Cat 5e vs. Cat 6

[gar]

080429-1110 EST

The terminations are defined as follows:

	   T-568A				   T-568B  
    Pin Number	Color		    Pin Number	Color
	
	1	white/green		1	white/orange
	2	green			2	orange

	3	white/orange		3	white/green
	4	blue
			4	blue

	5	white/blue		5	white/blue
	6	orange			6	green

	7	white/brown		7	white/brown
	8	brown			8	brown

 

[gar]

080429-1158 EST

The terminations are defined as follows:

	   T-568A				   T-568B  
    Pin Number	Color		   	 Pin Number	Color
	
	1	white/green			1	white/orange
	2	green				2	orange

	3	white/orange			3	white/green

	4	blue				4	blue
	5	white/blue			5	white/blue

	6	orange				6	green

	7	white/brown			7	white/brown
	8	brown				8	brown

The above layout hopefully allows you to see the pair mappings in a more graphical form. For those wires that do not map straight thru you can draw connecting lines to see their path.

If you wire an A to B it is called a crossover cable.

Getting layouts to format on a forum can be a challenge. You can not always know what will happen until it is posted. There is a short time limit here so I could not get things correct in post #21 and it needs to be deleted, but how?

.

 

[curt swartz]

Here is another article which is more current and has similar information to what gar posted.
http://www.nexans.com/US/2007/Category%205ev6-CBM-ReelTime-0907.pdf

 

[LarryFine]

To me, it's easy enough to remember that the B configuration has the blue pair in the middle and the orange pair flanking them, just like phone wiring (B as in bell), leaving the green pair and the brown pair.

Likewise, if you remember that it's the green and orange pairs that are swapped, you can remember that the blue pair is now flanked by the green pair, leaving the orange and brown pairs.

As long as both ends terminated the same way, it doesn't actually affect performance either way.

It's also easy enough to remember that the pattern of striped, solid, striped, solid, etc. continues all the way.

 

[gar]

080429-2157 EST

Larry:

Just terminating both ends the same way does not necessarily provide the same results as correctly terminating the ends. There are 4 twisted pair in the cable. Each pair in a good quality cable has a different twist pitch length and the pairs are bonded together to maintain a fixed relationship. This reduces crosstalk.

On a differential signal a twisted pair provides cancelation of many magnetic fields. Thus, one pair of differential signals should be on a twisted pair rather than crossed between two twisted pairs which would not provide the same cancelation.

.

 

[LarryFine]

You severly misunderstood what I was saying. Believe me, I understand how twisted-pair transmission lines work, and about common-mode noise rejection and differential op-amps. I also test every ethernet drop and cable I make.

I meant whether you use the orange or green pair on pins 3 and 6. In other words (from memory):

A: ("commercial")
orange/white
orange
green/white
blue
blue/white
green
brown/white
brown

B: ("residential")
green/white
green
orange/white
blue
blue/white
orange
brown/white
brown

 

[rexowner]

080429-0950 EST

rexowner:

Have you run comparative tests of bit error rate and thruput on CAT5E and CAT6 at 100 M, 125 M, 150 M lengths, and what were the results?

.

When I worked at a manufacturer of networking
equipment, engineers in another department ran
those tests. So, no, I have not personally run those
tests, but people in whom I have extremely high
technical confidence did. The consensus in engineering
was that Category 6 was snake-oil. If there is any
issue, it's going to be in crosstalk. BER
was astoundingly low. Throughput in either was
1 gigabit per second in the case of Ethernet.

The EIA/TIA standard for Ethernet is 100 meters. I would
not recommend anyone install Category-Anything out
of this standard, e.g. the 125M and 150M seem irrelevant.

I apologize if my criticism of Cat-6 was a bit of a rant.
Cat-6 is a pet peeve of mind, and I am sorry if it came
off wrong.

 

[rexowner]

Well, it worked on one customer of mine. He insisted on CAT-6 cable and jacks for his network, and got them. He also paid for the entire 1000' reel.

I've installed about 20 each of coax jacks, ethernet jacks, and phone jacks in various keystone-plate combinations in his home, over three different visits.

The house is three floors, including the unfinished-ceiling (thank goodness!) basement, but no attic. We can fish! Not one inch of cable shows anywhere.

If he insisted, you've got to give the people what they want.
If he wants to part with his money, that's his right.
Sounds like the extra money for the cable wasn't going
to break him in that house.

I have never actually terminated Cat-6, because with
my fat fingers it's hard enough getting the untwisted
wire to the termination down to 1/2" on Cat-5e.
Was it a pain to terminate the Cat-6? Would you charge
more per termination than Cat-5e?

 

[gar]

080430-1020 EST

Larry:

Sorry I misunderstood your intent.


rexowner:

The reason I tossed out longer lengths than the specification is because this helps identify marginal capability.

Operation does not just stop at 100 meters but gradually degrades. In poor quality components maybe before 100 meters. Any of these problems can be treated with signal detectability theory.

Just because you connect two devices together and transfer data does not mean that you have a good system. If I take an RF link and consider the thruput with and without a microwave on that is 100 ft away and thruput drops to 50% with the microwave on, then is that RF link a good system compared to a direct wired connection? I do not think so unless RF is the only possible communication path.

.

 

[rexowner]

080430-1020 EST

rexowner:

The reason I tossed out longer lengths than the specification is because this helps identify marginal capability.

Operation does not just stop at 100 meters but gradually degrades. In poor quality components maybe before 100 meters. Any of these problems can be treated with signal detectability theory.

Just because you connect two devices together and transfer data does not mean that you have a good system. If I take an RF link and consider the thruput with and without a microwave on that is 100 ft away and thruput drops to 50% with the microwave on, then is that RF link a good system compared to a direct wired connection? I do not think so unless RF is the only possible communication path.

.

The vast majority of problems with Cat-5(e) or Cat-6 are
termination, not cable quality. I used to support gigabit
ethernet products for a large manufacturer, and basically
all the physical layer problems were either poor
terminations or the electronics hooked up to the device.
Very rarely, there was a cable out of spec (e.g. didn't
meet Cat-5 specs.) I think Microsoft was mentioned
earlier on this thread as installing Cat-5e -- they
were one of the customers our company sold networking
gear to that was not baffled by the BS of the cable
manufactures -- they had smart IT people who understood
the issues and acted accordingly.

IMO, the whole Category-6 thing is a bit like if Cerro,
Southwire, etc. convinced a whole lot of electricians
to wire their tract home lite circuits with 10/2 and 10/3,
only worse. (I know a lot of devices only will take 12 or
14.) At least in the case of 10AWG, there are some
real corner cases, like lites a really long distance away
where voltage drop is an issue. There really aren't
any real world corner cases in the case of Cat-x,
Cat-5e will work just fine, but the the cable manufacturers
have just convinced people to install a more expensive
cable that they don't need by baffling them with BS.

 

[rexowner]

Here is another article which is more current and has similar information to what gar posted.
http://www.nexans.com/US/2007/Category%205ev6-CBM-ReelTime-0907.pdf

I get CBM and I laughed when I saw that article a while
ago. It is push marketing from the manufacturer trying
to sell a product that is unnecessary by using classic
"lying with statistics" tactics like the unlabelled or
partially labeled graphs. It really would be like Southwire
showing an unlabeled graph showing voltage drop
was lower by wiring up your lite circuits with 10/2 and
10/3 instead of 12 or 14 AWG. So what?

Also, I don't know where they get the # that Cat-6 is
35% more expensive that Cat-5e. It looks to be about
twice the price for the cable when I have looked at it.

Copper.org should hire the author of the article to
start pushing 10AWG for lite circuits.

 

[LarryFine]

If he insisted, you've got to give the people what they want.
If he wants to part with his money, that's his right.
Sounds like the extra money for the cable wasn't going
to break him in that house.

Fer sherr!

I have never actually terminated Cat-6, because with
my fat fingers it's hard enough getting the untwisted
wire to the termination down to 1/2" on Cat-5e.
Was it a pain to terminate the Cat-6? Would you charge
more per termination than Cat-5e?

Actually, I have big hands, too (I'm 6'3" and 300+ lbs), but I have to trouble manipulating small parts and wires. I grew up building electronic kits and soldering PC boards, as well as mechanical stuff and models.

I am anal about keeping the twists consistant all the way to the terminations. I even fret about having to undo a half a twist to land the wires. I've been working with LV work, phones, audio/video, etc., since I was a kid.

So, no and yes. No, I don't have trouble with either one. The only real difference to me is the terminal positions on the keystone jacks. But, yes, I do charge a bit more for the 'more demanding precision work' of CAT-6.

I say that if you can consistantly put plugs on UTP properly, you can do any of this work.