Diagnosing "split pair" 1

[theEclipse]

Hi cabling experts...can you help me understand this one?

My site has two buildings. We just moved in and I am charged with putting both buildings on the same network. There is a buried telephone trunk line between the two buildings (2x25 pair) which terminates into 66-blocks. Total cable length is probably 200'.

In an attempt to use part of the cable as a cat-5 cable I punched a patch cable into the blocks on both ends, on empty pairs. When I put my el-cheapo cat5 continuity tester on the run it lights up all wires in the correct order. When I put a fancier tester on the run it tells me that I have a split pair and signals 3-6 & 4-5 as places to look. 6 years of making cat-5 and I have never herd of a split pair until I used this fancy-schmancy tester (which probably means that there is a bunch of this out there...).

Some googling leads me to this: "A split pair is defined as a transmit or receive signal that is carried on a single wire, with no opposing wire to help filter out unwanted noise and crosstalk from a neighboring pair. It commonly occurs when a punch-down block is wired incorrectly or when RJ-45 connectors are crimped onto the wrong wires."

Which is great, except that I cannot fathom how this is happening. If my continuity tester gives me a straight-through reading on all wires, how can I have a split pair?
I checked and re-checked my terminated ends of my patch cables and the patch cable side of the 66 block -- all in order. I re-punched the patch cable and even checked for the correctness of the wires where the cable comes into the 66 -- all good.

My testers are (cheapo continuity tester)

http://tinyurl.com/6bgfpo

and

http://tinyurl.com/59lm9v

. The documentation

http://www.idealindustries.com/media/pdfs/products/instructions/62-200_instructions_v2.pdf

which comes with the tester has a diagram of a split pair but I can't figure out how that would work given the readings of my continuity tester.

How can I better diagnose this problem?
Is my continuity tester worthless against this foe?

Thanks

Robert Carpenter
Remember....eternity is much longer than this ~80 years we will spend roaming this earth.
ô¿ô

http://www.robertcarpenter.net

 

[lopes1211]

Can't say why/what the el cheapo tester is doing but your split pair is probably the riser termination, not your work which you have checked twice.

Punching cat 5 cable on the end of "telephone cable" is not going to give you a cat-5 100MB supported connection.

-CL

 

[lopes1211]

As far as finding it. you'll need a person on each end of the riser and an ohm meter/continutity tester. have your partner short wire 1 and 2 together. you test your end for continuity between those 2.

Hopefully wire 1 is good an repeat this process shorting

1&3
1&4
1&5
1&6
1&7
1&8

Somewhere in this process, your partner is going to short a pair of wires (lets say 1&3) but you'll see the continuity on a different pair (lets say 1&4). that will identify your split pair. Of course there would always have to be 2 split pairs as your correct wire must have gone somewhere!!

-CL

 

[wires]

"Is my continuity tester worthless against this foe?"

Yes. Your continuity tester will not find a split pair. All it checks for is continuity, not anything having to do with pairs.


"How can I better diagnose this problem?"

** First understand the problem.

A split pair is where the tip and ring conductors are not on the same twisted pair. There is a nice picture halfway down the page here:

http://www.idealindustries.com/media/pdfs/products/instructions/62-200_instructions_v2.pdf

A split pair or non-twisted pair system will cause problems with data as it does not reject noise and reject crosstalk as well as a balanced twisted pair system.

** Second understand how 100BaseT twisted pair networking works.

In a 100BaseT network there are two pairs, Transmit and Receive. These signals travel on pins 1&2 and pins 3&6 of a 8P8C (RJ45) connector. These are the orange and green pairs.

Pin 1 is TX+
Pin 2 is TX-
Pin 3 is RX+
Pin 6 is RX-

It makes NO DIFFERENCE if the pair on pins 1&2 is orange or green as long as you are consistent in staying with the same color on the same pin.

While the blue and brown pairs are not needed for 100BaseT it is SOP to connect them for physical strength and future use.

** Third after getting a split pair reading on a tester the easiest way to fix the problem is to physically inspect each termination point.

You could just keep trying wires on your tester and find a pairing scheme that does not produce an error. However if you are sure about you pairing a split pair reading may indicate a hidden splice of poor quality or the next problem.

Split pair detection is usually done by having the tester look for excessive crosstalk. In your case it could be that the pairing is correct but there is not enough twist to the pairs or there is cable damage that is driving up the crosstalk. Both of these will cause problems for a 100BaseT network.

For example if I connect my Microscanner Pro to a piece of flat silver satin phone cable it will show split pairs even though the pinout is the same on both ends. As a consequence 100BaseT does not work well on silver satin...


Good luck!

 

[theEclipse]

lopes - thanks for the idea. I did this and came out with no errors. This reinforces my continuity tester as being OK...it is supposed to just test the wires straight through.

wires - thanks for the details. I have been looking at the diagram you linked to and cannot figure out how my continuity tester would not show that error. My tester has lights on both ends (remote and base) and it lights up each wire individually and in order. I guess I don't know where it puts the current's return path as it goes through the wires...but I'd assume it puts it through the other wire in the pair. My usage procedure for my continuity tester is as follows: connect the base to one end and the remote to the other (things start lighting up...) check that both of the ends show the same wiring order (ie. the lights flash in order on both units).

I am going to purposly split a pair on a patch cable and put it in my tester.

Now, what you said about the nicer tester checking for excessive crosstalk and flagging it as a split pair. If this is the case, how can I solve this problem? I have about half of my buried cable not in use...should I just start putting wires in odd places until I can get a signal through?

Thanks for reading and responding

Robert Carpenter
Remember....eternity is much longer than this ~80 years we will spend roaming this earth.
ô¿ô

http://www.robertcarpenter.net

 

[theEclipse]

Also...*other than the number of wires in the cable* what is the difference between cat5, 5e, 3, 6? Is some just held to a higher quality standard? Is the rate of twist different? Is the conductor material different? Bah! so many possibilities...

Robert Carpenter
Remember....eternity is much longer than this ~80 years we will spend roaming this earth.
ô¿ô

http://www.robertcarpenter.net

 

[strmwalker]

give this a thought lets say you punched down a cat5
cable on the last four pairs on the first 100 pr cable
that will be pairs 97,98,99,100 thats v/o,v/grn,v,brn
and v/slt your cat5 has w/bl,w/o,w/grn,w/brn how did you
punch down the cat5 cable keep in mind the pinouts on
the cat5 patch cord depending on which configuration
you are using 568a or 568b that might be what is causing the
split pairs.

no problems only solutions

strmwalker

 

[theEclipse]

strmwalker- I can't quite tell what you are saying, but I put my cat-5 cable into 66 as if it were a cat5 connector.

yellow/slate - white/orange
slate/yellow - orange
violet/blue - white/green
blue/violet - blue
violet/orange - blue/white
orange/violet - green
violet/green - brown/white
green/violet - brown

violet/brown and violet/slate are actual phone lines going between the buildings.

----------------------------------------

Test Results
I wired up a cable to test the different mis-wirings. Its two patch cables connected by two wall jacks. The wall jacks allowed me to pull wires and rearrange them without sacrificing another cat5 connector.

I tested the following miswires: [ul]
[li]Swapping pin1 with pin6 (ie. tip-tip as the pdf from ideal's website diagrams as being a "split pair") [/li][li]Swapping pin1 with pin3 (ie. tip-ring swap across two separate pairs, which is written as description for a "split pair" and is the caption for the picture of the above)[/li][li]Swapping pin1 with pin2 (ie. a "reversal" of tip and ring)[/li][li]Swapping pin3 with pin5[/li][/ul]

All of the above resulted with the ideal tester indicating "Miswire" and highlighting the correct pairs as being miswired, except the reversal, which was also indicated correctly.

All of the above caused my continuity tester to blink out of order on the remote end - as I expected.

Robert Carpenter
Remember....eternity is much longer than this ~80 years we will spend roaming this earth.
ô¿ô

http://www.robertcarpenter.net

 

[LkEErie]

Wow, if you wired the cable as you see it, don't you see the split pair? You have split the blue pair and the green.
The proper order is 54123678. So the pairs go in blue/orange/green/brown order on the violet binder.

Don't expect 100 base T transmission on Cat 3 or less cable. A better plan would be to get an ethernet to single twisted pair transmitter such as the Netsys Direct Ethernet Extender boxes. <

http://www.huffingtonpost.com/2008/08/31/sarah-palin-photos-a-bust_n_12

2816.html>

Expect bo pay at least $300 for the solution.

LkEErie

 

[SYQUEST]

To verify the integrity of the cable between the buildings, try this:
1. Pull off the patch cords, and check the continuity of each pair separately. You can do this by putting a short on each pair individually at one end.
2. After you have the pairs shorted, check each for separate continuity and against the other pairs for a short. Make notes of your results. This should help narrow down the pairs affected.
3. One last thing to check are the pairs at the connecting blocks. Look at the COLOR code of the wires. The original installer may have split one of the pairs at the time of install or rolled the Tip and Ring. As you have the colors listed in your post, that is the correct order, Tip first, then Ring.(Y-SL/SL-Y,V-BL/BL-V, etc.)

Now I do see an error on your part in your post, the leads for your PATCH cord, W-GN & W-BL pairs are crossed or split the way you show them in relation to the feeder cable. You should be pair for pair consistent.

Hope this helps!

....JIM....

 

[wires]

I tested the following miswires:"

My guess is that you didn't swap the pins on BOTH ends. If you just swapped one end you will get a miswire, because that is what it is. To get split pairs the incorrect wire sequence needs to be on BOTH ends.

A split pair cable can look like this:

pin1 white-orange
pin2 green-white
pin3 white-green
pin4 blue-white
pin5 white-blue
pin6 orange-white
pin7 white-brown
pin8 brown-white

Make a cable with BOTH ends wired as above. You should get a perfect continuity test but split pair warnings on pairs 2&3 (orange and green).

Or this, all 4 pairs split:

pin1 white-orange
pin2 green-white
pin3 white-green
pin4 brown-white
pin5 white-blue
pin6 orange-white
pin7 white-brown
pin8 blue-white


In the end you might find that the cable you are testing is not sufficient to support modern ethernet due to poor noise and crosstalk suppression. If that is the case there are several solutions like Cisco LRE (discontinued) that create a "DSL" type connection over a dry telephone circuit. Here are some starting points:

http://www.zhone.com/products/SNE2040/

http://www.arcelect.com/patton_2168_ethernet_extender.htm

ethernetextender.com/ethernet%20extenders/865.php

http://www.motorola.com/business/v/index.jsp?vgnextoid=f5f9f6fd97b45110VgnVCM1000008406b00aRCRD

I know that several forum members have experience in this technology. If they don't reply to this thread you might want to start a new one with a title like "Connect two networks over buried telephone line".

Whatever you do with copper you will need to provide primary surge suppression on both ends of the buried cable.

If the cable is in a conduit then you should consider doing the correct thing and pulling a fiber to make the connection. The benefits (network speed, surge suppression, new vs old) far outweigh the long term cost. Fiber IS the way to go, reusing a buried telephone line is a hack that will come around to bite you.


Good Luck!

 

[wires]

LkEErie:

The unbroken url to the "WTF I might as well be VPOTUS even after my family releases these BS pictures":

http://www.huffingtonpost.com/2008/08/31/sarah-palin-photos-a-bust_n_122816.html

I guess you had the wrong url in your copy and paste buffer.

Here is a link for the Netsys part:

http://www.netsys-direct.com/proddetail.php?prod=NV-600EKIT

 

[LkEErie]

Thanks, but I was actually pointing to the home page that shows both units.

http://www.netsys-direct.com/

The paste moster got me

LkEErie

 

[theEclipse]

LkEErie -

Wow, if you wired the cable as you see it, don't you see the split pair? You have split the blue pair and the green.
The proper order is 54123678. So the pairs go in blue/orange/green/brown order on the violet binder.

I think I see the problem now. So instead of wiring up my patch cable to the 66 block as if it were an 8p8c connector, I need to wire it up so that all the pairs are consistent throughout the cable.

My guess is that you didn't swap the pins on BOTH ends. If you just swapped one end you will get a miswire, because that is what it is. To get split pairs the incorrect wire sequence needs to be on BOTH ends.

So a split pair is not where the wires are misconnected, but where the wires that are connected are not twisted together in the correct pairs. Wow...I wasn't even looking in the right ball park.

I did try that once...but I am thinking that my test cable wasn't long enough to develop crosstalk.

I'll try that tomorrow. If I can't get a signal through then I'll take a look at some sort of extender...though budget is probably prohibitive in that area.

Thanks very much for all your help.

Robert Carpenter
Remember....eternity is much longer than this ~80 years we will spend roaming this earth.
ô¿ô

http://www.robertcarpenter.net

 

[wires]

If I can't get a signal through then I'll take a look at some sort of extender...though budget is probably prohibitive in that area."

Remember about surge suppression.

What you are doing is making an electrical path between two buildings. Each building has it's own electrical system. During a lightning strike or other electrical surge the difference between these two electrical systems can be thousands of volts. Additionally thousands of volts can be induced into underground cables during a lightning strike. This is WAY more than any computer and network equipment can handle.

Apparently you are working on a limited budget. The problem is that if you make this copper connection without adequate surge suppression where will the money come from when the entire network is destroyed by a surge?

At the very minimum you will need primary surge protectors in each building that are DIRECTLY connected to the main electrical ground in there respective building. These devices may already be there. If so be sure they are good condition and the ground wire is splice free, corrosion free and tightly connected at both the suppressor and ground point.

Here are some links:

http://www.twacomm.com/catalog/jmp_ITW+Linx/dept_id_602.htm

http://www.twacomm.com/catalog/model_UP3B.htm?sid=4848156F0EF2AA1311F74D22EC6F7F2D

http://www.amazon.com/gp/product/B0009MU5M2

No surge suppressor works better than it's ground connection. BE SURE you have a GOOD GROUND!

However if you go with fiber you don't have to worry about ANY of this, and it is WAY faster!


Good Luck!

 

[wires]

So instead of wiring up my patch cable to the 66 block as if it were an 8p8c connector, I need to wire it up so that all the pairs are consistent throughout the cable."

Correct. White-blue should be on the same color wire for the entire run. Same with all the other wires.

One other consideration. 66 blocks are not designed for stranded wire. So if you are cutting a patch cord in half and using part on each end it may work for a while and then loose connection as the strands shift in the vee section of the 66 terminal.

If you are making your own patch cables from solid wire be sure you are using 8P8C plugs that are designed for solid wire. Solid wire 8P8C pins have a little "saddle" where they displace the insulation while stranded pins have a "spike" that stabs through the insulation into the strands. These days many 8P8C plugs have both and are rated to work with either solid or stranded. You need to check.

The issue with both the 66 and 8P8C connections is that they are both use IDC (insulation displacement connection) technology. This has fairly strict requirements for the type of insulation and wire that will work long term. I have fixed this sort of thing more than once where the previous tech had put stranded into 66 or 110 blocks or used incorrect plugs.


Good luck!

 

[theEclipse]

Success! Pairing the pairs correctly gives me a stable 100Mb connection. I connected my mac to it and connected to a network share and dumped files across the wire while watching the error count in network monitor. A full CD iso has gone across and my error count is at zero for both send/receive. Remote desktop to the server performs great (not that it requires alot of bandwidth, but still).

However if you go with fiber you don't have to worry about ANY of this, and it is WAY faster!

I wish. The cable between the buildings is directly buried...no pulling fiber through the conduit. Moreover, even if there was conduit, I don't think we could afford to have someone pull fiber and terminate it, let alone the actual equipment costs. Indeed, I wish.

Remember the surge suppression

Oooh...good call. I'm buying two of the lan surge suppressors.

Thanks to everyone for all your help. Without your help I'd probably have given up and put in a wireless bridge between the two buildings.

Robert Carpenter
Remember....eternity is much longer than this ~80 years we will spend roaming this earth.
ô¿ô

http://www.robertcarpenter.net

 

[SYQUEST]

theEclipse,

A limited budget is no excuse for violating the NEC in regards to proper entrance cable surge protection!

Another thing to keep in mind when testing cables with the current type of cable testers available today is: they use a technique for measuring the capacitive characteristics of the pair(s) end-to-end not the continuity in the way a VOM would measure a pair or lead. So this may be why you got different readings when using a VOM verses a Cable Tester. The tester measured all 4 pairs and was able to "see" the split.

Hope this helps!

....JIM....

 

[wires]

Pairing the pairs correctly gives me a stable 100Mb connection."

Great!

There IS a reason for those twists!

 

[aarenot]

This direct burial cable may have met the local, and NEC when it was installed for all we know. I do not know the NEC from 20 years ago in this area of subject matter. It may well have been in there for that long.