CAD/CAM discussion forum > 3D CAD/CAM > Flexible Subassemblies

Flexible Subassemblies

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 1 of 18

 Flexible Subassemblies
10-12-2009 02:25 . am | View his/her posts only
Hi Team,
Here is a link that explains Flexible Subassemblies.

Ideally I'd like to make a subassembly like a shock absorber or hydraulic ram etc., which has one or more flexible Alignments/Mates. Then when I insert the shock absorber into a larger assembly of other sub assemblies I want it to move within it's range inconjunction with the mounting parts it is aligned to.
Another example, same components, is the need for the subassembly to be flexible when the two rod ends are aligned to the parent part mounting points.

How is this done In VX?
Cheers

Rank: 1

cutter

Newbie

posts: 56

Registered: 2011-11-23

Message 2 of 18

10-12-2009 09:40 . am | View his/her posts only
The biggest secret here is proper constraints so that parts that need to rotate or move do so in the proper axis but stay in place relative to the assembly. In the assembly toolbar second icon in from the left has the command for stop at interference. There is no easy explanation here other than build something and start trying as it is not a topic easily answered by reading a post or help. When and not if you run into trouble but you have something to work with call Mike and he will get you pointed in the right direction. This can be complicated and to this day I still have an assembly I can't stop at colision and I dont know why. But I can at least move the assembly properly and freeze it where ever I want and have a close look at where parts are.

Rank: 1

Mike

Newbie

posts: 0

Registered: 2002-8-28

Message 3 of 18

13-12-2009 05:24 . pm | View his/her posts only
Thank you for your question and thank you for your faith in my direction capabilities.
The Merge Shape command basically moves the shape geometry in a component up to the active part. Well it also moves component in a sub-assembly up to the active part. It retains the assembly alignments to you now have a Flexible Assembly. See the second paragraph in the HELP link below.


On-Line HELP for Merge Shape command (Shape toolbar)

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 4 of 18

13-12-2009 07:30 . pm | View his/her posts only
Hi Mike,
looks like we are on the right track in concept at least.
Execution is proving to be another matter.

In the attached simple file (v13.75) - two working damper sub-assys are added to a swinging arm assy.

I am unable to get them to behave according to their respective subassembly alignments.
In the file I have removed my various efforts to Merge.
The suppressed alignment of the floating ends of the damper are present to show the alignment I am trying to create although I hope this is obvious.
It is unclear to me as to when I should insert the Merge command as nothing I have tried produces meaningful results.

I appreciate a corrected file with the dampers operating within their assigned range whilst connected to the swing arm assy. Please do this in V13.

Happy for this example to be added to the Wiki.


Cheers

Rank: 1

cutter

Newbie

posts: 56

Registered: 2011-11-23

Message 5 of 18

14-12-2009 05:41 . pm | View his/her posts only
Hi Paul,
Had a look today at your assembly and I have a suggestion. Your cylinders need to be more realistic as in the internal mechanism should reflect actual parts and not just a rod in a hole. When you have the internal plunger mechanism in place you can determine true travel ranges and use the surfaces on both sides of the plunger as your contact points for hard stops against cylinder bore inside ends in range of motion setups.

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 6 of 18

14-12-2009 10:05 . pm | View his/her posts only
Hi Dave,
IMO no more 'realism' is required.
The 'range' option coincident alignment between the bottom of the bore and the end of the rod is already set to control the max and min travel of the rod which works well in the sub assembly proper. It is a simple edit job to change the range. You should check out how this is done. It could save a bunch of modelling and speed up some jobs.
This way the software does not have to calculate interference, just limit the motion. I think this is computationally much lighter(quicker) than interference checking. I am ready to be advised otherwise if this is the case.
Cheers

Rank: 1

cutter

Newbie

posts: 56

Registered: 2011-11-23

Message 7 of 18

15-12-2009 04:50 . am | View his/her posts only
Hi Paul,

I will dissagree with you on this one. Complete modeling of all relative components gives you accurate data for clearances and travel. The attached is a device I just built again for a customer and because I did complete modeling of everything I did not have to tweak one part once to fit and I know the flapper in the chute at the bottom edge will clear all hopper sides with appx .25" clearance. I know that the range of travel through an assigned arc of movement by the typical operator will result in a device that works and complete clearance checks all the way around. Perhaps a down and dirty initial check could be done the way you have to save time on very simple parts but when you need to really know it must be modeled. I have seen no exception to this rule with companies that do MCAD for machine builds. If you will note in the YouTube video they have used stock parts that are acurate to ID and OD dimensions, such as the bearings for example, and have incorporated real parts in all places. For good data out good data must go in and thats why they did all parts for that conveyor. This also sets them up with a BOM and lets them do FEA to if they wish. Granted these are not concerns of yours right now but it is a big benefit on down the road if you start doing MCAD work. Talking to guys that do quite involved machine design I am told that typical parts found online like the bearing housings are quite often modeled completely in house so that the actual bearing movements, such as swiveling a bearing in a cast iron housing, can be added to models. And it has to be that way or else they have not truly modeled the part. SE claims to allow assemblies up to 100,000 parts. I dont know what VX claims but there new assembly stuff should reall help out here. Alibre is somewhere aroung 4,000 parts I think. And it takes computers with big cajonges to do big assemblies. No shortcuts here.

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 8 of 18

15-12-2009 02:17 . pm | View his/her posts only
Hi Dave,
I agree with you in regard complete modelling when it is required. I have failed to complete model on previous occasions and discovered the error of my ways

However, in regards the damper, (and other 'stock' parts), the internals are generally irrelevant. All the 'external' assembly form and function information can be provided with a representation that mimics the real part in geometry and motion capability. The skinnier the 'stock' part is, providing it meets the objectives, the better.
The live issue here is "How does VX allow the subassembly alignment information to be added to the assembly?
In essence a subass'y contains two classes of alignment info. The alignments that hold the subass'y together and the alignments that control its freedoms of movement. Not all subass'ys have a movement capability so it makes no sense to merge their alignment info.

The merge command seems the right one but I have not managed to make it happen.
I'd like to access higher knowledge on this matter - anyone....???

Re your ass'y, did you model the frame as individual components and weldments or as one part . If you did it as one part you have participated in the skinny modeling regime. I suspect you model the whole, then dimension a drawing showing the individual component sizes and locations then interpret this information on the shop floor. This is a short cut yet produces a good outcome so is good practice. Less is more.

Cheers

Rank: 1

Robert

Newbie

posts: 0

Registered: 2011-11-22

Message 9 of 18

15-12-2009 03:38 . pm | View his/her posts only
Hey y'all,

Seems there may be a bug in Merge Component that prevents the sub-assembly constraints from merging.

We're looking at it now...

bobf

Rank: 1

cutter

Newbie

posts: 56

Registered: 2011-11-23

Message 10 of 18

15-12-2009 03:40 . pm | View his/her posts only
Hi Paul,
Not sure of what you are really trying to accomplish with the merge command so I will have to ignore it. There are some shortcuts and you have mentioned one of them as the frame was a sketch extrude and corner round. I find for my purposes here since I am building it that a drawing of the complete frame with dimensions is the way to go and is far easier than pieces put together as an assembly. I don't have a BOM for instance on this as shop drawings are far more relevant for me and go with me to the shop for fabrication. The rest of the components are modeled and then assembled as seperate pieces using common sketches for hole patterns for instance.

Rank: 1

Mike

Newbie

posts: 0

Registered: 2002-8-28

Message 11 of 18

16-12-2009 08:52 . am | View his/her posts only
Let me see if I can summarize. When you Insert a Sub-assembly as a component in a part (we will call it the Top Level Assembly) the component is rigid, same as SW. In order to make it "Flexible" we use the Merge command (V14) or the Move up one level (V13). If you inquire on the components now you will have the piece parts not the sub-assembly and the alignments move up also. The problem we just found is that ALL the alignments move up including the "Anchor" constraints. So components that were anchored in the sub-assembly will still be anchored at the top level. Obviously, that is not the behavior we want and I will see that it gets fixed. In the mean time, you will have to build new sub-assemblies without the anchor.
See the SwingFrameAssy-Mike in the attached file. I removed the anchor constraints from the subassembly. This is not a perfect fix because it is hard to test the motion of the sub-assembly without an anchor so we will need the bug fixed.

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 12 of 18

16-12-2009 11:52 . am | View his/her posts only
Hi Mike,
you have it.
In reality not all the alignments have to move up in a true subassembly. It is a subassembly after all.
IMO the only alignments that need to move up are those that allow motion (degrees of freedom).

If you wanted to change the other alignments then this should be done in the subassembly.

Making all the alignments move into the upper assembly defeats the subassembly role... except if one is using it as a shortcut for modelling rather than a subassembly which would seem to be a bad reason for doing it.
Glad to see it will be fixed 'cos it's been an annoying problem for some time. I thought it was just me not understanding it.
Cheers

Rank: 1

Mike

Newbie

posts: 0

Registered: 2002-8-28

Message 13 of 18

16-12-2009 01:00 . pm | View his/her posts only
That is why we are going to drop the anchor constraint. All the other constraints eliminate degrees of freedom. The range values worked like a champ.

If this design is not proprietary I would like to use it for a training lesson???

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 14 of 18

17-12-2009 12:16 . am | View his/her posts only
Hi Mike,
very happy for the parts to be used for Help, Training, Wiki etc. The internals are proprietary!!!!!!!
I have no idea if the dimensions/range add up to anything useful as I was going to see what happened in an assy! You may want to tweak the range values and even the way it has been made - it's an experimental modelling method... just for the example. Simple as possible.

OK, you'll need to explain how you managed to get it to work. Preferable in V13 terms for now. Another good Wiki element...

Cheers

Rank: 1

Mike

Newbie

posts: 0

Registered: 2002-8-28

Message 15 of 18

17-12-2009 06:33 . am | View his/her posts only
My version of the assembly works the same way your original did except I had to remove the anchor constraints from both sub-assemblies. You insert the sub-assemblies using "Insert Component." "Move component up one level" (V13); "Component Merge" (V14). then align.

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 16 of 18

17-12-2009 03:40 . pm | View his/her posts only
Hi Mike,
I tried all that. The subassembly alignments are still vaporised. Do you mean align the entire subassembly again?
I need to see a file or a video if the answer to my question is No.
Cheers

Rank: 1

Mike

Newbie

posts: 0

Registered: 2002-8-28

Message 17 of 18

17-12-2009 03:59 . pm | View his/her posts only
The easiest way to do this is to rebuild the sub-assembly and do not anchor anything. Sorry it is a lousy workaround but it will be fixed in 14.3... I hope.

Rank: 7Rank: 7Rank: 7

Paul

Moderator

posts: 326

Registered: 2011-9-17

Message 18 of 18

22-04-2010 03:45 . am | View his/her posts only
Hi Mike,
can you give us all an update on this one.
I have just realized why my alignment scenarios are failing fail as I have a sub assy that is locked.

Flexible sub-assemblies are a must have.

Some time later.... like 8 months later.
What is the likely hood someone at VX can answer this request for an update on where it's headed with flexible sub assemblies? or will answer.......???
Still no better in Beta14.5 .....
See also