Permanent fiber termination polarity

[her209]

Permanent meaning the fiber that goes from fiber distribution panel to fiber distribution panel in the cross-connects.

What is typically used? I would think its should be the A-B configuration because it allows for use of A-B patch cords on both ends.

Any [good] reason why it would be terminated A-A?

 

[imagoon]

Long haul fiber is color coded to follow telephony colors + 2 extra colors and should be patched to the panel following the order. AB AA etc should be handled in the patches. Patch cables normally have a release pin so you can switch them for this reason.

http://www.telephonecentral.com/custom.aspx?id=29

 

[spidey07]

Always straight through, never crossed. You use the patch cables/jumpers to swap tx/rx. Corning has really nice ones you just unlock and twist. The only place it should be crossed is in a single patch cord to active equipment. This allows you to "cross connect" or jumper between fiber panels without worrying about polarity.

I'm pretty sure eia/tia standards are straight through.

 

[her209]

OK, let me illustrate in pictures:

MC = Main cross-connect
IC = Intermediate cross-connect
HC = Horizontal cross-connect

Given that the permanent fiber is A-A, you can use A-B fiber patch cords in all three cross-connects. You cannot use A-A fiber patch cords in all three cross connects. If you do use A-A, you have to use two.

Given that the permanent fiber is A-B, you can use A-B fiber patch cords in all three cross-connects. You cannot use A-A fiber patch cords in all three cross connects. If you do use A-A, you have to use two.

But if there is no IC between the MC and HC:

Given the permanent fiber is A-A, you have to use one A-B and one A-A fiber patch cord.

Given the permanent fiber is A-B, you can use A-B or A-A fiber patch cords.

Therefore it seems to me that A-B would be a better polarity choice. Am I missing something?

 

[spidey07]

You're missing that the blue fiber on one end should stay the blue fiber all the way through the route no matter how many jumpers there are. Keep it straight through, you way over complicate it if you don't. It's how all fiber distribution is done.

The only cross that should occur is at the active equipment.

 

[her209]

You're missing that the blue fiber on one end should stay the blue fiber all the way through the route no matter how many jumpers there are. Keep it straight through, you way over complicate it if you don't. It's how all fiber distribution is done.

The only cross that should occur is at the active equipment.

In order to keep the blue fiber all the way through end-to-end, you'd use the A-B configuration on the permanent fiber and the patch cables.

 

[spidey07]

No. You should always use straight jumpers, and then do a single swap at the active equipment.

You just don't cross fiber between panels. Don't do it.

Modern high density panels with 12 strand ribbonized fiber you physically can't do it.

The only crossing should occur with a single jumper at active equipment.

-edit-
Think of it this way, following the standards the blue strand is always in the first position, orange is 2nd, etc anywhere at any fiber interconnect shelf/panel. Be aware that there are applications where only a single strand is used for tx and rx.

 

[imagoon]

If you follow the industry standard color code, all of your permanent fiber will be straight though as it should be.

Like Ethernet and phone cords, crosses should only happen at the equipment. There is no good or valid reason to do it any other way. It is even less of advantage on fiber since the ends pop out of the cable retaining brackets so you can swap them. This makes "AA" or "AB" fiber universal because you flip it as needed.

 

[spidey07]

And one shouldn't be mucking with the back of a fiber panel at all to switch. That's the entire premise of fiber FICs.

 

[her209]

I think I might have found the answer I was looking for:

http://hqwww.panduit.com/panduit/groups/MPM-OP/documents/WhitePaper/CMSCONT_032945.pdf

Method A
Method A employs Key Up to Key Down Adapters to link straight-through Key Up to Key Down ribbon cables to fiber cassettes (see Figure 7). This method maintains registration of Fiber 1 throughout the permanent link: Fiber 1 in the near-end cassette mates to Fiber 1 in the ribbon cable assembly, which mates to Fiber 1 in the remote cassette. The fiber circuit is completed by utilizing one flipped patch cord, either at the beginning or end of the permanent link, to move Fiber 1 into the Fiber 2 position at the face of the remote cassette to ensure proper Tx-Rx orientation.

The method requires two Method A cassettes, a Method A ribbon cable, an “A-to-A” patch cord at one end of the connection, and an “A-to-B” patch cord at the other end. PANDUIT recommends that the end user introduce the pair-wise flip consistently at the beginning or end of the channel for easier management. Patch cords with duplexing clips that hold connectors in place may be used rather than permanent molded assemblies. These clips open at the sides to permit installers to easily reverse or “swap” the order of connectors, so that the “A” cable becomes “B” when necessary to maintain polarity.

Method C
Method C employs Key Up to Key Down Adapters to link a special Key Up to Key Down ribbon assembly to Method A cassettes (see Figure 9). In this method, polarity is achieved by using a ribbon cable that has the pair-wise flip built in to individual fiber pairs, and the fiber channel is completed by utilizing straight “A-to-B” patch cords at the beginning and end of the link. In essence, the only difference between this method and Method A is that the pair-wise flip occurs in the array cable itself rather than at the patch cords, so that oddnumbered Tx fibers leaving the near-end cassette are in even-numbered Rx positions when they arrive at the remote cassette.

Channels built with this method using one array cable may take the same time to deploy as channels built using other polarity methods. However, in longer channels that connect multiple array cords in sequence, the data center manager must keep close track of the number of array cables used in order to preserve the pair-wise flip, which results in more complex channel designs and longer maintenance times.

:hmm:

 

[her209]

If you follow the industry standard color code, all of your permanent fiber will be straight though as it should be.

Like Ethernet and phone cords, crosses should only happen at the equipment. There is no good or valid reason to do it any other way. It is even less of advantage on fiber since the ends pop out of the cable retaining brackets so you can swap them. This makes "AA" or "AB" fiber universal because you flip it as needed.

A host sends on pair 1/2 and receives on 3/6. A switch sends on 3/6 and listens on 1/2. Hence you had to use a straight-thru patch cord with straight-thru station cabling. Auto-MDIX has made that a moot issue.

In fiber optic transceivers, is B always transmit and A always receive? Is this true even on fiber optic switches with fixed transceivers or is it reversed?

 

[imagoon]

A host sends on pair 1/2 and receives on 3/6. A switch sends on 3/6 and listens on 1/2. Hence you had to use a straight-thru patch cord with straight-thru station cabling. Auto-MDIX has made that a moot issue.

In fiber optic transceivers, is B always transmit and A always receive? Is this true even on fiber optic switches with fixed transceivers or is it reversed?

If you stare at the fiber converters (when off so you don't go blind) with the pull bar down and the catch up, the left is always TX and the right is always RX. The pairs are "crossed" because of the way they are terminated. Look at a 3 foot patch. Both ends of the latch should be "up" (same side) and a straight cable. Imagine the two switches sitting across from each other. You will note that when connected TX (what you are calling A) will go straight across in to the RX (B) on the other switch.

Fiber switches follow the same rules. Left, TX (clasp up.) Right, RX (Clasp up.)

The ethernet thing with switches doesn't matter in fiber because there is only 2 physical... i dunno "carriers" which so keeping them straight as long as the cables are built correctly is easier and doesn't require the differing port standards.

 

[imagoon]

This youetube video shows you how to recross the fibers should you need to flip them.

http://www.youtube.com/watch?v=VTde2k-JcgY

Basically they are straight if you place the cable on the table, fibers uncrossed and the the couples are both up. If one end is up and the other is down, they are crossed.

 

[spidey07]

Yes.

Do believe what you read on the intarweb vs me doing it for 20 plus years.

 

[imagoon]

I am also looking at that doc... are you using panduit 12 fiber patching systems? If you are not then a) that doc doesn't apply and b)

This point of crossover is known as a “pair-wise flip” or, less commonly, as an “A-to-B polarity swap”.
Regardless of the term used, the flip must occur at some point in every serial duplex link. Polarity methods differ, in essence, by the location of the pair-wise flip: Addendum 7 recommends that the flip be built into the duplex patch cords (Method A), array cables (Method C), or through a combination of cassette and trunk fiber transformation (Method B).

Method a) is 'use a straight cable and flip the polarity with the patch cord'

I think that covers what spidey07 and I have both said.