run the cable had run out of one batch and > >> simply spliced on another. > > > > Our phone lines do that, thanks to a plumber putting his oxy-acetelene > > torch through the wires a bit near a pipe he cut through. > > If he was using oxy-acetylene, he did not know what he was doing or he > was working with welded iron pipe. Most plumbers use propane or > oxy-propane as it cannot overheat copper piping. > > > He could have > > > > lifted the wires away, but he didn't. And some of the pairs were so > > badly corroded, that it was impossible to re-splice them, no matter > > how far back we stripped the insulation. > > Replace the ENTIRE run. NEVER splice, and I mean NEVER splice, LAN > cabling. You might have crossover problems and you might not notice the > loss in bandwidth until you hit 20MB+ connection speeds and then things > get very 'interesting'. Yes, it is a PITA to do so, but in the long run > you will be happier. > > (BTW, who ran your cables, they should not be near plumbing anyway.) > > James McKenzie It must be time for my annual diatribe about broadband cabling. It may not look like it, but the characteristic impedance of a twisted pair is every bit as important for a piece of cat5/6 as it is for a length of coaxial cable. Any disturbance in the spacing of the wires such as created by a splice, represents a change in impedance at the point of the splice, which in turn will cause part of the signal to be reflected back toward the source. This _includes_ a cable with proper plugs on its end, being plugged into a splice block such as sold at the Shack. The plug itself is a horribly bad bump when looked at with a time domain reflectometer, and we get by with it only because it is normally less than 1 cm from the signal transmitter, or the receiver on the other end. Even that in a gigahertz circuit is important enough I have had to cut the plug off and redo the installation of another just because the wire was stripped back too far, leaving 1/2" of it loose out the back of the plug. All of this stuff IS a transmission line, and must be treated as such if you want it to work at all, let alone well. Anyone who considers splicing this stuff, really should have his punch down tool confiscated and be kept away from anyplace that allows him access to this cabling. They have no concept of a wires impedance or propagation velocity. And the fancier the diploma on their office wall, the harder it is to impress upon them that they do not know it all. That said, I am constantly amazed that a 120 foot run of now well aged Belden brand bright blue cat5, with about 45 feet it it swinging in the wind between my back porch roof and the peak of the roof of a shop building, is now in its 7th year out in the elements without a dropped bit that I have been made aware of. When I put it in, it was for bragging rights that I had the internet on my small milling machines control computer, fully expecting it to be destroyed before spring rolled around again. If it dies yet today, it doesn't owe me a thing and I will gladly string another piece. But I won't splice it ever, its all one run from the switch about 6 feet away, to a small 4 port switch out in the shop building, which I put in so I could run another 100 feet back across the back yard to the garage my alu siding won't let a wireless connection work in. Yeah, you could call my place wired... -- Cheers, Gene "There are four boxes to be used in defense of liberty: soap, ballot, jury, and ammo. Please use in that order." -Ed Howdershelt (Author) Dreams are free, but there's a small charge for alterations. -- users mailing list users@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe or change subscription options: https://admin.fedoraproject.org/mailman/listinfo/users Guidelines: http://fedoraproject.org/wiki/Mailing_list_guidelines
