| Forum Home | ||||
| Press F1 | ||||
| Thread ID: 32986 | 2003-05-03 14:30:00 | Wiring Ethernet CAT5e Cables... | JamesStewart (874) | Press F1 |
| Post ID | Timestamp | Content | User | ||
| 141081 | 2003-05-04 04:15:00 | > You MUST use the cable as pairs. Why? I've heard that before a few times, but nobody has ever been able to explain why. The cables work fine no matter what order you plug all the wires in, as long as the same wire is plugged into the same slot at the other end. So why then do you say we must do it in that particular order (1&2, 3&6, 4&5, 7&8)? I'm confused :) Mike. |
Mike (15) | ||
| 141082 | 2003-05-04 04:41:00 | The twisted pairs are actually RF transmission lines. They have a "characteristic impedance" of about 100 ohms. That is critical ... we are using RF at 100 MHz (or higher soon). If you use random wires to make up the pairs, the impedance won't be right. You also won't have the interference immuinity given by the twist of the pairs. You'll get reflections in the wires, because of the mismatch. Those are the standard connections used for each of the pairs. xxBaseT uses only the pairs on pins 1&2, 3&6. xxBase4 uses all 4 pairs. The cards use pins 1&2, 3&6 only. "Wrong" cables might "work". "Most of the time". But there are enough ways to have non-working networks without making the cables wrong :_| |
Graham L (2) | ||
| 141083 | 2003-05-04 04:49:00 | In simple terms, CAT5e is a high-speed version of CAT5 and both cables consist of twisted pairs . The twist keeps the impedance of the two conductors making up the pair in balance and ensures that the signal does not reflect back when it hits an impedance variation . This is one of the reasons why the untwisted length at the termination must be kept to a minimum . Twisting also ensures that any external interference affects each wire in the pair equally and thus cancels out . Breaking apart the twisted pairs might work over short distances but it slows the propagation of the data and you are very unlikely to get 100Mbit performance . At best you will see 10MBit . A recent poster had slow network that was traced to unbalanced wiring . All the associated problems will increase the data error-rate and make the computer work harder and slower . It is not worth the risk . Cheers Billy 8-{) |
Billy T (70) | ||
| 141084 | 2003-05-04 05:33:00 | Mike, Following on from Graham and Billy T, I'd suggest a little reading up on transmission line theory and characteristic impedance, I was going to say something about having to connect the twisted pairs the right way, but it is easier to just follow the correct instructions at that website I posted. Transmission lines can be imagined as an "infinite" series of inductors connected together with capacitors paralled across after each inductor, and when the line is terminated with a resistance equal to the characteristic impedance of the cable then to all intents and purposes the cable looks like a pure resistance at "all" frequencies. Of course that is the ideal. Transmission lines eg coaxial cable or twisted pairs, or parallel wires spaced apart with low-loss insulation are used for RF work including high frequency digital data transmission, eg connecting a transmitter to an antenna. If you are into tramping or SAR and ever use a mountain radio or a Codan, you will see that the wire dipole antenna is connected to the radio with twisted flex (they use "Remit" wire). This is essential to get as much signal into the antenna as possible without it being reflected back. |
Terry Porritt (14) | ||
| 141085 | 2003-05-04 06:24:00 | > The twisted pairs are actually RF transmission lines. > They have a "characteristic impedance" of about 100 > ohms. That is critical ... we are using RF at 100 > MHz (or higher soon). If you use random wires to > make up the pairs, the impedance won't be > right. You also won't have the interference > immuinity given by the twist of the pairs. You'll > get reflections in the wires, because of the > mismatch. > > Those are the standard connections used for each of > the pairs. xxBaseT uses only the pairs on pins 1&2, > 3&6. xxBase4 uses all 4 pairs. The cards use pins > 1&2, 3&6 only. > > "Wrong" cables might "work". "Most of the time". > But there are enough ways to have non-working > networks without making the cables wrong :_| Okay that makes more sense now :D <mind swells a little more> So does the wire colour matter? I mean I've read a couple of things that say certain wires go in certain places - or is it just the pairs have to go together, no matter what colour they are? Mike. |
Mike (15) | ||
| 141086 | 2003-05-04 06:27:00 | The pairs . | Pheonix (280) | ||
| 141087 | 2003-05-04 07:41:00 | There are standards there for a reason. But yes technically you just need to be consistent for a reason. There two ways of wiring them, 568a and 568b. It used to be standard to use one way, now the other way is the standard. It can screw up some people when they're looking at an older/newer network to what they're used to. A cross over has 568a at one end and 568b at the other. I'll try and dig out some good images of it later. |
-=JM=- (16) | ||
| 141088 | 2003-05-04 12:03:00 | ok... these "pairs" of 1&2 3&6 etc has got me REALLY confussed now. I am not doing what I was paining of doing (just got a 10 meter CAT 6 cable instead of crimping the 10 CAT5e blank one. BUT I still need to know what to do! | JamesStewart (874) | ||
| 141089 | 2003-05-04 20:58:00 | There is no need to be confused James, just read very carefully from that website reference I gave above, scrolling through page by page where it says Next at the bottom of each page, there are precise details and coloured pictures showing the twisted pairs with their colours and the connectors with the pin numbers etc. It could not be plainer. Cheers |
Terry Porritt (14) | ||
| 141090 | 2003-05-05 05:20:00 | You need 2 pairs of wires (one for transmit, one for receive) . Each wire needs one pin on a connector . The wires in a pair are twisted together, for "electrical reasons" . The pins allocated to a pair are defined by a number of standards . The only pins we are really concerned with here are those which are needed for the two directions of the signals . Those pins are: one pair -- 1 and 3, the other pair -- 3 and 6 . The two other pairs of wires go to pins 4 and 5, 7 and 8 . The colours defined for different pairs are specified in telephone company manuals . We can use any pairs we wish, because we don't work for the phone company . We just have to use the wires in pairs, as required by the construction of the cable . The comparitively slow serial standard RS-232 is complicated . :-( The only pin you could rely on was pin 7 (on DB25 connectors) . All the other connections were "negotiable" . :_| |
Graham L (2) | ||
| 1 2 3 | |||||