[solved] LinkTransform creates link at the wrong position

[uwestoehr]

I have a sphere primitive that I would like to create a link for. Since the link should be at the same position than the sphere, I
1. create a link
2. then I set in the "LinkTransform" to "true"

But then the link is still not at the same position than the original:

screenshot of the result

FreeCAD_7sAopfbHJF.png (49.56 KiB) Viewed 553 times

Testfile:

Link-bug.FCStd

testfile

(12.58 KiB) Downloaded 12 times

Do I make a mistake or is there an issue?

.

OS: Windows 10 (10.0)
Word size of OS: 64-bit
Word size of FreeCAD: 64-bit
Version: 0.19.23057 (Git)
Build type: Release
Branch: master
Hash: e2b58bf3d52e4f9b5eea8885d95e4a255d072ea2
Python version: 3.6.8
Qt version: 5.12.6
Coin version: 4.0.0a
OCC version: 7.4.0.beta
Locale: German/Germany (de_DE)

[Roy_043]

The LinkTransform works correctly I would say, but it does not take account of the 'final' global placement of the source object, which in the example is also determined by the placement of the Std_Part it is nested in.

[papyblaise]

the error was to create a Part (yellow staircase) instead of a Body (blue staircase)

[jmaustpc]

the error was to create a Part (yellow staircase) instead of a Body (blue staircase)

that comment is total rubbish.

uwestoehr has a lot of FreeCAD knowledge.

[jmaustpc]

I have a sphere primitive that I would like to create a link for. Since the link should be at the same position than the sphere, I
1. create a link
2. then I set in the "LinkTransform" to "true"

But then the link is still not at the same position than the original:

My understanding is that the position is correct it's just that the link follows relative to the Part container it is placed in or if not in one then global axis system.

This makes sense when you think about a typical use case. Lets say you have a bolt that holds two objects together in two different part containers, if the bolt goes through object B into a hole in object A in both cases, but in the second Part Container Object B is a different shape so its actually object C, but in both cases the Bolt is wanted to be in the same location relative to their Part Container since the hole in object A is the same, but not the same location relative to the global axis system. Because perhaps one Part container goes on the left of something but the second Part Contain goes on the right. Does that make sense?

In your example that is what has happened. The difference in Global position is due to the -4mm Z placement of the "Part" container.

[jmaustpc]

My understanding is that the position is correct it's just that the link follows relative to the Part container

I just check to confirm that it is placement relative to the Part container and not only it's offset from attachment.

This file is a modified version of yours, I put your link sphere in another Part, adjusted the second Part z to the same -4mm and the spheres were overlapping perfectly as expected.

Then I added the cube, and changed your original sphere's attachment to a vertex of that cube which moved it 10mm back, the link sphere in the second Part container (Part001) also moved the 10mm so that it still perfectly overlaps. This demonstrates that Link is following placement of the original object relative to the original objects Part container , and if the link output object is not in a Part container then its placement is relative to the Global axis system. When the link object is moved into a new Part container it automatically applies it's placement relative to that new Part container.

No doubt realthunder will advise if I have miss understood something. But If my understanding is correct then this is intended behaviour.

Jim

[jmaustpc]

Whoops, I forgot to attach the file. I had to go outside and check on my newborn kids, as in goats not humans! It's night here and its rained and cold, well wet hence make them cold, not cold compared to many of you guys. The mummy goat was cleaver, she found a nice warm dry spot under a shelter I had in the paddock, what a good first time mum.

Link-bug2.FCStd

(16.39 KiB) Downloaded 7 times

[uwestoehr]

No doubt realthunder will advise if I have miss understood something. But If my understanding is correct then this is intended behaviour.

But then I have a problem. My use case is this: I work a lot with assembly 4 files. They work this way that the different parts are put into a Std_Part container as links. I need to link these links to transform them further. I can and also will not make a link also to the Std_Part container, I only want a link to some of the parts in the assembly. I need these links to appear exactly at the same position as they are shown in the model.

@realthunder, do you see a way to extend the LinkTransform setting so that also links to objects in App::Part will appear at the same position? If not, what do you propose I should do instead?

[uwestoehr]

@realthunder, do you see a way to extend the LinkTransform setting so that also links to objects in App::Part will appear at the same position? I

I think an additional option "preserve global position" could be added. When used, it ass the placement of the App::Part container to the transformed link.

[Roy_043]

I think an additional option...

If you create a sub-link (Std_LinkMakeRelative) instead of a 'normal' link (Std_LinkMake) the link will follow the 'final' global position. So what you want is already possible.
See video (T=14:05): https://github.com/realthunder/FreeCAD_ ... /wiki/Link