Design intent governs –or should govern –every choice we make in CAD modeling. Design intent recognizes that the initial design phase is only a small portion of the life cycle of a product. We spend significantly more time modifying and updating models as requirements change. Therefore, we want to build additional information into our models so that changes propagate intelligently to related features and components.

Familiar Design Methodologies

When it comes to assembly design methodologies, bottom up design (BUD) and top down design (TDD) are the most popular and well-known.

In BUD, first you create individual part models, and arrange them together into lower-level assemblies. Those assemblies are placed in higher-level assemblies, and so on, until the top-level product is finished.

Bottom up has disadvantages, though. It does not work well for advanced products with large numbers of components and many levels of hierarchy. It especially doesn’t work well when parts and subassemblies have interdependencies between each other. Updating BUD assemblies tends to be tedious, overly manual, time-consuming, and error-prone. Even worse, it often results in cascading regeneration failures that frustrates the users and increases time-to-market.

In TDD, we first define our product structure – the organization of major subsystems, subassemblies, and other components—without focusing on the geometry in individual parts. We use special models to consolidate critical design information. Skeletons capture geometry that affects multiple assemblies and parts; notebooks do the same for dimensions and parameters. We communicate that design information from skeletons and notebooks to individual components. In this way, we can implement changes at the top level, and the interconnected components update in ways that we plan for and expect.

Introducing Middle Out Design

Smart, connected products introduce new challenges to our traditional approaches. Products have always contained commercial off-the-shelf (COTS) components, such as fasteners, cabling, and electrical components. When our products become smart and connected, this list expands to include:

• Sensors
• Human interfaces (tactile or voice)
• Processors
• Transmitters, receivers, and antennae
• Connection ports

We often need to locate these components first, and design the housing and/or rest of the product around them.

This particular technique –interior electronics then exterior housing –is middle-out design (MOD). This technique has already been used across many sectors, especially in products that involve the packaging of components in housings and boxes, sometimes known as line replaceable units (LRUs) in the Aerospace and Defense sector.

Which method to use when?

Some people see bottom up, middle out, and top down as separate realms, and never the “thrain” shall meet:

In the real world, though, products are designed using a combination of methodologies. The overall product may be designed using top down, but individual subassemblies may be created using bottom up. As necessary, some components and systems may be laid out using middle out, and then top down takes over. The situation actually resembles the following:

The development of smart, connected products exists at the intersection of bottom up, middle out, and top down design.

CAD Tools and techniques for SCP

With that in mind, what CAD tools and techniques support MOD in product development?

• Data sharing features that generate an envelope around existing components, to facilitate the design of housings and enclosures.
• Online libraries of 3D components for COTS and vendor parts.
• Technology that supports opening and integrating models from other software packages seamlessly without the need for conversion.
• Creation of ergonomic and/or aesthetically pleasing curvature-continuous surfaces that can reference existing geometry parametrically.

Using these techniques, and combining MOD with TDD, will improve your design process for smart, connected products.