rounding edges  1-20  21

Next
 From:  TOM (SIRTOM)
7844.1 
Hi,

I am having trouble rounding edges / fillet edges like
in the example attached - tips are very welcome.

PS the edges to be rounded are the yellow ones.

EDITED: 23 Feb 2016 by SIRTOM

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.2 In reply to 7844.1 
Hi Tom - selection does not persist through file saving and reloading so I'm not sure which edges you are trying to fillet (EDIT: oops I see it's the dark yellow style that you've set up, I will take a look).

But a couple of things to do - right now you have 4 separate objects and you probably need to combine them into one single object using Construct > Boolean > Union.

Also the filleter is probably not going to like the outer surface being split into different pieces rather than one single surface, I'd probably try to rebuild that outside surface as just one large surface instead of in separate pieces like you currently have.

I'll try to set that up for you.

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.3 In reply to 7844.1 
Hi Tom, I've attached a tuned up object that should fillet better for you.

I started by combining the separate objects into one by selecting them and running Construct > Boolean > Union.

Then I examined the control points of the upper and lower outside pieces by using Edit > Separate on them to break them into individual surfaces and then ran Edit > Show pt. I could then see that the upper surface originally extended all the way down to the bottom which is good. I then deleted the outside lower separate pieces, did an "untrim" of the upper surface by selecting all of its edges (select 1 edge then Ctrl+A), then hit delete. That restored the original full upper surface and I then used Edit > Trim to cut it with edges (and the top circle which I replicated by ctrl+c/ctrl+v before the untrim) to make it one single piece rather than segmented. Having it segmented introduces additional non-filleted edges into the filleting mix that will need to be cut by the filleter and that tends to make its job more complicated.

- Michael
Attachments:

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.4 
Thank you very much, Michael !

So is it necessary to follow a certain routine to build up a model to be able to fillet its edges ?

My way was to start with a frame of curves from which I created the surfaces which I
multiplied with array and then merged with boole > union - wtong ? (see attachment).


PS I have encountered with filleting on a variaty of models which are the results of
boole operations etc

EDITED: 10 Feb 2016 by SIRTOM

Image Attachments:
Size: 222.7 KB, Downloaded: 98 times, Dimensions: 1446x3432px
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.5 In reply to 7844.4 
Hi Tom,

> So is it necessary to follow a certain routine to build up a model
> to be able to fillet its edges ?

It depends on the particular situation, basically there are some kinds of edge arrangements that the filleter has difficulty with.


> My way was to start with a frame of curves from which I created the surfaces which I
> multiplied with array and then merged with boole > union - wtong ? (see attachment).

Well not necessarily wrong, but it is generally better to work more with solids from the start, like use some curves to construct an initial solid and then use boolean difference to cut away chunks of that, rather than doing a "patch by patch" type of fill in of a wireframe like you're describing.

Do you come from a polygon modeling background? I've noticed that there is a tendency for polygon modelers to try and do that "patch by patch" approach. There are some tips about building extended pieces and using cutting operations more in the discussions here:
http://moi3d.com/forum/index.php?webtag=MOI&msg=4865.2

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.6 In reply to 7844.5 
>>Do you come from a polygon modeling background? I've noticed that there is a tendency for polygon modelers to try and do that "patch by patch" approach. <<

Yes. I come from poly modelling and I enjoy alot working with the elegant minimalistic approach of MOI.

Being more familiar with MOI now, I still encounter some riddles like in the file attached whre I cant get a rounded / filleted edge with a radius bigger than 0,2. Again it concerns the yellow edges. Happy to get help
Attachments:

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.7 In reply to 7844.6 
PS It is the inner edge which does not get filleted

EDITED: 12 Feb 2016 by SIRTOM

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.8 In reply to 7844.6 
Hi Tom, so at a first glance the problem is likely to be related to this edge here (and also on the other side)



The surfaces on either side of that edge are not quite smooth to each other, they meet at a slight crease, you can see it more easily in a full shaded display:




Filleting generally doesn't like that kind of thing where 2 surfaces come fairly close to being smooth but are actually just slightly creased with a shallow 5 to 10 degree difference between them. Building things in a "patch by patch" manner can tend to make that kind of results more easily.

The problem that causes is that when surfaces are totally smooth to one another, the fillet surfaces themselves that are generated on each surface naturally line up with each other. When there are creases between surfaces it means that the fillet pieces will not touch each other directly and so a corner junction patch needs to be generated and it has to figure out how to intersect the fillets with each other and how to build a junction piece. But it becomes difficult to do this when the junction area is very compressed in size and is just a small slivery thing, which is what happens in shallow angled areas like you've got here. When the fillet radius is pretty small, the junction area may be tiny enough that the fillets are close enough to touching so that they glue together ok, and when you increase the radius it gets to be a more difficult slivery area to deal with.

So anyway, that's likely to be the cause of this particular fillet problem.

The solution will probably be to round off that edge first to avoid the "shallow creased meeting" type of situation. I'd probably try to fillet that edge with a large radius or if that doesn't work then cut some area off either side and put in a blend there to make it fully smooth. I'll try those and post the results.

- Michael

EDITED: 12 Feb 2016 by MICHAEL GIBSON


  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.9 In reply to 7844.7 
Hi Tom, also what radius value are you hoping to use there? It appears that some larger values do end up working with your existing geometry, like for example radius = 0.5 seems to work ok over here (there is one edge missing in your yellow edges, make sure it's also selected before doing the fillet).

To get the fillets to perform better would probably involve smoothing out that creased edge which will probably involve some trimming and blending work, let me know if you want me to do that.

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.10 In reply to 7844.9 
Hi and thanks for the quick reply. I appreciate your explanations as understanding the reasons
behind the problems helps me avoiding them in similar situations.

No patch tp patch was involved here - I rotated a curve and cut out a chunk with boolean op.

"either side of that edge are not quite smooth to each other"

Possible that the initial curve I used was nor perfectly smooth . I attached it below.

"Hi Tom, also what radius value are you hoping to use there?"

I wanted to try out fillets of different radius and then choose one if them
Attachments:

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.11 In reply to 7844.10 
Hi Tom,

> Possible that the initial curve I used was nor perfectly smooth . I attached it below.

Yeah, that's probably what happened there. The end direction of a curve is controlled by the direction of the last 2 control points, so for example if you extend out a line through these 2 points here you can see that it is slightly off of the neighboring line:






Anyway that's about the type of shallow sharpness that will make things difficult for filleting.


> I wanted to try out fillets of different radius and then choose one if them

Do radius values of 0.5 , 0.6, 0.65, 0.75, work ok for you over there? Are any of those adequate?

- Michael

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.12 In reply to 7844.11 
>>Yeah, that's probably what happened there.<<

Ill redo the curve and repeat the process - then filleting should work.>>Do radius values of 0.5 , 0.6, 0.65, 0.75, work ok for you over there? Are any of those adequate?<<

Ill give it a try now
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.13 In reply to 7844.12 
Sorry to report that it did not work yet. Attaching the last file
Attachments:

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.14 In reply to 7844.13 
Hi Tom, I loaded your last posted file and radius values between 0.55 to 0.9 seem to be working well with it now.

Will anything in that range do for what you're looking for or are you trying for a smaller value than that? I don't know at the moment why the smaller values aren't working, but I can investigate some more if you want.

After a quick look, it might be the same kind of problem on the other cutting direction, looks like slight creasing here:



- Michael
Attachments:

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.15 In reply to 7844.13 
Hi Tom, yeah it's pretty much the exact same thing on the profile on your cutting object:





If you tune up that curve so it's fully smooth with the lower line, and then make a new cutting block out of that, I'd think that should fully clean things up. A quick way to put in a smooth curve between 2 lines is to select the lines and then run Construct > Blend.

At some point I may modify the Edit > Join command to look for things that are nearly tangent and automatically adjust them to be fully smooth with one another but it's a bit of a delicate thing to do always because in some cases it can kind of change the shape too.

- Michael

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.16 In reply to 7844.13 
Hi Tom, I've attached a tuned up version here - I did an "untrim" on your main body to restore it to its original full shape, then deleted the slightly off round corner in the above screenshot and put in a curve blend in there instead and then extruded a new cutting block from that, then did the boolean.

With this attached version you should now have "all systems go" for doing any fillet radius along that spot.

That particular type of thing with things being only just a 5 degrees or so off from smooth tends to make filleting pretty cranky.

- Michael
Attachments:

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.17 In reply to 7844.16 
Thank you for the patience, your informations are really helpful. Impossible to come up with
a solution without it. Will have a look at the file - great !
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.18 In reply to 7844.17 
Everything fine now after reworking the curves.
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  TOM (SIRTOM)
7844.19 In reply to 7844.2 
Dealing with filleting I realized that it is not possible to round with a perfect circle - for example if
you have a distance of 2,00 it is not possible to fillet with radius 1,00 exactly but a bit smaller than
the radius needed for a perfect circle.

Is there a specific reason behind it ?

EDITED: 13 Feb 2016 by SIRTOM

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
7844.20 In reply to 7844.19 
Hi Tom, the fillet will fail if it would cause a face to be entirely consumed away as would happen in that case. The fillet engine does not currently know how to handle that type of case.

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged
 

Reply to All Reply to All

 

 
Show messages:  1-20  21