Cat 6 termination, A or B?

[JFletcher]

Ahhh! Now there's a breath of fresh air!
-Hal

You forget I was a regular at the Sundance forum for some years (as you were)... and I figured I would correct 'em before you got ahold of them.

 

[kwired]

That's what we thought too, until the cameras didn't work. When we tried one of the cable supplied with the cameras it worked. When we changed the pinning to reflect what the manufacturer specified because of the individual pair twists it worked.

is possible to have a wiring error and not use designated conductors of a pair, but you need to make same mistake on both ends of the cable for it to pass testing, then you still have proper continuity from one end to the other, just that not all "pairs" will be individually twisted like intended and it can reduce performance levels.

 

[JFletcher]

That's what we thought too, until the cameras didn't work. When we tried one of the cable supplied with the cameras it worked. When we changed the pinning to reflect what the manufacturer specified because of the individual pair twists it worked.

There is no standard for pair twist rates of different colors. One mfg might have the blue pair barely twisted (lowest rate), others might have it the tightest. It's likely you had a bad termination the first time and a good one the second.

I'd be willing to bet I could take a piece of cat 6 cable and wire it following 568B standard tho screwing up the colors and as long as both ends are identical, it would certify. Identical meaning brown on 1 and 2, blue on 3 and 6, orange on 4 and 5, and green on 7 and 8 (instead of the correct orange, green, blue, brown, respectively). The pair twist rates are important only in that they differ from one another.

One can take good cat5e cable that was installed to EIA/TIA spec (supported every 5', not kinked, not overpulled, under 100M in length) and put it down on cat6 keystones and patch panels and get it to certify cat6. Neat little trick should you ever get a customer that wants an upgrade but doesnt want to recable the entire building. Now 6a, dunno if it would pass that (Ive only ever done the aforementioned trick 'on the bench', mind you).

If you remember that camera mfg and stick by your story, please PM me the name - I never want to buy something that does not work to TIA/EIA specs.

btw, anything that works on cat5(e) will work on 6, 6a, and even 7 - and those cables have progressively tighter twist rates.

 

[JFletcher]

is possible to have a wiring error and not use designated conductors of a pair, but you need to make same mistake on both ends of the cable for it to pass testing, then you still have proper continuity from one end to the other, just that not all "pairs" will be individually twisted like intended and it can reduce performance levels.

If you punched down the pairs on pins 12, 34, 56, and 78, you'd wind up with 2 split pairs instead of 1 - green should be on 3 and 6, blue on 5 and 4. Even if identical at both ends, you'd be using one wire of the green pair and one of the blue as pins 1236 are used for signal transmission - that could get very ugly. Such a punchdown would never pass even a cheapie tester tho.

 

[kwired]

If you punched down the pairs on pins 12, 34, 56, and 78, you'd wind up with 2 split pairs instead of 1 - green should be on 3 and 6, blue on 5 and 4. Even if identical at both ends, you'd be using one wire of the green pair and one of the blue as pins 1236 are used for signal transmission - that could get very ugly. Such a punchdown would never pass even a cheapie tester tho.

You may have data speed issues, but if you accidentally swapped say the orange and blue, but kept the corresponding white that goes with each in the designated position, a continuity only test would still pass if you swapped them on both ends.

 

[gadfly56]

Ahhh! Now there's a breath of fresh air!



Lets not get carried away now. The USOC designations are still used although many designations like RJ-45 have no use today.

RJ-11 is your standard 1 pair phone jack or plug- 6p4c or 6p6c with T&R wired to the center pins 3&4.

RJ-14 is a 2 pair 6p4c or 6p6c jack or plug with the first pair on 3&4 as above and the second pair on 2&5. It could be used for two lines or commonly for office phones that use two pairs.

Extra credit: Does anybody know what a RJ31x jack is?

-Hal

Don't recall how it's wired off hand, but it's for line seizure by security or fire alarm equipment.

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

Whoops! Answered before reading the rest of the posts. Day late and a dollar short.

 

[luckylerado]

Neat little trick should you ever get a customer that wants an upgrade but doesnt want to recable the entire building.

Probably true 99% of the time but if your customer actually has or intends to on day upgrade to equipment that is capable of taking advantage of data transfer rates above 1 gigabit then CAT6 jacks and patch panels on CAT5E, regardless of how well installed, will not be sufficient. I would just make sure the trick is not pulled without that understanding.

 

[Fulthrotl]

I'd be willing to bet I could take a piece of cat 6 cable and wire it following 568B standard tho screwing up the colors and as long as both ends are identical, it would certify. Identical meaning brown on 1 and 2, blue on 3 and 6, orange on 4 and 5, and green on 7 and 8 (instead of the correct orange, green, blue, brown, respectively). The pair twist rates are important only in that they differ from one another.

One can take good cat5e cable that was installed to EIA/TIA spec (supported every 5', not kinked, not overpulled, under 100M in length) and put it down on cat6 keystones and patch panels and get it to certify cat6. Neat little trick should you ever get a customer that wants an upgrade but doesnt want to recable the entire building. Now 6a, dunno if it would pass that (Ive only ever done the aforementioned trick 'on the bench', mind you).

couple thoughts....

i read that the turn differences of the pairs were
a ratio to each other, not as absolutes on their own,
so that a manufacturer, as long as those ratios were
observed, could produce compliant cable.

i've got two cable certifiers that, while neither of them
is a $ five digit device, every cable i've ever tested with
them, at cat 5 or greater, assuming it's wired correctly,
has never failed a cat 5 cable or terminations at 1ghz cat 6.

 

[JFletcher]

Probably true 99% of the time but if your customer actually has or intends to on day upgrade to equipment that is capable of taking advantage of data transfer rates above 1 gigabit then CAT6 jacks and patch panels on CAT5E, regardless of how well installed, will not be sufficient. I would just make sure the trick is not pulled without that understanding.

Agreed. If the customer wants future proofing, we always recommend conduit vs cable-flavor-of-the-moment.

Some of the best Cat5e cable comes very close to meeting cat6 spec:

"CommScope Enhanced Category 5e (Cat5E) cables exceed ANSI/TIA-568-C.2 Enhanced Category 5 (Category 5e) and ISO/IEC 11801 Class D
performance requirements by significant margins, providing extra headroom for a more robust cabling system.

Re: top-level install, most cat5e cable I've seen installed is anything but top notch, a fair amount of it probably would fail a cat5e certification.... mostly due to poor terminations... max untwist at the jacks/patch panels is 0.5" - basically stitches right up to the IDC "teeth"

 

[JFletcher]

couple thoughts....

i read that the turn differences of the pairs were
a ratio to each other, not as absolutes on their own,
so that a manufacturer, as long as those ratios were
observed, could produce compliant cable.

i've got two cable certifiers that, while neither of them
is a $ five digit device, every cable i've ever tested with
them, at cat 5 or greater, assuming it's wired correctly,
has never failed a cat 5 cable or terminations at 1ghz cat 6.

Yeah I wanted a Fluke DTX 1800, glad I didnt plop down the 10k for it at the time, and another 10k for the fiber modules... I got out the commercial v/d/v just about the time we considered offering certification... most hotel owners dont even want to pay for conduit...

I'd have to look up the turn ratios to see what they are in relation to one another, tho lets say they are the following:

Blue pair: 1.08 turn per inch
Green pair: 1.35 turns per inch
Orange pair: 1.5 turns per inch
Brown pair: 1.2 turns per inch

The ratio of orange to green is 1.111..., ratio of green to brown 1.125, and brown to blue again 1.111.... Now, these arent measured specs, but they are close enough to see that the ratio of twist rate between pairs is roughly the same from one to another. Primary difference between cat6 and cat5e is cat6 is usually 23ga wire and 5e 24, and some cat6 has an internal spline divider.

Cat5e at the time I was putting it in, most of the good stuff was rated 350MHz even tho the spec is 100MHz... cat6 spec is 250MHz, tho the cable is, I believe 550MHz, and while that's not the only diff between them (all cross talk specs are tighter, overall max speed is higher), I agree that well installed cat5e on premium components will far exceed its minimum specifications.

 

[octavian]

Maybe not a big deal here

Maybe not a big deal here

The 568B standard is still more common in the US but I believe slowly moving toward 568A. For what it's worth in DOD work the rule is use A unless B is widely used at the present location. So with no other info use A.

 

[brantmacga]

That is the absolute truth, I've been preaching that for a long time. I've been putting plugs on cables for probably 20 years so you would think I would have it down by now. But I still have to cut some off and re-do them not to mention trying to figure out which end is at fault.

-Hal

It's especially difficult when you're color blind



Sent from my iPhone using Tapatalk

 

[gadfly56]

It's especially difficult when you're color blind



Sent from my iPhone using Tapatalk

Well, another career choice for my sons to avoid.

 

[Sesquipedalian101]

My Twisted Opinion

My Twisted Opinion

I'm new here; pardon me for just "jumping in"; however, I've been dealing with "data cabling" since before "AppleTalk" first started exploring twisted pair usage -- and "Ethernet" referred to some sort of "coax installation for your DEC computer."

Here are several of points that I did not see completely clarified:
_

1. Yes, the "twist rate" will be different between different pairs within the same cable and, it will generally vary between types of cable from the same manufacturer and between different cable manufacturers. The purpose of having a variable twist rate is to keep the cables from "nesting" inside of the overall cable housing. (Think of the way two threaded rod with the same "tread pitch" will "lock together" in their package.) When they nest together, you get increased cross talk. I have dealt with cable manufacturers who had dividers to keep the pairs separate and with manufacturers who used a left-hand twist on two pair, and a right-hand-twist on the other two pair -- making it necessary to have only two twist ratios instead of four.
​

_

While "twist is good" is a rule of thumb, it comes at a cost. The higher the twist rate, the shorter the distance covered by the same amount of wire (I've seen deltas of up to three feet in a 100 meter run). The higher the twist rate, the harder it becomes to terminate the cable -- whether using plugs or jacks (I've crimped thousands of plugs on to Cat 6 Cable, when doing it regularly my failure rate was 0.5%) -- you have to test with a high-quality tester every time.
​

There is no sure and certain guarantee that the "green pair" or the "orange pair" or any of the other pairs will be the "best" pair. In fact, at one point, Avaya claimed the "Blue Pair" was the "best pair" in their cables.​

_
_

2. The twist rate not only affects how much the cable pairs nest and how much they cross-talk, it also affects the cable's immunity to environmental noise. All things being equal, the higher the twist rate, the more noise immunity you tend to get. Thus, I have seen some devices on some cable that would work on one set of cable pairs, but not on another. Now this was in a situation where we were pressing some first-generation "Level 3" (remember, they called it "Level 3" before there was a "Category 3" designation?) cable that was being used for a 100 meg connection. So, we were "at the margin" to say the least -- technically, it should NOT have worked at all. But, in order to achieve functionality, we rotated the "Blue Pair" (see above) into use along side of the "Green Pair" to get 100mbps with an acceptable (but still non-zero) error rate.
​

_
_

3. But ALL of this begs the question of the difference between 568A and 568B relative to your camera installation. As has been stated before, in 568A places the "green pair" on pin numbers 1 & 2 while the "orange pair" goes to pin 3 & pin 6. By contrast, 568B puts the "orange pair" on pins 1 & 2 and lands the "green pair" on pins 3 & pin 6. The fact I did not see anyone raise (I apologize if I missed a comment), is that pins 1 & 2 and pins 3 & 6 are ALL part of the SAME bidirectional communications channel.​

_

From the DTE (Data-communications Terminating Equipment) perspective, pin one is essentially Tx+ pin 2 is Tx- pin 3 is Rx+ and Rx- lands on pin 6. What is "transmit" to one device is "receive" to the other -- which is why you may see the pins designated as "BI_DA+" and "BI_DA-" and "BI_DB+" and "BI_DB-" respectively. So, basically, you are not "improving" the channel by switching the pairs within the channel. You are simply changing which end is "driving" which pair. Assuming both pair meet specs, there should be no problem. If they do not meet specs, it's not like you were swapping a pair that was not part of the link into the connection. Let's say say NEXT (Near End X-Talk) is too high and the camera "hears itself" thereby causing phantom collisions. The inductive and/or capacitive coupling that is causing your issues is bidirectional -- if the green pair "hears" transmissions made on the orange pair, the orange pair will "hear" transmissions made on the green pair and the camera will still fail to transmit -- regardless of which pair it drives and which pair is driven by the other end.​

Oh, and one final point, the whole reason for specifying 568A as opposed to 568B has to do with backwards compatibility, not quality of operation.

The 568B option matches the old "Ma Bell" 258A color code (can anyone say, "WICO"?) and only matches single (one) pair USOC wiring schemes. Since the Ma Bell or AT&T (Alotta Time & Trouble) spec wiring was far and away the most common in private homes and private companies. This is because, at one time, the "phone company" had to do the wiring, or you would not get service from them. That restriction is why we occasionally had phone jacks screwed to desks -- wanna move your desk? gotta call the phone company! The 568B wiring pattern tends to flow "naturally" in those places.

Meanwhile, 568A provides "backwards compatibility" with two pair USOC -- which is why the Federal Government generally specifies 568A.

My $0.02 Worth
(and that's inflation)

 

[LarryFine]

Sesq, welcome to the forum. Here, we prefer to eschew obfuscation. Interesting to resurrect a seven-month-old thread, but refreshing to see someone else who understands twisted-pair theory, too. I have run and terminated plenty of UTP cable, and installed camera systems, too.

One more thing about twist rate: the greater the twists-per-inch, the higher the frequency (i.e., the shorter the wavelength) of interference that the receiver's differential amp can ignore (due to common-mode noise rejection) because the twists assure that both conductors pick up the interfering signal equally.

Yes, I'm primarily an electrician, but I've been interested in electronics and electrical theory since I was a kid. I built crystal radios when I was 6 (I'm 62 now), installed 8-track stereos when I was a teenager, modified audio equipment, and I am into A/V and home theater now. I remember token-ring networks using coax, too.

I read Popular Electronics magazine, and built stereo equipment from kits such as Dynaco and SWTPC, and read about, but never built, computers using the 6800 processor. i know more about hardware than software when it comes to computers. I prefer soldering irons and hand and power tools to keyboards.

 

[sameguy]

Rewired an elementary school years ago with cat 5, I'm talking to the Verizon guy punching down the pot head, when a teacher stops and said how nice it will be to have high speed internet so they can access the university library miles away. She walked off and we both burst out laughing.
What was the name of the shunted term blocks, supposed to have their tool to term with?

 

[sameguy]

Sesq, welcome to the forum. Here, we prefer to eschew obfuscation. Interesting to resurrect a seven-month-old thread, but refreshing to see someone else who understands twisted-pair theory, too. I have run and terminated plenty of UTP cable, and installed camera systems, too.

One more thing about twist rate: the greater the twists-per-inch, the higher the frequency (i.e., the shorter the wavelength) of interference that the receiver's differential amp can ignore (due to common-mode noise rejection) because the twists assure that both conductors pick up the interfering signal equally.

Yes, I'm primarily an electrician, but I've been interested in electronics and electrical theory since I was a kid. I built crystal radios when I was 6 (I'm 62 now), installed 8-track stereos when I was a teenager, modified audio equipment, and I am into A/V and home theater now. I remember token-ring networks using coax, too.

I read Popular Electronics magazine, and built stereo equipment from kits such as Dynaco and SWTPC, and read about, but never built, computers using the 6800 processor. i know more about hardware than software when it comes to computers. I prefer soldering irons and hand and power tools to keyboards.

Larry, I've got a Dynaco amp ~20-30w hit by lightning I need to rebuild one day; Knight kit, Lafayette Radio, like you I can build but never learned coding. Have a raspberry Pi trying to make security/pir/email/photo upload camera but don't know Python. )

 

[zbang]

Have a raspberry Pi trying to make security/pir/email/photo upload camera but don't know Python. )

You're lucky, I know enough python to rather despise it (a blot on the landscape of computer languages); AFAICT the only reason it ever got any traction was being new and not Perl.

 

[hbiss]

Rewired an elementary school years ago with cat 5, I'm talking to the Verizon guy punching down the pot head, when a teacher stops and said how nice it will be to have high speed internet so they can access the university library miles away. She walked off and we both burst out laughing.
What was the name of the shunted term blocks, supposed to have their tool to term with?

Krone perhaps? Love them for demarcs.

-Hal

 

[sameguy]

Krone perhaps? Love them for demarcs.

-Hal

Yes.