Explanation
Bottom-up is clear: each part is independent, can be checked separately and reused in other assemblies. Standard and bought-in parts are integrated this way anyway. Parts are positioned via joints or mates. The drawback: if an interface dimension changes, the mating part has to be updated by hand.
Top-down suits parts that depend closely on each other, such as a housing around a given circuit board or a frame around fixed main dimensions. A master sketch or skeleton model contains the leading dimensions and the parts refer to it. If a main dimension changes, all parts adapt. In return many cross-references arise, which become confusing in large assemblies and slow the model down.
In practice both approaches are combined: the concept is created top-down, bought-in parts and proven components are added bottom-up. Some teams break the references at the end so that finished parts remain independent.
In practice: Design & Make – Mecanum rover, Mecanum rover full assembly and wheel assembly
The Mecanum rover is built in Autodesk Fusion using the bottom-up approach. All parts exist as separate designs and are brought together with "Insert into current design". The base, inserted first, is grounded and serves as the reference part; all other components are attached to it with joints. The wheels sit in two mirror-image sub-assemblies, each with Wheel 1 and Wheel 2 twice. The project data pose the choice between top-down and bottom-up explicitly as a planning question.
To the training project Design & Make – Mecanum rover
Common mistakes
- Designing top-down with references running every which way. The model becomes slow and unpredictable when changed.
- Designing bottom-up without matching interface dimensions between parts. The parts do not fit in the assembly.
- Not grounding any component. The assembly has no reference and moves as a whole.