ACCEPTED SOLUTION
Accepted Solutions
And if the facets are not desirable... and you want a true round top, this was the best I could come up with:
The corners are blends, and the planar features are the flat triangles.
Did you want the Creo 2 part of either of these?
And if the facets are not desirable... and you want a true round top, this was the best I could come up with:
The corners are blends, and the planar features are the flat triangles.
Did you want the Creo 2 part of either of these?
The problem with your model appears to be in the boundary blend definitions. If you flatten just one of the triangle, and place a point along a long edge, you will see the end of the sheet is no longer tied to the point at the square. Lack of tangency control can do this.
Using the same curves, I made single blend with control points. Couldn't get a full single feature, but over half defined properly. The extrude is squaring-up trim. A point along the straight edge allowed for a reliable flatten quilt.
I am going to assume so. I am working the boundary blend with control points. I cannot get a 4-arc circle to join completely with the 4 straight sides, but I can get 3 to work. So I have to trim the boundary blend. But those are properly developed and tangent faces prior t the trim.
I suppose you could create the flat triangles with Fill feature.
That way you have something to make the corner boundary blend features tangent to.
I always forget about simply trimming suffaces.
Oh so many cats to skin 
The more I work with this geometry, more I see development throwing in random "factors" depending on the core geometry. In this case, I added fills with tangent boundary blends on the corners. When this is flattened, it adds radii at the corners and the round is "tighter".
For this object, no deformation is required, yet it puts it in there... and arbitrarily at that.
I see... then yes, the sheetmetal version will probably work best for you.
I cannot seem to modify the 2 flatten quilts in my Creo 2.0 as you have them defined.
Is this a Creo 3.0 file or WF5/Elements Pro? It opens in Creo 2.0.
I have noticed that flatten quilt builds a "plane" from the edge at or near the point used in the flatten quilt. If it is a good straight edge, it does pretty well. In the case of the fill/boundary blends, however, it takes some serious liberties with the tangent merge along the two surfaces.
Have you considered just forming a round tube with a square die 
The attached file is working pretty well. Is the same as the previous with the garbage removed and to re-forming.
initially i made it in creo elements/pro 5.0..but the latest one i uploaded is in creo 2.0.
i do not have creo 3.0...they stopped our maintenance mid-way..how cruel is that. 
PTC should really look at some of these saga challenges we come up with.
Including recording the time it takes an engineer to actually do the challenge, and not find some obscure work-around or tell us that this is not how -anyone- would ever do -anything-.
Finally got it using a two-stage flatten process. I thought I had it in one, but it left a triangle behind. For some reason, it doesn't recognize the split surface and maintain the rest as a single quilt.
I turned off the hidden flatten quilt UI so others can use this without it enabled.
PTC is like an Indian goverment office..they will only work if a lot of pressure is put on them..in this case a big customer asks them to put in some functionality..otherwise you have to do with what you have been given
Of course, after the mirrored group, you can flatten each half and merge them.
But most importantly, the feature creation method creates a reliable result apparently with very little distortion.
The edge is still 220 and the arc length is within .001 of the original.
And it works well flatten-quilt deformation if you needed to add something in the flat state.
