Test procedures measure fiber reliability in ferruleless interconnect

James Laumer and Todd Berger 3M

Standard Fiber Optic Test Procedures for fiber bending provide greater insight into the performance of a unique connector design.

The evolution of small-form-factor connectors for glass optical fiber has brought a number of technological innovations to market. But to be successful for fiber-to-the-desk applications, these connectors must be simple, low-cost, and easy to use, and meet industry-standard performance requirements.

Among the new designs, which are half the size of traditional SC connectors, one design can be considered unique because it eliminates the need for ferrules and the accompanying precision components they typically require. The VF-45 connector uses a V-groove alignment process that places a 45 bow in the plug fiber to produce the desired force for a low-loss connection. However, the use of a permanent 45 bow on the fiber is such a new concept in interconnect designs that no U.S. or international test procedures specifically address the effects of this application of optical fiber on interconnect performance.

To overcome this hurdle, standard Fiber Optic Test Procedures (FOTPs) EIA/TIA 455-62A, -28B, and -31C, initially designed to measure reliability and optical loss due to bending in cabled fiber, were adapted to also characterize performance and reliability for fiber in this application. The effects of induced attenuation due to macrobending (defined as an arc or bend larger than both the minimum bend diameter and critical bend diameter of the fiber) and added stress on the fiber were studied to predict the reliability of the interconnect with different fiber types.

Core-to-core fiber alignment

The VF-45 brand interconnect, openly licensed by 3M, consists of a full-duplex plug-and-socket configuration rather than two opposing connectors and an adapter (see Lightwave, May 1998, page 39, and July 1997, page 34). The V-groove design aligns and protects fiber cores independently, not jointly. The dimensional precision of the optical fiber is used to obtain accurate core-to-core alignment as the fibers mate in V-grooves. When the plug is inserted into the socket, cams on the socket door engage the plug door, sliding it to the side to expose the duplex fibers within the plug for receipt by the V-grooves. The plug fibers align and engage the V-grooves in the socket, glide down the V-grooves, engage the resident socket fibers, and align themselves along the V-grooves for core-to-core end-face alignment with the socket fiber (see Fig. 1).

At full insertion of the plug, a bow is created in the plug fiber to place a forward compressive force against the socket fiber end-face. The force maintains intimate contact at the fiber interface for optical performance. In addition, the continuous force enables each mated fiber pair to act independently in a multi-fiber interconnect, which maintains the connection during various environmental, tensile, and side-loading conditions. Protection of the components and interface is provided by the mated plug and socket housings. When the plug and socket are unmated, the fibers remain protected by separate doors on the plug and socket.

Several fiber designs are suitable for use with this interconnect. The FOTPs mentioned previously were adapted to test attenuation and mechanical reliability of the connector using both standard silica fiber and fiber that has a permanent polymer coating. The testing examined the use of a fiber with a 100-micron cladding and permanent polymer coat designed to withstand abrasion and attain the industry-standard 125-micron outside diameter. Standard 125-micron-clad glass fiber as well as high-durability fiber designed with an integral compressive glass layer may also be used within this interconnect. Therefore, additional tests included a 125-micron-clad high-durability silica fiber with an integral compressive glass layer. All the fiber types evaluated are available with a multimode core of either 62.5 or 50 microns and were produced with typical buffer coatings to a nominal 245 microns.

Loss due to bending

Extensive testing on both 62.5- and 50-micron multimode fiber was undertaken to verify the effects of induced attenuation due to macrobending. EIA/TIA-455-62A (FOTP 62), which measures critical bend diameter, formed the basis of these tests. The FOTP is designed to estimate bend-induced attenuation when laying cabled fiber in equipment racks or in splice housings and the effect of such scenarios on the overall attenuation budget. In test applications, these bend diameters are typically greater than 1 inch.

Using FOTP 62, the critical bend diameter is determined by wrapping fiber around a mandrel 40 to 100 turns. The range of mandrel diameters in the adapted test was 6.35 to 25.7 mm, whereas the default condition in FOTP 62 is 75 mm. For application of this test to the VF-45 interconnect, the number of wraps was reduced to a minimum of one-quarter of a wrap (90).

The attenuation due to bending very short lengths of fiber was measured and compared to the fiber`s configuration in the VF-45 interconnect. The critical bend diameter, or the diameter at which attenuation varies significantly from that of a straight fiber, is important because this interconnect design requires a small portion of the fiber to have a bend. This bend must have a diameter greater than the critical bend diameter to provide acceptable optical performance. Evaluating the attenuation at different bend diameters with 40 mandrel wraps established the critical bend diameter at less than 12.5 mm (see Fig. 2).

The fiber bend in the VF-45 interconnect is approximately one-eighth of a wrap at a 19-mm bend diameter, well above the critical bend diameter. For diameters greater than or equal to 19 mm, the contribution to the total loss of the VF-45 connector due to the fiber bending is less than 0.06 dB. Both the standard fiber and permanent polymer-coated fiber were optically equivalent in bending (see Fig. 3). The typical loss of the VF-45 interconnect is 0.25 dB with the maximum loss specified as 0.75 dB.

Fiber stress

To maintain the end-face compressive force required for the fiber to remain in the V-groove and provide low attenuation and reflection, bend stress in the fiber must remain above a minimum value. This stress is also important to maintain the fiber`s position through impact, twist, and flex testing.

The amount of stress required to break the fiber is measured using EIA/TIA-455-28B (FOTP 28). The inherent strength and the strength distribution of the test fiber can be derived from the resulting Weibull plot. "Inherent strength" refers to the intrinsic flaw distribution in the fiber, which is the product of raw materials and processes used to produce the fiber. Break strengths less than the inherent strength are due to larger extrinsic flaws in the glass. These flaws may have many sources, including fiber processing, proof testing, post processing (cabling), cable pulling, and the attachment of connectors, during which the protective coatings are mechanically removed.

The inherent strength of both standard and permanent polymer-coated fiber types was found to be nearly identical before stripping. However, after stripping, the strength distribution of the standard fiber showed the typical tail due to extrinsic defects caused by stripping. The fiber with the permanent polymer coat showed no reduction in the mean tensile strength after stripping, and no corresponding defects were created. The retention of the fiber`s inherent strength is due to the protection provided by the permanent polymer coating.

Fiber stress in bending

An adaptation of FOTP 28 is a 2-point bend test in which fiber is bent until breaking, rather than pulled until breaking. The additional test data show that fiber with a permanent polymer coating as well as unstripped standard fiber and unstripped high-durability fiber with a compressive glass layer, all remain identical in mechanical performance in the interconnect. But when the buffer coat on standard fiber is removed, defects due to stripping have an obvious effect on performance (see Fig. 4). The tests indicate that it is preferable, though not imperative, to use a buffered fiber or a permanent coated fiber rather than stripped bare glass.

Fiber stress in bending is directly proportional to Young`s modulus and the fiber diameter and inversely proportional to the bend diameter. Since bend stress is related to the fiber diameter, this relation must be a consideration during the design of the connector. The use of a permanent polymer-coated fiber in the VF-45 patch cord is an important method of reducing the fiber`s bend stress. This fiber has standard core diameters (62.5 and 50 microns) but a smaller cladding diameter (100 versus 125 microns), which reduces the bend stress at any given bend diameter (see Fig. 5).

Optical fiber is tensile-proof tested over its entire length to 100 kpsi to ensure mechanical reliability using test procedure EIA/TIA-455-31C (FOTP 31).

Low loss and mechanical reliability

Through adaptation of existing FOTPs, optical performance in bending has been demonstrated to be equivalent among the three fiber types considered. Loss due to bending is an insignificant part of the total loss in the VF-45 interconnect. Additionally, the mechanical reliability of fiber in the interconnect is dependent upon the applied stress due to bending and the strength distribution of the bent fiber. At any given bend diameter, permanent polymer-coated fiber, with its smaller cladding diameter, has a lower applied stress than either standard fiber or high-durability fiber with a compressive glass layer. Inherent fiber strength at the bend can be maintained by avoiding damage due to stripping in the portion of the fiber that receives bend stress. That can be accomplished by using a fiber with a permanent polymer coating or by leaving the buffer coatings undisturbed through the bent portion of the fiber path in the interconnect.

Based on these assumptions about fiber defects in stripping and applied stress testing, all three fiber types are expected to be acceptable for use in the VF-45 interconnect for static applications.

The second phase in this work will see the application of a model to calculate the failure rate of various fibers in bending. Following this work on static failure rates, the dynamic failure rates (mating durability) will be investigated to determine if a parallel approach can be undertaken. Finally, this work is also being extended to the application of the interconnect to singlemode fiber. u

James Laumer is a process development specialist at 3M Fiber Optics and Electronics Technology Center (St. Paul, MN). Todd Berger is a product development specialist in the Fiber Optic Laboratory, 3M Telecom Systems Div. (Austin, TX).