Quality Assurance Sourcing Strategies

Design discipline wards off obsolescence

Design discipline wards off obsolescence

As a fast-moving sector engaged in a strategy of continuous optimisation of design and manufacture, the electronics sector continually faces the prospect of older components and subsystems being obsoleted by newer products and removed from the supply chain.

Jean-Noël Mamousse, Design Office Manager, GAIA Converter further explores.

The replacements offered by suppliers may demonstrate similar capabilities, but the change will often need requalification in markets with demands for high reliability, safety and performance, such as military and aerospace. This process may be further complicated by changes in packaging or features that require a change in design. In the worst case, there is no replacement, demanding a larger change in design followed by requalification.

The timescales of military and defence projects can lead to obsolescence becoming a major problem. A study carried out by the US Navy in the mid-2000s on the design of sonar systems found 70% of electronic parts were considered obsolete before the system went into the field.

A 2021 survey performed by the US Defense Standardisation Program Office found a redesign triggered by an obsolescence event cost military hardware programs $1.2m on average. A complex redesign that resulted in large-scale replacement of components pushed this to more than $11.5m. The study found significant costs associated with less severe remediation processes. Even a simple component substitution could cost $14,200.

One mechanism for maintaining a supply of parts for the system’s full lifecycle is to perform a lifetime buy. In one case, the US Air Force needed to pay on the order of $2bn to secure sufficient microprocessors after a decision by the supplier to close the production facility responsible.

Component obsolescence during the lifetime of a defence programme is not inevitable. Obsolescence is frequently cause by a combination of market forces, legislative changes, and the availability of production equipment and key materials. Some of these events are difficult to predict years in advance by a design team. However, the suppliers have track records of how they react to changes. A key issue for the military and aerospace sector is that it needs advanced processors and other complex products to support its demanding requirements. These components are often designed primarily for commercial products that have far shorter design cycles and usage lifetimes but are repackaged in forms suitable for use in military systems. These products are often at risk of being withdrawn from sale earlier than expected if commercial sales are lacklustre. The supplier may choose to switch its focus to a later design that relies on a different semiconductor process and, in doing so, repurpose its manufacturing lines to make that new product.

Power systems can face similar issues. Suppliers can change focus on the types of components they provide. Manufacturers of power semiconductors, for example, are continually evaluating new processes that can deliver higher performance and lower losses. In recent years, devices based on gallium nitride and silicon carbide have become major competitors to conventional bulk silicon power transistors. These transitions can easily lead to obsolescence issues. Sometimes, the benefits of the newer processes fail to translate into market success. In others, the success of newer devices can lead to a change in focus that leads to production lines being converted from the older products. But these shifts are not universal.

There are many components that have been in production for many years because their suppliers recognise they have a viable long-term market position and also because they take care to source materials and production machinery that also have long-term guarantees of supply. This difference in approach to the market by suppliers provides the ability for design teams to manage component lifetimes successfully. It is one of the strategies adopted by GAIA to avoid obsolescence wherever possible.

Though contractors in military and aerospace programmes often use lifetime buys, it is possible to avoid the need for this in many situations. Close supplier relationships are far more important in determining how long-lasting component supply arrangements will be. This is hard for end users to achieve. However, experts in a particular domain, such as those involved in power-supply module design, can have detailed technical discussions with the design and manufacturing teams of key component suppliers. Often, during component design, especially if it is a complex part such as a pulsewidth-modulation controller, manufacturers will want to gain feedback from specialists in the field among their direct customers to tune their design. The module designer can advise them on timing offset and output behaviour that works best with the power semiconductors on which they rely.

For less complex parts, a key to long lifetime is to source products that are available from multiple suppliers. This is often achievable with discrete components, such as passives and individual transistors. This leads to another important strategy of using discrete functions where it makes sense rather than integrated parts from a single source. That more complex part is more exposed to obsolescence issues. Though it might simplify the circuit design, using discrete components instead provides a more effective way of avoiding obsolescence issues.

This approach extends to circuit topology choices. One contributor to high obsolescence risk is to have circuit designs that rely on a wide variety of components. The supply of parts on this bill of materials (BoM) is more likely to encounter end-of-life (EoL) situations because of the difficulty in determining the likely longevity of each of the parts. It is more practical to work towards a smaller BoM where discrete and other less-complex components are used multiple times within the circuit, and using a topology that favours this approach. The design team can leverage supplier relationships and knowledge of production techniques and material availability to ensure those multiple-use components will not cause obsolescence issues later. In addition, a continual tracking of industry trends helps ensure designers make the right decisions and avoid being trapped by changes in how components are made.

A less obvious strategy for minimising obsolescence risk is to focus on lifetime reliability. Designing for high mean time between failures (MTBF) is a successful strategy in this respect because it reduces the number of parts that need to be held to support repair operations. High MTBF is, as with design techniques to avoid obsolescence, a matter of design discipline: using circuits that do not push components into operating regimes where they are more likely to fail.

Conservative designers consider components that are expected to have long lifespans, or those based on emerging technologies that are known to be stable, perhaps as the result of long-life tests. Another key design approach is to use derating to extend lifetime. This involves choosing components that are rated for higher voltage, current and temperature than will be required in the target circuit. This reduction of operating stresses extends component life.

A similar approach to derating lies in temperature control. Extreme heat is one of the key reasons for early failure. Managing thermal margins is therefore a vital discipline and experienced design teams use extensive measurement of thermal performance to ensure they avoid hotspots forming in operation.

Sometimes, an EoL situation is hard to avoid, possibly because of circumstances outside a supplier’s control, such as a production facility being forced to close or a takeover changing their market strategy. Active resolution strategies help maintain a supply of components for existing products, which can include sourcing of existing stock from the distribution network if a lifetime buy option is not available. Another option is to accept the need for requalification if an alternative component causes a change in the design.

GAIA’s internal policies minimise the chances of requiring reactive strategies. The company encourages its engineers to question design choices and always plan for component availability. One example of the company’s approach to design is to perform a full obsolescence-risk analysis of each module. This includes an impact analysis and a statement of probability for each component. Such practices help embed obsolescence avoidance into the design and manufacturing process to ensure a long production lifetime for the company’s power modules. A further key to success is a focus on product innovation using core power electronics principles and pushing the boundaries in thermal management, magnetic optimisation, and power topology.

These strategies have enabled GAIA to maintain the production of its own power modules. They include the MGDS-04-J-C, created in 1996 and which is still available today in its original form as well as in a ROHS-compliant version developed in the mid-2000s. Newer products have increased efficiency, providing customers with options to improve system performance but still take advantage of a long-life guarantee. The original 4W series converter ranges provide roughly 80% efficiency. Its modern counterpart MGDD-08N, which was created in 2015, achieves 87% as well as delivering a higher power and an increased number of functions in a similar case size. However, the 4W converter has delivered reliable operation over the past 28 years, with extensive field data to confirm its reliability.

Though component obsolescence is a major concern for the military and aerospace sectors because of their long lifecycles, it is possible to manage these risks. One way to achieve a long lifetime for designed-in products is to use modules for power and other subsystems provided by suppliers with an extensive obsolescence-avoidance strategy. GAIA’s commitment to this issue ensures customers can confidently integrate power modules into their own long-lifetime product plans.