FreeCAD Forum

Hello

I want to ask for adding a both direction checkbox in Linear Pattern in Multi Transformation.

Only in MultiTransform or for LinearPattern in general?

With the feature freeze in effect it won't get into 0.20 right now. If I understand it correctly it also looks like a good "low hanging fruit" problem. Could you add this as an issue on github?

I sounds at first as a good idea, but is it really?

Both directions --> symmetric or not symmetric
if symmetric and the occurrences can not dived to 2 --> wont work
if symmetric --> mirror will always work and is the same work and already implement
if not symmetric --> use second linear pattern (is the same work as it would be in one task feature, and you do not need all field twice in the task feature, which only clutters the task feature)

Greetings
user1234

edit:
By the way, it is common in all CADs which i used, that a linner pattern works only from one side (corner).

user1234 wrote: ↑Sat Mar 26, 2022 5:20 pm I sounds at first as a good idea, but is it really?

I think so. Maybe a 'add a mirror to' checkbox could be an idea instead of adding a second transformation manually.

C4e wrote: ↑Sat Mar 26, 2022 5:32 pm Maybe a 'add a mirror to' checkbox could be an idea instead of adding a second transformation manually.

When you have an occurrences 4, then you get 7 patterns? This is misleading. When you mirror it, it takes literately 5 seconds and have an entry (feature) in the tree, which makes the whole thing clearly, that there is a mirror and where is the mirror.

Greetings
user1234

I tend to agree with user1234 here, there's existing features for this and this is probably not often done so not worth cluttering IMO.

user1234 wrote: ↑Sat Mar 26, 2022 5:45 pm When you have an occurrences 4, then you get 7 patterns?

That would of course be nonsense

if symmetric and the occurrences can not dived to 2 --> wont work

I don't see this implication.

Such a symmetric linear pattern should of course not do any mirroring. "Reverse direction" doesn't do this either. I would rate the following behaviour as sensible, which follows the idea of a symmetric pad (which could also be done by using a mirror).

Let's start with this model. The base plate is horizontally centered. The pad is formed as a half circle to see its orientation. This is what we can already do.
- 5 occurrences (forward): - 5 occurrences reversed (same length): And this is how it would be symmetric. Same number of occurrences, same length. It seems rather natural to me and works with odd and even numbers of occurrences:

chrisb wrote: ↑Sat Mar 26, 2022 11:26 pm

user1234 wrote: ↑Sat Mar 26, 2022 5:45 pm

When you have an occurrences 4, then you get 7 patterns?

That would of course be nonsense

if symmetric and the occurrences can not dived to 2 --> wont work

I don't see this implication.

Sorry i mean, if symmetric and the occurrences can not dived to 2 --> wont work, when if the number occurrences is practice in the model. For example when the occurrences is 4 and it is mirrored and practiced in the model, the origin feature is between two patterns and the real occurrences is 5.




It will only work, when the occurrences can not divide to 2.



Greetings
user1234

The meaning of a symmetric linear pattern would be like an offset of the participating features of half the length of th pattern. So an even number would indeed be as indicated on your image, including the peculiarity, that the original would vanish.

If that is not possible, I could live very well with the restriction to an odd number of occurrences for a symmetric linear pattern.