Wat. I can see why it might matter if it's open vs. terminated, but how is shorted just as okay as terminated, yet open is somehow unacceptable?
At least such advice tells you up front how valuable it is...
More particulars --
Let me state my use is receive for MW and HF up to 4MHz.
So, first of all, if you're using say 10m or less, it's basically just wire.. not a lot of transmission line going on here (electrically short, small phase angle). What you do with the other pairs, doesn't really matter. There's not enough to drive a lot of coupling between them, even if they weren't twisted at different rates.
And, if you need lower feedline losses say, you can happily connect pairs in parallel -- mind the impedance goes down of course, but the power capacity goes up. (Uh, hm, power capacity and attenuation aren't the same thing, I'm not sure if that would actually help with overall losses.)
A power combiner would be ideal, of course. You can connect pairs in series-parallel to keep the same 100 ohm total, but this does put fully half the signal, in common mode, on each pair-of-pairs. So the cable is balanced overall, but individual pairs aren't. Not ideal, but would probably still be fine. Note that because there's a TL impedance between pairs in the cable, the impedance of this configuration will be lower than the 100 ohms of ideal TLs in series-parallel.
Anyway, anywhere you're using multiple pairs in sync, mind that their velocities won't match -- a consequence of the differing twist rate. A few wavelengths' total length should still be fine. At some point (10s of wavelengths?), you'll find anomalous VSWR in the cable itself, due to circulating currents.
Two big advantages at least for directional antennas is your pattern is not altered by signal ingress on the feed line. The twisted pair rejects ingress very well. The second advantage is the extra wires for control and or power.
I mean... it rejects as well as anything properly balanced does, and it rejects as well as anything unbalanced does when properly grounded and shielded. You aren't granted any magical powers either way. Still need a good balun, and if you expect [common mode] feedline voltages/currents, you'll still need good CMRR in that balun, or the antenna itself, or the tuner, or transmitter, or anywhere along the signal chain that those common mode currents can go.
The consequence in terms of directional pattern would be, your feedline acting as a monopole or "long wire" or what have you, coupling in via poor CMRR somewhere, and summing with whatever the antenna's normal radiation pattern is. For example, a dipole mixed with a monopole would be... some kind of tilted or twisted torus, I guess?
"If the pairs are unused then they must either be terminated in 100 ohms at each end or shorted out. If any of the pairs are used for feeding power (two pairs are used for 12V feed in the active antenna) then a decoupling capacitor must be fitted at either end to short the pair to RF.
If the spare pairs are left floating then it really upsets the crosstalk performance and adds losses.
Already gave my thoughts on this, but I will note that the source and load for that 12V DC feed will probably appreciate the bypass caps (and the RF won't care, because, that's why you're using twisted pair in the first place, innit?
). Even more so: if there is CM coupling in the system, then adding a CM choke would be even more welcome! And maybe some caps to ground, ground early and often, y'know? For the RF line, that can be done using a CT'd choke, the CT providing a ground point. (A balun, that doesn't go anywhere. But it's also balanced only. So, a balbal. But, just a bal?
)
Now, my question, There is shielded and un-shielded Cat5/6, if I were to use a shielded Cat5/6,
How do I handle the shield, Ground at one end, ground at both ends, or leave it floating?
It seems to me the shield could actually be detrimental if not handled properly.
I'm wondering if I should just get un-shielded Cat5/6?
Why introduce another conductor you need to worry about? Wasn't that the point of getting UTP in the first place?
A shield probably doesn't do all that much good in a feedline context. It can provide a ground of sorts, but it is by itself a wire through space, so it's not going to earth anything over a distance. If it's run in a (metal) tower, you can ground it to the structure, but you can also use the CT choke with the unshielded pair to provide local CM grounding. Or use coax and ground the shield without using a choke at all (but, still needing a balun for the [balanced] antenna itself, of course). Likewise, where it enters the shack, you can ground it, but if you've not taken measures to also terminate the pair-to-shield (common mode) path at the antenna end, that same mode is piped straight through right up to your hardware.
The UTP/STP dilemma seems to fit into the "more is better" anti-pattern. If you can't reason out all the ways to use the additional features, and all the ways they might go wrong, it's probably better to just not use them in the first place.
Tim