Fiber Optic Patching question

[Wadoki]

We have a largeish facility with 4 buildings, each building has a D-Link DGS-3120 switch with a fiber module connecting the switches. One room is in the middle and has 3 fiber modules that connect to the other locations. The distance is less than 200ft per room, so I could even run CAT5 if I wanted to.

All is fine and dandy except we will be cutting the power in this building for a few days to do maintenance on the buildings electrical grid and it's not reasonably feasible to connect a generator and keep gas in it for that time period.

So I had a idea, that was to loop through a fiber patch cable in and out bypassing the middle for the other locations. That way I don't have to power the central area and still get data through. HOWEVER when I tried this I could not get the switches to see each other. I tried reversing the pairs but it still didn't work. I know the strains are good.

I don't know what to do.

 

[Dexman]

Are the light levels too low for the end devices to sync up to each other

I "FEATURE 00"

http://dexman.webs.com/

 

[mforrence]

Say the main location is A, and you have fiber from A to B, A to C, and A to D. If you take location A off-line, you can use couplers to patch B to C, B to D or C to D and it should work fine. Unless you utilize additional fiber strands and/or move GBICs, you can't connect all three together without location A in the middle. Cheers!

 

[Wadoki]

I will try the fiber couplers. I will get some on order and see how they work. Thanks.

 

[mforrence]

I will try the fiber couplers. I will get some on order and see how they work. Thanks.

How were you doing this without couplers? If you have a fiber box with pass-thru modules, they'll work as well.

 

[Wadoki]

Here some things that might help. 1. A very basic block drawing of how the switches are connected between buildings, and how they should be. 2. A photo of the BSB building, so you can see what I am talking about.

http://files.engineering.com/getfil...e=Ashampoo_Snap_2015.10.26_14h57m43s_001_.jpg

http://files.engineering.com/getfil...e=Ashampoo_Snap_2015.10.26_15h03m20s_003_.jpg

 

[libellis]

If you want to determine if it's likely to work before experimenting you would need to know:

1) Type of fiber (orange patch is usually for 62.5/125 um multimode). In the photo you say the aqua patches don't work (this color is usually used for OM3 or OM4 50/125um multimode fiber)
2) Total length of link
3) Number of mated connector pairs, excluding the connections to the switch ports at each end
4) Ethernet link speed (100Mbps, 1Gbps, etc.)
5) Determine the total link loss in dB and compare against the worst case limits for the type of fiber and Ethernet link speed that will be used (lots of information is available about these limits on the Internet).

NOTE: Assuming you don't have fiber measurement equipment, calculate the total link loss in dB:
Loss = (# of mated connector pairs) x (mated pair loss) + (fiber attenuation) x (link length)
= (# of mated connector pairs) x (0.50 dB) + (3.5 dB/km) x (link length in km)

NOTE 1: Assumes multi-mode fiber based on patch cable colors in your photo.

NOTE 2: Assumes mated connector pair loss of 0.50dB. Yours might be lower (and could be with current products and termination techniques), but standards still allow for up to 0.75 dB (no longer considered particularly good work), so yours could be worse than 0.50 dB depending upon who did the connector termination and when.

NOTE 3: Assumes no splices in the link. If you have splices, for example pigtail fusion spliced connectors, throw 0.10 dB per splice into the link loss calculation (standards allow up to 0.30 dB, but that is a lousy splice - even 0.10 dB isn't great).

NOTE 4: Assumes you can make a reasonable estimate of total link length.

NOTE 5: 3.5 dB/km is typical for 62.5/125 um multi-mode fiber. OM3/OM4 50/125um fiber is usually around 3.0 dB/km

NOTE 6: If you are using legacy 62.5/125 um fiber, then care has to be exercised if this is a 1Gbps link since - depending on length and link loss - you might have to use "mode conditioning" patch cords at each end.

If you've managed to read all the way through this lengthy response without having fallen asleep - good luck!

 

[libellis]

Forgot to clarify that 3.5 dB/km is the fiber attenuation for links using 850nm transceivers in the switches. If your link uses a 1300nm transceiver then the fiber attenuation used in the link loss calculation should be changed to 1.0 dB/km.

 

[robertjo24]

Your light blue patch cables are 50 micron MM, the orange are 62.5 MM. Not a good mix