rafter tie/ collar tie?

[dug]

First off, are rafter ties and collar ties the same thing? I thought rafter ties were to prevent the walls from spreading and collar ties were for strengthening rafter spans, but I'm not sure. Just wondering.
Anyway, I got my deck built for 20 by 30 1 & 1/2 story and am still trying to hammer out some design decisions as I go along. Rainy days are good for something!
So... loft will be 12 ft. by 20, leaving 18 by 20 where I want an open cathedral ceiling. Plans call for a rafter tie every 4 ft. at top plate in this section. Like many here, I am modifying my plans here and there and am considering collar ties apx. 2/3 down the rafter instead.  This would be for aesthetic reasons, I have seen some done this way on this forum, and I kind of like the look.  I realize that rafter ties attached to top plate and rafter is a stronger connection, but how big a difference would it be? I am using 2 by 10 rafters, and snow load here in So. New Mexico is practically non-existent. I'm not asking for any engineering advice, just your valued opinions.

[MaineRhino]

I figured the loft would act to tie two walls together (collar tie?) and then placed rafter ties (?) near the top, about 6" from the top on every rafter. That helped with the install of the 1x8 T&G too. Sorry, this pic does not show those...





Finished look......

[MountainDon]

A collar tie is usually in the upper 1/3 of the attic space. They are primarily used to resist uplift forces on the roof. They are not always used.

Rafter ties are normally at the upper wall plate but my be moved upwards into the lower third of the attic space. When they are moved upwards that can affect the sizing of the rafters, it changes the way the roof loads work, but with 2x10's there should be no problem there.

There may be other variables to be considered in some cases but that's it in a nutshell.

[rwanders]

Don, am I correct that if you use a appropriately sized and supported ridge beam instead of a ridge board the need for collar and rafter ties changes considerably? Especially helpful if you want unobstructed cathedral ceiling look.

[dug]

Don, am I correct that if you use a appropriately sized and supported ridge beam instead of a ridge board the need for collar and rafter ties changes considerably? Especially helpful if you want unobstructed cathedral ceiling look.

anxious to hear the answer as this was another question I had. I think this is an option for me if I understand the concept correctly, I have a solid center beam under my joists.

[MountainDon]

Don, am I correct that if you use a appropriately sized and supported ridge beam instead of a ridge board...

Absolutely! A properly sized ridge beam along with the required structure to carry that load down the end walls to the foundation will open up the ceiling space. You might still need some side to side ties to help with wind loads, but that's what engineers are for, unless one can find ready made plans to work from.

[Don_P]

A collar tie is usually in the upper 1/3 of the attic space. They are primarily used to resist uplift forces on the roof. They are not always used.

In the '06 and newer code collar ties are now required, I suspect this is in the aftermath of the midwest tornados in ('04?). Either way they are a good idea.
"Collar ties or ridge straps shall be connected in the upper third of the attic space...
Collar ties shall be a minimum of 1x4, spaced not more than 4' on center"

If you look at the bottom of the rafter span tables there is a formula to figure how much to reduce the table rafter spans according to how high you raise the rafter ties.

This is wandering off into more than anyone wants to know but it came to mind, so what the hey.
Strictly speaking a kneewall situation already throws this into an engineer's hands so go carefully as you raise ties, there is more going on there than most realize.
In old european timberframes a high posted cape ran into this. I've seen pictures and analysis of the kneewall posts bowing outward and splitting when they crossed the line. As they evolved they developed some unique ways of redirecting the forces. One approach was the French Jambe de force "leg of strength". Notice the legs running from the collarbeam down to the tiebeam. It acted as a brace and a compression strut. This both supported the roof and redirected a large portion of the thrust down to the tie area. Just more info for under the hat.

[dug]

I have been thinking a lot about this lately and have come up with a few ideas I'd like to run by all of you.

I case none of you have seen this it is an excellent post by DonP that clearly illustrates forces at work regarding roof thrust.

http://countryplans.com/smf/index.php?action=post;topic=8008.0;num_replies=6

This got me thinking about what could potentially happen if I raised my rafter ties a few feet. From what I can envision this is what might occur in a worst case scenario, greatly exaggerated of course.



So I thought, what if I doubled the rafters at those points, thereby doubling the strength of that weak point?

The reason I want to raise the ties is because I am trying to create a partial timber frame look and plan on using rough cut 4 by 8's or 10's. The sawmill nearest me. which is quite a ways away, will only mill up to 16 ft. lengths. Also it would work out better (with them raised) from a design standpoint because there will be a raised floor (15 in.) under this section.

So, thinking about how to connect this I pondered this



sandwiching  the beam between two joists. Seemed to make it easier to make a better connection, and also strengthen the weak point between the rafter tie beam and top of the wall.

Overkill? Not good enough? All opinions appreciated!

One more thing- If you were going for a modern look in a structure (I am not) would a tensioned cable work instead of a rafter tie? Just a thought.

[poppy]

dug, I thinks maybe this calls for the "hammer beam truss."  There are several ways to do it, here is one.
http://images.google.com/imgres?imgurl=http://www.heirloomtimberframes.com/3d/trusshammer.gif&imgrefurl=http://www.heirloomtimberframes.com/eng/shop/glossary.htm&usg=__LK4HQhi9t_Bcxu20YF42kKEa4Q0=&h=155&w=320&sz=10&hl=en&start=44&um=1&itbs=1&tbnid=8aTa2UGcfGQXhM:&tbnh=57&tbnw=118&prev=/images%3Fq%3Dhammer%2Bbeam%2Btruss%26ndsp%3D18%26hl%3Den%26rlz%3D1T4GWYE_enUS288US288%26sa%3DN%26start%3D36%26um%3D1
It was invented to get high ceilings without really long timbers.  I'll bet Don_P even has the design calcs. for it.

[dug]

I noticed the link I posted about rafter thrust didn't work so good. I'll try again-

http://countryplans.com/smf/index.php?topic=6732.0

Poppy- I would like to do something like that, as a matter of fact I would have preferred to build with a supported ridge beam but I am trying to strike a balance between my limited experience and my ideal design. Plus I need to get it built ASAP.

I do want somewhat of a timber frame feel however, and am thinking of also piecing in a strip if 2 by 4 rough sawn lumber at the peak to create a faux ridge beam. I also want a sound, sturdy structure so if I get the feeling that raising the rafter ties up a bit is too much of a compromise I will not do it.

[Don_P]

I'm sticking a gold star on my monitor  

The best advice I can give on hammerbeams is don't unless you have some way of buttressing the walls, they do thrust, much more than most designers have given them credit for. Great in a heavy stone walled church, dubious in a lighter framed structure. One timberframe engineer described them as a perfectly acceptable truss with the bottom chord cut out. There was a lengthy presentation on reasons why not to build a hammerbeam. If a house had 6' thick stone walls or wings on each side of a hammerbeamed greatroom, I would say go for it. If its just a rectangle with tall lightweight walls you'll be visiting a hammerbeamed roof later. Notice how many have a steel tension rod or cable either fitted or retrofitted (that answers another question dug).

If this is a cathedral area, I assumed the rafter design you drew is to be spaced 4' apart so it doesn't get too busy up there. Assuming 10psf dead load, the weight of the roof itself, and 10psf live load, the minimum allowed design live load. On a 12/12 pitch with ties across the walls at 4' oc I come up with about 400 lbs in the rafter/tie connection. Raising the tie about 2' off the plate to where a 16' tie works is below the 2/3 mark, so its code, it has around 520 lbs of shear in the connection. 1 rafter and 6 or more nails into the tie would work if all those assumptions were right. You can play with rafter spacing or the build to do some neat things if you want. Is the exposed tie the limit of the effect you desire?

[poppy]

Thanks Don_P for giving dug a more complete answer.

What I probably should have pointed out was that the link pointed to a modified hammer truss, rather than the traditional open design.

Another alternative is the scissor truss like in this link.  The scissor truss might make more sense for dimensional lumber anyway.
http://images.google.com/imgres?imgurl=http://www.charpenteriegs.com/photos/scissor%2520truss.jpg&imgrefurl=http://www.charpenteriegs.com/fermes_en.htm&usg=__GnRsssF4HYTUvJnsAZiS5OBv2Zo=&h=190&w=361&sz=9&hl=en&start=42&um=1&itbs=1&tbnid=wKGxNcR_g1Y3vM:&tbnh=64&tbnw=121&prev=/images%3Fq%3Dscissor%2Btruss%26ndsp%3D18%26hl%3Den%26rlz%3D1T4GWYE_enUS288US288%26sa%3DN%26start%3D36%26um%3D1

I'll give you 2 gold stars. 

[Don_P]

Dug's picture above brought an article on a timberframed kneewall cape to mind. The picture is of the "finite element analysis" of one bent of the frame under uniform loading. On rereading the article he points out some interesting things. In the FEA models the heavy lines are members in compression, the light lines are members in tension. Notice the raised collar "tie" is no tie, it's in compression.



In this model he used tying braces runing from the upper floor to the top of the kneewall. That diagonal brace is overstressed. He next experimented and removed the tying braces in an attempt to get the floor and raised tie to work together in tension to resist the roof's thrust.  In this case the raised tie remained in compression, the bending load and deflection in the wall post became unacceptable. He did get this into "marginally tolerable" range by increasing the roof pitch, lowering the kneewall, lowering the collar, upgrading and upsizing the posts and rafters.

Hopefully that helps explain a bit more from another perspective without muddying things too much.

Edit; there's some good articles on timber trusses, including scissor trusses here;
http://tfguild.org/publications/historictrusses.html

[Satinder]

Did you end up moving the rafter ties up? I am working on a similar project. These posts are turning out very helpful.