Surfacing problem

[marker4x4]

Hi,

I wonder if someone could enlighten me on the following issue:

I'm trying to create draft surface off of an imported model of sheetmetal part.
The objective is to create a tool that will form this part on the rubber press, thus the draft, or "springback" angle there.
Draft function wouldn't work because the surfaces are neither planar or ruled.

So I decided to copy the inside surfaces, extend them a little and intersect them with planes to create two chains of curves that are then used to create the Sweep. The Section of the sweep defines the 1.5deg. draft angle. Well - it ALMOST works...
As you can see on the picture, the drafted walls are concave which is not acceptable. It's the purple contour I'm refering to.

View attachment 2742


Anyway, I'd be grateful for any advice on how to fix it. Attached is the model in 2001 format, please feel free to experiment with it. The original geometry of the part is hidden on layers.

Thanks!

2006-09-28_125542_draft_surf.prt.zip

 

[jeff4136]

If you throw some planes normal to curve around one of your traj
curves and section the copy surface and the new VSS you'll see that
both are curved (it's inherent to the shape sectioned like that) but
the VSS has a higher curvature.


Your VSS deviates as it does because its element lines (the section)
are normal to the rib web and the original surfaces' element lines
are approx parallel to the skin roll axis, leading edge, spar
(guessing there's something on the tail of the rib). You can see,
mo'-o-less, the isocurves (which should correspond to element lines
if it is, in fact, a developable skin with an Offset Analysis. Shaded,
Mean Curvature analysis is also useful (look at the color band
boundaries).


How I think I might go about that ('til I convinced myself I'm
screwin up) would be to start with the master surface (the one the
skin and rib flanges are based on) or a skin surf. Offset an
appropriate distance (1.5 flg len * tan(1.5) or something of that
sort and create planar intersect curves for the overbent flange end
profile. I'd then trim these so I had the same angle relative
original element lines and rib web then create Boundary Blends using
flg heel and toe section curves, probably trying a Piece to Piece
point mapping.


A VSS might(?) also be created with a heel curve, an offset toe curve
and constant normal direction defined by a plane parallel to leading
edge.


I'm not really sure (and may be all wet). Would have to play a bit
and am short of time (and using WF2 so the best I could send you
would be a .neu).


An afterthought ... If you don't have access to the skin surfs or
their basis you might want to (on another copy) Edit Definition the
Import Feature, Heal Geometry, Manual, Edit Boundary and Delete the
trim boundaries on flange surfs, particularly the leading edge
portions. There's more there and you'll get rid of the corner
rounds.



- - - - - -
A little later...
The attached are application of the offset surf / new flange toe
curve / Boundary Blend theory. If I understand what I see
there's a variation in overbend angle ~1.5 to ~1.7 degrees (1.5
was the target). I'm wondering if a planar curve offset wouldn't
be a better choice for creating the new toe profile. This stuff
don't come easy to me and I'm out of time again. If you think
there's something of value but it doesn't make any sense I might
be able to talk you thru it. Ideally you'll be able to look at
the native WF2 and follow the feature tree, there's a bit of
surf extending, trimming, etc. involved. Even then it may be
hard to make sense of it. Not sure it isn't just jibberish
myself. <G>


2006-09-29_114736_draft_surf_base_wkg.neu.zip


2006-09-29_115005_draft_surf_base_wkg_wf2.prt.zip


===============================
Edited by: jeff4136

 

[marker4x4]

Thanks Jeff, I'm going through your tree right now (I'm on WF3) trying to understand your actions; the result looks good and the isoparams on the flanged sides go in the right direction which solves the original problem. I just need some time to digest all this!

I'm sure I'll be asking for explanation here 'n there so please bear with me

but in general I'm a happy camper again!

Thanks soooo much, have a nice weekend.

 

[jeff4136]

Super deal. One thing that might cause confusion is LE_FLG_IML
(f#15, fid_3020). It's an Independent Geometry feature (Insert >
Independent Geometry). The quilts were created in it (Edit Def)
using Geometry > Surface > From Surface and picking the original
import feature surfs with no need to remove the trim boundaries.
Then Edit > Feature Properties > Join Surfs and return to part
mode where Extend 1 thru 4 were created. The rest is pretty
straight fwd but holler if something doesn't seem copacetic, it
might just be the result of floundering around trying to figure
out how to go about it.

 

[marker4x4]

Super deal. One thing that might cause confusion is LE_FLG_IML
(f#15, fid_3020). It's an Independent Geometry feature (Insert >
Independent Geometry). The quilts were created in it (Edit Def)
using Geometry > Surface > From Surface and picking the original
import feature surfs with no need to remove the trim boundaries.
Then Edit > Feature Properties > Join Surfs and return to part
mode where Extend 1 thru 4 were created. The rest is pretty
straight fwd but holler if something doesn't seem copacetic, it
might just be the result of floundering around trying to figure
out how to go about it.

Hmmm.... I'm still new to WF and trying to find my way around it...
I knew LE_FLG_IML was done using equivalent of "old" Surface->Copy function, but I couldn't replicate it under WF. Also R/Click->Redefine of that feature doesn't bring anything - do I need adv. surfacing module for that??? Just curious.

Anyway, I went through the whole thing and I do understand now your concept; quite brilliant BTW

Being a lazy b****d I was trying to combine your idea of Boundary Blend with Sweep, hoping to get the desired result without going into calculating the offsets by hand, but Sweep wouldn't cooperate... I'll try again fresh on Monday and let you know.

Cheers,

Edited by: marker4x4

 

[jeff4136]

Re LE_FLG_IML:
> Also R/Click->Redefine of that feature doesn't
> bring anything - do I need adv. surfacing module
> for that???


Shouldn't (?). This is a hoot. I was anticipating having trouble
communicating because I've never used 2001, but I've never seen
WF3 either. `;^) I honestly can't say. There is no "Redefine"
on WF or WF2 RMB shortcut menus that I've noticed.
- - - -


> ... hoping to get the desired result
> without going into calculating the
> offsets by hand ...


I cheated on that. Sketch2 (fid_3077 in Group DATUM) is mostly
some scribbling to see what the part looked like but there's also
a couple of line elements (on ctrln) where the angle is defined
and a refdim (rsd12) measures the offset. The surf Offset features
are tied by relation to the refdim.


> I'll try again fresh on Monday and let you know.


Good deal.

 

[marker4x4]

Re LE_FLG_IML:.....
Shouldn't (?). This is a hoot. I was anticipating having trouble
communicating because I've never used 2001, but I've never seen
WF3 either. `;^) I honestly can't say. There is no "Redefine"
on WF or WF2 RMB shortcut menus that I've noticed.

Jeff, I give up

How did you create the LE_FLG_IML?????? I went through all the menus in WF3 I could think of, searched far and wide the manuals to no avail. Grrrr... I can't see anything that allows to manipulate imported geom. at all!!!!



Awaiting your life saving advice, thanks a lot!






A bit later: I think I found it - is that the Independent Geometry?And for the imported date manipulation, there's the hidden Geometry menu... Nice work, PTC

Anyway Jeff, thanks a lot for your help, the f/block turned out really nice - I owe you.

Edited by: marker4x4

 

[mechdes1a]

This punch geometry reflects the original piece-part fairly well - I think the worst deviation is about 2 thou...It's a lot of features. Problem is - the original copy geometry is corrupt in several places which prevents successful solidification I think(solidify is greyed out). These surfaces will merge but not solidfy. Go figure. Any help in resolving these surface discrepancies so that the punch can be solidified? Does anyone know if the surface model alone can be used for Pro/M or MasterCAMas is? I wonder if it will fail solidificationin CAM too.


I am not a sheet metal designer. As a side issue - does anyone know if there needs to be undersizing and over-anglingfor "creep" or "spring-back"in sheet metal for this punch? What percentage based on wall thickness if so?


2006-10-06_144708_draft_surf.zip

 

[marker4x4]

This punch geometry reflects the original piece-part fairly well - I think the worst deviation is about 2 thou...It's a lot of features. Problem is - the original copy geometry is corrupt in several places which prevents successful solidification I think(solidify is greyed out). These surfaces will merge but not solidfy. Go figure. Any help in resolving these surface discrepancies so that the punch can be solidified?...

Actually I didn't want to solidify the import, since that would add/subtract from the finished model of the tool that I'm trying to create here. The import is more for reference purposes, so I'm not worried too much about it being perfect (it's far from it, I know).

The springback depends on the material used for forming and it's condition (heat treated, annealed, etc.). There's engineering tables that specify those angles, try to Google it; I'm sure it's available somewhere on the net.

Thanks,

Edited by: marker4x4

 

[mechdes1a]

"Actually I didn't want to solidify the import, since that would add/subtract from the finished model of the tool that I'm trying to create here."


I'm not sure I understand your explanation, Mark. I am thinking you would need to copy geom the form surface geometry from the original reference part with the import featureinto another "form" part, make it independantand use it as a solidification to configure a tool to form this sheetmetal part. "Add/subtract" I don't understand. Any help?


You're welcome for the help. These kind of surfacing issues are interesting to me.they're a cut above the ordinary.


-chuck

 

[marker4x4]

I'm not sure I understand your explanation, Mark. I am thinking you would need to copy geom the form surface geometry from the original reference part with the import featureinto another "form" part, make it independantand use it as a solidification to configure a tool to form this sheetmetal part. "Add/subtract" I don't understand. Any help?

The use of the original, imported geometry is rather limited in this particular scenario; there's too many changes to be incorporated into the final design of those tools (hydroforming formblocks). There's offsets, springback angles, tooling holes, etc. - all that reduces the model of the actual sheetmetal part down to mere references. All you need really is the IML, bend radii, flange depth and angle. This sample fileposted at the start of this thread is quite simple; the only problem was to follow the isoparameters and that was solved by Jeff. Some of the more trickier ones can be quite hard to figure out, but in any case, the original model is there for reference only.





I hope it helps, cheers