04-13-2020
07:43 AM
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report
04-13-2020
07:43 AM
No downsides per se, multibody modeling is extremely powerful. But a few things to be aware of/keep in mind:
- Use master sketches and work geometry. As much as possible, build everything off of one or more master sketches and work geometry (planes, axes, points), rather than projecting geometry off of solids. I usually like to have all of my sketches and work planes at the top, then have extrusions, cuts, fillets, etc. after that. This is much, much more stable. Likewise, in assemblies, when you do need to constrain parts, place a copy of your multibody and constrain your parts to the skeleton sketches rather than constraining between parts.
- Use Sketch Blocks for structural/standard shapes. Dealing with structural shapes and standard shapes is one of the more frustrating aspects of multi-body modeling. Say you need something to rest on an I-beam. How tall is the I-beam? How wide is its flange? Usually, you would get stuff like this from the Content Center. But now you want the multibody part to be your "source" for every part. How do you get the dimension for all the structural parts you might use into your multibody part? Do you sketch them up every time? Update their dimensions every time there's a design change? This would be very tedious. The solution we've come up with that works very well is Sketch Blocks. We have our own templates for each type of shape (beam, channel, structural tube, etc) with a Sketch Block with the shape's profile. We use iLogic and an Excel table to drive their dimensions for the chosen size. So for each structural member in our design, we create a "Profile" IPT that is just used to specify the size for that member, and generate the correct profile in the sketch block. Then we derive that sketch block into the multibody part. Then we place that block wherever appropriate in the multibody, extrude it, perform any necessary cuts on it, and derive that out to produce the desired structural member. The "Profile" IPT is just used for just that -- the profile. It never ends up containing the full final geometry of the part.
- Multiple occurrences. This is probably the most frustrating aspect of multibody modeling. The great thing about multibody modeling is every part is already in its correct final position when you derive it out, so you don't need to constrain anything. Just leave it grounded and if its position changes in the multibody, it will change accordingly in the assembly. However, this completely breaks down when there are multiple occurrences of a part, whether patterned, mirrored, or arbitrarily placed. You could easily pattern, mirror, or copy and move bodies in a multi-body part, and derive those out -- and this would be fine if all you needed was for it to look right. But the big problem is that now each occurrence of a part becomes its own IPT file, with its own iProperties, and its own row on the BOM. This is unacceptable for most drawings. You could possibly give each occurrence the same Part Number and turn on Part Number row merging in the BOM, but this would be difficult to keep up with if you add new occurrences, and that also doesn't help the fact that you would have tons of extra IPT files. So, what do you do? Well, unfortunately there isn't a great solution. The best thing I've found to do is to just derive out one instance of the part (the one I modeled), and then copy and constrain it to the skeletal sketches in the assembly at each location where it's supposed to go. If it's supposed to be patterned, I'll create a dummy pattern in the multibody and associate (link) the assembly pattern to it. Another thing I've found is that often, if a part will be used more than once, it's often easier to model it at the origin in your multibody part (rather than at one of its intended locations) so that you can constrain it using its origin centerplanes in the assembly.
Sorry, that turned out to be a lot of information. I hope it's helpful though. Good luck and have fun learning the joys and frustrations of multibody modeling
and do let us know if you find any good strategies of your own!