Custom Roof Faces

[cwailes]

When creating custom roof faces, does it matter where I click to draw the roof axis? Will that cause my roof faces not to line up with one another? I created a complex roof and for the most part all roof faces line up correctly. However, I have a couple that if you zoom way in are slight misaligned. Is that because I did not use the same distance from edge when I clicked for the roof axis?

[CipesDesign]

Yes, it matters. The "Bearing Height" is calculated exactly at the line you draw to define the slope direction. For this reason I often draw the original shape without overhangs, thereby being able to draw my slope line exactly along the edge of the object during creation. I then add the overhangs as desired using Add Surface.

[cwailes]

Thanks. That is what I thought but wanted to make sure. I assume that is why some of my roof faces don't line up?

[CipesDesign]

Probably. Another cool trick: even if your roof is too complicated to create using one single "Roof Object", start out by using the "Create Roof" command, then ungroup the roof. You will end up with a bunch of roof faces, all perfectly aligned and ready to be added to, cut from, duplicated, rotated, etc. to achieve the final objective.

[Jonathan Pickup]

Also remember, you can edit roofs with the 2D reshape tool to remove, add or move the position of roof edges. This is one of the tricks in my new Vector-workout Guide to Productivity. http://www.archoncad.co.nz/training/vector-workout.php#productivity2009

[ionw]

The other thing the axis line determines is the edge angle on a square cut miter. Downslope of the axis line the edges are square, upslope they are vertical.

Cheers,

Ion

[CipesDesign]

Wow, thanks ION. You DO learn something new every day ;-)

[mike m oz]

I'd also like to add my thanks to Ion. I've also learnt something new.

[M5d]

Prior to the unified view the bearing height of a roof face corresponded to the design layer it was on, now in 2010 they're defaulting to the drawing's Z origin zero regardless of the design layer's Z height. The new behaviour seems to contradict the purpose of the design layer's Z height, is this right or have I overlooked something? Floors still behave as expected taking zero to be relative to the design layer they're on.

[CipesDesign]

M5d, I did a quick test with 2 design layers, one with the setting: z=o/delta z=8'; the second with: z=8'/delta z = 8'. I created a roof face on design layer 1, with a bearing height of 8'. I then moved the roof to design layer 2 and, as expected, it moved up 8'. So I am not seeing any change from previous versions. The bearing height of a roof face is still tied to the "z" values of the design layer on which it resides....

Did you double check that your layer "z" settings are what you expect??

Did you try it in a new clean file, as a test? This is always a good idea...

[mike m oz]

I've just checked this and M5D is correct. The roof face bearing height is being measured from the absolute z origin rather than from the layer relative origin. For example if the lowest layer has a Z height of 10 m and the roof face is created in a layer with a Z height of 16 m with a bearing height of 2.5 m the roof face will be created in that layer with a bearing height of 2.5 m relative to the absolute Z origin. The OIP actually says the Axis Z height is 2.5 m.

He should bug list it.

A roof object created with a bearing height of 2.5 m is created at the correct height in that layer.

[CipesDesign]

Aha, this is an interesting problem. The roof face seems to be at the correct relative (expected) height, at least if it is created and moved as per my previous post, however the OIP is reporting the absolute z value as opposed to the (expected) relative number. I will tinker some more (I bet Mike will to!). Obviously some sort of bug, but perhaps a fairly easy one to squash.

[CipesDesign]

OK, after a bit more experimenting I would say that there is bug exactly as Md5 described in the original post. Roof Faces are created at and report their bearing height in relation to absolute z as opposed to relative z. Easy to work around for the time being, but should be fixed (and I'm sure it will be fixed) soon.

The most interesting thing to me is that I usually set up my Roof Layer(s) to have the same Z values as the Plan Layer on whose walls the roof bears. Therefore I would never notice this bug in a single story building!

Also, try this: create a Roof (not a roof face) with a bearing height of 8', on a design layer that has values of (for instance) z=8'/delta z=8'. The Roof is created as expected (unlike the Roof Face would be, as we've discovered). Now, ungroup the roof. Note that while the resultant Roof Faces do not change height, the OIP reports their bearing height value as 16' (which is absolute) instead of the original Roof Object's 8' (which was - correctly - relative).

Very interesting.... I love problems like this ;-) Thanks Md5!

[M5d]

Thanks for having a look. I experimented with the Create Roof command and it's behaving in the same manner as the Roof Face command, unlike Peter's. The Australian edition is yet to see SP1, so maybe they're already onto it if it's not an intentional change. I'll do a bug submit anyhow.

[CipesDesign]

M5d, you are correct: the Roof Object is also reporting the wrong (absolute) value(s) for bearing height.... Thanks for bringing this to everyone's attention ;-)