Machine Design 101


Having CAD software makes you an engineer as much as having a calculator makes you an accountant: It doesn't.

Engineering can be broadly defined. Designing a part, or an entire machine, from scratch is only a narrow part of the field. It takes a specific skillset that is not common among all engineers. This skillset probably has more in common with artists, architects, and philosohers than it does with other fields within engineering. Not to imply any elitism: A very competent design engineer might suck as an application engineer or a project engineer because they lack the skills needed for those roles.

I think all skills can be learned and sharpened with practice. One's ability to effectively design solutions can be honed. Routines are good to guide you as you go, encouraging needed repetition.The following is my suggested routine for honing your design skills.

Design is an iterative process. Work through the following steps, do it again and again until you have a workable solution.


1. Define your goal.

What exactly are you trying to do? Define this as generally as possible with as few "Hows" as possible. Adding "Hows" at this point limits your creative possibilities and narrows your focus. If, for example, you say you are designing a fork truck to carry Load X from point A to point B, you have already limited yourself by defining a "How"- it's a fork truck and all that goes with it. A general goal at this point is to get Load X From point A to point B.

2. Identify your constriants.

What constriants are you limited by? Dimensions, weights, cost, environment, etc. Other factors like customer requests or regulatory requirments come into play here. This is where you may have to add some "Hows" to based on your constraints. In our example, maybe this is where it seems fitting, after reviewing your constraints, to insert the fork truck idea.

3. Layout Concepts

Now is the time to get some thoughts on paper (or a computer screen). Start with sketches and flesh out ideas as they come. Maybe you will end up with several very different concepts or several that aren't too different. It really doesn't matter. Some details will be important, but don't go crazy. Example: laying out a properly sized gearbox for an application is important, but probably not detailing the mounting of the gearbox.

4. Pick a way to go

Look at your concepts and decide which to pursue. Which are you most confident in? What are the cost comparisons? How much work will each require to manufacture? Do any meet your goal and mind your constraints better than others?

5. Detail, Detail, Detail

Once you pick the best concept, work on the details. Every part, every hole, everything. The devil is in the details. You will discover things that need to shift around as you go. Work with it.

6. Work with the end in mind

As you go, don't forget your goal and your constraints. It's easy to get engrossed in a design detail, forget what you are trying to actually do, and end up with something really cool but way too complicated for the task at hand.

7. Always ask, "What if?"

Ask a lot of questions.

Think through the build beforehand. Plan for the worst case. How will this be assembled? Can you get a wrench in there? What pieces will have to go together first? If this fails, how will it be repaired? Will any maintenance mechanics be swearing at me in the future because of this design?

Think through how the machine will be used. If you design something that can be used as a step, it better be tough enough for someone to stand on it. Where will the operator be most of the time? Will they be able to see the operator interface when they need to?

Practice makes perfect