They don't make them like they used to

Obsolescence is inevitable and it cannot be ignored. It is directly related to a documented part or service that is no longer available from the original provider. But it may be available from a different provider. Stock may still be available in the marketplace, it may be available under a different designation, or a new product may meet the required specification. Obsolescence is normally signalled when a provider has declared a given end-of-life period for a product or service and a notice to this effect has been issued. The obvious intention is to enable customers to react to the notice. Unfortunately those who are most concerned with the withdrawal of a product or service do not always see these warnings, or at worst, they are simply ignored or their importance goes unrecognised. Even with the benefit of the warning, it will be of little value unless the exact nature of current and future requirements is known.

Commercial off-the-shelf (COTS) products and project-specific parts such as new design tools and production processes, tend to have a much shorter life than those traditionally used. This issue of increasingly shorter lifecycles for electronic systems is mainly driven by the increasing rate of change of technology.

A major factor in this is the diminishing share of electronic component production that the industrial and military market is taking. Back in the 1960’s the defence market consumed 97% of the world semiconductor output. Today that same market accounts for less than 1% and the total output is rising at 16% annually, fuelled by the rapidly changing domestic market. Now, with the emphasis almost wholly placed on high-volume, low-cost, short- lived commercial devices,

building high-reliability and long life into a product is no longer economic for the majority of manufacturers. The option of trying to use COTS products as alternatives is therefore fraught with reliability and other problems.

MANAGING OBSOLESCENCE

Cost-effective obsolescence management is achievable only when the communication link includes all members of the supply chain – materials suppliers, software manufacturers, device manufacturers, device suppliers, original equipment manufacturers (OEMs) and users – working together in a collaborative workflow process.

Normally the obsolescence management strategies are a mixture of reactive and proactive strategies. However, an entirely reactive approach may be adopted if one or more of the following situations arise:

• The cost of planning is not easily affordable.
• The product has been procured to satisfy an operational need, has a finite duration and no further purchases are planned.
• The probability of obsolescence is very low, for example with low technology products.
• The product has a high reliability and can be supported throughout its service life from available spares.
• The system/equipment is composed of COTS computing equipment which can be supported throughout the service life of the system.
• Reliable supplier guarantees exist.

While the entirely reactive strategy implies no specific provision for obsolescence, there may be costs associated with following this option. For example, increased costs may be incurred for post-design services or for the purchase of replacement parts. An estimate of the costs involved should be included within the plan to feed into the lifetime management plan. The obsolescence strategy should be reviewed to take account of any changes in circumstances. Details of the strategy review mechanism should be stated in the plan.

The selection of the reactive strategy in the initial stages may limit the ability to apply proactive strategy in the future and may result in some items required to support the proactive strategy being unavailable at a later date. Items may be unavailable because of a lack of documentation or access to intellectual property rights (IPR).

It is important to note that the reactive obsolescence strategies have been with us for a long time. Product withdrawal notices are disregarded and the reaction comes only when a problem occurs. This worked reasonably well in the past where the pace of technological change was much slower, but incidences of obsolescence are now growing at an increasing pace and reactive solutions are tending to patch-up problems, not resolve them.

Whatever is achieved is likely to be a compromise, based on the problems involved, the way in which products may be affected, practical economics, and whether engineers and managers understand the problems and are able to seek out the answers (which may already exist either in whole or in part). Within any chosen strategy, there will be many solutions to help overcome existing obsolescence and these all need to be considered in order to establish supportable products.

Certain questions should be considered, such as:

• Is the device still available, but not to the correct specification?
• Is the device still available, but not in the right package?
• Has the device been ‘shrunk’, with some change to the specification or functionality?
• Is the device subject to a last time buy? And, if so, has the deadline been missed?
• Is the device not available in packaged form, but still available as wafer?
• Is the device not available from the manufacturer, but still existing in the supply chain?

Parts search

A parts search may be carried out either by the customer, supplier or by a specialist contractor. Methods include the use of proprietary databases, many of which are available on the Internet. It may be necessary to refer to the designers to determine compatibility or to carry out assessment work to maintain the qualification status of the equipment. In practice, the process is very time consuming and, on average, can take up to half a day per component to identify a solution and then confirm that it is viable. It may be possible to track down an obsolete part simply by inserting the part number into an Internet search engine. It is also often possible to source a product from an after-market manufacturer or an obsolescence tool provider.

In all instances, it is desirable to ensure traceability of a product to a recognised manufacturer or other sources. However, when dealing with the so-called ‘grey market’ such information may not be available and continuing under such circumstances requires a well-considered risk assessment. Subsequent validation and verification of the items may have a big impact on the overall costs of this as a solution.

Change of parameters

Search databases may identify parts which provide a limited parametric match, but which, after consultation with the design authority, may be deemed suitable for use within the applicable operational specification.

Cannibalisation

This is the process of using parts and assemblies taken from equipment within the service inventory to support other equipment. Other similar systems, which are no longer required, or are easier or less costly to replace, can provide additional spares once their duty or replacement is complete.

Die storage

Die storage is the storing of manufactured and fully tested products in part-manufactured form for later packing and test. Semiconductor die can be stored for many years without deterioration in the right environment. The cost of restoring any silicon chip design – even a simple transistor – is high in relation to the volume needs of the industrial sector, especially if an ‘old’ process has to be re-introduced. However, in certain circumstances this can be cost effective and is a common practice for custom parts.

Small volume semiconductor packaging and test

In some cases, a part may still be available as a silicon die, but not in the package variant that is required. Many specialist manufacturers exist to provide small-scale custom packaging and test to a specific requirement.

Up-screening

This involves applying specific screening tests to devices that were not applied by the manufacturer. The process is designed to add assurance that quality and reliability are suitable for a particular application but requires expert knowledge of the equipment involved.

Uprating

This requires the testing of devices to assess whether their operational performance can be stretched beyond the manufacturers’ guaranteed limits. Typically, this is employed for temperature extremes but could also applied to voltage or an environmental factor. In practice, this can be a flawed solution and any person undertaking such a task would need to be very familiar with the processes of the OEM. The majority of OEMs and suppliers would not support such an activity nor take any responsibility for any subsequent problems or failures that occur.

Partial equipment redesign

When it is impractical or impossible to procure a part which has become obsolete, it may be appropriate to invest in a redesign to procure a new design entity which gives a direct form, fit and function replacement.

Redesign (emulation)

In addition to after-market suppliers who purchase production rights (and IPR) to obsolete components, specialist contractors also exist who will design and manufacture new parts to order to replace obsolete items. The redesign may be carried out using the original specification, or from characteristics gained from an in-depth examination of a working example of the part to be replaced. This process involves significant cost and time, and can involve complex IPR issues.

Equipment replacement

When both procurement and redesign prove uneconomic, replacing complete equipment may be the only available option.

Technology transparency

This is a design methodology that depends on the specification of interfaces. The intended consequence is that any technology can be used in manufacture and support provided that the form, fit and function of complete assemblies are maintained. The concept may be extended by the use of open system architectures and standards. Technology transparency relies for effectiveness on the assumption that a component or module can be substituted provided that its interfaces are completely specified. This should be independent of the technology used within the module. Care should be taken because it is usually only when a substitution fails that the adequacy or otherwise of the interface definition is demonstrated.

Technology transparency is a concept that should be applied from the outset of a project. It is particularly appropriate for new projects and can be applied to legacy systems when they are updated or when modules are redesigned. Technology transparency is especially relevant for:

• Modular systems (a module is a discrete element of the system that performs a specific function).
• COTS items.
• Systems with a high probability of recurrent obsolescence.
• Components for specific applications. The design of a component for a specific application, such as a circuit board or an ASIC (application specific integrated circuit), may be considered for archiving as a high-level design description (for example, by using a hardware description language), to enable the component to be re-implemented at a later date in a contemporary technology. The necessary access to IPR should be acquired.

Reliability optimisation

The environmental conditions within the system (for example, temperature and vibration) should be optimised during design, commissioning and operation so as to maximise system life. An effective maintenance system needs to be installed before the start of commissioning so that any reliability problems can be identified and appropriate remedial action taken at the earliest possible stage.

Obsolescence monitoring

Obsolescence monitoring involves tracking the processes, materials and components used in design. It then involves taking action to provide alternatives when any of them approaches or reaches obsolescence, especially if that would prejudice support of the product. The appropriate action will often be some redesign and may involve planned system upgrades or lifetime buys (as described below) for all, or parts of, the system concerned. Ideally this process should be carried out by the OEM if it is still supporting the system. There are commercial organisations that collect information from manufacturers enabling the lifecycle of certain electronic parts to be predicted. Software tools exist which enable designers and production engineers to avoid the use of electronic parts that are obsolescent. These tools can provide data to assist in spares scheduling and the planning of updates. They can provide an analysis for legacy equipment to indicate the location and severity of likely obsolescence problems. However, there is no known organised commercial tracking of the technologies used in the production of mechanical and consumable materials. Obsolescence management in these areas may be achieved by identifying critical parts and consumables for independent investigation.

Obsolescence monitoring should be considered for use:

• For systems where the costs of obsolescence are high relative to the support budget.
• Where there is a single source.
• When the use of scarce skills is involved.
• Where the component performs a safety critical function.
• When there are large numbers of a particular system to be maintained in service, reducing the cost of obsolescence monitoring relative to the support budget.

The parts list supplied with a new project should make it possible to contract for continuous detailed monitoring of obsolescence. In the case of a legacy system without an LSAR (logistic support analysis record) it may be appropriate to carry out an obsolescence survey to scope the extent of likely problems such as unavailability of COTS and standard parts. Where problems are identified IPR issues may need to be addressed.

Planned system upgrades

This option involves predetermining points during the product’s life at which the design of all, or parts, of the system will be brought up to date and obsolete items replaced. These upgrades may or may not be synchronised with ‘mid-life updates’ which may enhance the requirement that the product is designed to satisfy. The system upgrade programme should take into account the need to minimise whole-life costs. Between the planned upgrades at least one of the other options for dealing with obsolescence will be needed. A lifetime buy will often be appropriate. A planned system upgrade is unlikely to be appropriate where it carries a high risk.

Planned system upgrades should be considered for use:

• For all new electronic systems.
• When the timescale for obsolescence can be accurately predicted.
• Under circumstances of rapid technological development.
• When a lifetime buy is inappropriate (for example, due to a short shelf life).

Continuous renewal by evolutionary development and insertion of technology (CREDIT) is a particular form of planned system upgrading in which the phases of research, development and use take place in parallel.

Lifetime buy

This involves purchasing the quantity of relevant parts predicted to be required for a defined period. It may cover the complete requirement for a production run and associated spares or it may cover only the items known to be at risk during support activities. A decision to undertake a lifetime buy should take account of possible upgrade slots. It will often become appropriate when a supplier makes known an intention to cease manufacture of a particular part if there is no known suitable substitute. A lifetime buy will not be appropriate when the shelf life under normal storage conditions is unsatisfactory. A lifetime buy may be considered by an individual contractor, on his or her own, or in cooperation with other contractors on the same project. Use of a lifetime buy avoids issues of IPR especially in complex components, modules or subassemblies. A lifetime buy applies equally to system modules and circuit boards as it does to components.

A lifetime buy should be considered for use:

• When there is a known or predicted obsolescence date.
• When the life expectancy of a system is short.
• When equipment is procured to satisfy an urgent operational requirement (UOR).

Cloning

The recreation of a component, circuit board or larger module from scratch is clearly possible and some companies are successfully providing this service. Cloning a complex chip such as a microprocessor is clearly an expensive task (typically $0.4-0.6 million). Cloning can be economic when sufficient quantities are involved and can save the costs involved in software rewrite.

MANAGEMENT TOOLS

A number of industry recognised obsolescence tools exist. Most tools concentrate on semiconductor obsolescence where the most dynamic changes occur. These tools only recognise industry standard items from the major players. It is normal for products from non-mainstream vendors or custom designs to be investigated manually in order to get a complete analysis. It is also worth checking whether an obsolescence tool provider has the capability of providing engineering back-up to assist clients in finding the best solution. Industry standard obsolescence tools can cost as little as a few hundred dollars to many hundreds of thousands of dollars for fully-integrated corporate-wide systems. The tools normally give weighting of obsolescence risk per part and per assembly – this enables a full product assessment to be made. Alternatives and equivalents are often included.

Ensure that you understand the responses that the tools give you and investigate the sources of information that companies use. It is not unusual to get conflicting information when using more than one tool. Other web-based component selection tools can also help and some of these offer additional aids in the location of obsolete stock and alternative products. Many of these web-based tools are free. There are, however, a growing number of obsolescence tool providers and their services will vary in quality and type. Do not be mislead by hype, cost and other factors which could lead to a situation that make things worse rather than better. Decide on your strategy, identify the kind of help you need, and then select the tool that meets (and does not exceed or drop below) your requirements.

Organisations with strong procurement or component engineering skills will be able to interpret and resolve the issues arising. Others may need sub-contractors or external help. Having a preferred supplier and good business relationships will also assist. You may still need to test the results you receive but you can prioritise the areas of risk. You should also question the environment and design criteria of your original design. It may be possible to review the scope for accepting alternative products.

Component Obsolescence Group

The software tools page within the Component Obsolescence Group (COG) website (www.cog.org.uk) contains a summary of industry recognised tools. COG is an industry-wide special-interest group set up to encourage the discussion of obsolescence issues, to share processes and solutions, to promote tool development, and to increase the profile of obsolescence management. Many different industries are represented such as the defence, aerospace, telecommunications, nuclear power, and medical sectors. Membership also includes tool providers, obsolete product resellers and interested industry organisations.

The level of membership is significant and comprises anything from small UK companies to some of the world’s biggest multinational operations. Similar bodies to COG, particularly in the USA, are also members. The ability of those involved in obsolescence to share their experiences with a wider audience is creating a broader pool of knowledge and a more powerful voice for the sector as a whole.

National Obsolescence Centre

In the UK, the National Obsolescence Centre has been created jointly by QinetiQ and COG, supported by funding from the UK government’s Department of Trade and Industry, to provide a centre of excellence for obsolescence management support for UK industry. Its primary aim is to provide low-cost obsolescence management support for small to medium sized enterprises. Its obsolescence management software tools are also fully scaleable to meet the needs of larger companies with multiple locations. And since the tools are web-based, they can also be used by organisations worldwide.

The NOC’s tools provide component lifecycle modelling of component listings incorporating life expectancy predictions, real time product change, and last time buy alerts delivered to the client’s PC in real time. The NOC’s services include a technical helpdesk providing impartial advice on alternative solutions such as die banking, component packaging, after-market suppliers, emulation, and technology insertion. NOC is also able to provide obsolescence management support on non-electronic and electrical components in areas as diverse as software, paints, protective finishes, polymers, adhesives, fuels, lubricants, and metallic and other structural materials.

QinetiQ’s alternative

Also available is an alternative software tool, QinetiQ’s ITOM (integrated technology obsolescence management), which provides all of the features of NOC’s tools, and in addition, unlimited client defined system hierarchy levels to enable proactive obsolescence management from platform and system, down to PCB (printed circuit board) and component level. Organisations requiring total flexibility, including turnkey solutions, are also provided for.

A CULTURAL CHANGE

There is much work being undertaken throughout industry to counter the concerns over obsolescence. Customers are expecting support strategies to mitigate against obsolescence and want to see evidence of the process in place. Resolution of issues is certainly helped by the tools that are available, but the real key to progress is the individual engineers and their familiarity with technology developments and the support available from the marketplace. Each person’s personal network of acquaintances will be a major help in overcoming problem areas. Special interest groups and the will to openly communicate can also make a big difference.

The management of obsolescence will need to become as standard a practice as safety assessment, assessing fitness for purpose, or quality assurance. Consequently good obsolescence management linked to product development may well begin to differentiate the capable from the average design team. Failure to recognise and manage the obsolescence risks can severely affect a product’s, or even a company’s, success. For long-life products where customers want the lowest lifetime costs, any calculation omitting an obsolescence element is unlikely to achieve this.

A cultural change is needed, within the industrial sector and the business world as a whole, to accommodate obsolescence management. This will be helped by the fact that physics-of-failure techniques have shown that failure mechanisms are far from random and it is possible to predict failure times with a high degree of accuracy.

Consequently, the management of component obsolescence will become a much more sophisticated process than would otherwise have been possible. This cultural change is gaining momentum although there are many legacy products for which significant challenges are yet to be overcome and in many industries, latent component technology concerns have yet to be properly tackled.

In the near future, tools, processes and the engineering/industrial culture will accommodate obsolescence management as normal practice within design reviews. Obsolescence will then be accepted in the same way as environmental legislation or electromagnetic compatibility testing has already been accepted. Obsolescence management is achievable and, despite the many risks it poses, we should not fear it and we certainly should not ignore it.

Author Info:

Taken from a paper given at NEI’s seventh meeting on ‘Plant Life Management and Plant Licence Extension in Nuclear Facilities’ (PLIM + PLEX) conference, held on 13-14 October 2003 in New Orleans, Louisiana, USA. Colin Fisher (Electrical and Instrumentation Project Sponsor) and Robert Wagstaff (Project Sponsor), Thorp Technical Dept., British Nuclear Group, B582-1FS, Sellafield, Cumbria CA20 1PG, UK