The designs that look good in the early stages often have many failures that tend to repeat themselves. I would like to know the ways in which engineers know which failure modes they can intentionally try to expose as early as possible before the design becomes costly to change.
Are there certain errors that you now intentionally test for early based on past experience (such as power sequencing assumptions, thermal margins, EMI susceptibility, connector orientation, tolerance stacking, or firmware coupling)?
More importantly, how do you set up early tests or reviews to intentionally look for these problems?
