Principles, not formulas
Design principles are habits of thought that tend to produce better products. They are not laws, and they sometimes conflict. A very simple design may not meet a demanding requirement, and a very modular one may cost more than a custom part. Principles help you ask better questions, but the final answer comes from the requirements, the constraints and engineering judgment. The step-by-step process in which these principles are applied is described in the engineering design process.
Keep it simple
Every part, feature and connection is something that can fail, cost money and need explaining. A simpler design often has fewer parts, fewer interfaces and fewer ways to go wrong.
- Ask of each feature, “What requirement does this serve?” If the answer is none, consider removing it.
- Prefer a single part that does two jobs to two parts joined by fasteners, where that is practical.
- Watch for complexity that exists only to work around an earlier decision.
Simple does not mean crude. A well-judged simple design often takes more thought than a complicated one. It also should not be confused with doing less than the requirements demand.
Modularity
A modular design is divided into sections with clear boundaries and defined connections between them. Think of a household appliance with a separate power supply, controller and motor assembly, or a program split into functions with clear inputs and outputs.
Benefits often include:
- Teams can work on modules in parallel.
- A faulty module can be replaced without rebuilding the product.
- Modules can be reused or upgraded.
- Testing is easier, since each module can be checked alone.
The price is that the interfaces must be defined carefully, and extra connectors or housings may add size, weight or cost. Most real problems appear at the joins, a point developed under system interfaces. Drawing block diagrams with clear system boundaries is a good way to plan modules before committing to details.
Design for manufacture and assembly
A design that works on screen but is hard to make is a poor design. Design for manufacture (DFM) means shaping the product to suit the production process, and design for assembly (DFA) means making it easy to put together. What is easy depends on the process, so involve people who know the process as early as possible.
Common ideas include:
- Reduce the number of different parts and fasteners. A product using one screw size throughout is quicker to build and easier to stock than one using ten.
- Make parts that can only be assembled the right way round, for example by using asymmetric features or keyed connectors.
- Avoid features that need special tooling when a standard one will do.
- Use generous, achievable tolerances. Tight tolerances usually raise cost, so specify them only where function requires them.
- Allow access for tools and hands. A bolt that cannot be reached with a spanner is a design error.
- Consider how parts are held during machining, moulding or welding.
Details depend on the process and material, so use guidance specific to the method you plan to use.
Maintainability and serviceability
Products need cleaning, inspection, repair and sometimes disposal. A design that anticipates this tends to cost less over its life.
- Place parts that wear or need checking where they can be reached.
- Use standard components where possible, so replacements are easy to find.
- Make fasteners accessible, and avoid permanent joints on parts that may need replacing.
- Provide clear marking, labels and documentation so the right procedure is followed. Good written procedures reduce errors during maintenance.
- Think about what a technician sees when something fails. Features that make faults visible, such as indicators or test points, save time.
This ties to the idea of the product lifecycle, where the cost of ownership includes more than the cost of building.
Margins, robustness and failing safely
Real conditions differ from the calculation. Loads vary, materials differ from the data sheet and users do unexpected things. Designing with margin and robustness means allowing for this.
- Leave a sensible gap between expected operating conditions and limits, as in derating and design margins.
- Read component limits carefully, because the absolute maximum ratings are not a target.
- Prefer designs that tolerate variation, rather than ones that work only if everything is perfect.
- Where a failure could cause harm, design so that a failure leads to a safe state where this is possible. How much margin is needed depends on the application, and in safety-related fields standards and regulations apply that vary by country and industry.
Make decisions visible
Good design is communicated. Drawings, diagrams and notes carry the intent forward to the people who will build, test and maintain the product. Recording why a choice was made helps those people avoid undoing it by accident, and decision records provide a light structure for this. Early comparison of alternatives is often supported by a decision matrix, which makes trade-offs explicit.
A simple illustration
Imagine a small electronics enclosure with a lid held by four screws. Ask a few questions in the spirit of the principles above.
- Can the number of screws be reduced to two, or replaced by snap-fit features? (simplicity, assembly)
- Are all four screws the same, and can a standard screwdriver reach them? (manufacture, maintenance)
- Does the lid fit only one way round? (error-proofing)
- Can the circuit board be removed without taking the whole product apart? (modularity, serviceability)
- What happens if someone drops it? (robustness)
The right answer to each depends on the product, its use and its cost target, but the questions are usually worth asking.
Using principles wisely
Principles work best as prompts during design reviews and in everyday decisions. Be wary of applying any single one rigidly. If two pull in different directions, return to the requirements and decide what matters most. Test your assumptions, and learn from past failures, both your own and those documented by others.
This article is general information, not professional engineering advice. Follow the standards and professional guidance that apply to your industry and location.