When updating a linked component, components in the design are altered

When updating a linked component, components in the design are altered

gentijo
Contributor Contributor
1,640 Views
11 Replies
Message 1 of 12

When updating a linked component, components in the design are altered

gentijo
Contributor
Contributor

In this example I have two designs, the head and the neck, the neck design links in the head as a reference model to verify mounting holes.

Reciently, when I went to the Neck design, and updated the head I found that the two components that were local to the neck design were modified in a material way.  The upper neck mount in particular, one body was split into two then relocated in the drawing. Here is a screen cast of the action. You will see that before the update the upper neck mount is a single body component consisting of 2 sketches, 2 extrudes and a filet.  After I update the linked head component, the upper neck mount becomes two disassociated bodies.

Screen cast

http://autode.sk/2EeeZ6m

 

Note: The best illustration of the problem is with updating version 47 of the design, 48+ is me trying to clean this up..

The design

http://a360.co/2moOccC

 

0 Likes
1,641 Views
11 Replies
Replies (11)
Message 2 of 12

jeff_strater
Community Manager
Community Manager

@gentijo,

 

Unfortunately, the current version of this design is 54.  And, I think that there is no way for me to download version 47 from the "share a public link" UI - that is bound to, I think, the current version.  So, for use to look at version 47, you would have to promote that version to be the latest, then share the public link, let me download it, then go back and promote version 54 back to being the latest, so you can continue to work on it.

 

However, what I suspect is that it is not the update that is introducing the error, but that the error is latent in the design before the update.  When you update a referenced design, it also causes all features downstream of that reference to compute.  So, sometimes you can see whether those errors are a direct result of the update by just doing a Modify->Compute All before you do the update.  I could be way off base here, sometimes an update can cause legitimate failures, and sometimes there are bugs.  This could be any of those cases.

 

Jeff

 


Jeff Strater
Engineering Director
0 Likes
Message 3 of 12

gentijo
Contributor
Contributor

Hi Jeff,

 

What I did was load version 47 and saved it as a new design, the new design file seems to have the same issues. 

http://a360.co/2G9xeqn

 

There are some downstream errors with joints as I would expect, I deleted all the joints before the update and I would expect the components to move around but not split up into multiple bodies but it did. 

 

Thank you for looking into this.

-John 

0 Likes
Message 4 of 12

gentijo
Contributor
Contributor

@jeff_strater was the link I provided illustrative of the issue ? That last link I sent illustrates the problem at the tip of the design file.

 

Thank you

-John

0 Likes
Message 5 of 12

jeff_strater
Community Manager
Community Manager

Hi @gentijo,

 

Sorry for the slow response, I'm in a training class all this week and next week.  Unfortunately, when I load the version 47 version of the design that you shared, I don't see any problem with it.  Perhaps I am missing what the problem is with this version of the design.  I don't see any failed joints in this model, and Compute All seems to work OK.  Perhaps I also need the later version of the "Dog Head" component, to simulate the update of that component.

 

Jeff

 


Jeff Strater
Engineering Director
0 Likes
Message 6 of 12

gentijo
Contributor
Contributor

Hi Jeff,

 

Thank you for getting back. The new link is version 47 saved as a copy so that version 47 was at the tip of the design file.

To see the error, the Dog Head 2 part needs to be updated, I did not want to apply the update and just leave the design

with in the error-ed state.

 

What I have done is loaded version-47 applied the updated and saved it as version 47-broken, so you can see the difference

before and after update in case you can't apply the update with the shared link.. 

 

The only part that breaks is a component defined in the Neck design, called Upper Neck Joint, there are only two 

components in this design, the rest of the design is comes from the linked "Dog Head V2"  design

 

This design has all the other components deselected so you can see what happens, to just that one component. 

To review, the component is made up of two sketches and two extrudes, the second extrude being a join. After

the update, the local component, "Upper Neck Joint", becomes 2 bodies vs 1 before the update, but the second

extrude is still specifying a join extrude and the color of the first extrude changes from red to clear/ wireframe.

 

Here is the pre update design

http://a360.co/2G9xeqn

 

Here is the post update design, where the only change is to allow the linked design to be updated,

http://a360.co/2o4SLcU

 

Please let me know if I can provide more info or do any other leg work to help you.

 

Thank you

John Gentilin

 

0 Likes
Message 7 of 12

jeff_strater
Community Manager
Community Manager

Thanks for the models, @gentijo.  I was able to reproduce the problem, but it's a bit complex to set up.  I had to export the "dog head" component from the "broken" version, then "upload new version" of that component to the "good" version, which then put it into an out-of-date state (you cannot export the full design history of a design, so you when you export, you just get whatever versions are currently active).  So, now I can reproduce the problem.  But, this is a large design, and it will take some time to understand the model, and this failure.  My guess is that it is doing exactly what it should be doing, and that some change in the new version is driving the design of the affected component in a surprising way.  But, I could be wrong about that.  Anyway, I'll see what I can find - it may take some time, so I appreciate your patience.

 

Jeff

 


Jeff Strater
Engineering Director
0 Likes
Message 8 of 12

gentijo
Contributor
Contributor

Hi @jeff_strater

 

Thank you for following up. I thin the problem is somehow attached to the fact that the bottom of the UpperNecKMount uses that inside plane of the head. 

To get there, initially I linked the Dog Head design, then build the UpperNeckMount by starting a sketch on that plane, then projected the mounting hole from 

the head.  To fix it, I removed the dog head, then after that the related sketches were broken, which I  reassigned to planes from the Origin. The I re linked the

head and aligned the mounting hole using a joint..  Now it updates fine,  So I think the sketch referencing a plane from a linked part, and possibly that sketch 

utilizing projected geometry is the cause of the problem.

 

Again, let me know how I can help.

 

-John

0 Likes
Message 9 of 12

TrippyLighting
Consultant
Consultant

I looked at your design.

 

You should familiarize yourself with Fusion 360's R.U.L.E #1. While that does not necessarily address your assembly problem, it will help you organize your timeline better.

 

Looking at the Dog Head subassembly I don't see any component Grounded, so its no surprise that something moves unexpectedly when inserted into another assembly.

Then I've taken a look at your stepper motor assembly. This is not how you would create a parametric model and I can only recommend that before continuing you watch all of the tutorials in the Support & Learning question so you have a better general understanding how to work with a parametric design software.

 

 

 

 


EESignature

0 Likes
Message 10 of 12

jeff_strater
Community Manager
Community Manager

@gentijo,

 

I dug a bit into this tonight.  I did find that the "Upper Neck Joint" component goes bad pretty early in its compute.  The second work plane in this component is "Next Top Mount".  Before updating to a new version of "Dog Head", this plane looks like this:

Screen Shot 2018-02-25 at 8.34.01 PM.png

 

after the update, it looks like this:

Screen Shot 2018-02-25 at 8.34.37 PM.png

 

I think this is the root of all the problems - why we end up with two bodies instead of one:

Screen Shot 2018-02-25 at 8.38.02 PM.png

 

If you have a memory of how this plane was originally constructed, that will help move us forward a bit.

 

Jeff

 


Jeff Strater
Engineering Director
0 Likes
Message 11 of 12

gentijo
Contributor
Contributor

Hi @TrippyLighting,

 

Thank you for replying.

 

When you mention the Stepper motor assembly, are you talking about the Stepper motor itself as a component ?  I grabbed that component out of the Gallery and really haven't looked at it too much, but I suspect it had some weakness because when ever I try to mate a joint to the shaft, the shaft is pulled out of the motor housing when it does that animation.

 

I didn't know you wrote up a Rule #1, but I have run into it several times, it would be super nice if you started a sketch or feature that was part of a body other than the active component that it would ask you if you would like to switch active components. Often I am double checking my work running my mouse across the timeline and seeing a timeline element, usually a sketch, associated with the wrong component and it's unlikely that you can move it from one component to the other. 

 

I fully expected some sketches to get messed up, but the real problem is that a component defined in that design, after the update to a linked component, not only does the component move but it's bodies are pull apart. That component only has a reference to the base of the head and the shaft of the stepper motor and only the head had ever changed.

 

I knew I could fix it fairly easily, and probably could of designed it better, but the error in the update action looked like a big to me, so I posted here. 

 

Thank you

John

 

 

 

 

0 Likes
Message 12 of 12

gentijo
Contributor
Contributor

Hi @jeff_strater,

 

I have done this design over a few times so I think I can recall what I did.. I had a basic idea of what I wanted as far as stepper motor arraignment to make the neck articulation work.

So I started a design, linked in the head to give me a spacial reference, then linked in the two stepper motors and placed/moved  them where I wanted them. Next I started to design components around the stepper motors and head to connect it all together. T

 

he first sketch projected the shaft profile from the top stepper. When I started the sketch, I wanted to center the stepper in the Head Front/Back Middle plane as a center reference, then instead of drawing on the face of the stepper, I created the Offset Plane (0mm len), that you are referencing as Neck Top Mount. 

 

What I am not sure about is that, the design may of initially had a position marker after the linked components and before the sketches, that I deleted after I secured the stepper with the joint from the stepper shaft to the upper neck mount. The position commands baffle me sometimes, so when I went to make some edits / clean up I deleted the position command from the time line,  I am kind of reminded of that because if you go to edit that plane on Rev #47, you will see that the motor moves from the construction plane. 

 

So I placed the steppers where I wanted them, then went to create the construction plane to sketch on, which was an offset plane from the face of the stepper motor. Before creating the plane Fusion360 probably prompted me with the dialog, "some components have moved, do you want to capture their position" and I said yes.  I created the plane, sketch, extruded the feature. After the components were formed, I assembled them with  joints so I could exercise the joints and analyse for clearance.  Some time later I was modifying something or cleaning up, maybe I created a sketch on the wrong component, violating @TrippyLighting's Rule#1, so I fixed that up and decided that all the components are held in place by joints, there was no need for the position command so I deleted it. 

 

Now, I went back and forth on the Neck/Head designs and the head always updated in the neck design just fine.  The at that rev, it decided to break. I think the problem came in after I linked in some of the other reference components in the head, but i am not entirely sure about that, 

 

-John

0 Likes