I’m a homeowner in the process of building a 26×48 workshop with a floor under the gable 12/12 pitch roof. My question is about framing the roof. I have read numerous opinions about the need for collar ties and am confused. If the rafters are placed along side the floor joists on the top plate then there appears to be no need for collar ties. However, if I build the second floor like a cape where the rafters sit on the floor system then how does that effect the need for collar ties? Any metal connectors appear to be more for uplift at the eave as opposed to restraining any tendency for spreading of the rafters. The walls below would still have the floor joists to resist spreading. Thanks for any opinions. We are located in Washington state with seismic considerations.
Replies
You need rafter ties at the bottoms of the rafters to keep them from spreading unless you have a bearing ridge.
Collar ties are farther up and are loaded in both compression and tension to help the rafters do their job of supporting the roof.
If your rafters are sized for the span and snowload, you don't need collar ties. [EDIT: oops, yes you do. Didn't think of the wind load the other way. Piffin learned me something again. ;-)]
If you have a non-bearing ridge, rafter ties are required by code. Attaching the rafters to the joists makes the joists acceptable rafter ties.
If you attach the rafters to the deck instead of to the joists, you need to make sure the connection can handle the thrust of the rafters. I don't frame that way so I can't give you good advice there. A couple of the guys more expert than me do it that way and I'm sure they will be along to give you solid advice.
Best of luck with it.
"Let's get crack-a-lackin" --- Adam Carolla
Edited 4/16/2006 10:47 am ET by philarenewal
The problem I see is that the floor of the loft will only hold the rafters as well as it it staked to the rafters. Collar ties are (supposed to work) in tension and need to be fastened that way. You state that the tiedowns are for more seismic considerations and that appears true. I suppose there could be something that does both but the inspector would need some convincing for that.
Are you building it all at once? or is the loft floor to be an "added later" type of feature? If so use really have to use the ties.
The problem is the walls bow, the ridge sags, and every fastener gets stressed. In fact if you are using hangers off of the ridge beam the toe (top) may make the bottom of the rafter cam out and slip down. It isn't worth the risk
Is there a certain place along the vertical path of the rafter that a collar tie must go to be effective, or is it within a range?
I have read articles and letters to the editor in Fine Homebuilding that debate this issue. Most taking opposing stances. For example FHB#140 has 2 letters to the editor with contrary opinions. One a structural engineer that says his analysis shows a collar tie is only effective within the first 1 foot or so of the plate on which it sits. I also have seen houses framed with the roof system on the floor deck(not sistered to the joists). My house in Connecticut was frames this way, as were all of them in the neighborhood with the collar ties at about 2/3rds height. I've also seen examples in FHB done that way. There must be a "correct" answer somewhere. I've also seen articles about using "finite element analysis" that showed ties under compression. What then are the forces at the rafter end?
The misunderstandings come in when there is a lack of knowledge regareding terms and definiitions and purposes.A Collar tie is by definition a horizontal member placed in the upper one third of the rafter asssembly. It it to tie the rafters at the collar - the upper portion where uplifting winds have been known to unhinge roof setions at the ridge. Southern states subject to wind damage require collar ties.The subject of this discussion is not a collar tie but a rafter tie which ties rafters to prevent them from spreading when attached in the lower one third of the length. Cieling joists typically fulfill the function of rafter ties. rafter ties are un-necessary when the rafters are supported by a properly designed structural ridge beam.So to answer the question where should it be placed to be most effective, one must know what is the right name for that memeber and what are you expecting it to do. maintaining the shape of the roof by keeping the rafters from spreading, the walls from leaning, and the ridge frrom sagging - by all means the rafter tie should be on the walls or at least no higher up on the rafter than the lowest third of it.. If you want to keep tornadoes and hurricane force winds from openning the roof at the ridge, use collar ties in that upper third, or hardware designed for the same purpose such as strap ties
Welcome to the Taunton University of Knowledge FHB Campus at Breaktime. where ... Excellence is its own reward!
Excellent response. Everything one would need to know about the situation in a few words.
Excellent explanation. What, in the frame design, makes a ridge load bearing ?
The absence of rafter ties would alllow the rafters to spread apart at the bottom, sliding on their connections or pushing walls out - say with a catherdral cieling design. Then the ridge must be a structural member designed to resist the load applied to half the roof, the other half loading onto the walls
Welcome to the Taunton University of Knowledge FHB Campus at Breaktime. where ... Excellence is its own reward!
OK, so if you use a wide enough ridge and/or maybe a gluelam and the ends are supported to take their share of the vertical loading of the roof then that's the difference.
Don't you still have to worry about the walls spreading if you don't have the joists to tie them together? I can see how the ridge must take the vertical load but how does the peak not spread as the walls are pushed outward? Is that the purpose of the collar ties?
Thanks for explaining this. "With every mistake we must surely be learning"
The only way for the walls to spread out is if tjhe ridge sags. A properlly selected / designed ridge beam wil not sag beyond its allowable limits
Welcome to the Taunton University of Knowledge FHB Campus at Breaktime. where ... Excellence is its own reward!
Piffin - From the way I read Mike's reply, I think the he is assuming that a ridge beam is a wide and heavy roof ridge, and not something lower that the rafters rest on.Tigger
Right. I wanted to know the structural difference between a loaded and nonloaded ridge and you guys made it clear. Yesterday I found and read a JLC book on structural aspects of framing and it also clarified it. Seems that if you have a large triangle formed by the two rafters and the ceiling joist then the ridge becomes a nonloaded member. The loaded ridge is practically a simply supported beam with a distributed vertical load; if it's not wide enough it will bend. When I get the chance I'll look into the basis for the collar tie location rule.
I suppose I should go to the basement and find my college statics book, but the book's moldy smell would make me feel too old.
If anyone wants to collaborate a FHB piece I'd be willing to help.
"With every mistake we must surely be learning"
" if it's not wide enough it will bend. "actuially, it is the depth, not the width that provides the strength to resist the loads.
Welcome to the Taunton University of Knowledge FHB Campus at Breaktime. where ... Excellence is its own reward!
Yeah. We're talking about the same dimension. The depth in the loaded direction. My mistake. (M x c) /I = deflection
"With every mistake we must surely be learning"
First, thanks for all the feedback to my initial question. One thing that seems abundantly clear is that nothing is abundantly clear on this issue. Does it need to be?There aren't a lot of house with walls bowing outward but in the aftermath of a hurricane ther appears to be plenty of analysis that says construction methods and design were at fault. Let me just note that I'm a transplant to Washington from Connecticut and thats why we're doing it all over again.
Let me recap my plans. 26x48 garage/shop. 10 foot ceilings with 14" I-joists to give a clearspan on the first floor. The second floor is fully covered with 3/4" subfloor to create the diaphragm needed to transfer forces to the shear walls. The rafters are 12/12 resting on an additional 2x6 plate and meeting at a ridge board(not a bearing ridge). Collar ties at about 8' above the floor. This is essentially a copy of how my garage with living space above was built by the builder back in CT, although on a smaller scale(24x24). I thought I wouldn't reinvent the wheel and use the same plan on this project. Living in Washington State has made me look more closely at all the seismic and wind requirements and question that design. Interpreting the code and how to apply it is akin to reading the tax code--10 accountants would probably have 10 different amounts on your refund or how much you owe. On my previous workshop I wired the place myself with a 100 amp subpanel. I had an electrician come in for an opinion before I started and he used terms like "I think you can get away with--". I'm pretty sure there may be more than one way to do it right but his response was less than confidence building. Back to original problem--I've seen roofs point loaded with tons of weight when the roofers had all the shingles loaded at the ridge--it survived that stress test. The forces created by a distributed and asymetric loading such as snow and wind create different problems. Given what seems to be a simple structure- a triangle-the analysis seems immense. FHB should do an article on this with a variety of inputs. I'd like to see real world experience and scientific method combined.
If the ridge is self supporting, then the rafters are the thing that actually hold the walls from spreading.
You should get an engineering drawing or numbers done on the type of ridge and its size and the rafter size for the span of everything. They should have the charts at your local lumber yard where they in turn get them from the company that supplies engineered lumber. Better yet, you can pull everything right down from the internet, just go to the company's website. I think I used Truejoist the last time I did this. I recommend an engineered ridge if it's to be self supporting.
A couple other things, you should use "hurricane connecters" where the rafters meet the top plate for a better shear connection, I think they are required by code in a lot of places for this type of framing. Also if you're insulating this, check the required R-value for cielings and use the proper width of rafter to accomodate it, even though it may be structural overkill. Around here it went from 30 to 38 in the last few years.
One minor point of clarification: if you have a ridge beam (structural ridge) as you say, the ends are supported - thereby forming point loads at each end of the beam. These point loads must be carried all the way down to the foundation and are supported with solid material.
This is not to say that you coundn't have a window or door centered on the gable end as in this case, the point load would rest on the opening's properly sized header and then the load transfered to the header's jack studs (probably more than 1 per side of the header) and again, below these jacks solid material all the way down to the foundation.
Maybe we need to come up with a FAQ-type thread about collar ties vs. rafter ties one of these days.Kinda like the "Floor Vibration" thread I refer to all the time...
If a case of the clap spreads, is it then considered a case of the applause?
With that sugggestion, I've marked this to my subscrinbed list so I can find it again someday
Welcome to the Taunton University of Knowledge FHB Campus at Breaktime. where ... Excellence is its own reward!
I was thinking about a thread with some pics/sketches. Heck, it might make a good article for FHB. Maybe I'll work on some sketches when I have some spare time. Like next winter if the guys down in Costa Rica have run me out of a job...
If your brain itched, could you scratch it by thinking about sandpaper?
Excellent description. But hey, you taught me a lot of what I know.
There is one more purpose for a collar tie in compression. It actually will help to transfer loads in the vertical load path between the opposing rafters. (both sides want to deflect "inward" -- the collar tie in compression will help resist that).
"Let's get crack-a-lackin" --- Adam Carolla
Piffin - In my 'almost' house the ceiling is open. LVL beams will be supporting the roof load at the top. But because I'm in hurricane territory I also have to have collar ties even though it's all strapped. I would like to have put them about a 1/3 the way down but yesterday my engineer said I needed to stay withn 2 feet of the top. So I guess I have the answer even if it isn't the one I wanted. "Oh, sigh."
Tigger
There is metal hardware straps designed to take the place of collar ties at top. See if your engineer is aware of them
Welcome to the Taunton University of Knowledge FHB Campus at Breaktime. where ... Excellence is its own reward!
Piffin -
"There is metal hardware straps designed to take the place of collar ties at top. See if your engineer is aware of them"He's specked SIMPSON LSTA18 straps across the top connection but he says that FLa code also requires the collar ties. I don't have a problem with that I just would rather have had them a bit lower to hide some ducts.
Edited 4/16/2006 5:58 pm ET by thetigger
See if he will allow metal straps under the ridge for "collar ties" and then put the wood ones down 1/3 for decorative purposes.
Or, what about installing 2 sets of collar ties. Call one set collar ties, one set ceilings joists.
This is where we depart from the text...as they say in "Goodnight Moon."
Yep my construction prints and a bunch of sources put it 1/3 of the way down from the top. I guess I don't know why.
Speaking from a physics perspective the fasteners that connect the collar tie to the rafter would be the least stressed when attached at the bottom of the rafter where it meets the wall. If all the collar ties were trying to do was stop the spread of the rafter bottoms that would be the location to install them.
Obviously that spread isn't the only reason collar ties exist.
Maybe to hold the top of the opposing rafters (at the peak) in connection.
Maybe to strengthen the jointery overall. ie to keep the triangle (rafter pair) from being affected by uneven loads of snow or wind.
Maybe it got decided in the bar based on the most beguiliing length of a skirt.
As Piffin describes so well, the use and the description depend on more than just a rafter in space. The ridge beam is a mighty factor. I can only assume that there are varying degrees of load on the ridge beam as well.
Was the original poster building this from some plans or winging it? I'd love to take this to the mat. Tradition makes for short conversation. Facts or TrVth last forever.
Jack of all trades and master of none - you got a problem with that?
Edited 4/15/2006 7:54 pm by booch
Unless there is some compelling esthetic reason not to install collar ties, I'd go ahead and put them in. If you did every other rafter pair, and pushed them nearly all the way up to the ridge board, you would be looking at maybe $20 in material and 1/2 hr labor with a nailgun, step ladder and a circ caw. A lot of the time there is enough scrap laying around the job to take care of it meaning material cost could be zero.