Bad Five-Pointer?

[robcat2075]

Does any one know why the Group "Bad 5" can't be made into a 5-pointer?

Divider001.mdl
 

[Tom]

I think I just closed it.

Divider001Fixed.mdl

[robcat2075]

11 minutes ago, Tom said:

I think I just closed it.

Wonderful! What did you do differently than I did?

If I undo it, and try to re-make it, I still can't close it.  😒

[Ganthofer]

So, what I found was that the circled intersection is not 2 splines intersecting, but 3. The highlighted (yellow) and the one coming up from below are not a spline, but 2 end points of separate splines. 

I found 2 ways to make the 5pointer work.

1. Detach one of those 2 end point splines and re-attach, making then a continuous spline. And re-peaking it if needed.
2. If for some reason you need/want to have the 2 end point splines, then detach the cp from the spline coming up from the bottom (that seems to be the cause of the problem - an extra CP ?). Make the 5pointer (now possible). Add a CP to the spline from below, re-attach the new middle CP and then delete the end CP.

I haven't been able to recreate the problem. Having the 2 dead end splines attached to a 3rd didn't seem to reproduce it ( for me).

F.Y.I. - I've gotten in the habit of using the TAB key to check if, (1) splines are connected, (2) finding the oposite end of a dead end spline  so I can use the Y key to extend it with an additional CP.

 

note: 19.5e

 

 

[Tom]

I have attached a screen shot showing a five point spline and every point has a number corresponding to a spline entering that point. Some have four different  possible points of connection and some have three.

When I "fixed" the 5 pointed patch and made it whole all I did was try different combinations as to how the points were selected and what order they were selected in. It makes all the difference in the order regarding if the patch will be filled and I frankly have no idea why selecting the points in one combination works and why selecting them in another order doesn't work.

Obviously there must be some mathematical rule at play underneath it all but I will leave that to someone smarter than me to figure out. But I just "trial and errored" picking those points in different combinations until it worked...took about five minutes of fiddling.

[Ganthofer]

I totally agree. That usually works. Maybe I missed one of the combinations when I tried that. 

I also find, that if you select them in any order, hide everything else and then use the box or lasso to select the five pointer, it always works. 

Until this one.

[Rodney]

Hitting of period key twice (once to invert the selection and the second time to revert to the original 5 point selection) works for me with 5 pointers 90% of the time.

Not in this case however.

 

Interestingly, copying these five points and pasting them as a new set of 5 CPs and then attempting to close them works.

Obviously that wouldn't solve the problem because it'd be a floating 5 point patch unconnected to the rest of the model.

 

It does suggest that while before there was something invalid in the 5 CPs copy/pasting resolved that.

 

I post this to basically say that when the 'whack the period key twice' approach fails I almost always know I need to go in and rework the splines.

Hitting the period key twice effectively reorders the CPs... allowing them to be seen as valid for closing as a 5 point patch.

(but again, not in this instance)

 

 

Aside:  I noticed I did have to Flip Normals to get the patch to display correctly.

[serg2]

It is better to close such patches with non-continuous splines and monitor the overall structure of the normals of all patches.

Divider003.mdl