Considerations in Hardware Design: The Don’ts

1) Don’t opt for a complex solution when a simple one will suffice.

There can be a temptation to overcomplicate things and opt for the “big and shiny” solution. Resist that urge. When deciding on an approach to a problem, always ask yourself if there is a simpler solution that achieves the same outcome. If the priority is to get from point A to point B, all other factors aside, a Civic can do the job as well as a Lamborghini.

A straightforward example of this might be considering the use of a DC/DC buck regulator to derive a 1.8 V rail from a 3.3 V rail. The temptation to choose a more efficient DC/DC solution over a simpler, lower efficiency linear approach is there. However, if the 1.8 V rail only requires a few milliamps in total, a linear approach may be perfectly suitable and offer lower cost and complexity.

And if you’re worried your client might think the solution too simple, don’t be. They’ll be glad the design works and that you’ve provided a cost-effective and reasonable solution to meet their needs. They’re not going to ask you to complicate things just for the sake of complicating them (or at least, they shouldn’t!).

2) Don’t get fixated on a single approach to a problem.

Being single-minded about things may have its place, but that’s rarely in electronic product design. Instead, always keep an open mind when confronting challenges, and consider multiple options for how you might address problems.

Let’s say you need to put an applications processor like the NXP i.MX 6 in your design. The product needs the performance but you only have one week to incorporate the processor into the hardware – hardly enough time to properly perform the schematic and layout work for the processor, DDR3, flash, etc. If your design isn’t as cost sensitive as it is time sensitive, consider using a premade module for the i.MX 6 instead of a custom chip-down design. This could save weeks’ worth of design and verification time, and give you a quick, low-risk solution to satisfy the requirement.

Also, don’t be afraid to gather ideas and suggestions from peers. Your colleagues can be a great resource for new approaches you haven’t considered and may have ideas or relevant experience that can save you time, money and effort. Don’t forget to ask the junior and senior team members alike – they don’t all think the same and can all offer good suggestions. A good suggestion can save days of struggle.

3) Don’t let feature creep take over the product.

No product can be all things to all people. So don’t force yours to be. In our desire to make the best possible product we can sometimes overcomplicate matters and spend time and effort including features that people might want, while losing sight of the features that people need. This ‘feature creep’ is a real danger and can lead to cost overruns and missed deadlines.

To avoid this, stay focused on the minimum viable product (MVP) and the original scope of the project you set out to design. Don’t think of all the things you “could have,” but rather on the items you “must have.” And don’t worry: the market will decide and tell you what features it wants in the next iteration of your device.

4) Don’t forget to conduct peer reviews.

Performing peer reviews is critical to developing your design. Having others review your design and provide feedback can help identify potential issues, catch design errors early in the development process and increase the chances that your product will work as expected.

The peer review process need not take long either. Schedule a few hours for your team members or your technical lead to look through your design and supporting materials and give you their initial feedback. If they need additional time, they’ll let you know. Help them out by getting them the documentation, schematics, etc. that they’ll need well in advance so they don’t have to wait around when the time comes for their review and can focus their full attention on your project.

Ensure your design is reviewed from both the high-level and the low-level perspectives. High-level review items could include things like the system architecture, power budget and interconnections. Low-level review items could include things like component symbols, footprints and I/O voltage levels.

5) Don’t forget about certifications.

Electronic products must meet a variety of regulatory certifications before going to market, and any one of these can cost you time, money, or potential redesigns to get certified, so make sure you’re giving consideration to them early in your process.

Identifying the required certifications for your product, contacting certified test labs for quotes (both for cost and time required), and identifying any required additional support for testing (e.g., prototypes, test fixtures, etc.) need to be budgeted for and scheduled in your delivery. Keep in mind that certification requirements can vary significantly between countries.

It is often a good idea to limit your product launch to one or two key markets initially, to minimize the time and money invested in certifications. This especially applies for products with wireless connectivity. Focus on getting to market first, then expanding to new markets once your idea has been validated and people are buying your product.

From relatively long development cycles to 2,500-page datasheets, from the growing importance of the Internet of Things (IoT) and cloud-based computing, the future of electronic hardware design lies in supporting increasing system complexity. With the requirements of ever higher speed and more complex circuit board designs matching or outpacing the evolution of design tools, it will be more important than ever for designers and engineers to take this growing complexity into account in all aspects of their design.

This blog was first published at Predictable Designs.