can't get tangency at seam of opposing cylinder extrudes

[pimm]

I have a fairly simple parting faces quilt that I need to finish out but I can't get the boundary seam of the opposing cylinders to become tangent.  After completing the parting faces quilt I need to use the thicken option to uniformly thicken these faces to build a gutter model; as is I can only thicken the quilt to about .125 where it will need to be .500 in thickness.

Shown below is the area I am struggling with.  The highlighted edges show the area that I am working on.  I have penciled the non-tangent seams in red.

I have attempted different options from Boundary Blend to the Style function.  I have attempted this with and without curve rails in the red highlighted areas.   I know that there naturally is an area of twist due to the conflicting up facing and down facing cylinder extrudes.  Might there be some kind of option that would iron out the natural seams to allow tangency throughout? 

(I also tried this as a simple 4 sided boundary blend but the tangency toggle failed on the left side which has 3 edges)

[BH_12246893]

Can you display curvature of the curves/edges that are highlighted in red, particularly the left side?  Are those three edge tangent and curvature continuous to begin with?  Any chance you could share a stripped down model (even a .stp or .igs)?

[pimm]

The curvature plot does show the areas of concern.

With the above example I did not build this with any style curves but the surface set on the left of the cylinder transition was made by boundary blend and the surface on the right of the cylinder transition was made by style surface.

I believe that I have attached a STEP model of the above quilt.

[Van_AG]

The green edges are not tangent at the highlighted points

And there is a 9µm gap at the point

You should fix these issues

Then surfaces with tangent joints can be created.

[pimm]

I have been working at this for a lot more time than I feel comfortable with.☹️

Thank you for pointing out issues with my tangencies.

As far as the gap is considered I believe that this might have been introduced by the conversion of the geometry from Creo to STEP and back into Creo again.

I do believe that the error of tangencies must be from my construction technique.  At 1st consideration I thought everything should have tangency because each of the faces inside of the rectangular highlighted region (shown on my initial post) was created using tangency.  1) I built a boundary blend between the 2 opposing semi-circles using tangency on the 2 edges. 2&3) I built the side faces in the rectangular region by using tangency on all of the edges.   What I didn't see initially and see now is that the edge joints that you have circled did not have a tangency curvature that follows the straight through edge continuation.  I would likely not have that issue if I used a Style curve network which would take that straight through lined up with the beginning and ending edge with true tangency.

Even though I would like to solve my submitted problem I am seeing that this likely is an exercise in futility.  Even if I could get a nice tangency fed through every face within the rectangular region it would have to suffer at the transition zone where the upward curved faces meet the downward curved faces.  Moreover, because of that I will not be able to use the Thicken function to get the needed .50 thickness.

If there was a way to minimize the transition distortion that would be great, but I do think that it is time for me to punt with my original concept. ☹️  Sometimes the issues of doing something a certain way is not apparent until after fully going down the rabbit trail.   If someone can prove that this is possible that would be awesome, but I won't hold out any hope that this is possible.

[BH_12246893]

I played with your model yesterday and before I spent too much time, I noticed that the tangency issues were not that hard to address; however, the resulting curvature would not have allowed the offset/thickening that you wanted.

[pimm]

Yes; I was hoping that there might be a way of removing tightness from the side to side transition, but the more I looked at this the more unrealistic this appeared.

[Van_AG]

Please see the attached model (Creo8)
The surface can be thicken up to 5mm into any side
I slightly rebuild imported STEP

[tbraxton]

Using the STEP data, I am able to create tangent connections across the inflection of the "cylinders". There are issues with the geometry in the STEP file in the context of what you need for the design. The best option would be to create the geometry incrementally while checking the ability to offset it to the needed thickness.

It appears that the inflection of the cylinders is not the issue but the topology of "base" surface. This may or may not be true for the native Creo source geometry.

In this image tangent edges are displayed as dimmed, so all of the grey edges are G1 or higher continuity connections to adjacent surfaces.

Based on the STEP geometry this boundary (in red) is particularly problematic (the quilts do not intersect/touch).

[pimm]

In the Creo geometry I do get a seal in the pink highlighted area so unfortunately the STEP model has loosened that edge.  I was able to thicken this area of the model when doing a solidify remove to my needed size.  (Even though the curvature is not 100% smooth on that edge in Creo).

I personally have had more issue with the tight convergence of the rectangular boundary with the opposing curvatures as it tightens too much with offset.

I am taking a different approach which will hopefully give the needed thickness and blends that is a lot different than my original submission as it negates the need to criss cross opposing surfaces.