Bundling DC with Cat6 and Cat3

[David Emiley]

I've just moved into a new home (new for me, it was built in 1953). The phone wiring is crap (some of it fabric coated) and my plan is to rip it all out and replace it with CAT6 and CAT3 for gigabit Ethernet and 2-line telephone (which I won't use and maybe no one ever will, but I believe in doing things right).

I then got it into my head (shower idea) to run USB power to each of these boxes. Was going to use 2 wire insulated 18-gauge solid copper, pushing 5 volts and 1 amp (5 watts) from the source at most 50 feet, providing 4.5 volts 750 mA (3.375 watts) minimum at point of use.

NOTE: I don't want to use PoE if possible, because of some of my other goofy-ass networking plans. I'd rather keep the three systems (data, phone, power) separate. And because if everything works I might jack the power up to provide 5.5 volts and 3.1 amps (17 watts) at point of use.

It occurred to me however that the DC line might cause interfere with the data/voice lines. What are the guidelines for running CAT6 next to 110AC?

 

[mnewsham]

Generally you'll want to run the lines at least 2-4" apart when using shielded cables such as CAT6a since it's got more robust shielding from outside interference than standard CAT6. CAT6a also ensures you can get 10Gbps speeds on your entire network (CAT6 supports 10gbps up to 37m, CAT6a is up to 100m)
If you do go for an unshielded cable like standard CAT6 then i'd recommend a clearance of 6-8" or more.

The CAT3 will be the most susceptible as it's the oldest and has the least shielding by design, but it's also probably not going to matter that much since it doesn't sound like you'll be using phone lines anyway. If you were truly concerned you could run CAT5e or even another CAT6a run for the phone lines, it'd just be overkill.

 

[David Emiley]

A good point here is that buying cable in bulk, CAT5e might actually be CHEAPER than CAT3 which is used less these days.

Generally you'll want to run the lines at least 2-4" apart when using shielded cables such as CAT6a since it's got more robust shielding from outside interference than standard CAT6. CAT6a also ensures you can get 10Gbps speeds on your entire network (CAT6 supports 10gbps up to 37m, CAT6a is up to 100m)
If you do go for an unshielded cable like standard CAT6 then i'd recommend a clearance of 6-8" or more.

The CAT3 will be the most susceptible as it's the oldest and has the least shielding by design, but it's also probably not going to matter that much since it doesn't sound like you'll be using phone lines anyway. If you were truly concerned you could run CAT5e or even another CAT6a run for the phone lines, it'd just be overkill.

 

[JackMDS]

And, using Cat5e gives spare Giga capable wires too.

Running POE can be very useful.

USB? The unpleasant part is the USB cables from power source to the USB devices. Plugin USB PSUs are negligible in size and price.

In contrast finding yourself stretching USB cable, or being forced to used USB diveces near few static outlets in the wall can be really painful.

Part of our technological "Invocations" are more some persons finding an idea to make money from rather than real useful needs.

 

[repoman0]

What are the guidelines for running CAT6 next to 110AC?

Are we talking 110VAC or 5VDC? Network cables shouldn't be run parallel to 110V power lines due to interference, but they can cross perpendicular. DC power cables should have no effect on data cables since they don't emit a changing EM field. No idea what the official code / guidelines are though.

I'm about to run Cat6/6a through an even older 1915 house. Plaster and lath walls, difficult to navigate basement ceiling ... still might just pay an electrician to do it.

 

[David Emiley]

The plan is 5v 1a, up to 5.5v 3.1a, but as I doubt there are extant technical recommendations for this, I figured I would use the official recommendations for 110v AC as a guideline.

Are we talking 110VAC or 5VDC? Network cables shouldn't be run parallel to 110V power lines due to interference, but they can cross perpendicular. DC power cables should have no effect on data cables since they don't emit a changing EM field. No idea what the official code / guidelines are though.

I'm about to run Cat6/6a through an even older 1915 house. Plaster and lath walls, difficult to navigate basement ceiling ... still might just pay an electrician to do it.

 

[repoman0]

The plan is 5v 1a, up to 5.5v 3.1a, but as I doubt there are extant technical recommendations for this, I figured I would use the official recommendations for 110v AC as a guideline.

They are fundamentally different technologies. The recommendations for AC are in place because the changing current induces relatively strong fields around the wire, which then induce fields in the ethernet wire. SNR is decreased and the cable may no longer operate or only be able to operate at a slower data rate. On the other hand, PoE can transmit 25W DC down the same line as a 10gbit cable with relative ease. There's no problem at all, engineering and physics wise, to run a Cat6 cable next to a cable that might be carrying 17W DC at 5V. That said I have no idea what kind of building code recommendations are in place. I'd probably just run PoE and in fact that's what I plan to do. USB is too volatile a standard for me to bother running through my walls.

 

[David Emiley]

Not all of the jack drops have AC near them, so I want there to be USB power on hand. Also, one of my plans is to install USB powered wireless access points at each drop, specifically NEXX WT3020F's, which are OpenWRT capable and offer wired, wireless, and NAS all in the size of a cigarette pack.

And, using Cat5e gives spare Giga capable wires too.

Running POE can be very useful.

USB? The unpleasant part is the USB cables from power source to the USB devices. Plugin USB PSUs are negligible in size and price.

In contrast finding yourself stretching USB cable, or being forced to used USB diveces near few static outlets in the wall can be really painful.

Part of our technological "Invocations" are more some persons finding an idea to make money from rather than real useful needs.

 

[repoman0]

Could just spring for PoE access points, like the Ubiquiti 802.11AC stuff. Then you have the advantage of running 48V down your cabling and have a lot less power lost as heat in the cable and in your walls. 5V power over USB is not meant to travel a long distance. Use the savings of not having to buy a spool 18ga copper

 

[David Emiley]

Think I will probably go with PoE and spring for $20 drop boxes that split it into ethernet and 2 usb ports, though that means I'll need another panel for the phone jack :|

NEXX APs are the way to go though, as 1. They run OpenWRT, 2. They have a built in USB NAS, and 3. You can run a channel 14 network off them.

 

[thecoolnessrune]

As already mentioned, 5V would have huge voltage sag at any meaningful distances. PoE is 48V, and even then it's limited to 30 Watts.

My recommendation would be Passive PoE. That way you can use 48V power down Ethernet cabling, while aggregating power input into DC. There are 16 port power injectors that would give you 240 watts of power budget (30 watts per port using have the ports) like this: https://www.poetexas.com/collections/all-products/products/gpoe-16-48v240w

You'd need a PoE splitter on the remote side for each port, so it would be more expensive. But Passive PoE splitters and injectors like the above are substantially cheaper than Active PoE. You just have to be very careful you don't plug in a non-PoE device into a Passive PoE system, because you can burn it out (don't get your cables wrong in other words).

If you want to do more than that, well, there's ways to do it, and less than 48V systems have fairly simple NEC codes, but you'd have to review the codes for your local area and follow most of the guidelines that are in place for Solar installations in regards to grounding and distribution. That would still likely be the cheapest route, but the 48V -> 5V converters would eat up a lot of the money saved on PoE gear.

 

[razel]

I was going to reiterate the above. Trying to push 5v through the tiny gauge wiring in relatively long LAN cables will just warm the wire insulation wasting that energy. You can gang the 8 wires to increase gauge or actually gang 2 sets of 4 wires for negative and positive 5v, but really I wouldn't recommend it. If you have to run one CAT6 it's really not much more effort to run 2 CAT6 cable or run the another required cable needed to accomplish what you want.

 

[Red Squirrel]

Low voltage DC probably won't really cause issues, it's AC that causes issues because of the constantly changing magnetic field. I would run some 14/2 wiring for the DC that way it's rated for in wall use and rated for much more power than you need, so it is future proof, and easy to get. I would also run 6 volts instead of 5 just to have a bit of head room for voltage drop.

For phone lines I'd run cat6 too and do home runs. That way you can convert to ethernet later if you want. Just need to change the jacks.

 

[JackMDS]

LOL in the wall there is only wires.

Whether it is 4 VDC, or 5 VDC, or 6VDC, it is the same wires The power supply that has be install in one end would establish the level of the Voltage.

 

[VirtualLarry]

3. You can run a channel 14 network off them.

Just a note, in the USA, we only have 11 2.4Ghz band Wifi channels, 1-11. Using the higher channels intended for other countries, could get you an FCC fine, and your equipment confiscated. Not worth the hassle. (Yes, they do have "vans".)

 

[David Emiley]

The entire spectrum belongs to the public, and public can make use of any part of the spectrum, so long as that use does not interfere with licensed use. The "vans" are only used to track down known interfering signals, not as a policing patrol. My channel 14 network is an internal mesh within my home, and not detectable outside of it, and also as a focused directional antenna down the steep riverbank to an island I like to spend time on. Unless someone else with a channel 14 satellite device lands on the river island to do some work, my use of channel 14 is legal.

Case-on-point: FM transmitters that consumers use to beam audio to their car stereo using commercially licensed frequencies.

Just a note, in the USA, we only have 11 2.4Ghz band Wifi channels, 1-11. Using the higher channels intended for other countries, could get you an FCC fine, and your equipment confiscated. Not worth the hassle. (Yes, they do have "vans".)

 

[David Emiley]

My final plan is to run two CAT5e cables. First cable will be for Gigabit Ethernet. Second cable, 2 pairs will be for 2 phone lines, and the other 2 pairs will each carry 5v DC and will be wired in parallel at source and destination. Start at 5v 500mA on each pair, then increase to 5.5v 1A if tolerances hold. Thus any interference, which is unlikely, would only effect the phone lines, which are not in use.

 

[thecoolnessrune]

My final plan is to run two CAT5e cables. First cable will be for Gigabit Ethernet. Second cable, 2 pairs will be for 2 phone lines, and the other 2 pairs will each carry 5v DC and will be wired in parallel at source and destination. Start at 5v 500mA on each pair, then increase to 5.5v 1A if tolerances hold. Thus any interference, which is unlikely, would only effect the phone lines, which are not in use.

But it won't hold? You can look up the voltage sag yourself on any basic calculator. At 5V/500mA on each pair, you'd be looking at a 25% Voltage drop. http://www.calculator.net/voltage-d...ance=50&distanceunit=feet&amperes=1&x=42&y=19

You can see very clearly why low voltages over DC with high gauge wire is not a productive idea. Compare it to 48V DC which drops 0.5% over the same distance and power output. 5V power transmission is just burning up power.

 

[David Emiley]

Only 2 drops are at 50'. The NEXX APs run on 500mA. 5.5v 500mA => 4.86v at 50'. 2.75 watts in, 2.43 watts out. Negligible. Monthly energy lost < .5kWh = 6 cents.

The other drops are 20'-30'. So worst case scenario, I lose 10 cents of energy a month.

 

[thecoolnessrune]

Only 2 drops are at 50'. The NEXX APs run on 500mA. 5.5v 500mA => 4.86v at 50'. 2.75 watts in, 2.43 watts out. Negligible. Monthly energy lost < .5kWh = 6 cents.

The other drops are 20'-30'. So worst case scenario, I lose 10 cents of energy a month.

You said you were going to run 500mA per wire pair, in which case it most certainly is not just a drop to 4.86V. If you are running each AP on both pairs of conductors, then your figures would hold. Otherwise you should be good except for housing codes. According to some of the NEC policies, it looks like you'll be required to fuse on the main side, then you'll need to have a fuse at every endpoint. I'd definitely check around for any other building code requirements applicable to your area before running something non-standard like that.

 

[David Emiley]

5.5v 500mA on each pair, wired in parallel for a potential output of 1 amp to a single USB power port. Thus the port could in theory be used for charging, though my intention is to use them to power a device. I intend to off-the-shelf brick PSUs as the power source, which have single use non-serviceable fuses built-in, at 6 volts 1 amp.

My actual end result, for the 2 farthest drops, 6v 1A source run over two parallel pairs will give a 500mA operating voltage of 5.36v and a 1A charging voltage of 4.72

Now looking at 18 cents a month in wasted energy.

 

[thecoolnessrune]

5.5v 500mA on each pair, wired in parallel for a potential output of 1 amp to a single USB power port. Thus the port could in theory be used for charging, though my intention is to use them to power a device. I intend to off-the-shelf brick PSUs as the power source, which have single use non-serviceable fuses built-in, at 6 volts 1 amp.

My actual end result, for the 2 farthest drops, 6v 1A source run over two parallel pairs will give a 500mA operating voltage of 5.36v and a 1A charging voltage of 4.72

Now looking at 18 cents a month in wasted energy.

Keep in mind the USB Spec for high power devices (which is anything over 5V/100mA) is even narrower than the low-power spec at 4.75V/5.25V. While your APs probably won't care about the wonky delivered voltages (they've probably got their own regulator built in), USB charging does care, and you'd be out of spec on both sides. If that was your plan you'd have to put a regulator at each tap to bring it down to spec, but that would give the side effect that you would need to ensure you're always delivering >5V to the tap (unless you want to get into buck / boost regulators, which can get pricey per unit), and in the case of linear step down regulators, they'd be sucking more power out of your power network at all times.

Considering your farthest drops would already be out of spec, your shorter runs would simply be even more out of spec.

 

[David Emiley]

I've had cheap USB cables that drop the voltage down to 4.5v, and devices still accept them.

I suppose an easy solution would be to install a stepdown at the source on every line (I've got tons of these left over from when I was using 18650 3.7v batteries to run AA/AAA devices), run each drop, stick a USB voltmeter into the port, and then tweak the source stepdown until I get exactly 5v output at the drop. Sounds complicate but isn't. Plus it will look really cool in my wiring cabinet. 6 circuit boards with red LED displays.

Keep in mind the USB Spec for high power devices (which is anything over 5V/100mA) is even narrower than the low-power spec at 4.75V/5.25V. While your APs probably won't care about the wonky delivered voltages (they've probably got their own regulator built in), USB charging does care, and you'd be out of spec on both sides. If that was your plan you'd have to put a regulator at each tap to bring it down to spec, but that would give the side effect that you would need to ensure you're always delivering >5V to the tap (unless you want to get into buck / boost regulators, which can get pricey per unit), and in the case of linear step down regulators, they'd be sucking more power out of your power network at all times.

Considering your farthest drops would already be out of spec, your shorter runs would simply be even more out of spec.

 

[thecoolnessrune]

I've had cheap USB cables that drop the voltage down to 4.5v, and devices still accept them.

I suppose an easy solution would be to install a stepdown at the source on every line (I've got tons of these left over from when I was using 18650 3.7v batteries to run AA/AAA devices), run each drop, stick a USB voltmeter into the port, and then tweak the source stepdown until I get exactly 5v output at the drop. Sounds complicate but isn't. Plus it will look really cool in my wiring cabinet. 6 circuit boards with red LED displays.

If you have a proper isolated stepdown, you shouldn't need to adjust anything as long as your Voltage Source is within a broad range (and any switching regulator worth it's salt on the market should be able to do at least a 3:1 range in its Voltage input). Obviously though if this is some hand-made device without supporting components, you might very well need to fiddle with the source to get usable results.

 

[David Emiley]

These are adjustable stepdowns that take up to 12v input and then output is selectable in .5v steps using dip switches. Thus AFTER running the drop, I can set the source voltage for each line to exactly what I need to produce 5v output on the drop end. All adjustments done within the wiring cabinet, keeping the drop hardware as simple, uniform, and modular as possible.

If you have a proper isolated stepdown, you shouldn't need to adjust anything as long as your Voltage Source is within a broad range (and any switching regulator worth it's salt on the market should be able to do at least a 3:1 range in its Voltage input). Obviously though if this is some hand-made device without supporting components, you might very well need to fiddle with the source to get usable results.