dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
18
share rss forum feed


jmich
Premium
join:2001-08-28
Toms River, NJ
reply to voxframe

Re: Cable Nazis... Grab a tissue

Very nice. Vendors are now saying that lacing and piping cable amplifies cross-talk potential and that random cabling is better for performance. Discuss.

tomdlgns
Premium
join:2003-03-21
Chicago, IL
kudos:1
with random cables going every direction there is less distance between them which means less of a chance for crosstalk.

when the cables are laying on top of e/o and touching, it allows for a greater chance of crosstalk

that is how i read that post.


alphapointe
Don't Touch Me
Premium,MVM
join:2002-02-10
Columbia, MO
kudos:2
reply to jmich
Bollocks... If there is any more xtalk, it's a couple hundredths of a db. If this was a real problem, Bell wouldn't have done lacing, etc. for the past hundred years with no problems.

tomdlgns
Premium
join:2003-03-21
Chicago, IL
kudos:1

1 recommendation

i agree. no matter what, someone will always argue the opposite.

Hahausuck
Premium
join:2003-12-14
kudos:2

1 edit

2 recommendations

reply to tomdlgns
I think that's a lazy excuse because the installers are generally lazy.

The last job I worked in the field on, the data comm guys stopped and watched me several times while I laced. They never had seen it before. The electricians had seen it only in black and white photos while in trade schools. There was one hvac guy who had been around the block a few times in his days and he knew exactly what it was. He commented how good it was to see someone who still did that sort of work and it had been over 20 years since he'd seen it done.

There are a few of us who still do it. It is seriously time consuming but when your cabling has to be perfect, it's the only way.
--
"Saying something in another language that you don't think the other person understands is just saying that you're a pussy and are too afraid to say it in English." --Harddrive

cramer
Premium
join:2007-04-10
Raleigh, NC
kudos:9
reply to jmich
potential, yes. Reality, not so much. Cat6 has issues at 10G over longer runs, and yes, bundles amplify the problem. That's why there's a cat6A -- there will be no meaningful crosstalk at rated bandwidth over full 100m runs. (increased pair spacing, and thicker cable shell... voila the cables cannot get close enough to create a problem.)

If you're really worried about it, fork over the cash for shielded cable. For the record, I've only had to use shielded cable twice... once in a radio station, and once in a UL Labs test cage. (we still failed, btw.)


DarkLogix
Texan and Proud
Premium
join:2008-10-23
Baytown, TX
kudos:3
reply to alphapointe
said by alphapointe:

Bollocks... If there is any more xtalk, it's a couple hundredths of a db. If this was a real problem, Bell wouldn't have done lacing, etc. for the past hundred years with no problems.

If it were a problem then just use the unofficialy named CAT7 cable.

5 sheilds, 1 around each of the 4 pairs, and 1 over all 4 pairs, plus a proper STP RJ45 end not the normal UTP RJ45 that you normally see.

then with the shields grounded theres no chance of xtalk between cables.

However I'll say this phone lines 100 years ago were less likely to have issue with xtalk than gig or 10g over copper


DarkLogix
Texan and Proud
Premium
join:2008-10-23
Baytown, TX
kudos:3
reply to cramer
said by cramer:

potential, yes. Reality, not so much. Cat6 has issues at 10G over longer runs, and yes, bundles amplify the problem. That's why there's a cat6A -- there will be no meaningful crosstalk at rated bandwidth over full 100m runs. (increased pair spacing, and thicker cable shell... voila the cables cannot get close enough to create a problem.)

If you're really worried about it, fork over the cash for shielded cable. For the record, I've only had to use shielded cable twice... once in a radio station, and once in a UL Labs test cage. (we still failed, btw.)

If you wanta go to the insane level then use CAT7 + 1 cable per steel conduit.

And if RF is still something to worry about go fiber.


jmich
Premium
join:2001-08-28
Toms River, NJ
Reviews:
·Verizon FiOS
Agreed. The instructors from Fluke and ADC both mentioned this issue in recent VAR classes. I agree that well-installed networks work better and there's a great deal of pride to be had in doing a nice job...time allowing. Another issue they mentioned from bundling is heat potential. Some PoE devices do draw a decent amperage. Not an issue in a DC but enterprise wiring.

cramer
Premium
join:2007-04-10
Raleigh, NC
kudos:9
Interesting points. PoE would be a special case. Using the first chart google points out [link], 100m of 24AWG (worst case, cat6 is supposed to be 22) wire would have 8.41976ohms of resistance. (PoE uses 2 parallel pairs.) 802.3af specs 15.4W (350mA) and 802.3at 34.2W (600mA.) So, a max "af" device would have just over 1W in the cable, and "at" just over 3W. If you put 48 runs in one conduit, they were all pulling full power, and they were all 100m... (an unlikely case) that would be a 50W or 146W pipe. (31 and 92 for 22AWG) That'll get toasty if there's no air moving around it. (BTW, I've seen power conduits like that, but never any data wiring.)

[PS: It's worth noting your switch or midspan would need a 750W or 1700W power supply *just for PoE* to run such a setup.]

TheMG
Premium
join:2007-09-04
Canada
kudos:3
Reviews:
·NorthWest Tel
said by cramer:

If you put 48 runs in one conduit, they were all pulling full power, and they were all 100m... (an unlikely case) that would be a 50W or 146W pipe. (31 and 92 for 22AWG) That'll get toasty if there's no air moving around it.

That works out to 1.46W per foot, which is unlikely to create any problems. Might feel slightly warm to the touch but that's about it.


DarkLogix
Texan and Proud
Premium
join:2008-10-23
Baytown, TX
kudos:3
reply to cramer
said by cramer:

[PS: It's worth noting your switch or midspan would need a 750W or 1700W power supply *just for PoE* to run such a setup.]

BTW some POE switches do have that

The ones where I work have dual 930watt powersupplies
so in theory 1860watts, I think when one PS fails it cuts the available PoE power down to what one powersupply can do.

cramer
Premium
join:2007-04-10
Raleigh, NC
kudos:9
Yes, but they are exceedingly expensive, rather large, and generate far more heat than all the wiring in the building.

Hahausuck
Premium
join:2003-12-14
kudos:2
HIPOE is awesome, but you're right it is horribly expensive even in lower density options.


DarkLogix
Texan and Proud
Premium
join:2008-10-23
Baytown, TX
kudos:3
reply to cramer
I really don't foresee any issue with POE heating cables.

Ya the heat coming off the switches is far more than all the wiring in the place.

I was just pointing out that some switches do have that kinda wattage, though how much is for POE and how much is for the switches to do normal stuff I'm not sure but I'd bet the bulk is for POE.