Cable Length 2

[hosikuma]

What is the maximum lenght of Cat5 that can be put between cascading switches? I have two wiring closets in this building. Half the oulets go to one closet and half the others to another patch panel in another closet. I would like to join these two segments together...so i would like to put another switch (hub)in the second closet and running a cable from the switch in closet one to a switch in closet two. the distance between the two closets is about 250 feet.

Would anything be different in terms of cable lenght in terms of running a lenght of Cat5 from a switch to a router?.............thanks guys!!

 

[franklin97355]

The spec says 100m total of which 10 can be jumpers and patch cords.

 

[daronwilson]

You should have no problem with a 250' run. Generally you should limit the length of installed permanent cable to 90 meters (about 295 feet) and allow for 10 meters of jumpers total between the two ends.

Of course anytime you are talking a backbone, you may want to consider fiber for future growth if it is going to be a permanent backbone installation.

Good Luck!

It is only my opinion, based on my experience and education...I am always willing to learn, educate me!
Daron J. Wilson, RCDD
daron.wilson@lhmorris.com

 

[mrpeabody]

I would suggest using a Cat6 cable. It has a little less resistance over that distance, and when a Cat6 switch becomes available/affordable, you can upgrade for more speed.

 

[buitenhek]

Cascading hubs, switches or whatever is a great way to go with a few exceptions. You are limited to 5 segments and 4 hubs between end points. 2 of these hubs cannot be populated.

Plus if you cascade one port to another hub or switch, if that port fails you have xx people down. I suggest running 2 CAT5e (min) lines between the prime and cascaded hub. Keep one cable as a backup in case of a problem…(do not hook up this 2nd line or you will have oodles of problems). This way you can diagnose the problem whilst the customers are up and running.

Final issue is if you have a hub (or switch) that has (usually port1) available for extension. You would obviously need a cross-over cable.

Regards,
Peter Buitenhek
peter@profitdeveloper.com

 

[Robhub]

You need to be aware that cascading hubs/switches, etc., isn't an approved EIA/TIA-568B cabling method. While it may be OK as far as IEEE standards are concerned, it's going to take you outside the accepted cabling standards for commercial/industrial applications. If this is your own building you may not have a concern along those lines, but if the project is for a customer who expects to have his system 568B-compliant, you're going to have a problem. Good luck.

 

[hosikuma]

Robhub......sorry, don't quite understand. Are you saying that one should not cascade hubs or switches? If I wanted to expand an existing hub/switch how would I do it to comply with 568B standard?.....thanks!

 

[daronwilson]

I'm confused on a couple comments.

Cascading hubs, switches or whatever is a great way to go with a few exceptions. You are limited to 5 segments and 4 hubs between end points. 2 of these hubs cannot be populated.

Where is this specific information coming from? While there are very specific requirements for hubs and segments, I don't know that they blanketly apply to switches. I've never heard the 5/4 rule applied with the requirement that 2 of the hubs not be populated. Sorry I don't have my ethernet books with me this week, but can you point me to someplace that is explaining this?

You need to be aware that cascading hubs/switches, etc., isn't an approved EIA/TIA-568B cabling method.

Can you explain this? I dont see where 568B prohibits connecting two switches together. Rather than me go through the entire standard again, can you just help us out by pointing to the section that you feel does not approve of cascading switches?

Thanks!


It is only my opinion, based on my experience and education...I am always willing to learn, educate me!
Daron J. Wilson, RCDD
daron.wilson@lhmorris.com

 

[Robhub]

Daron,

In your post, you cite 568B as allowing 2 switches to be connected together. There is no problem with that scenario. I was addressing the issue of having 5 hubs and 4 switches between end points. This seems to me to be clearly outside the 568B standard.

 

[jimbopalmer]

I think the 5/4/3 rule was 10base2 ethernet, not 10baseT or its sucsessors

I tried to remain child-like, all I acheived was childish.

 

[daronwilson]

Rob,

Hooking two or more switches together is generally what I would refer to as cascading switches. Unless there are specific higher speed ports on the switch for connecting them together, you are likely going to cascade them via the 10/100 ports.

The 5/4/3 rule was indeed part of the old 10Base2 and 10BaseT design for hubs. I do not believe this rule had anything to do with 100BaseT or switches.

So, again, please point me to where you are getting this information:


You need to be aware that cascading hubs/switches, etc., isn't an approved EIA/TIA-568B cabling method.

I've got my entire up to date telecom standards on CD here and I can't find that.

Buitenhek,

Cascading hubs, switches or whatever is a great way to go with a few exceptions. You are limited to 5 segments and 4 hubs between end points. 2 of these hubs cannot be populated.

Help me out, where are you getting this information? Again, you are close to using the 5/4/3 rule but that applies to hubs, not switches, from what I have read.

Share your sources!



It is only my opinion, based on my experience and education...I am always willing to learn, educate me!
Daron J. Wilson, RCDD
daron.wilson@lhmorris.com

 

[franklin97355]

You need to be aware that cascading hubs/switches, etc., isn't an approved EIA/TIA-568B cabling method.

I've got my entire up to date telecom standards on CD here and I can't find that.

I think that's what he meant. You can't find it in the spec. so it's not an approved method (it's not dis-approved though)

 

[daronwilson]

You need to be aware that cascading hubs/switches, etc., isn't an approved EIA/TIA-568B cabling method.


I guess my point is more that the 568B standard doesn't cover cascading hubs, that I can find, nor is 'cascading hubs/switches' a cabling method. The comment infers that if you cascade switches, your installation will not be 568B compliant. For that to be the case, there must be something in 568B that says you can't cascade switches.

Maybe I'm not understanding his approach, but as I read it the 568B stardards don't deal with connectivity of the active components required to make your network work.

It's making me read though!


It is only my opinion, based on my experience and education...I am always willing to learn, educate me!
Daron J. Wilson, RCDD
daron.wilson@lhmorris.com

 

[buitenhek]

Thanks for the support to all who wrote me personally. I too dusted off some old networking books. To the only junior expert detractor; it is physically impossible, for me at least, to get my head in there in the first place.

Cascading:
1. This is an electrical issue not cabling related and thus not “compliance” addressed (yet) in 568B standards. It is perfectly legal nomenclature for Ethernet; provided the max distance is not exceeded from the hub port to end user (delayed packets due to cable length causing electrical reflections and collisions).
2. 5/4/3 rule comments were very observant and correct on Ethernet 10BaseB2, BT and may I add 10Base5 as well. The number of repeated signals (latency) was the driving factor. The same factor considered box layout for both a pass-through (hub to hub) and cascade (hub port to new hub). If you remember the early hub mfg’s employing the passive backplane to get around the 5/4/3 rule of cascading hubs in a pass-through environment.
3. Note original question was distance between switch (hub) and 2nd location of 250 feet. The follow on question was cabling differences between switch and router.
4. Although not addressed in the original question, the issues surrounding copper/fiber cable length limitations in a 1000BaseT were very informative and helpful. I can see many knowledgeable and expert people monitor this Tek-Forum in silence mode. Perhaps a solid bit of advice for me!

End-of-line

 

[daronwilson]

I guess we should start a new thread perhaps. My goal was to avoid giving people misinformation here when they ask for help. I certainly don't have all the answers, but in this particular post we've given out some wrong information. If I have offended you buitenhek I appologize, and hope you don't think I suggested you put your head up there. I'm simply on a quest for accuracy, and if I have provided misinformation, please help me get it straight so that everyone learns.

1. You need to be aware that cascading hubs/switches, etc., isn't an approved EIA/TIA-568B cabling method. Well, having a switch by the bathroom door for the vent fan is not an approved EIA/TIA-568B cabling method either. Why? Simply because the scope of the EIA/TIA-568B does not address this issue, nor is it concerned with how you cascade hubs. Cascading hubs will not provide you with a non-compliant EIA/TIA-568B installation.

2.Cascading hubs, switches or whatever is a great way to go with a few exceptions. You are limited to 5 segments and 4 hubs between end points. 2 of these hubs cannot be populated. Cascading hubs and switches are NOT the same thing. Hubs are ALL on the same collision domain, i.e. the all are on the same 'channel' and collisions, latency, etc. are a very big issue. Switches are NOT all on the same collision domain, so a collision on port 1 won't even be heard on port 12 or whatever. You cannot blanketly apply the same rules to both. I cannot find anything about the 5/4/3 rule that says 2 of the 4 hubs cannot be populated, and I believe this is incorrect.

3. To answer the question with a standards based solution, I refer to EIA/TIA-568B.1 (5.5) Backbone cabling distances. This section allows backbone Cat5e twisted pair to 90m, allowing for 5m patch cables at each end. This same length is allowed for any ethernet device between two points. If you wish to take the LAN port of your router and drive it 90m to your switch, that would be acceptable as well.

If I am in error, please let me know where, and I'd be happy to learn more about it.






It is only my opinion, based on my experience and education...I am always willing to learn, educate me!
Daron J. Wilson, RCDD
daron.wilson@lhmorris.com

 

[jimbopalmer]

I cannot find anything about the 5/4/3 rule that says 2 of the 4 hubs cannot be populated, and I believe this is incorrect.

Daron, I think this is a misinterptation of how you could cascade 10base2 (and perhaps 10base5) 2 of the 5 repeeater segments could not have end devices. Even if true, it is not currently useful, as twisted pair 'won' the inside wiring 'war' at least a decade ago. Very little co-axial ethernet is still in use, I hope.

I tried to remain child-like, all I acheived was childish.

 

[sweetiepie]

Hi mate, this is my first time on this site, great is'nt it, i am always getting stuck on my p.c at home, i have been a Cable Engineer for 10 years now so i know my stuff the wiring side, but not so good on the hardware side, the answer to your question is 90 meters maximum on cat 5 regulations, but give or take a few meters would'nt matter at all, hope this helps, Chris, Brighton, U.K