Reliability

When a optical fiber cable is bent, it induces compression and tension forces on the surface of the glass fiber. This does not present a problem to the integrity of the fiber unless a microscopic flaw, which is in all glass fibers, is large enough and located precisely in the outer portion of the bend. This could cause a flaw growth over a period of time.

There are two basic ways to determine the reliability of fiber when two controlled 180-degrees, 0.5" diameter bends [0.25 inch {6.4mm} radius] are applied to a patch cord.

  1. A probability analysis, based on scientific data applied to multimode and single-mode fiber cables, was analyzed by Corning, Incorporated. The findings were based on using a Two-Region Power Law* model for predicting crack growth. The calculated failure of probability is 2x10e-6; that corresponds to a two parts per million-failure rate over twenty years.
  2. Actual data from previous field installations of products that utilize 0.25" bend radii. The 0.25" bend radius has been used for decades in fiber network products, called attenuators. One such attenuator is the Fico, Inc., 3-Step, which has an impeccable track record for performance and reliability in thousands of installations. The SGA10025 applies three 0.125" bends in single-mode fiber cables (see figure 1).

*For information on the Two-Region Power Law model, see Hanson, T.A., Glaesemann, G.S. (1997). "Incorporating multi-region crack growth into mechanical reliability predictions for optical fibers." M. Mater. Sci. 32, 5305.
Figure 1
Attenuators have decades of proven reliability performance at 0.125" bend radius and less
Attenuation

The restricted launch condition or limited phase-based launch condition (LPS) tables below indicate measured attenuation loss values when inducing two controlled, 180-degree, 0.5" diameter bends (0.25" radius) into a three meter, 62.5um and 50um tight-buffered multimode fiber cable (as used in patch cords).

The worst case measurement of attenuation loss from the LPS launch tables was 0.312dB at a 1300nm wavelength for 50um tight buffered fiber cable. This value was used with an extra 0.188dB added (margin) to arrive at 0.5dB total loss value for a 0.25" patch-cord adjuster pair.

The 0.5dB patch-cord adjuster pair attenuation loss should be calculated as shown in the equation below and factored into the channel loss budget just as 0.75dB attenuation loss is for mated connector pairs (see example for calculation of backbone channel loss budget).

Attenuation Loss of 62.5um & 50um Multimode Fiber Deployed with Two Controlled 180-degree 0.5" Diameter Bends (0.25" radius).
LPS Launch
62.5um Tight Buffered ATTENUATION (dB)
  850nm 1300nm
Measurement 1 0.051 0.071
Measurement 2 0.048 0.065
Measurement 3 0.060 0.079
Mean 0.053 0.072
Std. Dev. 0.0062 0.0070

Testing Performed by:
Corning Incorporated

Corning, NY
50um Tight Buffered - Table ATTENUATION (dB)
  850nm 1300nm
Measurement 1 0.256 0.292
Measurement 2 0.261 0.312
Measurement 3 0.258 0.300
Mean 0.258 0.301
Std. Dev. 0.0025 0.0101

Testing Performed by:
Corning Incorporated

Corning, NY
(.25 Patch-cord Adjuster)) 0.5dB = 0.312 (worst case from tables) + 0.188dB margin

NOTE #1: When under an overfilled launch condition (LED), typically used in hand-held field testers at this time, it is recommended to follow the Field Testing with an LED Source (see below) as used in the proposed TIA-568-B.1 document.

Field Testing with an LED Source
To remove high-order mode transient losses from multimode optical fiber measurements when using a source which excites these transient high-order modes. The reference jumper shall be wrapped in five non-overlapping turns around a smooth round mandrel (rod) during the reference calibration of the source to the detector and for all loss measurements. The mandrel diameter depends on fiber core size and shall be as specified in table #1.

Table # 1 Mandrel diameters for multimode optical fiber core sizes
Fiber core size (um)
Mandrel diameter buffered fiber
[mm (inches)]
Mandrel diameter 3.0mm jacketed cable [mm (inches)]
50
25 (1.0)
22 (0.9)
62.5
20 (0.8)
17 (0.7)
Calculating premises cabling channel loss budget
Fiber-optic networks are always specified to operate over a range of loss. When determining the cabling channel loss budget for a fiber-optic network, you must know the supportable cable distance. The following equation should be used when designing a centralized, backbone or horizontal channel in the structured-cabling system.
 
 

Channel Attenuation Maximum Channel Attenuation (from table 2)

Channel Attenuation = Cable Attenuation + Connector Attenuation + Patch-cord Adjuster Attenuation + Splice Attenuation

Channel Attenuation = [Cable Attn Coefficient- Table 3 (dB/m) x Length (m)] + [# Mated Connector Pair x 0.75dB] + [# Patch-cord Adjuster Pair x 0.5dB] + [# of splices x 0.3dB]
   
 

Table 2
Supportable Channel Attenuation for Optical Fiber Applications by Fiber Type
Application

Wave
Length
(nm)

Maximum Channel Attenuation 2 (dB)
62.5 um
50 um 1
10BASE-FL
(Ethernet)
850
12.5
7.8
Token Ring 4/16
850
13.0
8.3
Demand Priority 3
(100VG-AnyLAN)
1300
850
7.0
7.5
2.3
2.8
100BASE-FX
(Fast Ethernet)
1300
11.0
6.3
FDDI (Low Cost)
1300
7.0
2.3
FDDI (Original) 1300
11.0
6.3
ATM 52
  155
  1558
  622
  6228
1300
1300
850
1300
850
10.0
5.3
10.0
5.3
7.2
7.2
6.0
1.3
4.0
4.0
Fibre Channel 133
 
266
 
2668
 
5318
 
531
 
10628
 
1062
1300
1300
850
850
1300
850
1300
6.0
1.3
6.0
5.5
12.0
12.0
8.0
8.0
-
-
4.0
4.0
-
-
1000BASE-SX8
(Gigabit Ethernet)		 850
3.29
3.99
1000BASE-LX8
(Gigabit Ethernet)
1300
4.09
3.59

1. A worst-case source coupling loss of 4.7 dB is used for 50 um relative to 62.5 um for LED based applications. This coupling loss is based on the theoretical maximum coupling loss. 10BASE-FL specifies 5.7 dB maximum coupling loss into 50 um fiber. Token Ring, FDDI (Low Cost), FDDI and 100BASE-FX specify 5.0 dB maximum coupling loss into 50 um fiber.

2. "NST" (non-standard) entries indicate where the standard does not specify support for the media, but where equipment is commonly available to convert the native application signals to a form compatible with the non-native media.

3. Application specifies 62.5 um fiber with 200 MHz-km bandwidth at 850 nm.

8. This is a laser-based application. When not so noted, multimode applications are LED-based.

9. Maximum channel attenuation based on channel insertion loss plus unallocated margin from IEEE 802.3z.
 
 
Table 3
Maximum Cable Attenuation Coefficient
Wavelength (nm)
62.5 um
50 um
850
.0035dB/m
1300
.0015dB/m
 
Premises Design Seniors
 
Example: Calculating Backbone Channel
Attenuation Loss
 
NOTE #2: 0.75dB-mated connector loss value has been determined by overfilled launch condition as used in LED-based systems. The restricted launch condition, used in laser-based systems, is currently under study to determine the proper attenuation loss value.The restricted launch condition (laser) is expected to result in lower attenuation then the 0.75dB used in LED.
Conclusion

As indicated in the Corning analysis, the reliability of two controlled 180-degrees, 0.5" diameter bends (0.25 inch radius) applied to a patch cord have an insignificant effect on the life cycle for patch cords.

When using a 0.25", patch-cord adjuster pair you can easily apply a 0.5dB attenuation loss value just as an attenuation loss value of 0.75dB is applied for a mated connector pairs in calculating the channel loss budget. This provides the designer with an option, when utilizing patch-cord adjusters in their design, to reduce the fiber patch cord lengths, and eliminate patch-cord slack thus increasing the management system's efficiency and aesthetics.