Method 1 has the Layout sketches in the Assembly.
Method 2 has the Layout sketches in a Part, which is the first Part in the Assembly. That Part has only sketches and planes, no features.
What do you think is the best approach and why?
John -John Layne wrote:
What do you think is the best approach and why?
I'm not brave enough to use SP0's anymore so will not be using 2008 until Late 2008.Peter Yodis wrote:
I am working on a design of a four bar linkage system in SolidWorks 2008,
Pete
Jason,
'Method 3' is my favorite method.
I use it in Inventor, and it works like a charm.
I'm having problems with the custom properties imported from the inserted part in SW.
I find this especially useful when making tanks with many nozzle openings in them.
I don't like weldments for welded tanks b/c it has a number of cross-link problems, not to mention the custom properties and the final drawing BOM
Cheers ... Chris
To summarise so far:-
Method 1 Layout sketch is in the assembly.
Method 2 Layout sketch is in first part in the assembly.
Method 3 Layout sketch is in every part in the assembly.
I suppose Method 1 is what the creator intended that we use; and that the other methods were dreamed up (nightmared?) by users.
I accept that the other methods work in specific cases, but I have not so far understood a justification for preferring them to Method 1.
It seems that the killer argument against general use of Method 3 is that it prevents reusing in the assembly, parts that were created in another context, and thus do not include the specific layout sketch.
It seems that Method 2 violates the KISS principle. My miniscule experience of editing parts in the context of an assembly is that SW "shells out" to a part document. While this should work, this forum is filled with agonised cries about things that should have worked.
It is my impression that where the layout includes multiple non-intersecting profiles, and the user crudely "Fixes" (nails) each profile to document space in order to achieve Full Definition of the sketch, instead of rigorously constraining each profile to the others, then on inserting the layout part into another document, the profiles will be unconstrained and free floating because the nails remain with the original document. (I will test this if I do not receive confirmation or rebuttal)
I frequently use a variation on method 2. The layout sketches are not in the first part, they are in a skeleton subassembly. This is because my layouts are intensely related to the customer's part model, which is prone to changes. So, I can swap new part models into the skeleton subassembly without much trauma. The reason I do this in a subassembly instead of in the main assembly is that I frequently need the layout sketches in other subassemblies, and it just makes a mess if the subassemblies have to reference the layout sketches in the context of the main assembly. KISS dictates this method for me.
In the past, I have used different methods, all variations on methods 1 and 2. The bottom line is that you should be able to use whichever method is best suited to the project at hand.
There isn't one right way, but there is only one wrong way: don't understand your tools, don't think about what you're doing, and have no plan.
Sorry Derek, You have lost me.
My understanding is that references between documents are created when something is inserted, which was created, and can be edited, elsewhere. The references are between documents, not between components of documents.
You seem to be describing a sketch part within a sub assembly within a top assembly.
I have not figured out what you are trying to tell me, and I think I would be foolish to ignore you rather than admit my ignorance.
At the simplest level where you have a single part sketch in an assembly driving other parts, the relationship between it and say sketches in the other parts would be in context external sketch relations. The parts other than the driving sketch would still be mated to the main assembly. Like you say they would still have a file reference link to the assembly, but would have no file reference link to the other parts.
It is pretty abstract when you discuss it like this. When you use it on a job it is lot simpler. I have used it in the past on plant design involving conveyors, hoppers and other mining plant. The main plant layout is a driving sketch in plan. Conveyor profiles are driving sketches in elevation referencing the plant layout driving sketch by sketch relations. Conveyor headchute assemblies can contain the conveyor profile sketches to indicate conveyor angle and pulley diameter. The chute driving sketch can reference these so it updates when they change.
Personally I don't use the assembly layout sketches. The part driving sketch is a bit more clunky but the application is more powerful and general.
I have recently been using a technique that seems to be a cross between Method 2 and 3, however I am finding it difficult to implement on a new project that involves moving parts and the need to release components for manufacture before the entire assembly is designed.
That is, I design the entire system as a single multi-body part. I have three sketches to define components in the x,y,x planes. I build up all of the components. When I "finish" the design of all the components I go through and rename all of the bodies with part numbers and then use "Save Bodies" to export all of the bodies into new parts. This method seems more stable than inserting the part into new part files and deleting the other bodies as "save bodies" seems to track body references better - I don't have to go open individual files and edit the delete bodies feature to fix them if i make changes in the original part.
I, unfortunately, have three problems/issues with this method.
1. I cannot release parts to manufacture before I've finished designing the rest of the model because I have to wait until the end to split out the bodies into individual parts. I'm concerned about losing body references in the new parts by saving out bodies too soon. (Perhaps this is not an issue but my experience with body references is that in the past they can become confused as mentioned above)
2. Re-use of parts. In an assembly i would just insert a part multiple times. However when designing the multibody part, I don't necessarily want to rebuild the component since it could cause errors, so I end up having "empty" spaces in teh multi-body part. Someone looking over my shoulder tends ot freak out until the end when i can rebuild the full system as an assembly.
3. I am now working with moving parts and testing their range of motion is necessary. While I can move components in sketches, rotating sketch planes where sketches are driven with equations is computationally intensive and singularities seem to cause fairly disruptive errors.
I am now considering using Assembly Layout Sketches to drive a new system that involves moving parts. Can anyone comment on the effectiveness of layout sketches or perhaps another method in addressing the issues i mentioned above? I'm planning on experimenting on this next design round, but I am cautious of digging a couple weeks into intesive design only to find myself hitting a wall.
Thanks.
Matt Carney - If you create cofigurations in the part window and then "create assembly from part" the assembly created will always reference that config in that part. I use delete bodies and save out each part as its own assembly - if you update the base configuration and propagate the changes through the rest of the configs, all the parts in the assembly automatically update too. This works well up to about 3-500 parts and then it better to have a part with just sketches in the assembly (#1)
For most parts I hash out the base geometry on the RT plane and then using a 3d sketch I determine the widths at various point along the part, I try to only insert planes if they are driven by a sketch, they rarely fail at that point, and sketch dimensions can be set to different configs, offsetting a plane cannot be changed between configs.
-IMO- I build furniture and small industrial items.
John Layne wrote:
I have been having a discussion about Top Down design and Layout sketches/ skeletons with a colleague, wondering what the group's opinions are?
Method 1 has the Layout sketches in the Assembly.
Method 2 has the Layout sketches in a Part, which is the first Part in the Assembly. That Part has only sketches and planes, no features.
What do you think is the best approach and why?
Pros and cons:
Method 1)
Pros:
a) No extra parts
b) no extra mates.
Cons:
a) Remembering to make sure ALL driving sketches are at the top of the tree and above ALL components
b) Can't re-use the same driving sketches in other places (other assemblies) without adding a bunch of derived sketches
c) If you need to add actual solid/surfaces ( I have run onto one occasion) to drive from, this cannot be done
d) I had another issue with this method and I can't remember what it was. Something to do with something that I wanted to drive with was being forced to the bottom of the assembly hierarchy and I couldn't move it.
Method 2)
Pros:
a) Can use the driving part in multiple assemblies
b) Solids and surfaces can easily be used as drivers
Cons)
A) see pros for assemblies.
Method 1 is preferable since I am the one hunting for other people's missing parts. Its just one less part to shepherd. "But my parts never go missing" Ok, then. ***backing away slowly hands up***