View Single Post
  #11   Spotlight this post!  
Unread 14-02-2008, 21:04
Tristan Lall's Avatar
Tristan Lall Tristan Lall is offline
Registered User
FRC #0188 (Woburn Robotics)
 
Join Date: Aug 2001
Rookie Year: 1999
Location: Toronto, ON
Posts: 2,484
Tristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond reputeTristan Lall has a reputation beyond repute
Re: How do you manage larger assemblies?

Quote:
Originally Posted by Tom Bottiglieri View Post
Abstraction is your friend. The software is designed to be parametric and its a great help to use it that way.
That actually brings up an interesting question: when should you parametrize a model, and when should you keep it separate?

I find that it's quite helpful, every time I make a feature, to try to imagine what variations on the feature might come to exist, what other features might depend on this feature, and whether or not the part will be interchanged between assemblies.

For repeated parts, tempting as it might be to link their features to another assembly member, you may encounter difficulties when the first instance of the part (from which the dependencies are derived) is changed or removed. Also, for parts that can exist in different assemblies, you might inadvertently create a dependency to an outside assembly. This can be rather painful, especially if you don't have a robust data management system in place (either software-based, or just good control over the directory tree).

Then again, for parts which can't reasonably be expected to be reused, dependencies can make it much easier to pattern features and assure yourself that geometry is aligned. It makes assemblies that much easier to work with, and simpler to maintain (because when you change the parent feature, all the children are automatically regenerated in the new state).

But to answer the original question, yes, use subassemblies! In addition to all of the reasons listed previously, you'll have advantages like reduced memory usage in an editing session, if only some of the parts need to be modified (especially a problem in Inventor, it seems), and you'll be able to more conveniently make assembly drawings.

And like others have mentioned, when complex models are needed, I tend to use Pro/E instead.