Reliability engineering helps compare the final end-use conditions of surface-mount-technology terminal blocks and through-hole-reflow technology terminal blocks.
By Michel Hodak
The quality and reliability fields define reliability as “the duration or probability of failure-free performance under stated conditions,” or “the probability that an item can perform its intended function for a specified interval under stated conditions.” Reliability is often the determining factor when customer satisfaction is objectively evaluated. The total cost of a purchased product is the purchase price plus its process failure costs and its in-field failure costs at the end-user. In other words, you get what you pay for.
In reliability, we often talk about failure with respect to time or cycles of use. Time or cycles to failure, failure probability, and reliability are important numbers to evaluate with respect to field environmental conditions, materials used, human factors, customer requirements, and health and safety issues. Some telecom industries have categorized required life expectancies of five, ten, and 25 years, depending on applications and field-service conditions. Heating, ventilation, and air-conditioning (HVAC); military; security; medical; industrial control; and appliance applications have their own criteria.
The type of reliability testing to administer depends on the product. A depluggable terminal block inserted onto a genuine surface-mounttechnology (SMT) pin strip achieves reliability by assuring robust component lead coplanarity throughout the manufacturing, transportation, and assembly processes (see Fig. 1). A pick-and-placeable through-hole-reflow (THR) pin strip of the hang-through type can mate with a variety of plugs (see Fig. 2). Another SMT terminal block obtains its infield reliability through two SMT anchoring elements at each end and an elevator-style design that protects the leads from screw-tightening torques and forces (see Fig. 3). An SMT header design with a removable pick surface gets its reliability from its adaptability to non-planar PCBs and elimination of coefficient of thermal expansion mismatch fatigue-induced failure.
This complex question remains: if we must assure in-field reliability of five years, must we test for five years? The answer is a combination of yes, no, and in between. Applications that pertain to human life have enormous total system costs, are located in an inaccessible location, hold reputations at stake, or involve years of preparation would require components that have been extensively tested under real-life conditions. Such a customer would require historical data and records showing that the processes (people, methods, materials, designs, and equipment) are qualified, proven, and validated. Naturally, extensive audits and quality, reliability, and system testing would also be imposed. Many commercial or consumer applications need not go that far. Most newly created or designed products are similar to other products and use tooling, materials, and personnel that have done these things before in other similar applications.
Accelerated testing
A great deal can be inferred from the past. A field called accelerated aging, accelerated testing, or (highly) accelerated life testing enables parts to be qualified in case historical data are incomplete, or schedule constraints do not allow enough time to do tests lasting many years. During accelerated tests, applied elevated stresses produce in a short period of time the same failure mechanisms that would be observed under normal conditions. The objective is to identify these failure mechanisms and eliminate them as a cause of failure during the useful life of the product.
One challenging aspect of this field is establishing agreement between the supplier and the customer about what the accelerated tests infer about reliability in real-life conditions. For example, tests can simulate daily temperature cycles by cycling every hour or so, but not by cycling every second. The test could involve higher temperature ranges to imitate the severity of the cycle, but not over or under some specific material transition temperatures. Arrhenius-type equations are often useful to estimate failure rates versus temperatures when chemical and diffusion mechanisms are involved in materials or electronic components. Acceleration factors are calculable based on the ratios of these rates. A rule of thumb is that reaction rates double for every 10-K increase in temperature.
To prevent product failure, designers must always consider requirements including criticality, accessibility, maintainability, and availability. Failures in applications involving systems that have traditionally been accepted as requiring frequent maintenance, adjustments, or seasonal upkeep are viewed differently than systems that require the highest levels of availability, uptime, and reliability.
Defects can be classified by degree of seriousness. Defects do not cause the same degree of customer dissatisfaction, nor do they have the same consequences on product fitness for use and safety. For example, a critical defect will surely cause personal injury or illness and render the product totally unfit for use. In addition, a critical defect will cause a high degree of dissatisfaction in the value of the product, causing the company to lose customers and suffer losses greater than the value of the product.
From a statistical point of view, reliability measurements often follow a Gaussian or bell-curve distribution. In the evaluation of clearance, interference fits, or simple stress versus strength, the mathematics are quite similar. Stress and strength in these cases can sometimes be temperature, force, torque, voltage, or another physical characteristic. For example, the strength of a material can be measured on a number of samples, from which a mean (average), a standard deviation (sigma), and a variance (sigma squared) can be calculated. Similarly, the stress imposed on the system has its mean, standard deviation, and variance. We can evaluate the stress-strength probabilities of pass, failure, clearance, or interference. In calculating probabilities of interference or clearance, different suppliers can be compared for a fit into an assembly. Supplier “A,” with two distinct normal curves, would fit 100% of the time with a probability of failure of 1 part per million (see Fig. 4). This can be compared to other suppliers who may have a higher inherent variation in their process or may be off-center from the targeted values. The normal curves would then be close together, overlap more, and show probabilities of failures of 1%, 10%, 50%, or higher.
 FIGURE 4. Supplier “A” would fit 100% of the time with a probability of failure of 1 part per million (PPM). The probability of the required clearance fit is adequate. |
The same technique can be used to evaluate connector strength versus installation forces, resistance to temperature versus environmental conditions, and other pertinent characteristics. From this type of information the yearly cost of failures can be calculated as follows:
Estimated annual volume × failure probability × failure cost = annualized failure cost.
End-use conditions
The end-use conditions are crucial because they have a profound impact on the material requirements and the overall costs of the terminal block, the components and materials used in the circuit-board assembly, and its enclosures. For example, designers must choose the thermal, electrical, chemical, and mechanical properties of the materials (solder joints, leads, moldings, encapsulants) to survive the imposed thermal, electrical, chemical, and mechanical stresses. The evolution or deterioration of these thermal, electrical, chemical, and mechanical properties with time, age, or cycles must not cause early failures or unduly high maintenance costs and downtime. The materials must stay reliable.
The designer must understand and plan for the magnitude, extremes, ranges, means, cyclic properties, and statistical distribution of these thermal, electrical, chemical, and mechanical stresses. For example, a slow change in temperature is different than a quick change and repeated changes are different than a one-time change. The compounding effects of these thermal, electrical, chemical, and mechanical stresses can hide, accelerate, or change failure modes and mechanisms in difficult-to-predict ways.
The application’s location, such as interior, exterior, tropical, marine, stationary, mobile (in a vehicle passenger or engine compartment), or within an enclosure, can be used to quantify or categorize the environmental conditions, such as exposure to water, salt, dust, lightning, vibration, and pollution. Temperature and humidity extremes can affect material flexibility, ductility, and hardness, and as a consequence, affect durability.
Designers must determine beforehand the exposure to contaminants, dust (conductive or non-conductive), pollution level, and corrosive gases and liquids. Certain metals and resins can handle certain conditions well while others quickly fail.
Exposure to vibration, impact, penetration, voltage spikes, and current overload determine the size, shape, and type of materials used. The local temperature within the enclosure during use and proximity to heat sources as electrical circuits cycle on and off is much more important than many people realize and can be the source of unexpected failures, even after testing.
Derating principles
As environmental conditions become more severe, engineers should derate the components used. Derating should consider human factors involving field wiring and the torques, forces, and abuse during installation and maintenance. These affect the type of wire clamping in different ways. Are the wires held via screw tightening, spring forces, insulation displacement, or depluggable two-piece type connection? Does the enclosure protect the general public from shock hazards? The local laws and regulations may require UL, CSA, EN, IEC, CE, or other product certifications. These likely involve flammability, creepage, clearance, dielectric breakdown, impulse withstand, shock hazard testing, and performance to a standard.
The customers themselves have explicit and implicit requirements. The public expects that electronic devices and electrical equipment will not cause fires, electrocutions, or injury. Thus, failure modes that are dangerous must be prevented and due care must be taken by the manufacturer to assure this. Product liability is defined as the liability of the seller or manufacturer for damages caused by its allegedly defective product. Lawsuits or legal actions can be based on the legal theories of negligence, strict liability, fraud, misrepresentation, warranty, and so on.
The intangibles of customer loyalty, customer retention, and customer satisfaction must take into consideration the degree of dissatisfaction caused by the field failure of a product. Everyone wants quality, reliability, safety, and the best price possible. A low purchase price may, in fact, be costlier than a higher price when all costs are accounted for.
MICHEL HODAK is quality manager for WECO Electrical Connectors, 19900 Clark Graham, Montréal, Québec, Canada H9X 3R8. Tel: (800) 724-2928; Fax: (514) 694-0956, email: mhodak@weco.ca.
