Something that has puzzled me for a while:
If you have two colinear and coplanar lines end to end that resulted from various editing operations, they don’t always want to weld together. I don’t understand why. Here is an example in which it is clear that the Y and Z coordinates suggest they should be weldable. But they aren’t. The only way I know of getting rid of unwanted endpoints is to delete the line sections and redrawing.
Have I missed something obvious?
All the coordinates have a tilde in front. Have you checked what you see if you max the unit precision settings? Possibly they are ever so slightly not colinear.
Welding does nothing to decrease the number of segments. It just combines them into a “curve”. The way to unsegment is to draw an edge from the point where two colinear edges meet and erase it.
Not sure that’s it. Here are a few more screenshots.
The tilde shows that the measurements are rounded approximations. Although I see there are none in you screen shot. If the errors are that small it’s likely that they are beneath the “color by axis tolerance” so that will not necessarily be a reliable way to inspect if the lines are on axis. With very small errors color by axis can return a false positive. I see one of your lines looks red and blue?? That’s odd, is this imported from a dwg, is all very far from the drawing origin? Can you upload the file so we can inspect?
Oh that segment is probably just selected in that screen shot.
Yes, to show that it is a line broken in two.
Well, here’s an odd thing. To show you an actual file, I copied the part of the drawing shown above to a new file so that I could upload it. The line representing the top tread of the stairs was still broken in two but Weld was able to unite it. More weirdly still, the line representing the inner face of the external wall that you can see broken into four sections above, imported into the new file as a single unbroken line.
My impression has always been that this behaviour arises as a result of a process of creating and erasing lines, using push/pull, etc, and whether the lines self-heal is hit and miss. But when they miss, the only way to correct it oftentimes is to erase and redraw.
None of this is a huge problem but I guess it could give rise to anomalies and inaccuracies in a drawing that later on becomes hard to pin down.
test plan.skp (180.0 KB)
I am trying to get my mind around this. Are you saying that the welded lines become a “straight” curve?
Here is the outside face line of my wall, which you can see is broken in two. It is unweldable.
I then draw a line at right angles starting at the intersection.
I then erase it. The line is still broken and unweldable. You can see all this in the uploaded file in my last post.
I see. I say I see but I wasn’t able to zoom in as close as you. But I could get close enough to determine that the main lines are offset by about 1/10th of a mm (which I think is what you show in your entity info box).
Good. Suggests the behaviour is not as random as I thought.
This is a 3rd edge - 0,09070 mm length. You need to select all 3 edges to be able to weld. Otherwise the
weld command makes two curves - each contains one segment- and the 3rd will remain as individual edge as it was.
Otherwise, it is best to delete those 3 edges and redraw them as one…