Best approach to designing complex assemblies

bert_fourie

I come from a free form modelling background, trying to convert and develop a parametric mindset. My question is about the best design approach. In free form modelling I would start building my design and add parts as I go along. In parametric, I want to retain design intent so the design has integrity allowing me to make easy changes later on. I now find that the approach of building the entire assembly in one Part Studio ends up very complex and not that easy to change later on, simply because there are so many operations in one studio and it becomes confusing. Please advise.

bert_fourie

I have experienced the slowing down issue but of course did not realize why it was happening.

Edward, your suggested approach sounds sensible - I shall derive into new part studio's parts/assemblies that can be developed further on their own, and then build a model in assembly to provide an overall perspective of the design.

For me design is an intuitive process and I do not always have a firm design concept as I progress, trying this way or that way with different ideas and concepts. I need to find the way that suits this process best while retaining the advantages of parametric modelling. One of the challenges is that it is so much quicker to sketch up ideas and try them in free form modelling, therefore I tend to get a bit frustrated at the lack of tools in parametric modelling - the disadvantage of free form modelling of course only takes effect later

bert_fourie

I may add that I understand I can use multiple part studios, but that seems cumbersome to navigate while you are building a design and the endpooint is not clear yet.

Do parts across multiple studios retain relationships?

traveler_hauptman

@bert_fourie More mature parametric CAD has ways to chunk the complexity (grouping operations so that you can label and think about the group instead of all its parts) so that it is less confusing. This is missing in Onshape.

Using the  feature filter can help,

My general rule is that if the parts have a shared mechanical interface that is changing, they go in the same studio. If there is a unidirectional relationship (the dimensions of one part always depend on another) then it's a good time to start another part studio if the current one is getting complex.

bert_fourie

Thanks, understood. With time will come competence!

edward_petrillo

As a parts studio grows in complexity you will pay a price in performance as the overhead of navigating the feature tree and waiting for the model to refresh increases, not to mention uncovering any weaknesses in the graphics performance of your platform. After learning this lesson the hard way, I will be on the lookout for opportunities to partition a model into smaller sections.  One possible tactic is to choose a section of the model that can be derived into a new studio to serve as the basis for further elaboration of the model.  The resulting associativity between the studios mitigates some of the disadvantage of breaking up the model.  I've found that this approach can be more effective than joining independent parts studios into an assembly. 

A model that has many parts arrayed around a plane of symmetry poses additional challenges.  Parts can only be mirrored in a parts studio, but mirroring a complex structure doubles the size of the studio in one fell swoop.  I'd welcome any suggestions on how to manage this particular design challenge most effectively.

shashank_aarya

One approach could be to create number of sketches in single part studio at locations according to the part position in assembly. They will have all the basic profiles which will belong to all the parts required in the assembly. After that particular sketch can be derived in new part studio to create the part accordingly. This will probably maintain some parametric association with respect to parts.

bert_fourie

I have experienced the slowing down issue but of course did not realize why it was happening.

Edward, your suggested approach sounds sensible - I shall derive into new part studio's parts/assemblies that can be developed further on their own, and then build a model in assembly to provide an overall perspective of the design.

For me design is an intuitive process and I do not always have a firm design concept as I progress, trying this way or that way with different ideas and concepts. I need to find the way that suits this process best while retaining the advantages of parametric modelling. One of the challenges is that it is so much quicker to sketch up ideas and try them in free form modelling, therefore I tend to get a bit frustrated at the lack of tools in parametric modelling - the disadvantage of free form modelling of course only takes effect later