Where to source quality bulk ethernet cable?
December 24, 2017 8:16 AM Subscribe
I'm helping to wire a home for Ethernet. We'd like to have the cabling be of high enough quality to eventually support 10Gb Ethernet, which means Cat6 or Cat6A. Unfortunately, it looks like a lot of network cabling is mislabeled and not up to spec. Who are some trustworthy suppliers of Cat6 and Cat6A bulk cable?
I don't think any of the runs will be longer than 160 feet, but I'm still leaning towards installing Cat6A over Cat6. Since this stuff will be permanently in the walls, it sounds like a good idea to use the best quality possible (for future proofing).
Also, what are the implications of going with shielded cable? I've been reading that installing it can be tricky (with grounding, ground loops, etc.), and it probably won't provide much benefit in a home setting. However, if we do it right, maybe could help with noise from power wiring, etc. Are there any trustworthy guides on how to properly do the installation, with grounding, etc.?
I don't think any of the runs will be longer than 160 feet, but I'm still leaning towards installing Cat6A over Cat6. Since this stuff will be permanently in the walls, it sounds like a good idea to use the best quality possible (for future proofing).
Also, what are the implications of going with shielded cable? I've been reading that installing it can be tricky (with grounding, ground loops, etc.), and it probably won't provide much benefit in a home setting. However, if we do it right, maybe could help with noise from power wiring, etc. Are there any trustworthy guides on how to properly do the installation, with grounding, etc.?
Monoprice or Graybar
posted by iamabot at 8:32 AM on December 24, 2017 [1 favorite]
posted by iamabot at 8:32 AM on December 24, 2017 [1 favorite]
Best answer: I am a certified Leviton installer and the company issues warranties on certified installation based on the brands of cable use. A lifetime warranty is available when using Berk-Tek cable. A 15-year warranty is available when the following manufacturers' cable is used: Belden Cable, CommScope Cable, General Cable, Hitachi Cable, Mohawk, and Superior Essex. Leviton issues no warranty if any other brand of cable is used. At my job I mostly use General Cable.
Shielded cable offers a higher level of protection from electromagnetic and radio frequency interference and from crosstalk from other cables and is generally used for healthcare, manufacturing, gaming (as in casinos), military, government, and financial institutions. BUT. In order for shielded cable to be effecting it has to be properly gounded at your main telecomm room and shielded patch cords have to be used throughout the cabling system. So, yes, unless you're going to do the extra work to bond the network racks to gound and buy and use all shielded patch cables, there's not much benefit in a home setting. If I were pulling cable in my own house at my own expense I wouldn't go shielded.
posted by glonous keming at 8:54 AM on December 24, 2017 [8 favorites]
Shielded cable offers a higher level of protection from electromagnetic and radio frequency interference and from crosstalk from other cables and is generally used for healthcare, manufacturing, gaming (as in casinos), military, government, and financial institutions. BUT. In order for shielded cable to be effecting it has to be properly gounded at your main telecomm room and shielded patch cords have to be used throughout the cabling system. So, yes, unless you're going to do the extra work to bond the network racks to gound and buy and use all shielded patch cables, there's not much benefit in a home setting. If I were pulling cable in my own house at my own expense I wouldn't go shielded.
posted by glonous keming at 8:54 AM on December 24, 2017 [8 favorites]
Response by poster: Thanks glonous keming for your answer, it's exactly what I wanted!
So, to properly ground shielded cable, you only do it at the central patch panel and leave the ends unconnected to (local) ground? I've seen that described elsewhere, but not anywhere I'd consider authoritative. Do you know of any manuals that describe the procedure?
Would the ground connect into a consumer device, potentially causing a ground loop (e.g. with audio equipment)? I've had problems with that in the past with things connected to different grounds.
I agree, shielded cable doesn't sound like a good idea for a home, but I just want to understand it and the issues.
posted by cosmic.osmo at 9:27 AM on December 24, 2017
So, to properly ground shielded cable, you only do it at the central patch panel and leave the ends unconnected to (local) ground? I've seen that described elsewhere, but not anywhere I'd consider authoritative. Do you know of any manuals that describe the procedure?
Would the ground connect into a consumer device, potentially causing a ground loop (e.g. with audio equipment)? I've had problems with that in the past with things connected to different grounds.
I agree, shielded cable doesn't sound like a good idea for a home, but I just want to understand it and the issues.
posted by cosmic.osmo at 9:27 AM on December 24, 2017
Best answer: Grounding and bonding gets in to some hardcore alphabet soup stuff but I'll try to give you a brief overview. Standard disclaimers apply: IANYCablingTechnician. IANYElectrician. This is meant as a general informational overview and not professional guidance. Telecomm gounding and bonding should be done by properly licensed professionals.
> to properly ground shielded cable, you only do it at the central patch panel and leave the ends unconnected to (local) ground
Correct. No ground connection is required for the data port at the user end where the computers or access points or whatever plug in since the grounding is accomplished through the use of shielded patch cords.
> Would the ground connect into a consumer device, potentially causing a ground loop
I haven't dealt with ground loops in the field so I have to fall back to training. My manual states that all bonding systems shall be bonded together to achieve 0 potential, but grounding alone does not eliminate the potential for ground loops, which are defined as unwanted ground current flowing back and forth between 2 devices that are grounded at two or more points. This occurs when grounding points are not at the same voltage potential. So if the telecomm bonding system is connected to the electrical service ground, then ground loops shouldn't be an issue, since all equipment is ultimately connected to the same single spike driven into the Earth.
So basically all the network racks, cabinets and patch panels should be connected to the grounding system through bare metal-to-metal contact. This can be correctly-gauged wires connected to each piece of equipment or to a rack grounding busbar (RGB) that is is connected into the grounding system. You can see some example diagrams on page 3 of this overview PDF.
All those bonding connections attach to the Primary Bonding Busbar (PBB), which is connected to the electrical service ground.
You can look at a basic overview of terms and concepts for bonding and grounding here, from Electrical Construction & Maintenance Magazine.
Finally, here is the Leviton Basic Residential Installer Guide [PDF].
posted by glonous keming at 10:39 AM on December 24, 2017 [1 favorite]
> to properly ground shielded cable, you only do it at the central patch panel and leave the ends unconnected to (local) ground
Correct. No ground connection is required for the data port at the user end where the computers or access points or whatever plug in since the grounding is accomplished through the use of shielded patch cords.
> Would the ground connect into a consumer device, potentially causing a ground loop
I haven't dealt with ground loops in the field so I have to fall back to training. My manual states that all bonding systems shall be bonded together to achieve 0 potential, but grounding alone does not eliminate the potential for ground loops, which are defined as unwanted ground current flowing back and forth between 2 devices that are grounded at two or more points. This occurs when grounding points are not at the same voltage potential. So if the telecomm bonding system is connected to the electrical service ground, then ground loops shouldn't be an issue, since all equipment is ultimately connected to the same single spike driven into the Earth.
So basically all the network racks, cabinets and patch panels should be connected to the grounding system through bare metal-to-metal contact. This can be correctly-gauged wires connected to each piece of equipment or to a rack grounding busbar (RGB) that is is connected into the grounding system. You can see some example diagrams on page 3 of this overview PDF.
All those bonding connections attach to the Primary Bonding Busbar (PBB), which is connected to the electrical service ground.
You can look at a basic overview of terms and concepts for bonding and grounding here, from Electrical Construction & Maintenance Magazine.
Finally, here is the Leviton Basic Residential Installer Guide [PDF].
posted by glonous keming at 10:39 AM on December 24, 2017 [1 favorite]
Best answer: In order for shielded cable to be effecting it has to be properly gounded at your main telecomm room and shielded patch cords have to be used throughout the cabling system.
This is exactly correct. If you want to use shielded cable, the entire system needs to be designed around doing that. The grounds need to be low (zero) resistance, and more importantly low impedance and all tied together to common earth ground.
Not doing this work correctly leads to issues with signal propagation on the wire, and those effects can be far worse than the interference. The reasons are... difficult to explain in depth in this comment field*.
Broadly speaking - shielded cabling is just not worth the expense and difficulty, and I wouldn't bother unless you know you will have significant EM interference (say a HAM radio shack). And in that case, I'd maybe look at doing fiber where it makes sense to because it's far less fussy and troublesome. UTP is pretty amazingly resilient to interference to begin with, its part of the design, and you shouldn't have any issues with performance in a SOHO setting.
You should consider using plenum rated if you're going through open areas (crawl space, basement, drop ceiling) - it may be required by code where you are at.
* I have 20+ years networking experience, degree in electrical and computer engineering, HAM and FCC licensed radiotelephone operator. There is a lot to it.
posted by Pogo_Fuzzybutt at 10:50 AM on December 24, 2017 [2 favorites]
This is exactly correct. If you want to use shielded cable, the entire system needs to be designed around doing that. The grounds need to be low (zero) resistance, and more importantly low impedance and all tied together to common earth ground.
Not doing this work correctly leads to issues with signal propagation on the wire, and those effects can be far worse than the interference. The reasons are... difficult to explain in depth in this comment field*.
Broadly speaking - shielded cabling is just not worth the expense and difficulty, and I wouldn't bother unless you know you will have significant EM interference (say a HAM radio shack). And in that case, I'd maybe look at doing fiber where it makes sense to because it's far less fussy and troublesome. UTP is pretty amazingly resilient to interference to begin with, its part of the design, and you shouldn't have any issues with performance in a SOHO setting.
You should consider using plenum rated if you're going through open areas (crawl space, basement, drop ceiling) - it may be required by code where you are at.
* I have 20+ years networking experience, degree in electrical and computer engineering, HAM and FCC licensed radiotelephone operator. There is a lot to it.
posted by Pogo_Fuzzybutt at 10:50 AM on December 24, 2017 [2 favorites]
You could consider installing in conduit. That way upgrading to quantum wormhole cable (or whatever they use in the future) will be easy. Don't forget to put a pull wire in each conduit. And label, label, label everything.
Yes, more expensive, and more work.
posted by H21 at 12:42 PM on December 24, 2017 [2 favorites]
Yes, more expensive, and more work.
posted by H21 at 12:42 PM on December 24, 2017 [2 favorites]
Best answer: Ok, I thought I'd give a try at explaining some of the issues with shielded TP wiring and why doing it wrong can be worse than not doing it. This is all very basic, and missing many details. But it should be enough to get you started.
UTP works by using two wires that are twisted which couples them electrically. They don't work like audio cables in that the signal goes down one and returns down the other - rather, *very* basically the signal is on one wire, and the opposite is on the other wire. The difference between the two wires is the signal, so in theory, any interference should affect both in the same way - say you have 2 volts one wire and 4 volts on the other, and add 10 volts interference, you'll get 12 volts and 14 volts, but the 2 volt difference is unaffected. That's not precisely how it works, but that's the principle. Point is - UTP has a good deal of resilience to interference baked into the design of it.
But, lets say you have a shielded wire, connected to an electrical system. The shielding wraps around the wire in a tube - but over a long enough run, this approximates a wire - and a wire connected at (at least) one end, approximates an antenna. To the extent that this antenna is resonant, EMR will couple into it, and to the extent that it is impedance matched to the electrical system, that EMR will resonate around system. This means that a poorly done shielded system can actually be more sensitive to EMR than an unshielded one.
Now, as I say, the shielding is a tube - and it is conductive. A metallic tube is, to a rough approximation, a waveguide. This waveguide will couple capacitively and if not properly designed and grounded can present an impedance to the signals on the wire. A mismatched impedance can actually suppress the signal, and to the extent that it is also acting as an antenna, the network is actually more susceptible to EMR than the UTP would have been because it is both suppressing the signal and incorporating interference into the line. This applies to conduit as well - so you should prefer to use PVC conduit if you go that route.
Point is, designing a generic system resistant to a generic "interference" is difficult - and in that case, I'd ground the shielding at the rack and the receptacle and the device. This is my Ground The Shit Out of Everything Method (GT-SOEM). If you know what the source may be (radio antenna, MRI machine, etc.) then you can optimize for that - but I'd call in professionals in that case.
Again, I'm really, really, hand-waving away a lot of details, but hopefully you get the idea. Go UTP. So much easier.
posted by Pogo_Fuzzybutt at 1:18 PM on December 24, 2017 [4 favorites]
UTP works by using two wires that are twisted which couples them electrically. They don't work like audio cables in that the signal goes down one and returns down the other - rather, *very* basically the signal is on one wire, and the opposite is on the other wire. The difference between the two wires is the signal, so in theory, any interference should affect both in the same way - say you have 2 volts one wire and 4 volts on the other, and add 10 volts interference, you'll get 12 volts and 14 volts, but the 2 volt difference is unaffected. That's not precisely how it works, but that's the principle. Point is - UTP has a good deal of resilience to interference baked into the design of it.
But, lets say you have a shielded wire, connected to an electrical system. The shielding wraps around the wire in a tube - but over a long enough run, this approximates a wire - and a wire connected at (at least) one end, approximates an antenna. To the extent that this antenna is resonant, EMR will couple into it, and to the extent that it is impedance matched to the electrical system, that EMR will resonate around system. This means that a poorly done shielded system can actually be more sensitive to EMR than an unshielded one.
Now, as I say, the shielding is a tube - and it is conductive. A metallic tube is, to a rough approximation, a waveguide. This waveguide will couple capacitively and if not properly designed and grounded can present an impedance to the signals on the wire. A mismatched impedance can actually suppress the signal, and to the extent that it is also acting as an antenna, the network is actually more susceptible to EMR than the UTP would have been because it is both suppressing the signal and incorporating interference into the line. This applies to conduit as well - so you should prefer to use PVC conduit if you go that route.
Point is, designing a generic system resistant to a generic "interference" is difficult - and in that case, I'd ground the shielding at the rack and the receptacle and the device. This is my Ground The Shit Out of Everything Method (GT-SOEM). If you know what the source may be (radio antenna, MRI machine, etc.) then you can optimize for that - but I'd call in professionals in that case.
Again, I'm really, really, hand-waving away a lot of details, but hopefully you get the idea. Go UTP. So much easier.
posted by Pogo_Fuzzybutt at 1:18 PM on December 24, 2017 [4 favorites]
You can go Cat6 or 6A if you want, but you'd be better off installing 5e UTP (we use Belden mostly) in conduit so you can pull fiber to do any 10GB you find necessary in the future. 10GE over copper sucks on every way. It wastes power and is always more finicky than you think it will be. It's like trying to run 1GbE over cat3. Also the equipment is still far more expensive.
Inside a single room you're way better off using twinax and for longer runs fiber, IMO.
posted by wierdo at 1:15 PM on December 26, 2017
Inside a single room you're way better off using twinax and for longer runs fiber, IMO.
posted by wierdo at 1:15 PM on December 26, 2017
« Older Recipes for a 3-qt. slow cooker? | Can a Chromebox work sucking data through a tiny... Newer »
This thread is closed to new comments.
posted by zsazsa at 8:24 AM on December 24, 2017 [1 favorite]