May 22, 1997

Evaluation tools help you design reliable surface-mount attachments

Douglas J Leonetti, Aromat Corp

It's not enough to use reliable surface-mount components. You also have to properly solder them to the pc board. The right tools can help you understand the causes of board attachments' long-term unreliability.

Manufacturers of surface-mount-technology (SMT) components now provide devices with life expectancies greater than 20 years. However, equipment using such components, such as telecommunication gear in outdoor enclosures, often fails much sooner because of solder-attachment failures. System manufacturers have had to specify high-reliability components and properly solder them onto pc boards, only to experience returned failures with SMT components dangling from their solder pads.

OEMs such as telephone companies have reacted by demanding 20-year solder-attachment reliability as an integral factor in their IC specifications. This situation has prompted SMT-component vendors to closely examine more than just the internal structure and assembly of their devices. For example, using several design-evaluation tools, including finite-stress analysis, lead-compliance investigation, and design-for-reliability (DFR) calculations, SMT relays meet the objective of 20-year solder-attachment reliability over a ­40 to +85°C span.

Stress starts with SMT assembly

Although these design guidelines apply to SMT devices in general, this example focuses on a general-purpose polarized relay with two Form-C contacts. This SMT relay resides in a 10-pin dual-inline package, with terminals arranged in a 0.100-in. grid using L-shaped leads for solder attachment. The temperature profile for infrared assembly imposes extreme temperature demands on the tiny SMT device (Figure 1). For more than 200 sec, the component experiences 150°C, with a 30-sec period at 235°C. Spacing of only millimeters between the device and the pc board aggravates this harsh temperature stress.

The design uses as few elements as possible, enhancing the reliability of the relay device. Multifunctional-component design reduces the assembly to the permanent magnet, the coil, and the body block, all resin-molded into one unit using a low-pressure molding technique. The temperature cycling during assembly alters the pressure within the sealed structure; in turn, this pressure changes the current and voltage levels and so influences relay triggering. The component uses a high-resistant liquid-crystal polyester (LCP) to provide increased mechanical strength and to lower the component's susceptibility to deformation.

A key step to improving the long-term device-terminal-to-pc-board attachment involves finite-element analysis of terminal-stress distortion. You can calculate the percentage of distortion or expansion of the solder joint that occurs after you solder the relay to the board and expose the relay to 10 to 45°C ambient temperature changes. Increasing the lead-terminal height while staying below the 10-mm height limit for automatic board insertion significantly improves attachment reliability. A heavy concentration of stress distortion occurs directly below the terminal with short leads and is more broadly distributed below longer terminal leads (Figure 2).

Evaluate torsional stiffness

Corner-lead compliance evaluates torsional stiffness, and lead compliance of SMT device terminals is critical in the reliability of the solder-lead attachment. Major telecommunications companies and system houses have made major efforts to provide a lead-compliance-evaluation tool that quantifies lead stiffness by five independent directional-spring constants (Reference 1). With these constants, you can establish the effective flexural and torsional lead stiffness. The corner-most leads on an SMT device have the highest risk for solder-joint failure, and the diagonal-directional-spring constants for these leads measure effective lead stiffness. These spring constants are available in easily applied algebraic format, convenient for rapid computer analysis.

Device-substrate, thermal-expansion mismatch, and circuit-board warping induce a lead-foot loading system (Figure 3). Five independent spring constants characterize the lead compliance that is responsible for the major modes of lead deformation. The three forces P_x , P_y, and P_z act in the x, y, and z directions, and transmitted moments M_xz and M_yz act in the xz and yz planes. Forces P_x, P_y, and P_z activate flexural spring contacts K_x, K_y, and K_z, and moments M_xz and M_yz activate torsional spring constants K_xz and K_yz.

A separately acting axial force, P_{i(x, y, z)} causes a lead-foot displacement, i, in the i direction, and a separately acting applied moment, M_{j(j=xz, yz)}, induces a lead-foot-displacement component, O_j, in the j plane. You can determine the effective flexural spring constants by the ratio P_i/i and the effective rotational spring constant, expressed by M_j/O_j. The directional stiffness then includes both the flexural and the torsional spring constants. The flexural spring constants are

and the torsional spring constants are

g constants for J-lead, gull-lead, butt-lead, and S-bend designs are available in a format suitable for computer analysis (Reference 1).

The corner-solder connections experience the largest deformation during thermal cycling, because they are the farthest from the device's neutral center. Position 1 (Figure 5) refers to the corner lead on the long side of the surface-mount device, and Position 2 refers to the farthest lead on the short side. By combining the spring-constant data with the aspect ratio of the surface-mount component, you can derive the diagonal-lead stiffness, K_d, for the critical corner lead (Figure 4). The diagonal force, P_d, is directed at angle O, with tangent O=L/W, where L and W are the overall length and width of the component; the L/W factor is the package aspect ratio.

To derive the diagonal-lead stiffness, K_d, you use the flexural stiffness values of K_x and K_y and the lead position on the package (either Position 1 or Position 2). For Position 1, the diagonal-lead stiffness, (K_d)₁, is

Torsional forces strain the solder joint at the foot of a compliant lead.

and for Position 2, (K_d)₂ is

The diagonal-stiffness parameter significantly aids predicting surface-mount solder-attachment reliability, because this parameter controls the mechanical interaction between the component's corner leads and their solder joints during thermal cycling. It is also a key factor in applying the figure-of-merit (FM) design tool to assess the long-term interconnection reliability of leaded SMT components (Reference 3).

Design and assembly factors

Reliability experts at Bell Laboratories designed an FM DFR tool that links crucial SMT-design and -assembly factors to provide realistic estimates of attachment reliability (References 1 and 4). They derived the process from extensive data from fatigue-life modeling; large-scale, accelerated thermal cycling; and lead-compliance calculations. Other factors of the technique include package design, assembly technology, and the operational thermal environment.

The FM technique minimizes the need for time-accelerated thermal cycling, because you can quantify interconnection reliability in cumulative probability of failure per component and failure rate per component at the end of its product service life. Both the lead-compliance and FM DFR tools find extensive applications in a variety of military and commercial applications.

The factors affecting SMT-attachment reliability are mismatch between the coefficients of thermal expansion for the component and those of substrate; the operational thermal-temperature environment, including temperature ranges; cyclic frequency and power cycling; the SMT-component dimensions and lead-compliance figures; and the effective height and area of the solder joint. Other factors include empirically verified models for fatigue damage; fatigue damage per operational cycle, quantified in terms of component- and assembly-design parameters; test-to-use fatigue-life extrapolation; and Weibull statistics that relate the product-design life and the acceptable cumulative probability of failure per component.

By combining these inputs, you can produce a number of easily applied, dimensionless FM values for various stages of component and assembly development (Reference 1). The key equations for FM factors are component design/selection (comp), component-to-substrate assembly (assy), product thermal environment (env), and, most important, attachment reliability and product-design life (rel), as follows:

should exceed 20 years for typical outdoor applications.

where

• L_l=distance from the device's neutral center to the farthest solder joint,

• K_d=package diagonal-lead stiffness,

• A=maximal load-carrying area for solder joint,

• h=effective solder-joint height,

• (uppercase delta)(lowercase alpha)=absolute device-substrate coefficient of thermal-expansion mismatch,

• uppercase deltaT=cyclic temperature range,

• N=cycles of operation,

• F(N)=cumulative probability of failure per device after N operational cycles, and

• B=Weibull shape factor.

By applying various design guidelines and evaluation tools, for example, Aromat (New Providence, NJ) has developed a high-reliability TX-SMD with an FM DFR reliability figure that exceeds the 20-year attachment-reliability design goal (Figure 5).

References

1. Kotlowitz, Robert W, "Practical methods to design-in and predict surface-mount attachment reliability," seminar developed and presented at Lucent Technologies/Bell Laboratories (Whippany, NJ) (no implied endorsement by Lucent Technologies).

2. Kotlowitz, Robert W, "Compliance metrics for surface-mount component lead design," Proceedings 1990 IEEE Electronic Components and Technology Conference, pg 1054.

3. Kotlowitz, Robert W, "Comparative compliance of representative lead designs for SM components," Proceedings 1988 International Electronics Packaging Society Conference, pg 908.

4. Clech, JP, W Enelmaier, Robert W Kotlowitz, and JA Augis, "Surface mount solder attachment figure-of-merit design-for-reliability tools," Proceedings 1989 Surface Mount and Related Technologies Conference, paper SMT V-48.

| EDN Access | Feedback | Table of Contents |