KISS and avoid obsolescence

21 April 2008

A basic FPGA with area available to add a replacement device

Keeping design simple can help avoid the pain of obsolescence and extend the lifecycle of a product. Complexity however, can be unavoidable, but not be at the cost of flexibility or function

It is a reality that the investment needed to launch a new product increases with every generation. It therefore follows that the simpler a product is to design and manufacture, the more cost efficient it will remain during the course of its life. It also follows that by extending its useful life, the product will become even more profitable. These simplistic economics belie the fact that complexity is often unavoidable, due to the demand for more sophisticated products with an ever extending feature set.

Predicting the future
However, if that demand is anticipated at an early stage in development and the product design is approached with the intention of extending the feature set, then as new technologies become available, complexity can be kept to a minimum without sacrificing flexibility or functionality.

Although this may seem to make obvious engineering sense, it only makes economic sense in certain markets. For instance, the consumer market once supported this model with the home PC but it is much more affordable for us (and profitable for manufacturers) to replace rather than upgrade. As a result, the PC market is now much more synchronised with other consumer devices, such as media players and mobile phones. The business model for this class of product relies on a comparatively shorter lifecycle where, thanks to demand for the latest technology, they are replaced before they are redundant.

This ‘throw away’ society may suit a handful of large multinational companies that sell their products in millions and can afford to spend on product development every year, but what about the majority of electronics companies that do not have that luxury?

Form, fit and function
Outside of the consumer market – and perhaps the telecommunications sector – it is clear that replacing technology on a biyearly basis is unacceptable. Time-to-market is an often-used term. Less frequently used, but no less important, is the phrase ‘time-inmarket’ because it is through longevity, rather than short bursts of high sales activity, that most SMEs remain profitable.

Ensuring a product has a long shelf life requires continuity of supply. However, it isn’t uncommon for manufacturers to be faced with the obsolescence of critical components in a product that has years of potential sales ahead. In these cases, even the loss of supply of a seemingly inconsequential component can trigger a potentially costly and unplanned re-design. Faced with this decision, some companies may submit to the premature retraction of a product rather than devote valuable engineering resources to replacing it. However, this could also be seen as an opportunity to upgrade a product without forfeiting its
original form, fit and function.

Directing resources here not only has the benefit of overcoming an obsolescence problem, but also of adding the latest features and extending the product’s time-in-market.

The practice of designing to avoid the impact of obsolescence and to maximise useful life is a speciality of design consultant ML Electronics, and one that can be emulated by following simple guidelines.

All important KISS
The electronics industry is swamped with acronyms but there is one that is particularly relevant; KISS (keep it simple, stupid). The simpler a product is to understand, the easier it will be to maintain and upgrade in the future.

This is important not only in combating component obsolescence, but also when key members of an engineering team with specialist knowledge leave the company. Another service offered by ML Electronics is reverse engineering a product when the engineering resource responsible for it is no longer available. This is a much more costeffective process if the original design is kept as straightforward as possible.

With a little effort towards developing a design strategy that promotes longevity, it becomes clear that the KISS approach makes sense and is easily achieved.

The first recommendation is to odularise. Through the use of daughter boards, it is much simpler to partition a design and mitigate the impact of obsolescence by offering an inherent upgrade path by design. Using daughter boards may seem to add complexity when compared to a single board approach, but it repays that by way of increased flexibility. For example, a single daughter board could be used for the main microprocessor of a particular system. The rate at which microprocessors evolve provides frequent opportunities to improve the performance of an end product by simply replacing the microprocessor. With a modular approach, this no longer requires an entire PCB to be redesigned, allowing for more regular upgrades. This is comparable to laying out a PCB with dual footprints. However, this still carries the risk of offering just two alternatives. The use of an intermediate platform, such as a daughter board, removes this risk.

Interface standards
Interfaces between modularised components are now highly standardised, but that doesn’t mean they are not subject to change. The second recommendation is to adopt a proprietary interface standard for daughter boards, so as to give more control in products that are intended for many years of service.

While it is true that ASSPs can offer amazing levels of integration and potential savings in board space and component count, they also represent a risk in terms of their availability. If a large part of a product’s functionality hinges on the availability of single ASSP, it becomes a significant risk. Therefore, it is imperative to identify components that may represent a problem and minimise their use.

Also consider using programmable or configurable devices. The cost per FPGA gate reduces as each new fabrication process ramps up to production, while their functionality increases commensurately. This means it is now commercially and technically possible to use an FPGA in place of an ASSP in any number of applications.

While the flexibility of a configurable platform is unparalleled by any other class of device, it does come at a cost; design time. The use of an FPGA demands a greater design investment than using an ASSP and this forms the fifth point; the hardware/software partition. It is easy to believe that software developed in-house is
free, but the design cost should not be under-estimated. A software configurable platform offers amazing flexibility, but it should always be weighed against the design cost.

The final point in the KISS list is contingency. By designing a ‘creative white space’, future features can be accommodated without running into power problems. It is inevitable that the markets where products enjoy a longer time in service are not the same markets where stringent regulatory approvals are enforced; such as medical equipment.

The approval and regulatory process tends to elevate the cost of design and present a barrier to entry for many companies. This has the effect of reducing competition, but also limits the number of new products entering the market, and offers a longer time-in-market for the right product. However, that shouldn’t invite complacency. Through regular upgrades or design enhancements, margins and demand can be maximised. Often the process of upgrading a particular component can also lower the overall cost of manufacture, whilst adding new features could clearly allow a supplier to command a higher selling price.

This potential increase in margin needs to be weighed against the cost of submitting a product for approval. Every product developed by ML Electronics carries a TCF (technical construction file) which can be used to minimise the cost of going through the initial approval process, as well as for subsequent product enhancements. Through the judicious use of a TCF, design decisions that counter obsolescence for components and products can be justified and recorded, which significantly eases the approval process.

Creating additional modules
Modules are traditionally defined around product functionality. An alternative in long life products, especially where high-risk single source components are identified, is to create additional modules around the critical component. These can be specified within the TCF and in doing so, it is possible to create a design specification and a means of verification should the component ever need to be replaced. This eases the process of future compliance verification with hardware that has to be modified due to the replacement of obsolete components.

There are many reasons to adopt a strategically simple approach to design. Not only does it lower the risk associated with product development, but it also offers an opportunity to extend and ensure the lifecycle of the end product.

NICK PALMER is technical director, ML Electronics

Contact Details and Archive...

Print this page | E-mail this page