As an electrical engineer, it is easy to see solvable hardware problems where there aren’t any. It’s the classic case of when all you have is a hammer, and everything is a nail. Sadly, however, it is rare that a hardware solution is truly the most appropriate, and when it is, there is often existing hardware that can do the job sufficiently well. So I’ve been thinking of how to decide on a solution. When to pull out the wallet for an existing solution, and when to pull out the EDA tools and get down to work. When to pull out the microprocessor and when to pull out the OpAmps and analog components, even when to pull out the big guns and go to a chip-level solution. These are the ideas I’ve come up with so far.