TRD a CAT-5 Cable ?

[w4dnr]

I have one CAT-5 cable/jack in a medium size office that refuses to connect to the office router.

Wiring is hidden in metal raceway at the walls/ceiling junction.

Any suggestions as to how I might use the 4170 to guesstimate if the cable is pinched/shorted between the desk and the router that is apx 100 feet away ?

Thanks

W4DNR

[smokchsr]

Don,
I haven't done it but you should be able to test one pair at a time easy enough. The only difficult part will be making up the connection. I'd start by measuring a known good cable and go from there. Although first I'd just use one of the ethernet testers available at the home big box store, since it test all 4 pairs at once and give a quick pass fail.

[w4dnr]

Thanks for the reply.

Yes, I have one of the two part CAT-5 testers.

I will test a good cable with the AIM and see what it looks like.


W4DNR

[wa9vez]

A more formal way would be to sweep each of the four twisted pairs open circuit at the far end and then short circuit at the far end. You can save the graph which also gives a CSV file of the data. I generally open the csv file in a spread sheet and cut and past the frequency the real resistance and the imaginary reactance onto a single page. The spreadsheet has a set of complex number manipulations. Convert the separate real and imaginary into a single A+jB complex number for each of the impedances. At each frequency multiply the OC complex impedance with the SS complex impedance using the complex multiply function. The complex square root of the product of the OC and SS is the cable impedance and the atentuation constant. You don't need all the frequency data points but they should all give about the same number thru 100 mhz.

The alternative method for calculation is to multiply the Mag and add the angles and take the square root of that but I have always preferred the direct Re +jB notations. . .just old fashioned. Quattro works about the same as MathCAD.

Up to 100 mhz, the cable pair impedance should be 100 ohms +/- 15%. At some low frequency point that falls apart for a one port measurement.

I magine that you already know CAT-5 pairing at the plug and jack is 1-2, 3-6, 4-5, 7-8.

Also, you can hang a 100 ohm resistor across the far ends of the pairs and see if they looks like a flat 2:1 SWR over the frequency range on the VNA. If you parallel a 100 ohm resistor across the VNA end of the line and the cable is OK, the 100 paralleled with 100 should look like 1:1 SWR.

You can also try a band sweep and see if there is a signal intrusion into the Cat. that could be goofing up the data.

However, before you attach the VNA, check out the pairs for any "surprises" from unknown voltage/current sources. Who knows, maybe a critter has gnawed insulation of wires and there is 120 VAC on the wire.

Would definitely use a 9V battery to power the VNA. . .

I sort of glossed over details that may be useful, like a ferrite ground choke (unbalanced to balanced issue), but with a battery and the wiring floating, it shouldn't make much difference. It doesn't on my antique Belden 300 ohm transmitting feedlines.

While you have the VNA attached, you can set the VNA to a fixed frequency and use an AM radio to chase the wire in the wall.

My first guess is that one of the pairs is not a "pair" at one end or the other. Or I have seen CAT patched with grey wire twist "nuts" and Scotch 66. It works for short runs. The worst was when my son in law did not follow the pairing and it was pin to pin but not in the transposed pairs. Lots of cross talk and the wire was short in the boxes. Horrendous AM radio interference.

[w4dnr]

Thanks, James.

To expedite things I purchased a SC 8108 Cat-5 tester for $30 and it is job-specific, so it finds reversed pairs and out of sequence pairs as well as the length of cat-5 in either meters or feet. It has a far end termination and beeps when the master is connected to the termination for ease of finding the cable. It can be heard in the next room so it helps a single person operation like me. I like your BAND SCAN for interference. I will use that !!

Don W4DNR

[wa9vez]

Hi Don,

Hey that is a great price and it saves the problems of splitting out the small wire pairs. Thanks for the info and I will try to remember it for my "next time".

I am retired and have a lot of time to theorize (excepting days propagation when is good) how to extract an exact solution while running CAT over the floor to get things (my wife's scanner, printer, computer, DSL, and a router) running. In the old days, I would have tried a grid dip meter. . . after a Simpson 260. And then Hi-pot the pairs to see if I could hear snapping in the wall.

With the 4170 it is a case of trying to use it for everything above 5khz.

Jim 'VEZ