EMI & Cat5

It's a bad idea, because you can induce enough voltage in the wires, to cause a hazard. It generally won't bother the NICs though, as the frequencies used to carry the data are well above those used for power and they're also supposed to be able to withstand several hundred volts, without creating a safety hazard. Now if you were to touch the cat6 conductors, while also touching some grounded metal, you might see why you're not supposed to do that.

How does anything get induced into the CAT5/5e/6 wire? The pairs are twisted and carefully balanced to meet CATx specs and certification. When CAT3 was new there was lots of discussion about EMI and if any was found it was only in the most extreme situations. One way CAT5e wire differs from CAT3 is carefull symetric twists, which cancel out induced currents from external magnetic fields.

For that matter, for the OP, 440 volts is a "don't care" or even a plus since it halves the amperage and it's amps, not volts that create a magnetic field.

IMO Follow construction codes and be happy.

There's both common mode and differential mode coupling. Twisted pairs reduce only the differntial mode coupling. As for coupling, there's both inductive and capacitive coupling between the power and lan cables. There will be some of both.

Small amounts of common mode currents/volatges are blocked by the transformers used to couple the cable to the circuity.

This document is interesting. (The text I've quoted is about 3/4 down)

My reading of the document says that in the normal home/office/light manufacturing environment EMI is a non-issue with an ethernet plant built to proper CAT5/5e spec and proper grounding of the hubs, switches and computers. The Allen-Bradley paper also addresses heavy industry applications where I imagine the ground currents could make common mode issues interesting.

Shielded cable comes off worse than UTP.

"Testing was performed on a system where the communications conductors (four-pair Ethernet cables) were placed in wire ladder-ways. These cables were placed along side conductors carrying 480 V drive-control voltages connecting the controller and a high-horsepower motor. The differential voltage coupled into the Ethernet cables varied, depending on the type of cable tested.

The picture in Figure 12 shows the cable routing with respect to the drive conductors. Table 3 shows the differential voltages measured on the communications cables. The additional CMRR is referenced to the shielded cable. The shielded cable is used as a reference because this cable provides adequate performance in environmental testing for Ethernet modules.

It is worth noting that there were no physical layer communications errors during this sequence of testing" [....] "The four-pair Ethernet cables were tightly strapped to the high-current secondary of the welding conductors and routed in close proximity to the weld tip. This setup provided both magnetic and radiated noises during the welding event. Figure 13 shows the routing of the cables on the robot arm. A telecommunications analyzer was used to generate 100Mbit NRZ data that was sent to a pair of Fast Ethernet transceivers for encoding and decoding into MLT3 type data. The telecommunications analyzer provided bit-level error detection. The performance statistics were recorded for each of the five cables tested. Surprisingly, zero data errors were recorded for each of the five cables tested during the welding events."

Yes, but _some_ isn't necessarily significant or even noticable.

Voltage will drive capacitive coupling. There are fractional picofarads per meter in a twisted pair. A foreign conductor (inside another sheath) will haveseveral orders of magnitude less capacative copling. The twisting will make that coupling roughly identical to each leg of the pair. I wouldn't worry about less than kV levels, and then first for safety. I wouldn't be surprised to hear of ethernet running fine parallel to clean 15 kV.

Current produces magnetic fields. Voltage doesn't matter. Changes in magnetic fields induce current. That makes transformers work and AC preferred for power distribution. But 60 Hz is way below the sampling/detection frequency for ethernet so won't interfere with anything. On top of that, the twisting evens out induction.

-- Robert

I was thinking more of a safety hazard. As I mentioned earlier, interference to data is likely not an issue. Still, it's a good idea to maintain separation.

If you go back, you'll find I mentioned that interference is not likely. A more real concern, is safety. You cannot guarantee that insulation will not fail. You cannot guarantee that the power wiring is up to code with no failures. Therefore, the best policy is separation of data cables, from AC power. 99 times out of a hundred, there'll likely be no problems, but there's always that one time.

From a safety standpoint CAT5 is the same as any low voltage wiring and if you install to local codes you'll be safe.

thought you

The reality is that some places limit you to running the datacomm/telecomm cabling in the same raceway adjacent to high current cables. I should stress that the voltage is not as important as the current because the coupling is mostly by electromagnetic induction, not by electrostatic induction.

My main concern about running them side by side is not day-to-day interference, but when high fault currents go thru the power cabling, usually from lightning, possibly causing the induction of high currents into the datacomm cabling, and possibly damaging the equipment at the ends.

It's mostly inductive, because the signal cable isn't running in the same conduit, but adjacent to the power lines, which in most cases are shielded from the signal cable with a metal conduit. If the conduit is metallic, then there's no capacitive coupling, because the metallic conduit is grounded.

Well, generally speaking that's why the power cables are often in EMT metallic tubing or flex, which is grounded so that an insulation failure won't cause shock hazard or fire.

Murphy's Law.

any won't _any_ metalic conduit attenuate an AC magnetic field by induction and resistive loss? I have no feel for what magnitude the effect is.

Another way to take into account the concern that the wire insualtion can fail it to use insulating conduit. For example in Finland many builings you can find electrical tubing made of plastic material and wires in them. I wire insulation fails, the tube still works as insulation, so no fire or shock.