high ceiling building

[struengineer]

Question,

I am designing a residential house and owner wanted a high ceiling study room. Study room is 16’ long x 16’ wide. Collar beams (connecting both side of rafters) are supposed to be located at 2/3 roof high. How do I considerer the lateral load due to raise of collar beam? Can the vertical bearing wall be used to resist the lateral load? If so, how to analysis this structure? Can anybody help me?

Thank you.

 

[msquared48]

The ceiling structure does not resist the lateral force from the wall, the roof diaphragm does. The rafters and collar ties may define the ceiling, but the roof diaphragm is the lateral structural element here.

The placement of the collar tie on the rafters has nothing to do with the lateral capacity of the structure. The collar ties only help the roof structure carry the vertical loads from the roof.

Mike McCann
McCann Engineering

 

[AlmostPE]

Go back to the drawing board. By seeing you saying collar tie makes me think you dont have a ridge beam ( only a weak ridge board to nail rafters on). Todays residential design, you either need to specify trusses (not the historical collar ties trusses) OR stick frame with ridge beam.

 

[apsix]

The typical method to deal with this situation is to provide a ridge beam supporting the upper end of the rafters.

 

[struengineer]

Thank you guys.

I did think of both solutions (ridge beam and truss) but have questions for both.

Ridge beam:
The mid point of my 16’ ridge beam will allow l/240= 0.8” settlement. Because roof is 5:12 slope, 0.8” settlement in the middle means the wall in the middle of building will move out at least 0.8”. This is not acceptable. Even I use a very big ridge beam with L/480 settlement, the wall still move out about 0.5”. How do I handle this problem?

Truss:

I believe that truss acts similar to a ridged arch. 2 pin connections at the top of both walls.

My collar beams are only raised 2’ (1/3 roof high). Should the structure (rafters + collar beam) work the same way as truss if rafters are big enough? However, the structure analysis for this structure seems to be very difficult. Is there any short cut for this kind of analysis? How do I decide the rafter size?

 

[apsix]

I'm confused about your geometry:
- The room is 16' wide
- Roof slope is 5:12 (vertical:horizontal?)
- Collar ties are at 1/3 roof height = 2', so roof is 6' high.
- Previously collar ties where stated to be at 2/3 roof height.
- Ridge deflection = wall lateral deflection; which is only true if roof slope = 45deg (assuming both walls move out).

Not all of the statements are compatible.

If roof slope is 5:12 the wall lateral deflection is about 40% of ridge deflection; you should be able to make it work.
If ties are at 1/3 height there is a greater chance rafter truss action will work; contact your local truss manufacturer.
Alternatively model in your favourite analysis software, checking first for deflection and then looking at strength. Could even look at 3D structure combining ridge beam and truss action.
If the roof is steep you could look at diaphragm/truss action in the roof planes.

I hope I have helped and not confused.

 

[AlmostPE]

Maybe I've been doing it wrong but I never consider lateral deflection on the wall if I use ridge beam. If you use ridge beam then theoraticaly it shouldnt kick the wall out. Just design the ridge as L/360 and call it good. You shouldnt really use collar tie. Either specify scissor truss to get your ceiling height (let the manufacturer worry about the lateral deflection) or use the ridge L/360 ridge beam I suggested.

 

[struengineer]

Sorry I did not descript my geometry clearly. I try to explain more clearly here.

The room is 16’ x16’. Roof is about 7’ (16x 5/12, say 7’) high. Collar beam (same direction as rafter)is 2/3 of 7’ from the peak and 1/3 of 7’ from the top of wall. It is about 2.3’ above the top of the wall.


The amount of wall moves out (s) is calculated as the following:

(H original high, L original ½ span, d: deflection):

Ridge deflection d = 16’ X 12/240 =0.8”

(L+S)^2 + (H- d)^2 = L^2 + H^2 (Rafter length remain the same)

L^2 + 2LS + S^2+ H^2-2Hd+d^2= L^2+H^2

S^2 and D^2 very small, say=0

S= H/L *d = 7/8 * 0.8 = 0.7”

By the way, I do have Staat Pro. However, I don’t know how to model this situation. I try to regard all member as an elastic member with E= 1.3x10^6, use pin connections for all connections. The result is also unacceptable. I don’t know what is wrong.

 

[AlmostPE]

You are making it more complicated than it is suppose to be. You really dont need any computer model to do ridge beam and rafters. I feel like you need someone to teach you how to design a stick framed house. Like I said before, you shouldnt specify a collar tie. The only time i do collar tie is when I use a heavy timber truss to support the ridge beams (for the look).

 

[struengineer]

Thanks AlmostPE.

The collar beams are needed for the vaulted ceiling so that gym. Board can be secured on them. If I used ridge beam, collar beams do not have their structural function. However, if I treated rafters + collar beam as a truss, I believe that it will act as a horizontal member to stop outward movement.

You suggested selecting a strong ridge beam so that deflection is L/360. Even so, my top of wall still move out ½”. That is noticeable for an 8’ wall. Isn’t it?

I just like to find out what is wrong with me. I saw a church near me with ridge beam span over 25’, the wall is not out of plumb, what is theory behind it?

 

[AlmostPE]

If the room is 16 ft long, the ridge beam is 16 ft long. If the ridge beam deflects L/360, there is no way the 16ft wall will deflect more than the ridge beam. You should be looking at horizontal wall deflection if you use ridge beam. Just make sure the ridge beam has L/360 or less and call it good.

The horizontal deflection of the wall caused by the deflection of the ridge beam is called p delta effect (2nd order). You shouldnt analyze p delta for simple stick framing. Plus, I dont think you calculated it right. If your ridge deflects L/360, there is no way a wall with same length will deflect more than L/360 in the middle.

 

[AlmostPE]

correction, "you shouldnt be looking at horizontal wall deflection if you use a ridge beam"

 

[struengineer]

AlmostPE said: you shouldnt be looking at horizontal wall deflection if you use a ridge beam

But why not? If ridge beam deflect, the only way this structure deform accordingly will be this way: peak of roof settle and the top of wall move outwards horizontally. There is no any force to prevent this deform (if we do not consider the collar beam). Deflection (L/360) is 16X12/360=0.533”, the top of wall will then move outward 0.533 x 7/8 = 0.47 “.

Is it possible the ridge beam will never deflect L/360 if we consider the roof diaphragms work together with ridge beam? If so, how do I evaluate the real deflection? Can I select a smaller beam with deflection L/240?

 

[AlmostPE]

I dont know whether your calc is right or not but based on your result you have less deflection on the wall than the ridge. That means your horizontal wall deflection is better than L/360 which I think pretty darn good.

 

[msquared48]

Typically, I have found that you will get much more lateral deflection at the exterior wall top plate if you use a collar tie arrangement than a ridge beam. In fact, uyou may have to increase the size of the rafter, or lower the collar tie to control it.

This small deflection should not be an issue with the ridge beam arrangement. None at all.

Mike McCann
McCann Engineering