In the case of pre-wiring with fiber, it seems there are two options: 1) use a proprietary solution such as Celerity, or 2) run general purpose fiber such as OM3 multimode. What's your recommendation here for a residential pre-wire where the goal is to have fiber in place for a couple of key rooms?

If all four pairs were created equal, then I imagine it would have no bearing. However, as I understand it, different pairs have different characteristics. In that case, it would make sense that matching specific pairs to specific pins in the HDBaseT scheme could impact transmission quality.

Run conduit.  Fiber should be in conduit anyway so just run it up front.

In my non-scientific observation, different brands seem to twist different pairs tighter or not so tight. Some may have the browns rather loosely twisted, and others will be more loose on another pair.

Bulk Fiber. There are some extenders out now..there will be more. OM3 is not expensive run it....proprietary will be ...well you know. You can keep it dark for now but you are putting it out there future proof

But it's not likely actually that.

We all know that CAT cables are sets of twisted pairs of wires where the rate of twists per inch is different for each of the four pairs in order to minimize crosstalk from one pair to any other. So all CAT cables are the same.

But I've seen different models of wire where which color is twistiest, which is least, etc, is not the same as with another model or brand of wire. (I'm finding this hard to express in a simple clear way.) The four pairs do not all carry exactly the same information as one another, so the performance of the cable will vary in some way from model to model, because of which pairs get which parts of the signal.

This means it's possible that the particular wire that you use has rates of twist that work better with 568B than with 568A, but another brand or model of wire might not.

Note that cables are usually specced for a 100 meter length, meaning every pair will function properly at that distance. Any differences due to twist rate should not show up over distances significantly below the spec limit.

Incidentally, it makes sense that a wire manufacturer, having once chosen which colors to twist how much, will keep setting up their manufacturing process using the same rates of twist for the same colors as last time they ran this model.

All this means it totally makes sense that once you've settled on a brand and model, you can expect performance to be the same with that wire, but don't assume wire from other manufacturers will perform the same.

My recommendation is to buy bulk fiber and run it. Product is much more robust then it once was, and while it can be damaged, I've seen it tied in knots and still work. There are several purveyors out there that work with the product; we use Techlogix. Very satisfied, and they offer tutorials online on termination plus lots of ancillary hardware to meet most of your needs.

To echo my Miata driving friend, it's tough to get 18 g worth of a cable that's rated for 12 g or less...

Fiber solves the problem today and tomorrow.

"Twistiest"? Is that the long-rumored follow-up 45 RPM to Chubby Checker's "The Twist" and "Let's Twist Again"?

I terminated cables for more than 15 years before EZ ends. I see no reason to go back

Skew. [Link: flukenetworks.com]

Same exact observation I have found, thus further strengthening the belief A vs B is hog wash.

Pairing must be maintained.

Huh? All pairs are twisted.

Remember the idea of twisted pairs is to reduce crosstalk (egress) and interference (ingress) from other pairs, so we have to maintain the pairings.

The positioning of connections on the male and female connectors is also optimized to reduce crosstalk from pair to pair.

So you could make up your own pinouts, and have a degradation of performance.


That's "experienced persons' tale" now.

Right, Im talking about how much they are twisted. The pairs vary in twist and are not the same per pair in each cat, nor the same across the various manufacturers.

And LOL on the PC comment.

When you said "what pairs are twisted," I saw no connection between that and "how much they are twisted."

They vary in twists per inch, and they only vary by as much as is required to meet the crosstalk spec.

As the amount of twists increases, the actual length of the pair increases, so adding more twists delays propagation differently per pair... so the more twists, the better the crosstalk spec, and the fewer twists, the better the time alignment of the four sets of signals... so a balance must be struck between a better crosstalk spec and a better time alignment spec.

The reason that twists vary from manufacturer to manufacturer is that a performance spec must be met regardless of color and there's no master spec god telling everybody what color must have most, what the least, etc twists per inch.

Going back to our Aussie friend, better performance as 568A than as 568B implies that the wire is being used at the extremes as its spec, for more than 100 meters in length.