Join the party to join the eyebrow!

[Gene Davis]

These are tough for me. Getting the curved swoop to join into the flat pitch is just not easy, as Chief is unpredictable doing this.

But, maybe there is a preferred "best practice" that makes it fall into place. I hope so.

Two pics are attached, plus the Chief X2 file. As can be seen in the Chief render, the three curved sections are built and joined to each other, but are not joined to the 6.5-pitch flat plane that surrounds them. The other pic attached, shows the roof solution, as done in Sketchup.

If you are good with this stuff, download the X2 file, join things up, and in your reply with your completed model file attached, tell us the principles of doing this roof join.

[Glenn Woodward]

Gene,

How does this look.
It still needs a bit of clean up either side where the 3 curved roof planes meet (the apex of the small concave roofs)- but it's 95% there.
In plan view, I edited the opening in the skillion roof by eliminating 2 of the corners down from the top and then snapped the apexes together.
I then just used the Join Roof tool to join the roof planes.

I think the problem with the junction of the lower roofs is that they are drawn icorrectly. ie, the apex of the small concave roofs doesn't meet the other roofs at the correct location.
If you go around and click each roof plane the roof meeting point crosses are all over the place meaning that the roof planes are drawn incorrectly - or maybe the roof planes will never meet because of the geometry.
It looks like you have drawn guide lines at where you believe the roofs intersect.
The more accurate way is to let Chief do it by clicking on 2 roof planes and then looking for the point that Chief placed. This indicates the intersection of the 2 roof planes - and it doesn't coincide with your drawing.
You can then use Chiefs auto temorary points to snap the roofs to them.

[ebdesign]

Good night............

[ebdesign]

Playing w/ this some more this morning..........
As soon as I angle the baseline, it becomes virtually impossible to get a "perfect" union display in both 2D plan & 3D cameras. Doesn't seem to matter whether using a single roof plane for the "hump" or two different roof planes.
Lowering the facet angle to "3" (lowest it will go) gets it real close, but.............not perfect; which is what Gene is after, I guess. I suspect that only a facet angle set to "infinity" would produce a perfect join.
This whole exercise was "educational". I hadn't done a sloped eyebrow like that in a while; plenty of "flat" ones......Anyway, back to the real work.
Jim

[Gene Davis]

Wow! Thanks, Jim!

It is interesting, looking at your solution, how it seems necessary to digress from precision, in finding a way to get this group of roofs to join smoothly and render well in 3D.

By "precision," I mean that the fillet arcs would have a "start bearing angle" of zero, where the fillets come tangent to the level fascia lines of the surrounding 6 1/2:12 flat roof. Those fillet arcs would then be tangent to the large top central arc of the 'brow, thus the bearing angles where the fillet arcs end, and where the central arc begins, would match.

Precision would also require that the fascia height at the low end of each fillet would exactly match that of the level fascia line of the surrounding 6.5-pitch roof.

Your workout has small digressions, with the fillets jumping from flat at almost 2 degrees, and with the fascia tops down 3/16" from the level ones adjacent. Your fillets join the top arcs not in tangency, but with angles that differ by almost 2 degrees, again. You split the top into two, and those two aren't tangent at top. Almost, but not.

I'm sure you didn't jump to these deviations from precision right off the bat, but used iteration to arrive at your solution. Would you mind saying how you worked this out? How did you find that the two-degree non-tangency would work? How about the fascia tops not matching at point of tangency, but differing by 3/16"? Did you have to go both up and down before you found out what worked?

I find, as you did too, that eyebrows join with pitches when the brow arcs have level baselines, but when you go and tilt the baselines, as we do here, things get interesting and unpredictable.

Three pics are attached here. One shows your solution, and you can see that the top arc is two sections, with the joint visible at top. Another pic shows mine, before I saw yours, and you can see the problems I encountered when springing the fillets at zero bearing and with matching fascia tops.

This is a case where imprecision is best. Sort of counterintuitive, but what the heck.

The other pic shows how this roof is used in a shed dormer, with the workout done in Sketchup. I know you can use Chief for just about anything, but when you want to warp the ceiling underneath this brow so there is no conic-section "valley", and do it with trusses, it was easier to go SU. At least for this non-pro Chief user. Construction docs will include details done in SU and copied and pasted into Chief.

[ebdesign]

Well, Gene, you're right. I didn't take the time to make tangent the short "wing" CAD arcs in the elevation. I checked in after the football game, saw your post & played w/ it for a little less than an hour.
The "hump" is perfect. 104" baseline/fascia, 112" ridge & both have a 148" radius. Those roof planes were "joined" at the ridge w/ the join roof tool. All the roof planes were "joined". But the short negative radius wing roof planes..... I just "eyeballed" those arcs in the elevation. You could make those tangent & use that radius & they would be "more better".
To keep things simple, I changed your base flat fascia to 100" w/ 104" & 112" the other key numbers; & laid it out per the CAD work in plan & elevation.
I rarely use the 'angle at ridge" & "angle at eave". I get all my roof properties from the CAD work. The roof planes snap to the CAD work & then I enter the properties I want.
w/ flat baselines, the roof fascia/ridge hgts would have all been exactly per the CAD work, but w/ angled baselines, I suspect this causes the slight change in values (99 13/16", etc); also the less than perfect roof joins; probably because of where the baselines start/end.
The only times I use the "angle at ridge/eave" is when I have to join two different curved roof planes, like a full roof return (24" wide or so) across a gable wall to a full roof plane. Then I would just match the "angle at eave" of the roof return to the main roof plane. Otherwise I stick w/ entering radiuses per the CAD.
One thing that makes this particular exercise a little more difficult is that the base slope is the same for both roof planes on each side- 2/12; even though I never once entered a value in the slope field.
Here's a tip for rejoining curved roofs a that already have the arc-ed valley lines. Straighten both edges out before rejoining them. The line-to-arc/arc-to-line tool will be available in the edit toolbar if that roof edge is curved. (see attached)

Jim