[Jazzbass]

I'm doing some renovation work on my study and installing new CAT6 in the basement and running up into my office. As far as I know, there are only three NEC code issues I need to be aware of:

1. Don't run low voltage and high voltage wires through the same holes into the walls
2. If installing in a multigang box, the low voltage and high voltage sides must be separated
3. Fire stop all holes drilled from the basement into the first floor walls (esp interior, non insulated walls)

Running these wires, a lot of the places that is the most convenient is along the beams or joists where 110VAC NM is already run. I've never thought of it being an issue running UTP cable next to (even zip-tied to) NM cable. Yeah it can induce 60Hz noise into the network cable, but that's why it's a twisted pair balanced line, right? Noise cancellation and all that.

Out of curiosity I did a Google search and was surprised at how many people warned against running CAT6 next to A/C lines. Always with the same info - keep them 6" apart or the A/C will induce noise into your network and f**k things up. Here's the interesting bit - most of the guys that say to keep them separated have some story about A/C induced noise screwing something up. On a speaker wire, coax, etc., but never a story about it messing up a UTP network run.

I know we have a lot of network engineers and EEs here, so I was curious if you guys had any thoughts/real info on this, beyond the typical "you can run speaker wire next to A/C, so keep the CAT6 away as well" type stuff I've found on the interwebs.

[joep]

Not an EE but based on first hand experience, but for short runs anyway I wouldn't think its a problem. CAT 6 has more twists for reasons such as this I thought.

Bear in mind UTP wire is basically a digital signal vs the analog speaker wire or old fashioned coax too so the Networking software in the comps will compensate for this interference but at the cost of max capacity. You may never notice the difference until you ran the wires elsewhere.

[Lupin..the..3rd]

Balanced differential signalling works great for common mode noise rejection. The pairs are twisted to ensure that any noise is truly common-mode, even when the noise source is very close (as in a parallel cable run). I don't see a problem.

Edit: You could also use shielded twisted pair for that segment, instead of UTP, if it's still a concern.

[Jazzbass]

Originally Posted by joep

Bear in mind UTP wire is basically a digital signal vs the analog speaker wire or old fashioned coax too so the Networking software in the comps will compensate for this interference but at the cost of max capacity. You may never notice the difference until you ran the wires elsewhere.

I know, that was the problem with all the googling I did. Any time there was any anecdotal evidence is was always with analog signals, not digital, hence the reason I asked. None of their anecdotes made any sense.

Originally Posted by Lupin..the..3rd

Balanced differential signalling works great for common mode noise rejection. The pairs are twisted to ensure that any noise is truly common-mode, even when the noise source is very close (as in a parallel cable run). I don't see a problem.

Me either, to be honest. But I will admit I did drink a lot sophomore/junior years at Tech and might have missed some things in Fields, Communications or Electronics, so I thought I'd see if anyone knew something I didn't.

Originally Posted by Lupin..the..3rd

Edit: You could also use shielded twisted pair for that segment, instead of UTP, if it's still a concern.

I thought of that, but it's more hassle than it's worth. Besides, I still don't get why a 60Hz induced signal on a UTP line would affect it at all.

[smdubovsky]

*IF* you could get a 60Hz component induced on the pairs it will saturate the ethernet transformer magnetics pretty easily. The transformer cores are TINY (10MHz+) and just wont take the volt*seconds. This was one of the reasons that PoE has such an odd startup sequence. They were worried there was some legacy system w/ odd magnetic configuration where even a voltage between SEPARATE pairs would saturate/damage something.

Im in the camp it will work just fine. I've seen ethernet ziptied to AC cables for long runs and no problem. Wired ethernet 'just works'.

Edit: Note modern AC systems may be VFD driven motors. If the VFD for the outside unit is in the house then there can be a lot of EMI on the cable (though the outside compressor motor should have its VFD located out there too.) But then they wouldn't pass radiated EMI standards. A lot more stuff wouldn't work than just the enet cable run beside it.

[Lupin..the..3rd]

Originally Posted by smdubovsky

Edit: Note modern AC systems may be VFD driven motors. If the VFD for the outside unit is in the house then there can be a lot of EMI on the cable (though the outside compressor motor should have its VFD located out there too.)

Be sure to check both the high and low side pressures before sending any TCP/IP packets. Also, see if your compressor motor supports jumbo frames.

[cmartin]

Good point. Adjust MTU as needed to compensate for encryption overhead.

I have cat5 all over the place, next to just about everything you can imagine and it works fine. To me it's one of things you read about all the time but I just havent seen any issues. Not sure if that means anything.

[jsheiry]

We run CAT5E most of the time and find most homes will never exceed the needs for CAT5E. The only time we use shielded wire is when we are distributing HDMI over baluns through a home and noise from dimmers etc.. have proven to cause problems. Have not experienced problems with un-shielded twisted pair cabling in a data-environment. Also having said that, we have not metered the band-width of home installations or had to certify the bandwith (these meters are still fairly pricey) I think you avoid long parallel runs when possible and I would not "hang" the low-voltage wiring to the high-voltage cabling just for convienence. Do not zip tie the bundles of low voltage tightly and avoid sharp bends/radius and you should be fine.
John
Home Technology Source
www.hometechsource.com

[Fritz]

I've run updated wiring through 3 houses now and have never noticed any issues being close to the A/C wiring. I do try to avoid it out of principle, but don't go out of my way if it proves difficult. Best approach for multilevel houses is to run all leads up the AC shaft to the attic and distribute down through the top of the walls. You can easily see where the AC is located and just run the new drop a stud or so away. And I have one of those 56" drill bits if you need to borrow it.

[tbernard]

Well, the only solution is fiber to the desktop.

Keep your runs in length spec and you'll be fine. 10/100 uses 2 pair and 1000 uses all 4 pair.