Flexibility not working properly in inventor - BUG that has been there for ever

[Cris-Ideas]

Hi,

I am trying to use flexibility feature but every time I give it a chance it fails me. There is a bug somewhere in inventor that makes it not capable of properly solving assemblies with flexible components.

This time I have run on to this issue with a very simple assembly, had been able to reproduce this buggy behaviour  tens of times, and had made video for you.

After unsuccessfully trying to post this I thought I will make another video for you so you could clearly see what is the difference when assembly is flat.

Here it is.

Here I have uploaded data set for you to play with :  http://a360.co/2fmTsvD

And in case you also think this is not working properly you can support idea to fix this: Fix flexible assemblies !

Cris.

[johnsonshiue]

Hi Cris,

For [CaseNo:12531239.] # 104 bug, I am still trying to understand the model and the design intent. I notice that if you ground the first part (adaptive part) or clear the adaptive status. The position will be correct.

Many thanks!

[johnsonshiue]

Hi Cris,

I am look at "[CaseNo:12342298.] Bug 114 - flat assembley - constrain not possible to apply." The movement behavior of Component pattern should be fairly similar to subassembly. I means all the component pattern occurrences should move together. In this case, the components are severely under-constrained and the 4 Component Patterns are partially linked together. You can simply demote each Component Pattern as a subassembly and see how they work. You will find out the demoted Component Patterns are locked. There is no DOF. The required movement does not make sense.

One thing I would like to point out is that 3D Constraint Solve is quite different from 2D Constraint Solve. In 2D, most of the cases can lead to unique solution. However, in 3D, most cases, particularly under-constrained ones can lead to multiple solutions. The art here is to find an acceptable solution fast. Unfortunately, sometimes the solution may satisfy all the participating constraints but may not be the desirable one from the user's perspective. As a user, you will need to help reduce DOF reasonably and logically to obtain the desirable result.

For this particular case, I do not see a defective behavior. If you know any CAD package can handle this exact situation better, please let me know.

Many thanks!

[johnsonshiue]

[CaseNo:12344625.] bug 118 - inconsistent POS REP for FLEX component: The attachment does not tell how to reproduce the defect. I am contacting our Support team to get more information.

[CaseNo:12531250.] # 125 bug - constrain not possible to apply for flat assembly with patterns: There are two issues. The first issue is the same as "[CaseNo:12342298.] Bug 114 - flat assembley - constrain not possible to apply." The Component Patterns have cross constraints. They are basically locked.

The second issue is about the bolt not being able to inserted to Component Pattern. This has something to do with solve order. If you try to insert the bolt to the Component Pattern source (the other side), it will work. I think there is room for improvement here. In theory, the Component Pattern is grounded already. The bolt should simply snap to the hole. I will find out the defect ID and work with project team to resolve the second issue.

Many thanks!

[johnsonshiue]

[CaseNo:12531275.] # 129 bug - POS REP update bug: This is a limitation in PosRep. The issues are two folds. When PosRep:5.0 deg is active in Ramie Lancucha 8_FL_L .iam, PosRep:0 deg is active in Lancuch 8_kl 10_ 2.5t_10 ogniw_skretny.iam. Everything works until you change the number of chain links in Component Pattern. The offset distance is properly calculated by using an equation considering number of chain link (d73 = -d60*d58+16mm). Everything works so far. When the number of chain links is changed, the offset distance is changed in PosRep:Master. Then you activate PosRep:5.0 deg, the offset distance remains unchanged (based from the offset distance of longer chain links). 1) It does not know it needs to be changed according to the chain length. 2) The occurrence offset only takes absolute value. It does not dynamically update if an equation is entered.

PosRep operates under the assumption that there is no geometric change between two given PosReps. The components should be the same but their position has been changed. This is not a bug to fix. To make it work, it will be a project. Please report an idea on Inventor Ideas.

Many thanks!

[johnsonshiue]

Hi! I was trying to look at "[CaseNo:12531361.] # 132 bug - constrain solver and pattern bugs" but the dataset contains corrupted file names and it seems to be incomplete. Could you zip up using 7zip or rar?

[CaseNo:12546519.] # 151 bug - POS REP update bug: This is indeed a bug. It was reported as Bug-172822. The solve result is correct but the scene is not updated when the workplane is invisible.

Many thanks!

[WHolzwarth]

Hi Johnson,

it's good to see, that you're actively looking at these problems. I'm having similar issues as Cris, and finding them is no easy task.

Let's hope for improvements in this segment.

Thanks, and good luck

Walter

[johnsonshiue]

Hi!

[CaseNo:12614127.] # 153 bug - DOFs Anlysis failing = Constrains meaningless: It has been logged as Bug 173196. It is indeed a bug in DOF analysis. There should be only rotational DOF.

[CaseNo:12615337.] # 155 bug - POS REP bug - grounded component flipping: This is a good catch! The PosRep carries outdated cache leading to ground component to move. I will work with our Support team to have it reported.

Many thanks!

[Anonymous]

We have not had much luck with Flexibility since it came out.

Our biggest issue is after we get everything set and working right, Inventor will all of the sudden throw a fault (red cross) stating that there is an issue with one of the flexible assemblies.....we go in to the assembly, hit rebuild and the red cross goes away for a time...sometimes it comes right back, sometimes it doesn't.

This has been with pretty much the same with every flexible assembly for years, we just get used to it...it would be great if it actually got fixed!   Not a huge deal but I'm glad to hear it's not just us. Lol

[johnsonshiue]

Hi! I am not denying there are issues in Flexible, particularly with Positional Representation. However, each case is different. Some are defects but some are geometric issues. It is hard to tell where the problem is without seeing an example. Could you share an example exhibiting the behavior here or send it to me directly (johnson.shiue@autodesk.com)? I am more than happy to take a look and understand the behavior better.

Many thanks!

[Cris-Ideas]

Hi Johnson,

I see you have been working on my examples and from just a quick look I see you have confirmed some bugs already.

Since I am not receiving forum notification e-mails for some unknown reason I was not aware of the progress in this thread.

I will go through your posts and see what you have managed to establish.

I wonder what is the progress with fixing those bugs you have already confirmed.

Hoping for some good news.

Cris.

[johnsonshiue]

Hi Cris,

Flexibility and PosRep are indeed not as reliable as we hope them to be. It is a multi-faceted issue. There is dependency on constraint solving, cached override positions, inherent workflow issues, and some limitations. We have been making some if not a lot of progress on improving it. This particular topic has been indeed a focus within the team. I cannot make any guarantee but we are making progress.

Many thanks!

[Cris-Ideas]

Hello All,

I have another example of DOFs analysis failing and constrain solver bug showing it self again.

This time it is one of assemblies I am currently working on so it is real life project.

This video documents what is causing the problem exactly, so I hope Autodesk developers will be able to track this bug down and remove it.

As always data set available for download from: bug # 175 model

Or model and video: bug # 175 model & video

And for all of you interested Autodesk Support Case No.:  14381427

Hope for some response.

Cris

[Cris-Ideas]

@johnsonshiue

Hello Johnson,

I have documented another example of bug in POS REP and assembly solving.

It is too big for forum.

Autodesk Support case ID: 14398660. You will be able to download data set and video.

Cris.

[johnsonshiue]

Hi Cris,

Just to give you a heads-up, I did receive your emails but I have not taken a deep look at the issues yet. I will take a look and understand the behaviors better. Then I will work with the project team and see what we can do.

Many thanks!

[Cris-Ideas]

Hi All,

I was contacted by Johnson and advised that this problem is caused by the fact I have not grounded any of the components.

As I believe this is not the problem because I have many times showed this has nothing to do with badly solved assemblies I have once more made an attempt to follow this approach.

It didn't work obviously. So I still insist: Grounding or not grounding component is not the cause of the badly solved assembly and bad behaviour in case of mechanisms in assemblies as well in case of using flexibility.

Below I have placed video documenting this assembly with grounded component. Constrain solver and DOF's analysis behave exactly the same. Assembly CAN NOT be solved.

Cris.

[johnsonshiue]

Hi Cris,

You initiated the contact. I did not. I only replied to your emails with my explanation, which I think is common business etiquette.

Regarding this particular issue, I do believe it is related to lacking a grounded component. Have you grounded at least one component in the assembly shown in the video? Certainly, not every constraint solving issue can be solved by grounding components. But, it is indeed a good practice to ground components when the given component is stationary. Also, from constraint solving's perspective, grounding and fully constrained are two different things. Grounded component reduce degrees of freedom of a given component. Although fully constrained component cannot move at the end of the solve, they can be repositioned during the solve, as long as the constraints are honored. This is why directed/indirected angular constraint can flip.

Many thanks!

[Cris-Ideas]

Hi Johnson,

Yes I have grounded component. The same as you did I believe.

I clearly demonstrated this in the video above. Please watch from 1:30 .

The assembly behaves exactly the same as when there was no grounded component.

I ground component on video and even delete redundant constrains to be sure nothing is in the way of solving process.

As for contact remark I do not quite understand it. Is not your e-mail form of contact? Is there something wrong in saying "I was contacted" ? I speak English but it is not my native language and I may misuse some words.

Cris.

[johnsonshiue]

Hi Cris,

I took a look at the video and the assembly. I cannot say I fully understand this particular assembly design but I manage to find some interesting facts. Here are the steps involved.

1) Go to Tools -> Application Options -> Assembly -> check "Enable relationship redundancy analysis"

2) Ground frame 1_schodnia:1 like you did in the video.

3) Rebuild All. You will see Flush:1 and Mate:8 are marked as redundant. Suppress or delete them.

4) Now suppress Angle:2. It seems to be conflicting to angular constraint "fold."

5) Drive "fold."

The upper frame will rotate correctly.

Could you try it and see if you notice the same behavior?

Many thanks!

[Cris-Ideas]

Hi Johnson,

I have already did that and you can watch how this works in the video above.

So angle 2 is no in conflict with fold angle as it constrains grounded component with top level assembly coordinate system.

I have suppressed it as well as deleted it.

please watch video from 2:10

And

It is not the upper frame to rotate but upper frame is to stay grounded and vertical frame with horizontal up/down moving extension is to rotate about upper frame.

Cris.

[Cris-Ideas]

Hello All,

While working on my assembly I am finding more and more inconsistencies in DOFs analysis.

This moment I only have time to do an upload this short video presenting how ridiculously this works in Inventor.

This is generally the same data set as before.

Cris.