Ethernet and Power Cables

There's another very good reason for not putting data together with mains, it's called safety..... :-S Minimal distance is 30cm if no wiring tubes are used over the individual lines.

And indeed, voltage spikes on mains *might* momentarily disrupt data traffic, or even damage databuffers due to induced spikes

Is this true? If tit is, is there a way to have the power cord shielded so that it does not bother the Ethernet cable or vise versa? Can I buy cables that already include a shield on them?

Thank you.

Electrical code does not permit power and low voltage signals to run in the same conduits or channels. Shielded or not.

Interference? Not much, I would guess.

My ac cable and the utp ethernet cable are running one upon the other, although this is not considered a good practice, but believe me; I never noticed any disturbance yet.

Then go for a stp cable.

Not even if power is only 12 volt DC?

Don't even worry about shielding. That is 60 Hz field - very low frequency - induced on tightly twisted pair wires. Almost no coupling exists. Furthermore any minuscule currents induced on the CAT 5 wire are made irrelevant by baseband signaling equipment (NICs) that operate in megahertz frequencies. What he failed to provide are numbers - the perspective that makes the problem "insignificant".

Others have def>> That is fine how ever, pulsing loads on the DC line could

That is fine how ever, pulsing loads on the DC line could induce cross talk problems in crappie communication cables like CAT5 and the like over a long hall.

"Power" means "electrical dangerous voltage-levels", in Europe the "dangerous level" is set to 48V. Usual indication of "power lines" in this context is "mains supply"...

Not even if power is only 12 volt DC?

Even if the CAT5 cable is shilded?

Changing currents of a specific frequency on one line will induce induce identical AND harmonic frequencies onto the other. The level of induction depends upon the distance between the lines (reverse-square of the distance) The intent of "twisted pair" is to have an induced current to be identical in both wires, thus (theoretically) eliminating a voltage-difference between both wires of one loop. Thus the level of induction depends upon the size of the "spike" in one line AND it's actual frequency AND the "not-perfectness" of the twisted pair AND the distance between the line and the twisted pair....

One can of course take the datasheets of a specific twisted pair and begin calculations until your hair turns grey, but that was not the point of this thread. The point was to inquire if power lines could be aside datalines and what were contra-indications to do so, nothing else... Personally i prefer to leave the math to the designers of the cables, and just apply "rules of best craftmanship"....

Chang> That is fine how ever, pulsing loads on the DC line could

well if its shielded then you should be ok how ever, most cat 5 i used is not shielded but that doesn't mean you don't have a good supply of the shielded type.

Rene wrote:

I guess working 21 years in a lab performing Lost, Skin, Radiated, and velocity tests in a work area that has made electronic wire and cables for 50 years isn't enough ? To let you in on something, our facility was one of the first to start massive manufacturing of CAT 4,5 and the like along with the foam pairs bundled with twisted pairs of control wires etc... high levels of DC currents pulsing over long runs bundled in the same race way with no form of EMF shielded does un-balance the Twisted pairs there by influencing the signal.

w_tom wrote:

Jamie posted:

bandpass, cross over, burn test for smoke, copper purity, rip cord break with out deforming pairs. Practical application testing, one of them just happens to be a a test that requires a length of a 100' to be laid along in side of a race way with mounts that have a high current lines. the CaT is placed in a holder that places it 1 " away from the high current lines which AC and DC currents are generated up to 500Amps while monitoring the twisted pairs.. this test tell us how consistent the twisted pairs are and how well they are bunched together. twisted pairs that are not uniform at the correct the lay will show uneven readings under AC currents.. etc.. P.S. we are currently making some CAT wire with combined Optics and rein forced binders made of Kevlar along with some high current control wires as a composite cable to be used in future homes and businesses. should be interesting. we have never combined those types of components in one cable before. so far, testing on the samples appear to be passing very well on the CAT and control wires. i think this maybe due to the fact that we have put lightly braided shield around those.

Btw. putting Cat wire in a race way that houses high current wires for drives DC/AC has prevent to not work well for our electricians. it took them aprox 3 mounts of playing around with a machine that was having a very slow response rate reading data from controllers via a Cat 5 . it turned out that, many errors were being generated in the interface and thus the network connection was doing alot of resends or waiting for updated readings thus causing the machines performance to be unpredictable. running the CAT 5 in its own race way solved that problem.

Jamie,

I hope you still re-visit this old post from time to time. I'd like to learn more about the tests you are referring to. What kind of currents you are talking about? What do you refer to by pulsing DC in this case, switching loads? What frequency? What type of load? What was the distance between the cable with DC current and the twisted pair? What was the length of the samples? What kind of raceway used? What was the UTP terminated on, and how did you measure the mis-balance? From your post I got a feeling that you were talking about tests conveyed quite a few years back, so I'm sure you'll open no trade secrets by posting here (or in comp.dcom.cabling to that matter). Very interesting stuff, please continue this thread.