Cabling, particularly in flexing systems, has been a problem for system designers. System designers must provide the reliability required to maintain productivity while responding to the increased use of automation. Flex cables, often misperceived as low on the technology food chain, have a direct impact on the reliability of equipment. Their failure can instantly neutralize all electronic and mechanical functions, bringing entire production lines to a standstill.

Because of the importance of constant uptime to lean production scheduling, system designers are applying more sophisticated evaluation techniques, such as Weibull charts, to their flex cable design. Although designers often use these charts to illustrate fatigue failure in electronic components, you can apply them to flex cables to go beyond typical "flex-cycles-to-failure" data. The charts more accurately pinpoint the expected failure levels of a given flex cable. Understanding failure profiles and their underlying basis allows you to take appropriate precautions in the design stage. You can, thus, avoid production-stopping cable failures.

Designers often use reliability statistics to characterize the lifetime of complex electronic components that are subjected to stress, such as ICs and pc boards. Also, designers routinely use Weibull distribution plots to calculate the reliability of ICs but have never applied them to flex cable systems. By using this more sophisticated analysis, you can gain insight into expected failure plateaus for cabling components.

Fig 1 shows a Weibull distribution for a flex cable with the cumulative failure fractions over a range of rolling flex cycles. Once you establish the distribution, you can calculate the system lifetime to a certain level of confidence. For example, from the plot in Fig 1, you can see that more than 95% of the cables will last a minimum of 500,000 cycles.

Traditional data for flex cables provide only a minimum bend radius, with no average lifetime or distribution. Therefore, you can't make accurate predictions about reliability. A statistically valid Weibull plot for flex cables should contain data from at least 20 actual stress-to-failure tests. Designers have not used stress-to-failure tests because of the expense. However, a serious analysis of performance demands this commitment level. Fig 1 shows 38 separate runs of the stress test.

By applying this analysis and other advanced statistical tools to the design of flex cables, you can get a more accurate determination of MTBF, which affects service scheduling. You can apply these models to a number of design parameters. The combination of modeling and detailed design knowledge results in a significant increase in the reliability of flex systems.

You can segment the flexing of cables into the three basic motions (Fig 2). Each time you bend or flex a cable, you stress the copper conductors and shields. Copper has poor resistance to repeated stressing and can fatigue and fail even if the stress is kept below its ultimate yield point of 15% elongation. Copper has low resistance to shear stress because of its different crystal structure. The metal can deform even if the stress is below the plastic yield point. (In contrast, steel does not deform because of its crystal structure. Instead, steel springs back to its original position each time it is stressed below its plastic yield point.)

Cable that is subjected to flexing can fail for three reasons: cable and conductor insulation degradation, conductor and shield fatigue in flex area, and conductor and shield fatigue at point of termination.

Insulation and jacket degradations. Cable insulation or cable-jacket failure is caused by constant abrasion between the cable and other cables, hoses, or cable-management hardware, such as cable tracks. Metal or plastic chips, solvents, and lubrications can also contribute to rapid degradation of the cable insulation. When the cable jacket fails, it exposes the cable interior. Any liquid that is present works its way into the cable and eventually causes a short circuit between the conductors. Abrasive particles can also attack the conductor insulation and cause it to fail. If the cable has an overall shield, it opens to ground. Table 1 shows the jacket materials resistant to abrasion.

Another physical property to consider is tensile strength. You can use a thinner jacket if the tensile strength is high. Jacket size is an important factor to consider when trying to reduce the size of the cable bundle.

Temperature extremes and low atmospheric pressures (vacuum) weaken or embrittle the cable jacket. You should select a jacket material that is rated for the operating temperature and does not contain plasticizers. Plasticizers evaporate over time or at low pressures and cause the jacket to embrittle.

In addition to meeting the same requirements as the cable insulation, conductor insulation must resist crushing. In a typical round cable, clamping or flexing a cable in a track with many cables and hoses exposes the conductors to high compression forces. Insulation with high tensile strength and low elongation, such as polyester, is best for most industrial applications. If you're managing a large number of conductors in a small area, a thin-wall, high-strength conductor in-sulation reduces the size of the cable and has the greatest effect.

Conductor and shield fatigue in flex area. The most common mode of cable-flex failure involves the eventual fracture of the conductor in the flex area (Fig 3). The conductors are still functioning when the shield fails, so the operation continues. The cable is susceptible to EMI/RFI and emission, however. This situation creates errors and false signals with a source that is difficult to identify--typical of EMI/RFI problems.

To understand the mechanism of conductor and shield failure, look at the basic concepts of stress analysis. A rigid body's resistance to bend depends on the material, shape, and area of the cross section and radius of curvature of the bend. This stress, s, is represented mathematically by:

where M=bending moments, c=distance from the neutral axis of the body to any fiber in the cross section, I=moment of inertia of the cross section, and ø=stress in the fiber at distance c.

For a typical flex cable, the geometry of the bend is fixed by considerations that include mechanical-design constraints and package layout. You must work within these constraints and minimize the conductor stresses that reduce flex life. To reduce the fatigue on the copper conductor and shield and, therefore, eliminate wire breakage, you should make the bend radius as large as possible. The diameter, or thickness, of the cable should be as small as possible.

The most important factor in determining flex fatigue life is the maximum stress in any part of the cable. Assuming the bend radius does not go below a minimum value, this maximum stress, R_MIN, is given by:

where E=modulus of elasticity in psi (17million for ETP copper), C=maximum distance from the neutral axis to any fiber, and R=bend radius. Notice that this relationship holds for any cross section because the moment of inertia, "I," does not appear.

You can minimize stress by decreasing the cable thickness, or diameter, C_MAX, or by increasing the bend radius, R_MIN. You can also select conductor and shield materials of higher tensile strength than copper.

Flexing tests show that the resistance of the copper conductors and shield increases as the metals work harder under flexing. The harder you work the metal, the more brittle it becomes. Faster cycle rates of the equipment generate greater temperatures in the copper. A small bend radius also creates a higher temperature as well as a higher degree of fatigue. The increased temperatures can create insulation softening. The insulation softening changes the insulation's physical properties, such as reduced abrasion resistance, decreased cut-through resistance, and decreased tensile strength. Therefore, all of these changes can cause premature cable failure.

Conductor and shield fatigue at the point of termination. Bending stresses and vibration from moving cables cause connectors and crimped or soldered cable terminations to break. Unsupported cables prematurely fail because of fatigue at the connector interface. You should always clamp flexing cables to a rigid surface at the point of termination to remove any stress from the termination points. Your choice of the proper strain-relief method is as important as the proper selection of insulation and conductors. Cables can suffer "whiplash" from fast-moving carriages, which cause the cable to change direction quickly and snap. In all high-speed flexing applications, a cable that is stiff offers resistance to bending. A stiff cable flexes much longer than a limp cable, which offers little resistance to bending and results in the whiplash phenomenon.

To increase flex life in cables, you need to consider size carefully. By reducing the diameter of the cable, you achieve an exponential increase in flex life when the bend radius is held constant. Using standard copper conductors and reducing the size and weight of the cable can increase flex life (reliability) and reduce costs. By starting with a conductor insulation that has high dielectric strength and good tear-resistance characteristics, you can use insulation thickness 50% thinner than current insulations. Standard insulations that are voltage-rated at 300V rms for 0.004-in. wall thickness are available. (One example of this type of insulation is the MIL-ENE from WL Gore & Associates.) Reducing the conductor insulation thickness decreases the overall diameter of the finished cable.

You should also consider the lay-up of typical industrial cables, which designers often ignore. Each manufacturer has a process that seldom varies. Nevertheless, lay-up is one of the most critical processes to increase the flex life of round cables. The vendor can adjust the lay of the cable, the twists per inch of the conductors, and the conductor lay to optimize cable reliability for flex applications. Cable manufacturers with proven understanding of flex cable requirements develop cable processes with optimum lay-up to maximize the cable life. If done properly, this process has a significant impact on reliability.

Reducing the diameter of the cable results in an exponential increase in flex life at a given bend radius. The cable shield is often the first component to fail. The shield is at the greatest distance from the neutral axis of the cable and, therefore, experiences the most stress. You need two elements to correct this problem. First, replace the standard braid with a double-served wire shield. As a result of many years of testing, this shield type has proven optimal for flex life and shielding effectiveness. You must isolate the shield from the conductors and the outer jacket to reduce friction in the shield wires. This friction generates heat and reduces flex life. Second, you need to use material with a low coefficient of friction (both static and dynamic) to isolate the shield from the cable. Fluoropolymer-based materials, such as GORE-TEX Expanded PTFE, have the lowest coefficient of friction of any cable material. Designers have used these materials for applications ranging from coaxial dielectric to ruggedized outer jackets on limp gimbal cables.

The cable jacket protects the shield and conductors from the environment. If you properly clamp and terminate the cable, the jacket can increase the pull strength and flex life of the cable. You should keep jacket thickness to a minimum to keep the cable diameter down. Therefore, materials with high tensile strength and tear resistance are desirable. Your cable also needs resistance to hydraulic, cutting fluid, and solvents. An excellent material for cable jackets is polyurethane because it is flame-retardant, resists most industrial fluids, and has excellent abrasion resistance.

As cycle times reduce, the weight of the cable and cable-management system become limiting factors. In these applications, you can use a ribbon power cable instead of a standard power cable. The flex life of the ribbon cable is 100 times greater than the round cable (Fig 4). The ribbon cable is one-fourth the weight of the round cable. Ribbon cables help you reduce the mass of the moving cable bundle, allowing greater acceleration, less vibration and oscillation, and decreased wear.

You can often flex and move a ribbon cable without the use of a cable chain. The ribbon cable is self-supporting and, with the proper clamps and guides, can be used in most rolling, torsional, and tick-tock (back-and-forth) flex applications. The cable can also feature mounting brackets molded onto the jacket, which results in significant cable installation labor savings.

The ribbon cable offers greater flex life, lower weight, and other advantages over round cables and cable chains in flex applications. Unfortunately, you can't easily add or replace wires in ribbon cables. For this application, you can use a cable chain. New cable chains are available that are 50 to 75% lighter than existing standard cable chains. These chains have no moving parts, and you can simply snap in round or flat cables. The segments can be riveted together in any combinations of sizes. By contrast, standard cable chains have molded links that snap together. The links rub against each other, generate particles, and make noise. The molded parts have sharp edges and mold lines that can abrade cables and tubes.

Paul Warren is a product manager at the Electronics Product Division of WL Gore & Associates, Manor, TX, where he has worked for 13 years. Warren received a BSME from the University of Texas, Austin, TX. In his spare time, he enjoys off-road motorcycling, biking, and tennis.

Tom Rosenmayer is the product development leader at the Industrial Electronics Business Group of WL Gore & Associates, Austin, TX, where he has worked for three years. Rosenmayer holds a BS in metallurgical engineering from the University of Missouri, Rolla, MO, and an MS and a PhD in materials science from Rice University, Houston. He enjoys golf and computers.

Moll, Kenneth W and David R McCarter, "Flex Life in Cables," Electronic Packaging and Production, June 1976.