We present here a few methods that have helped us keep systems and products working, manufacturable and relevant to users. These methods start from the design phase and continue over the entire lifecycle of the system. Nobody likes to think about the end of a marriage while they’re planning their wedding, but a smart, disciplined, well-prepared businessman will insist on a prenuptial agreement anyway.
Good designers design systems that work, while great designers design systems that work and are also manufacturable for a long time. A part of this craft only comes from experience, of knowing which parts are likely to go obsolete and which parts have a long life. For example, displays and their drivers often go obsolete within a couple of years; so experienced, great designers use displays that are being used in the automobile industry.
It is important to capture ALL communications relating to the system – formal and informal This includes formal design and test documents, as well as informal emails, chats and meeting minutes. In our experience, formal documents rarely capture all the issues and informal communications provide the context in which key design decisions were taken.
Both build and test environments should be virtualized as far as possible, and the virtual machines archived. Later versions of compilers and test tools will differ from the original and may introduce variations that will be difficult to track down. If the machines are not virtual, components necessary to rebuild the system should be archived. This may include hard drive clones, compilers, test jigs and test programs.
When the design is complete, it may be advisable to hand off the design to an obsolescence management team at the same time as the design team is being given to manufacturing. The obsolescence management team can be made up of individuals from the design team or be a separate obsolescence management laboratory. In our experience, top designers want to do another design and not be tied down to maintaining the design.
It is important the obsolescence management team is able to test inexpensively and test quickly. External test laboratories and testing organizations tend to be slow and costly. In our experience, they should only be used for system acceptance testing. Earlier test phases should be carried out in-house using less expensive and more effective methods. An obsolescence management laboratory should have the ability to carry out all relevant tests relating to:
- Electromagnetic compatibility
- Electrostatic Discharge
- Temperature cycling
- Humidity
- Vibration
Proactive parts and sources monitoring should be done to ensure their availability. It is important to know where to get parts. There are specialized distributors of hard to find parts that can be used. It is also important to know how trustworthy the parts obtained from the distributors are. Some distributors pull expensive parts from discarded equipment. These parts while being indistinguishable from new parts often fail in the field.
It is also important to know which parts to use and substitute. For example, automotive grade parts often have better specifications and availability and are cheaper than military grade parts. They are also less likely to go obsolete!
Requirements obsolescence is another factor to keep in mind while doing obsolescence management. User requirements ALWAYS change. Sometimes the requirements change so rapidly that design work on the next version must begin while the current version is still being taken to manufacturing. In other cases, requirements change more slowly. This case is much harder since the design team and their knowledge is long gone when design changes need to be made.
It is important that the obsolescence management laboratory be able to carry out design changes cost effectively. Sometimes the part required is not available or the requirements change so that a design change is often the only way out. The obsolescence management laboratory should be able to carry out the necessary modifications whether they be mechanical, electronics, hardware or software.
The obsolescence management laboratory must also ensure that the modified design meets the requirements – functional, environmental, electromagnetic immunity, power budget, thermal budget, and so on.
Obsolescence management may seem like a costly and involved process. However having a centralized obsolescence management laboratory may help. A centralized obsolescence has well defined processes for checking parts for availability. Since it maintains many designs for obsolescence, the effort of checking vendors and sources for parts can be amortized over multiple products. The same design and test facilities can be used for multiple designs. The individuals at a centralized obsolescence management laboratory are more productive since they since they do the similar tasks day in day out and build up their expertise. This results in a cost savings for each design.
