search

martinproks / cosmeticthread3d

FreeCAD tool - cosmetic thread for 3D geometry - experimental
GNU Lesser General Public License v2.1
4 stars 0 forks source link

Part & PartDesign attachement helper are inconsistent #32

Closed pierreporte closed 5 months ago

pierreporte commented 5 months ago

Consistency is very important in UX. Both the cone and LCS are fine but it should be the same thing for both implementations so that the user knows that is is the same operation. I would suggest keeping the cone or showing some sort of arrow.

martinproks commented 5 months ago

I know. I want to have cone/arrow as in Part, it is perfect solution I think. But it is not as easy in PartDesign as in Part. I'm stuck on it for 5 days... So i decide to public it just for now by LCS +Z axis...

martinproks commented 5 months ago

Hi,

Try this version please. The helper has been changed in Part and PartDesign to the same visual look. It is based on Draft.makeWire().

arrow-DraftWire

pierreporte commented 5 months ago

It works perfectly. Let’s hope that https://github.com/FreeCAD/FreeCAD/issues/11214 is fixed sooner than later to have something better than a real object in the tree.

I tried with a large thread (by default the cylinder diameter is 20 mm). The object would be more visible if its size depended on the selected circle diameter.

martinproks commented 5 months ago

I tried with a large thread (by default the cylinder diameter is 20 mm). The object would be more visible if its size depended on the selected circle diameter.

Did You try Apply button? apply (edit citation)

martinproks commented 5 months ago

It works perfectly. Let’s hope that FreeCAD/FreeCAD#11214 is fixed sooner than later to have something better than a real object in the tree.

In the case of Pads or Pockets, the situation is different. At the beginning You have sketch, you can "guess" direction and value and this could be presented as ghost geometry. In the case of champfers of blends You have value and You are selecting edges to apply the value. It could be calculated ghost champfer or blend quit well too (usually - if the values are in reasonable relation to faces).

But in the case of thread at the beginning You have to attach to some hole/rod. From the selected edge guess thread type, major diameter and length and at this moment You can have ghost feature. But in the attachment moment You have nothing. I can create directly the thread feature (in the 0,0,0 and direction 0,0,1 with some random/small parameter). At the other hand till You have complete attachment and select at least approximately correct values, the ghost feature is totally misleading. The direction arrow is better solution in the first moment I think.

You can try it actually. You can create non attached default thread from python console (follow doc//How-to-use-it-from-python-console.html). It creates non attached thread with default parameters in 0,0,0 with direction 0,0,1 and then You can manually attache it. Try it on small threads (M2), "normal threads" (M8) and big threads (M100).

pierreporte commented 5 months ago

Indeed, the Apply button does the job. Since it works without it, I didn’t even try. The thing is, the helper is at the right place once the concentric entry is selected, so it already knows some information about the circular edge.

martinproks commented 5 months ago

Well for now, I'm closing this issue as solved.

Question/request to apply ghost thread in the attachment phase instead of direction arrow is another issue to discuss and may be enhancement.