FOR FREE MAGAZINE

Obsolescence: not the end of the world

Electronic Component life-cycles—the time between market introduction and formal end-of-life—are shortening. A large part of the world’s semiconductor demand is driven by consumer electronics, and this market typically has shorter and shorter product lives. Component obsolescence is affecting more companies, more regularly than ever before. Manufacturers in the automotive, industrial, medical, transport, aerospace and military markets have product lives which now far exceed those of the components which go to make them.

When long-term product availability is vital, companies need to ensure a reliable source of components is in place, even after the component is made obsolete. This means companies need to plan and manage obsolescence strategically. Failing to do so could lead to:

•Line-stops

•Unnecessarily large financial commitments tied up in last-time-buy stocks

•Long-term storage costs

•Forced product re-designs

•Premature product end-of-life (EOL) and reduced service lives

Here are the top tips to minimise the cost of obsolescence.

Obsolescence management starts at the design phase
Poor component selection in development can lead to premature product re-design and re-qualification. We have all heard of stories of products being launched with obsolete components inside. This is particularly relevant for customers with long development and qualification cycles. A new car, aircraft or industrial controller will typically spend up to five years in design and qualification. Add to this: five to seven years of production; and seven to 10 years of after-sales support and it is not uncommon to need a total of 20 years of component supply.

Choosing the right component technology and supplier can have a dramatic impact on long-term availability. Lowest cost may not be best choice for long-term supply. The first question to ask your supplier is ‘what are their commitments to long-term availability’. It is always difficult to obtain absolute guarantees of availability over the long-term. There are unforeseen natural disasters, market instability and company acquisitions: all of which cannot be predicted years in advance. But can your supplier demonstrate a controlled transition process through the EOL and into long-term authorised supply or even long-term production?

Other important questions to ask your supplier. Are the heart-beat components of your design, the software packed microcontrollers, FPGAs or ASICs comprehensively documented? Can the true design files (VHDL, Spice models, test-vectors) be retained and archived at the design phase to offer a chance of re-build if the unexpected happens? Do they contain proprietary intellectual property? If so, the ability to port such designs when/if the components are made obsolete, will be compromised or at least subject to re-licensing/royalties.

Understand the total cost of obsolescence
Does the company understand and model the cost and risks associated with obsolescence? Component obsolescence is not just a purchasing problem to be addressed as an after-thought. Does the project plan need to include anticipated product re-designs during its life? How will the capital locked down in long-term component storage be accounted for? How will component obsolescence impact on after-sales service commitments?

Plan for obsolescence and resource to management it
If your equipment has a long qualification, production or in-service life you will face component obsolescence. Customers who are surprised by component obsolescence and treat its resolution as an inconvenience to be overcome as cheaply as possible, will ultimately pay a heavy price in terms of disruption, cost and risk. The best-in-class devote skilled multi-disciplined resources to the task of obsolescence management. Preventative planning by purchasing, component engineering, design and programme management can reduce or eliminate the cost and risk. The devil is in the detail and analysis must be line-by-line. Unexpected obsolescence of a 1 cent transistor can stop a programme as easily as the obsolescence of the main microcontroller.

Pro-active monitoring of component life-cycles
Regular component monitoring allows a user to anticipate problems before they occur. There are some excellent tools such as IHS Parts Intelligence and BOM Intelligence which track a component’s life-cycle, lead-time and specification changes during its life. Such tools provide a life-cycle prediction, and alerts can be triggered when product discontinuation notices (PDNs) are issued.

Be aware of product discontinuation notices
Many component management databases can provide PDN notification services. This can be generic, where you are shown everything, or specific where you load BoM structures into the database, and it matches and highlights PDNs which affects the products. Each manufacturer has its own unique PDN format so it can be time consuming to assess and log all the part numbers affected. There are attempts to standardise formats with initiatives like SmartPDN but this will take time.

It seems bizarre but it is increasingly difficult for customers to know which PDNs affect their products. Increased system integration, and the use of embedded processing mean that sub-tier suppliers control these BoMs. A poorly managed component obsolescence in either of these areas can still trigger a forced re-design for the owner of the overall system. Will your sub-tier suppliers share their BoMs? Do your sub-tier suppliers have adequate obsolescence management processes in place?

While many CEMs offer pro-active component life-cycle management, some are completely re-active. PDN notifications are typically only aimed at the direct purchasers of the component in the last two years. Intermittent or irregular production, or low level after-sales service support, may not trigger the PDN notification. Does your CEM have an adequate obsolescence management process?

Last-time-buy: What to forecast?
Forecasting is not an exact science. It is almost guaranteed that your forecast will be wrong. If production forecasting is difficult, after-sales needs are a nightmare. Underestimate your needs and you risk prematurely killing a product and losing sales. Overestimate your needs and you will tie-up unnecessary capital in stock, whilst paying excessive storage costs. Should you plan a re-design in the future to limit your LTB? The design and re-qualification costs, plus the opportunity costs of using precious engineering resources looking backwards rather than forwards, need to be factored in.

Whilst there are few options except to place a traditional LTB order, a supplier with an established EOL transition path offers at least the hope of risk-free ongoing authorised stock and production. If demand rises, re-designs are delayed, or in-service commitments extended, then these aftermarket partners will be able to support. They provide an extra layer of security to the forecasting process.

Purchase from authorised sources
There is a common misconception that once the original manufacturer stops making a component, unauthorised/grey market sources are the only option. Nothing could be further from the truth. The zero-risk option of an authorised after-market supplier like Rochester Electronics, should always be the number one choice.

The risk of counterfeit and simple poor-quality product from unauthorised sources represents a significant risk to production yield and mean-time-between failure rates (MTBR) in the field. Inferior or substandard ‘testing’ by unauthorised third parties gives a veneer of confidence that ‘goodness’ can be tested. In truth, the testing is visual or x-ray or a poor partial copy of the original manufacturers test processes. Full tri-temp testing can rarely be offered, and the risk of commercial grade components being re-marked as industrial, automotive or military parts has never been more real.

Unauthorised component risks include:
Poor handling. Resulting in ESD damage and the destruction of the device. Externally there will be no indication that a failure has occurred poor storage. Excessive heat, cold or moisture during any part of its storage life. Leading to: external lead corrosion and failed solderability; or moisture ingress into the plastic devices and a catastrophic failure when subject to reflow temperatures.

Fake documentation.

Recovered, re-marked or re-packaged components claiming to be something else.

There are also documented quality problems related to foreign chemicals. Cleaning chemicals used to recover, wash and re-mark used components, slowly migrate into the products, shorting and corroding bond wires and pads alike. Superficial testing will not guarantee to find these faults. Recovered components may pass these tests and survive for a period in-service. However, the ultimate failures will destroy MTBR figures, and result in reduced reliability and damaged reputations.

Original component manufacturers do not provide guarantees for products purchased through unauthorised channels. Many explicitly prohibit the sale of components to unauthorised sources. Sources like Rochester Electronics receive their stock exclusively from manufacturers.

Components never leave the authorised bubble so Rochester can offer the original warranties and guarantees. Increasingly, Rochester can offer ongoing build from known-good-die and test product according to original test procedures. Rochester produced parts are current date code with no solderability risk and marked with the original manufacturer’s part number, 100 per cent compliant with the original specification. In some cases, Rochester continues to build components first made EOL by the OCM 25 years ago.

Rochester is the trusted authorised partner for most of the world’s leading semiconductor industry. All product from Rochester is 100 per cent authorised.

www.rocelec.com