Decisions, Decisions

Twenty years ago, a colleague devoted many hours to developing an all-inclusive computer program that would convert a twenty-question form into finished detail drawings of a gearbox — with little or no operator intervention. He was inspired to do this after writing a fairly successful cost estimating program for custom speed reducers. This was long before the powerful computers we enjoy today, and no linkable 3-D CAD software was available. A challenging undertaking now — even more so in 1995. Not a popular activity within the engineering department — but attractive to management for its potential to reduce highly paid and, frequently, highly opinionated staff. We had some difficulty in agreeing on how to design a simple part like a bearing retainer. Coming up with a series of computer instructions for designing a fabricated steel housing was several levels of difficulty higher. The gears are actually the easiest parts of gearbox design, probably because we have detailed standards and lots of experience wading through those formulas. That is, after you decide on what type of gears to use; how many stages there should be; how to divvy up the total ratio amongst the stages; and what the face-to-diameter ratio limits should be. A major roadblock was the need to iterate designs to use off-the-shelf tooling. Getting “so close” and then having to start over can be frustrating; damn the need for integer tooth numbers! What brought the whole house of cards down was bearings. So many different types and arrangements might work but, but despite the combined efforts of international standards organizations and lots of wonderful bearing companies, the ratings don’t follow a curve-fitable pattern. Management finally ran out of patience when the program insisted upon designing gearboxes with overlapping bearings once too often. Looking back upon the experience, I see real value in developing design rules for individual components. I like having guidelines for other “proportion” decisions. To me, design can never be just right or wrong; it must fit the circumstances. What is “good” design for a steel mill might be completely wrong for a food processing plant. It takes experienced people to make those decisions — people who deserve to be fairly compensated — and celebrated — for their contributions.