Moving assembly Top-Down design best practices?

[marcrichard]

Hi all,

After many searches on the web, i can't find a good technique for Top-Down modeling of assembly. I want to create all the parts in context of the assembly (so an update to one part is propagated to other parts) but i would like after that to be able placing mates between the parts to check the assembly motion. When parts are created in-context, SolidWorks automatically places an In-Place mate which cannot be supressed without having big troubles. The only technique i use is to create another assembly and insert all the parts created in context and mate them together so i have 2 assemblies: 1 "master" assembly with all the parts fixed and 1 assembly for motion. Is that a correct technique? How do people do in industry?

Thanks,
Kind Regards,
Marc

 

[ctopher]

I model approx 90% from bottom up due to not many other users I work with understand (or want to) the top down approach.
I create the assemblies and edit in-context when needed.

Chris, CSWA
SolidWorks 14
SolidWorks Legion

 

[Timelord]

I've done some top down and the way I handle it is as follows: Leave the in-place mat alone until you have finished creating the part. Then just delete the in-place mate and add your own back. Be aware that if the part is not totally restrained relative to other parts referenced during creation, things can go wonky if the part moves.

Timelord

 

[TheTick]

When I model top-down I usually have two assemblies: one for layout and context, and one for final product.

The "layout and context" assembly is the master, with any and all in-context features. Sometimes it has duplicates of parts or subassemblies as required for motion or alternate configurations. It may also have additional components needed to define or document the design process. It is NOT the final product, but it is the master document for documenting the design process. It is used for design context, but not to document final product assemblies and BOMs.

The "final product" assembly is just that: the final product. It is a bottom-up assembly of parts created in the "layout and context" assembly. It is used for final product drawing and BOM.

I find this scheme has a number of advantages. Design context is not usually compatible with product documentation. Also, many times there are multiple products emerging from a single design context (e.g. a dispenser and multiple refills). Doing a final rebuild of a product from the bottom-up also serves as a good checking tool.

Treat the "design and context" layout the same way one would treat the negatives for making photos or the tools and fixtures for making parts.

 

[SBaugh]

I create all the parts first, then I make an assembly and try to determine all planes I would need to control the assembly.Then I make an Design Table (DT) within the assembly and control each of those planes within.

Now I place each part into the assembly using the mates to locate, then I in-context the sketch and features to the planes in the assembly.

Now that that you have a DT controlling the planes and the parts are in-contexted to the planes you can move the planes to control the parts.

I would recommend taking the extra time to create an interface within Excel to make the control of the assembly much easier. Something like height width depth or if you have configurations in your parts then you can use Vlookup's to control configurations, etc...

You might also use equations as another part of control over parts\assy.

In-contexting can get confusing very quickly so take your time and pay close attention to what your doing and why your doing it otherwise once you get a circular reference you will start fighting the software. My first attempt years ago was a 9 month long flop. The next job too 2 months and cut the design time down to just a few hours versus several weeks in AutoCAD.

I also made all the drawings for all the parts as well so when the assembly was made to desired specs the drawings just needed to be cleaned up and they were ready for print and Manufacturing.

Hope this helps,

Scott Baugh, CSWP
Gryphon Environmental

http://www.2gryphon.com

http://www.scottjbaugh.com

"If it's not broke, Don't fix it!"

faq731-376

 

[drawoh]

marcrichard,

I do top down modelling as much as possible.

Usually, I create the first part, then I build the assembly around it.

I create sketches and pseudo-models to define the system geometry. The sketches in particular give you the chance to use design tables and the equation editor. I use these to locate main system components.

I create the subsequent parts and I insert them into the assembly, and I use the in-context links to make everything fit. I start my fabrication drawings immediately. This gives me one more visualization tool, and I can apply critical dimensions and tolerances. The sooner you realize your part cannot be fabricated, the better.

Anticipate how you will modify your model. You cannot easily model for all eventualities. If you do this wrong, everything will break on you.

When you are done, localize all the in-context features on your fabricated parts. An in-context fabrication drawing can revise itself depending on what else you have loaded when you load it to print. Manufacturing cannot cope with this.



--
JHG

 

[TheTick]

One more thing...

When designing moving parts, I do not use in-context relations on a moving instance of that part. All in-context relationships are on an instance that is fixed or fully mated. For movement, I add a second (or more) instance to go through the movement. That way, I can check movement without changing any in-context relationships.

A simple example is this hinge design in

http://esoxrepublic.com/models/,

model "Hinge.zip". The moving leaf is modelled in a fixed, closed position (as an envelope component in this case). The movable instance is separate, and has no in-context relations.

 

[marcrichard]

Thank you very much all for your answers!

TheTick, i use the same methodology than yours. Actually I wonder what is the pros and cons of using that technique (one " in-context master" assembly and one "final product" assembly) VS the multibody modeling technique (build the assembly first as multiple solid bodies in one master part, then make each solid body a part of an assembly). An example of this technique can be seen in this video: Go to module 3 (in the menu on the right) and see the video at 13:49 (link).

I have found information about the multibody technique from Inventor tutorials, it seems we can do the same in Solidworks. I understand well the multibody approach when you are designing a product in which parts share the same shape (e.g. mouse example or this other video), in the other cases i am not sure.

Kind Regards,
Marc

 

[marcrichard]

Any other advice?

 

[marcrichard]

Anyone using multibodies parts for creating mechanisms?

 

[PatCouture]

I've used in the past what a call the "Skeleton" part. The mechanism would be first designed in a 2D sketch in a part and we would have all the defining dimensions in this part using sketches, planes, axes and any references Tools you can think of. In every sub-assembly we would insert this "Skeleton" part first, then build the other parts using in-context relations with the "skeleton". The advantage I like about this is that you can have many sub-assemblies that are related from the "Skeleton" but each have their unique in-context relation so it's easier to manage. Also when you need to mate the master assembly, you can either use fixed mates from the "skeleton" or use different mates for the parts to move without having to screw around with the driving dimensions of the "Skeleton". The downside is that you may have a couple level of parent/child Relationship that may be hard to fully update after some changes.

Patrick

 

[smcdesigner12]

I use Top Down assembly sketches and in context mates, for our start assembly file, to drive all the sizes of components which will be custom machined and have created components libraries (with configurations of all the sizes, driven by a design table) of purchased components. (We build a lot of similar assemblies) The purchased components are mated to planes (bottom-up assembly) and points and have pocket cut sketches. If there are more than one of the same components on the same plane, a sketch driven pattern is used. If I need a motion study, I will use the new exploded view feature under the configurations tab, to create any movement needed, then I can open the motion study, and click on the animation wizard and bring in the new exploded view, just created and fine tune the motion study. I have made videos from the motion study to show our end customers, how a complex assembly will sequence, and to verify movement and clearances.

Phil M
SMC Ltd
SolidWorks 2015 SP2.1

 

[marcrichard]

Thank you Patrick and Phil for your feedback!

Kind Regards,
Marc

 

[drawoh]

marcrichard,

Parts that I design and that have to be fabricated typically are not suitable for multi-bodies. I use SolidWorks to do silkscreen ready panel artwork. It is possible to do this without multi-bodies, but a major pain. When I need to model a coil spring, I model a set of hoops, rather than a helix. Helices slow down CAD updates.

I cannot remember modelling an actuator with pseudo-parts as a multi-body, but probably, I have. All sorts of problems can be solved this way. Normally, I delete all such pseudo-parts before registering my drawings and models in PDM.

--
JHG

 

[marcrichard]

Thank you drawoh for sharing your experience.

When I model top-down I usually have two assemblies: one for layout and context, and one for final product.

The "layout and context" assembly is the master, with any and all in-context features. Sometimes it has duplicates of parts or subassemblies as required for motion or alternate configurations. It may also have additional components needed to define or document the design process. It is NOT the final product, but it is the master document for documenting the design process. It is used for design context, but not to document final product assemblies and BOMs.

The "final product" assembly is just that: the final product. It is a bottom-up assembly of parts created in the "layout and context" assembly. It is used for final product drawing and BOM.

TheTick, do you use SolidWorks professionally? If yes, do you use this modeling approach in your professional projects? If I understand well, you use the same modeling approach than me, I want to make sure it is an acceptable practice in industry (to validate my learning on SW).

Any other engineers/designers using this technique?

Thanks. (Sorry for my imperfect English)
Best Regards,
Marc

 

[TheTick]

Yes, I use this technique in a professional setting, and I often recommend that others do the same. This arrangement evolved over many years of design in many different settings.

 

[marcrichard]

Thank you TheTick for these precisions. In what industry are you working if i can ask?

Marc

 

[TheTick]

Currently I'm a consulting engineer for a product design firm in San Jose.

I started making assemblies like this when I worked for an automotive supplier that made custom hinges. Obviously, all those designs are articulated. There was always shifting input from customers, and it was important to find a way to keep context while being agile enough to respond to changes.

Since then, I found this technique works quite well for nearly everything. It may seem redundant to have a layout AND a final product assembly, but it saves much more heartache than it causes.

 

[marcrichard]

Thanks for your answer.

Marc