I know the rule but what is the longest CAT 5 run you have seen working? I once strung together over 1000' of IBM token ring wire, daisy chaining building runs together, and it worked. I was showing the guys how TDR worked and I was trying to get a lot of length, wire types, baluns and connections on the line.

Sounds like you need a fiber backbone there. Cat5 will work for a good distance but the performance will degrade as you go farther and farther. You will have to go pretty far to have it just flat out not work. I made a 1000' patch cord out of a box of cable and just for an experiment used it from my PC to the wall jack and it worked. This is 1000' and approximately 190' of additional horizontal cable to the hub in the closet, it was pretty slow but it did work. Sounds like you need a fiber backbone there.

Actually "there" no longer exists. I strung together a bunch of lateral runs from the distribution panel to get that ~1000'length (out one, daisy chain to another one, back to the panel and out on another one). I was playing with TDR so I wanted lots of ugly stuff, switching from STP to UTP with baluns and such. I even had some damaged cable in there.
I couldn't resist hooking it up to a file server and moving big files around. It still worked.

I have only seen a 5000' reel once of cat5, in the early days of it. (Cheap import knock off stuff) And never more than a 1000 of 5E.

What TDR do you have? I think these are that coolest thing out there. I have a Psyber (m.s.) (Numeric), and the shop has a Megger (Graphic). Both are really usefull for troubleshooting all kinds of stuff. Especially if you sample VoP's of larger wire, like 14 - 2 in conduit or cable. Use it to find screws in cable, and on occasion hidden j-boxes, and splices with the graphic.

Longest CAT5 4 pair 24 AWG UTP run that I saw in use on a 100baseT Ethernet LAN - and had no packet errors, etc. (at least for the time period I was there), was around 2,400 Ft. from Switch / MDF, to the Station.

Saw an even longer one on a LON (Local Operational Network - as in Building Automation Control Systems), which was used for Programming the MBC from a "More Convenient Location".
This was a good 3,000 Ft. - or more!

Here's a point to ponder:

Ever seen a CAT5 or CAT5e cable spliced - using normal Wirenuts (little Blues or Little Grays), and there is no data loss situation?

Seen it at 3 different locations.

Who'da thunk it?!

Scott35

We were doing TDR on a Tektronix 475 scope.
That was why I wanted a lot of wire out there (5ns/cm). We could walk down the trace with the "display after time delay" knob and fairly well determine which segment of wire we were looking at. I know these days TDR has become a magic box that gets you within inches of the failure. This was the pioneer days. It was a kerosene powered scope ;-)

When my last electrician got back from learning to install Cat 5 cable and the requirements on bending radius, he told me how he had tied some cable in knots and then proceeded to hammer the lie (meant to type life, but lie looked good here) out of the cable until the insulation was breaking up. He then placed this on a tester which found the cable at a 100 percent pass rate. I am unsure of the length.

Shane

We had a guy run AS/400 data down a barbed wire fence once to prove the twinax rules were silly but I have seen minor bruises on a cable grind the system to a halt.
As in most things like this YMMV

The spec on a 10/100-base-t run is ni more than 100 meters, 320' or so. If I remember correctly that is part of EIA/TIA 568A. Does that mean it won't work at longer runs? Of course not, but how many times have you seen an appliance powered from 16awg zip cord, it works but that does not make it right.

Actually, it comes from the design of the CSMA/CD protocol used by Ethernet. That stands for "Carrier-Sense Multiple Access with Collision Detect." If you visualize the original Ethernet which had transcievers tapped off of a coax backbone cable, it works like this: When a controller has a packet to transmit, it listens for a carrier on the coax. If there is no carrier, it transmits immediately. Otherwise, it waits for the other transmission to complete, waits the specified inter-packet time, and begins transmitting. While transmitting, it listens for others attempting to transmit at the same time. If it detects this condition, it continues to transmit for the specified minimum packet length, and then terminates the packet.

At that point, it begins to execute the Truncated Binary Exponential Backoff algorithm. Which basically means it waits a random amount of time before attempting to retransmit the packet.

The issue with distance is the collision detection scheme. At the bit rates Ethernet operates at, you've transmitted a lot of bits before the first bit you transmitted reaches the other end of the cable. If you were to transmit a rather short packet, you might complete transmitting the packet before the first bit you transmitted reaches the other end of the Ethernet. Someone at the other end could then, seeing a clear ether, begin transmitting. He would collide with your packet, destroying it. But, since you already finished transmitting it before he stepped on it, you won't realize that your packet wasn't delivered. Your packet is therefore lost.

Lost packets are a really bad thing. Higher-level protocols, like TCP, are designed to be able to recover from lost packets, but a large amount of bandwidth is lost in the process.

So, Ethernet was designed to prevent undetectable collisions. It does this by defining a "slot time", which is essentially the round-trip delay through a maximum-length Ethernet LAN. The spec then defines a minimum packet size that is longer than a slot time. This guarantees that a controller on one end of a network will see a collision caused by a controller at the far end of the network, thus allowing it to enter the backoff algorithm, and assure that the packet is eventually successfully transmitted.

By exceeding the maximum cable length, you may create a situation where collisions result in lost packets, resulting in a substantial loss in network performance.

But wait, the situation is even more complicated than that...

A hub essentially emulates the old Ethernet coax, so my foregoing description applies to anything connected to a hub.

A switch, on the other hand, operates on completely different principles. The connection to a switch is an unshared, point-to-point connection between the switch and the device at the other end. Most switches use a full-duplex version of Ethernet that doesn't involve collision detection. So what I wrote doesn't apply.

Unless... the device connected to the switch is a hub. In which case it does apply.




[This message has been edited by SolarPowered (edited 04-22-2005).]

Thanks for the clarification. So if the long run is a leg on a switch collisions are not a problem? If I am just running on a hub it is a problem?
Otherwise propagation loss is the only issue?

If it's between two switches and/or routers, that should be the case. Between a switch and something else it depends on whether the something else is operating in full-duplex or half-duplex mode. Since you can't always tell what you've got (and folks may change stuff around after you leave), I'd say it's best to just follow the spec.

[This message has been edited by SolarPowered (edited 04-22-2005).]

I think we will all agree if you are selling this installation it should be spec compliant.

SolarPowered;

Thanks for posting the CSMA/CD information!

Scott35

True, 100m is the 'maximum', but ideally you'd have 10m of that reserved-- 5m at the head between the patch panel and network equip, and 5m between the work outlet (user end) and PC (etc).

LOL errr maybe.

I am just a hack when it comes to this stuff.

We where remodeling a large dept store while it was open and 'owned' maintaining all systems.

The demo guys would work nights and end up cutting the POS cabling. It was not Cat 5 but some strange two pair twisted shielded solid hard to work with stuff.

Anyway we had good luck repairing these lines with small grays.

One day we even connected two registers to one home run. As far as we know it work fine.

Well it lit the 'communicating' light on the register and we did not here any complaints from the store people.

Do not worry by the end of the job we had pulled all new POS cable and professionals came in and terminated it.

Bob

You could use wirenuts or pink ladies on a CAT5 and probably get away with it. Just remember to keep the twists in each pair maintained as long as you can and to make a "Y" at the connections with the wirenuts pointed in opposite directions to avoid as much cross talk as possible. It would get you through the weekend but they may not be getting the full speed on the connection.
A slow connection is better than no connection at all when you have a store full of customers.
In real life I can't see why a carefully made splice shouldn't go full speed but I would schedule a proper fix.

BTW Bob that sounds like Token Ring cable you were working with. It really only goes 16mz so you can get away with a lot. If it does a step down it will be running at 4mz and that will darn near run on Romex or thermostat wire. We were running baseband LANs on telco "silver satin" desk set cords at Burger King until we got the right patch cords.

Ya, this sounds like "IBM - Type 1" Cable.

It's "STP" (Shielded Twisted Pair), with two #20-#22 solid twisted pairs, shielded by foil and an Aluminum/tinned Copper braid - all covered by a black PVC jacket.

Like Greg said, this was most likely used on an IBM Token Ring LAN (Local Area Network).

Also mentioned, was the speed of the Ring possibly reducing to 4 MBPS, from 16 MBPS.
That most likely happened, and the POS users on that "Ring" (or out of that MSAU) did not even notice the difference.

I dealt with a lot of Token Ring LANs at Banks, which were of the IBM System 4700 Architecture.
Teller Stations were simply Terminals ("Dumb Terminals" to be exact), which were driven from the 4701 / 4702 / 4705 Controller.
If a teller Station's Terminal changed speed, no one would even notice - mostly because the workload (Data) was so very basic!

As to the Wirenut splices on the Ethernet; there was one location where the Pairs were not twisted, and another location where the Pairs were twisted about 1 twist in +/- 2 inches.
Looked like they stripped back an 8" section of jacket, then untwisted the pairs to make up the splices - but the pairs naturally re-twisted during the make-up work.

These Stations weren't doing anything that could be considered "Power Use", only Logon and pull Templates off the Server, then save things at the end of the sessions.

Don't even know if they (the users at the Workstations) really needed to Logon at all!
They may already had the Template Files on their Machines, and as far as file saves, this may have been done daily / weekly by Administrators, or an automatic archive application!

The reason I mention this (as many of you that deal with Office Workers know), is the Workstation users simply - by default - Logon to the LAN before doing anything... and I mean ANYTHING!!!

The funniest things to hear, when for some reason Server Access is unavailable are:

* The Network is down, so I cannot:

• Work on the file(s) I saved yesterday to my machine,
• Send the document I am looking at right now, to the Printer next to me (local printer),
• End this game of Solitaire,
• Do anything!

As annoying as things may become in this part of the field, I really miss it a lot!
Actually, I miss all the stuff I was involved with, "back in the day!" (prior to 2004).
Now, I rarely deal with anything more than "simple LAN work", and certainly nothing with the complexity of the "Active Remodels" on Bank Branches, Loan Centers, and Corporate Campuses!

Makes one feel kind of used up and worthless.

Sorry for the sob story, it's bed time for me!

Scott35

Thanks Greg and Scott for the info.


I don't think so as the cable was red and much thinner overall than the black IBM cables we used to install for another retail customer.

If I remember the IBM 'plugs' where fairly bulky square and black.

This place used another type of connector, I can not describe it as I did not have much to do with it.

The specifications required we leave 15' of the cable in each floor box! (That did not happen even with the large floor boxes we where using (4" round x 6" deep)

The idea was they had many locations where they could place registers through out the store as needed.

Glad most of our customers have gone wireless with portable cash registers.

If it helps the stores where owned by 'May Company' if anyone has worked for them in the past.

Token Ring will run on any 2 pair shielded cable or even 4 wire UTP with baluns (cat3 or 5). The type one cable and ICS connectors only come if you buy the IBM cabling system.
I have also seen oversized RJ plugs similar to a phone plug, D shell connectors in various sizes or pin and sleeve Molex type plugs. In older bank stuff you even had gold 1/4" phone plugs.
There are also "bank loops" "store loops" and RS422 connections that run on 4 wire twisted pair.

Oh ya, seen lots of that stuff too!

It was used with the COINS, COLTS, &/or BATS Systems.

The connectors were 1/4" Phone Plugs + Jacks (Panel Mount type Jacks), with the "TIP, RING, SLEEVE" configuration.

Data Cable was a single shielded pair, with a jacket that made it look almost like "Zip Cord"!

On the Type 1 terminations, I have also seen several connector types used here and there.

Still have a bunch of patch cords with a DB-9 or DB-25 connector on one end, and the "typical" big-old black Amp connector on the other end. Patch Cable is Type 1 STP (with a CMP rating too! Never thought of putting a Workstation in a Plenum Ceiling; but if the opportunity ever arises... ...)

Seems odd, but I ran across more LANs with Type 1, or some similar type of paired Cable (the "Zip Cord" like cable of "COLTS/BATS" systems, or UTP in CAT 2 and 3 flavors), than with "Thin-Net" (RG-58 / 59 / 62u, or simply COAX).

LANs using "Thick-Net" were scarce - except at Corporate Campuses.
Had to bring Garlic with me, to scare off "The Vampires" - which were "Tapped" into the LAN Cable, at the NICs of each Workstation


Excuse me, I meant the:

"IBM Token-Ring Network PC Adapter"



Scott35

My office was wired with type 6 cable. It was fatter with the token ring set and 2 twisted pair for the phone. This terminated in the combo plate with the ICS connector and an RJ14 phone jack.
There is also a silver plenum rated token ring cable with a clear teflon jacket over the silver braid. I still have a bunch of that. I have made long PC keyboard and mouse cables with it. Works great for line level audio too.

I have a 50ft long piece of Cat5e running from my router, under the carpets, along the baseboard in the hall, to my filer server iMac. It has about five electrical tape spilces in it, from where my dog chewed it up and I didn't have any wirenuts. I un-twisted the pairs about 1", and I've seen nodegradation of performance. At one time I had used miniature terminal blocks and simply twisted the wires around the screws. The only time I have/do see a performance decrease is when the Cat5 runs parallel to a electrical cord. Oh, and did I mention, I also have a piece of Cat5 with RCA connectors spliced on for audio use. 25ft and it sounds fine, even up to 512kbs MP3s!

Ian A.

Ian A.

There is a big difference between 10BaseT and 100BaseTX. The old 10BaseT standard was restricted mostly by signal strength and attenuation. At 100mbps, though, the timing is SO tight that if the cable is really ANY longer than 100m, the bits will not reach the recieving terminal in the allowed time, an error is detected, and collisions occur. 100BaseTX can start to have problems with as little as 102-103m! Doesn't matter how high quality the cable is, or whether it's Cat5e or Cat6, as the wall is timing related. Luckily for us, most modems are 10/100 auto-sensing and switch to 10BaseT without the user even noticing And 10BaseT is FAR more forgiving- if high quality shielded cable is used and connectorized well, 500-600'+ is possible, though not really recommended Theretically, I suppose 10BaseT could work up to 1000m if the signal is strong enough.

I've seen 10BaseT work over cat5 as long as 500', but with issues.

[This message has been edited by SteveFehr (edited 08-19-2006).]

I know this thread is a bit old, but I wanted to add a couple of things.

If the Ethernet link between devices is running full duplex, this disables CSMA/CD and since no collisions can occur, there is no distance limitation.

Any limitation on the cable length would be the result of physical properties of the cable such as attenuation, crosstalk, etc.

[This message has been edited by 1000BaseT (edited 03-08-2007).]

It's been my experience that autosensing 10/100 devices will not automatically switch to 10mbps mode when there is a cable issue preventing successful operation at 100mpbs. They're continue to operate at 100mbps and little or no traffic will get through.

You either have to fix the cable issue (which may not be possible, if for example it's a run of Cat3 that can't be replaced without a lot of work) or switch the device to operate in 10mpbs mode (which may also not be possible if there's an unmanaged 10/100 switch at each end).

If there's a PC on one end it's easy enough to force it to operate at 10mbps instead of autonegotiating and ending up with a 100mbps link that isn't working.

What Brian said makes sense to me. As I remember the autonegotiation process, it doesn't attempt to measure the quality of the line, it just exchanges data that says, "I can run a these speeds; what can you run at?" And then it picks out the fastest speed.

Fiber would have been ideal but I saw one in an office suite use about 1100'.....was he getting 100 meg? Close he was getting spikes of 25 to 75 steady but I would never try that myself for a client.