Cable Length

[RobPa]

I know theres a max length for twisted pair cable.

Is there a minimum length?

I need to connect 2 switches together that are mounted on top of each other.

Thanks

Rob

 

[Kyusaku]

Nope, there is no minimum length, I have made patch cables as short as 3 inchs and they worked just fine. Just stick to whatever standard you are using and your boat should float.

Kyu

 

[hexn]

There really is no minimum length for an ethernet cable, but I have run into instances where I couldnt get a Cisco switch to work, and so I called Cisco Tech Support, and they told me my ethernet cable had to be at least 1 meter long (3 feet). I didnt believe them, but made another cable just to humor them, and it actually fixed my problem.

They also pointed me to the IEEE website that says the standard on minimum length is 1 meter for CAT5 cable running Ethernet.

But I have cables that are around 1 foot in length and they work just fine, but if you want to go with the standard (so that tech support is forced to help you) then make them at least 1 meter. ~~~~~~~~~~~~~~

tech@shnypr.com

http://shnypr.com/tech

~~~~~~~~~~~~~~

 

[turborob]

Hi,
While you are on the subject of cable length, what cable or cables would be ok for 130 mtrs or over in length.
ta, Rob....

 

[hexn]

Well, an ethernet cable would perhaps do it, although you are really starting to push your luck. Your best bet is to split that length in half and put a repeater in the middle.

Otherwise, you are probably looking at fiber.

 

[jimbopalmer]

10Base5 can go 500 Meters
10base2 can go 185 Meters
10baseF, and 100baseFX can go 2km
1000BaseSX can go as long as 500 Meters in Multimode, but 260 meters is MUCH more common
1000baseLX can go 550 meters in Multimode, but its range in Single mode can be 20 km The one thing you can't give for your heart's desire is your heart. - Lois McMaster Bujold

 

[getimran]

hie,
Here is one more qurestion. How can i check what Base my network is on. I mean 10Base2, 10Base5 ...

one more question please. pins setting on my pc is

brown-white
brown
blue-white
blue
green-white
green
orange
orange-white

I didn't number them cause i don't know from where to start numbring. However what does this mean ???

bye

 

[jimbopalmer]

10base2 is a large Co-axial cable (one wire inside a braided ground)

10base5 is a thinner co-ax cable about the size of TV cable

you have a 4 pair Unshielded Twisted pair, so it is 10baseT I tried to remain child-like, all I acheived was childish.

 

[Svensson]

I have something to add to the cable length discussion. I have a 170 M long UTP cable that connects two hubs. It works fine and have done so for the last four years.
The max lenght of 100 M for 10BaseT just guaranties that it will work with some interference. If there is no or just a little interference it will probaly work with a lot longer cable.

Daniel Svensson

 

[noproblems]

I have found that a molded twisted pair cable (2m) between a pc104 ethernet connection and a Cisco BR 350 severally limits transmission rates from pc104 to base computer (components are contained in water tight enclosures to be used outside). We performed several ftp get/put tests, always with the same results. I added a 3' cat5 cable to the "outside" Cisco junction box and this fixed the problem. I appears, at least in this instance that there was a minimum length.

 

[daronwilson]

Well I left my reference books at the office....but, I believe there is a standard that indicates the minimum length of ethernet jumper should be not less than 18". I have seen shorter ones and they seemed to work. As for the length of cable, the 100m rule can certainly be pushed further. However, without actually checking the packets to determine the amount of retries, etc. one can't really know if the link is working well. The problem with cheating on the standards, is when you do develop a network problem, you probably should eliminate those parts that are outside the acceptable standard in order to narrow down and locate the problem. So, why put them in place to begin with?

There are rules in cascading hubs, and even more stringent rules in cascading switches. There is a length limitation on the one allowable cable between two switches, of course I cant remember what it is, but it seems to me it is relatively short (3m maybe?). I have seen networks crawl or fail when violating this rule.

The main reason you experience failure when violating the length and cascading rules is the amount of time required to push the packet to its location and receive an acknowledgment. Remember if the transmitting station doesnt receive an acknowledgment it will resend. If the cables and hops are too long (translation "too slow&quot, the packet will be sent repeatedly because it doesnt wait long enough for the acknowledgment.

I appologize for not being armed with the reference material, I get frustrated with people who post 'rules' with no back up. I'll try and remember to look it up next week and be sure I'm correct.

I believe the rules change considerably for switches since there is a different amount of latency due to the switch actually examining the data packet and routing it where it belongs. This process takes time, where hubs just rebroadcast blindly.

For pin numbers and wire colors. Looking at your NIC or the ethernet jack on the wall, the narrow notch at the bottom and the wide part of the jack at the top, the pins are numbered 1-8 starting at the left. Looking at your modular plug, with the cable exiting the plug to your left, and the pins facing up, they are from the top down 1-8. 568B would look like this:

1 - white/orange
2 - orange/white
3 - white/green
4 - blue/white
5 - white/blue
6 - green/white
7 - white/brown
8 - brown/white

white/orange is a white wire with an orange stripe. When I make these plugs, i always look down the pins for white - color - white - color - white - color - white - color to make sure none of the pairs got rolled when i slipped the wires into the plug.



Daron J. Wilson
Telecom Manager
LH Morris Electric, Inc.
daron.wilson@lhmorris.com

 

[StyroFoam]

Issues you'll run into with too long cables are late collisions and such- the ethernet preamble is calculated to just long enough that the signal can make it the entire length of a maximum cable length and back to allow enough time for every workstation on the network to realize that somebody is going to start transmitting. If you have a cable longer than this, you may run into instances where one station at the end of your segment will start its preamble, think it's safe to start transmitting a packet, and then have a machine on the other end of the segment start to transmit its preamble as well, causing an illegal collision. This isn't necessarily going to bring your network to its knees, but if you have two very busy segments connected by an overly long hub-to-hub connection, there might be a performance decrease as a direct result.

If your long segment is a Full Duplex one (switch to switch), I imagine you'd be able to get away with this, and only have to worry about any interference...