Fiber question

Got a job coming up soon where we need to pull a 24 strand 50 micron cable from point a to point b.

There we need to distribute 12 of those 24 to two seperate locations (6 strands a piece) to two areas.

Is it necessary for us to fusion splice these strands together, or can we just patch them together using a fiber patch panel with patch cords? Is the loss considering good terminations negligible?


Reply to
Loading thread data ...

Not negligible but usually tolerable. Expect the patch to yield some extra

0.6 dB. You will need about 10 dB for 10(0)BaseF(X) and you are not attenuation bound for Gbit+. Thus, as long as you are within 1.5 miles you should be ok. N.B. always consider to pull single-mode as well.
Reply to
Manfred Kwiatkowski

You will have to consider few factors before you can make a decision:

#1: do you anticipate any need to patch between the strands at the branch-out location? If the answer is yes, then the splices are out; you'll have to terminate and then use patch cords. In a situation like yours it is always the question of how far in the future you can look to anticipate anything. I've seen quite a few times people installing switches and PBX remotes and all kinds of fiber equipment in closets that were originally designed to be only splice points. It may not be as simple of the question: having terminations and patch cords at the branch-out location may give you more flexibility AT THAT LOCATION, but it will complicate management of the whole system from the central point.

#2: if there will be no cross-connections at that spot, then you can either fusion splice it or mechanically splice it. Fusion splice is preferred as it has very low reflection and pass-through loss values, but mechanical would do just fine on multimode. You can expect about 0.3 dB per a decent mechanical splice.

#3: If you splice, you'd need a fiber shelf that is capable of housing the splice cassettes. It is a common mistake that people just leave splices laying on the bottom of the regular termination shelf. You can bet it will get damaged next time someone opens the shelf. With the splice cassette compartment the shelf occupies more rack (wall) space, make sure you can accommodate for that.

Good luck!

Reply to

Hi Manfred,

0.6 dB per cord is rather on a very low side: every coupling (and every cord's got two) can add up to 0.75dB as per the standard, although a decent cord may have less. Besides, cords are left without dust caps for long time, get dragged on the dusty floor, over-bent and scratched every way imaginable. Therefore, a good margin for loss deterioration should be figured into the system when it gets designed. I would not go for less that 2.5 dB - 3 dB margin per a cross-connect point considering future deterioration.

Gigabit Ethernet is VERY attenuation-bound. It is not the only parameter you have to worry about (dispersion, too, to a great degree), but it is extremely important. As a matter of fact, it is much more strict than

100BASE-FX's 11dB. You can only have 3.5dB attenuation for a 1000BASE-SX on a (regular quality) 50/125 multimode fiber. Even LX needs not more than 4.7 dB at best.

So, if Gigabit is anticipated, I would get rid of all the unneeded points of termination.

Reply to

Strange. Last time I looked, the budged was 7.5dB per spec. In reality I usually have 10dB as well but never managed to get Gbit running at such an attenuation as the distance was 2km+ then.

My longest 1000BaseSX run is 1.2 km over two intermediate locations (= 4 Patches). To make up for future degregation and better sleep I test such runs with an additional patch inserted.

Reply to
Manfred Kwiatkowski Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.