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.

[Cris-Ideas]

Hi,

I thought this thread would use new refreshing example after latest discussion.

Let's let inventor speak for it self.

This is one of the most basic situations when using flexibility. There are only those constrains that are required to get desired elements relations (if this worked).

So 2 parts, 1 flex assembly (also 2 parts), 3 constrains inside flex assembly, and just a handful of constrains on the top level (After all how many constrains can you use on 3 components).

And a Total Disaster with solve.

Last time I showed this I was tolled there is no bug :-).

How do you think?

[video]

data set available: https://autode.sk/2L0IyXJ

Cris.

[smokes2998]

I would not call it a bug  it just that the acis /pagemanger Kernal design is not designed for dynamic/automatic solving, When you use CATIA v5  Like I do you can see all the work flows really restrict dynamic or automatic solving which keeps CATIA stable.  As it will only solve if the user initiates it  if  the solution breaks it  keeps the part in the same place  before the solution failed and highlight the failed solution in the constraint tree. 

[Cris-Ideas]

Hi,

But there is no dynamic solving.

It is just simple 3D static geometry.

Every solve for any assembly has only two theoretical outcomes. Either it is solved or it cannot be solved because constrains applied are in conflict.

There is no third option. So in any case properly functioning solver must wither solve or bring up an error message indicating which constrains are in conflict. 

Buggy solver is capable of placing components against constrains that are base for the solve it performs. So in reality this kind of solution is not even consistent with boundary conditions used for its formulation.

Thanks for bringing Catia up. I will have it in mind.

Cris

[Cris-Ideas]

@johnsonshiue

Hi Johnson,

In relation to one of recent support cases this article came up again Link

Could you comment on the tips given, as your advice usually is to ground components, and also that flexibility can be used and should work, even in multi level assembly.

I was also thinking if it would be possible so someone from solver team give us a comment and some more deep explanation why should given approach be used and how it helps in getting stable solve from the solver side.

Cris.

[johnsonshiue]

Hi Cris,

This issue was reported as INVGEN-5699, which is also fixed in our internal build. If you have access to Inventor Beta, you can try it on the non-install Alpha build.

Many thanks!

[Cris-Ideas]

So I have heard.

But what is more of interest to me is to know what exactly was wrong and what was fixed so to know what else may work.

As we already know not all issues will be fixed in next relies.

And how about dealing with flexibility in general.

Is there any stable workflow that is verified to work?

Cris.

[Cris-Ideas]

Hello All,

it is documentation case 200 for my current Autodesk account.

but before:

@johnsonshiue

I have tried case from post no. 123 in current beta build - it does not work. Sorry.

I have documented this and posted on beta-forum, also you should have an e-mail with appropriate link.

As of # 200 bug

another super solve form Constrain solver in presence of Flexible component

@SteveMDennis  I need to emphasise this once more:

Constrain Solver forces solution against defined constrains

and insists ALL is fine. Assembly solve validation is WRONG.

as usual data set for download: https://autode.sk/2Ob6sC0

Cris.

[johnsonshiue]

Hi Cris,

I never claim that every bug you find is fixed on the latest build. It is possible some of them may not be addressed. Unless I misunderstood how NDA works, Beta testing discussion should be bounded within Beta forum, not on public forum. I will take a look and get back to you on the Beta forum.

Many thanks!

[Cris-Ideas]

I will only make brief comment over this one:

7 of 9 constrains present in the assembly are NOT FOLLOWED

How can you clam Flexibility working?

data for download: https://autode.sk/2CP9fj5

Cris.

[johnsonshiue]

Hi Cris,

I watched the video for #201. The behavior at 1:00 was wrong when you turn on Flexible. The little rod should stay put. And, it was the slippery slope after that. All constraint behaviors since that point become flaky until you turn off Flexible.

I have tried the case on our internal build. The behavior no longer happens. One thing you may want to consider is to avoid using "Ground" status in the Flexible sub. I recalled we discussed it before. The "Ground" in Flexible sub can lead to confusing behaviors. I could be wrong as always but I am not seeing the same behaviors shown in the video on our internal build. You can try it on the Beta build.

Many thanks!

[jtylerbc]

---

@johnsonshiue wrote:

One thing you may want to consider is to avoid using "Ground" status in the Flexible sub. I recalled we discussed it before. The "Ground" in Flexible sub can lead to confusing behaviors.

---

I haven't checked out the specific files involved here, but from a general practice standpoint will confirm this.  At my company, we used to struggle a lot with crazy behavior from Flexible assemblies.  We still have trouble occasionally, but we eliminated a large portion of the instability ourselves by changing our practices to eliminate grounded parts in subassemblies that are going to be used in a flexible situation.

Be aware that the grounded part issue is not just a one-level-deep problem.  If you have a Flexible subassembly that contains a subassembly that contains another subassembly that contains a grounded part, you've still got a problem.  In my experience, there can't be any grounded parts anywhere in the structure of a Flexible sub without introducing instability. 

I'm not saying it should be like that - I consider the fact that grounding doesn't work properly in Flexible applications to be a bug.  But it is a part of the Flexible instability problem that you can start avoiding now, simply by changing your own practices a little, without waiting for any Autodesk fixes.

[Cris-Ideas]

@johnsonshiue

Hi Johnson,

thanks for trying.

However, you lately reminded me of NDA in terms of discussing internal new build features, and you now bring this subject again your self.

It is the case I have tried this already.

Please tell me now how can I answer you as this is not the case what you say. I have also tried this and I have placed appropriate report in beta forum along with data set and video. (no answer for two weeks now)

Since you brought this up and said "it works" I feel entitled to replay - it does not in fact.

I will stop here an await your input in beta forum. I do much to provide autodesk with simple examples easy to debug.

I have been tolled, much more than few times, "this works in internal new build" which turned out not to be the case after new relies was available.

As for using grounded component in flexible assemblies. I am happy there is finally a information I can confirm.

Using constrains rather than grounded components proven to be (for me) more stable. This is also the point I was making from the very beginning and that was argued many times.

I am happy that finally we agree on that.

Waiting on beta forum regarding this case.

Cris.

[Cris-Ideas]

@jtylerbc

Thank you for input.

Can you share other tips you use?

As for grounded status, I was also using constrains instead of grounding, but was advised many times by Johnson and others that grounded is the proper and more stable way. So I was consistently trying to prove it is not the case.

And finally  this is no longer an issue of argument.

Cris.

[Cris-Ideas]

# 204

This one is not as severe in consequence as previous once, but still:

• Painful,
• Annoying,
• unjustified.

as always data for download: https://autode.sk/2p80RBO

Cris.

[johnsonshiue]

Hi Cris,

This case is very interesting. It actually does show the ground status matters. If you add ground status to the flexible subassembly itself, the rotation will work after you add the limit Mate constraint. I cannot explain the behavior but this case does support my perception that you want to free up components within the flexible sub as much as possible, but also you want to ground the flexible subassembly itself.

My theory is that by grounding the flexible subassembly, the sub coordinate system is fixed in the space and the DOF is much more straight forward compared to when sub coordinate system is allowed to move too.

Many thanks!

[Cris-Ideas]

Hi Johnson,

I would not agree this case supports  anything but the point this is not working properly.

As for grounding components in sub assembly we have finally get to the point this is no longer being advised.

Now to the grounding of flexible assembly it self.

This is not possible to use technique. If flexible assembly moves all whole this is not possible.

Also if sub assembly container is constrained to parts(components) that are in it its position is fully defined whenever position of its components is defined in context of top level assembly.

But the most important point is that grounding of flexible component is just not possible in real life. Specially if flexibility should work in multi level.

We have discussed this approach before in this thread.

Another thing is you did not explain anything really in your last post, except that you cannot explain this behaviour, and that it "seems to prove grounded state matters".

This answer really gives us nothing. We need to know WHY things happen rather than they do. If something does matter (regardless of whether it should or not) we can see just from trying it. But until we understand WHY we cannot do anything with it. Especially in situation like her when observations lead us to conclusion that following approach that seem to work is not possible in real life designs.

Check around post #20 in this thread

Also have ou noticed that even without grounding anything when I drive constrain handle rotates as it supposed to? So in face flexible assembly is recognised correctly if solve is invoked by driving a constrain, and it is not correct if I try to drag.

Can you please explain how exactly flexible assemblies are considered and WHY thing are happening?

Cris.

[Cris-Ideas]

@johnsonshiue

I have realised something else:

We have agreed on following:

1) flexibility should work any level deep

2) inside flexible assembly components should be constrained and grounded status should not be used.

Is this two are correct that you cannot ask to ground flexible component it self as it would violate second point in more than one level deep structure.

So grounding flexible assembly has not sense.

Cris.

[Cris-Ideas]

---

@Cris-Ideas wrote:

@johnsonshiue

I have realised something else:

We have agreed on following:

1) flexibility should work any level deep

2) inside flexible assembly components should be constrained and grounded status should not be used.

Is this two are correct that you cannot ask to ground flexible component it self as it would violate second point in more than one level deep structure.

So grounding flexible assembly has no sense.

Cris.

---

@johnsonshiue

Hi Johnson,

Any answer?

Do you agree or have another argument to support grounding of flexible components?

Cris.

[Cris-Ideas]

# 205

Another example of flexible assembly and a kind of buggy behaviour.

This time some constrains are followed or are not flowed. And they obviously should be followed in all cases.

data set for download: https://autode.sk/2xCAHej

Cris.

[Cris-Ideas]

I see no answers to latest examples.

But even despite that I will post next.

# 207 - very simple, almost real life example how flex could be used if only it was working.

data set form download:  https://autode.sk/2QYMPiD

Cris.