*
Andrew:
Hate to disagree, but collar ties are more effective in the
upper 1/3 of the rafter, at least as far as space
consideration vs. structural strength go in protecting the
outward push of the sidewalls. I know it seems backwards, but if
the collar ties are securely screwed or well nailed, the top
1/3 area works out best.
Dog
x
Replies
*
Andrew-
I don't want to argue this point because in the past I
thought the same thing as you but I did some research on
this for a job where the ceiling needed to be removed in a
bedroom and raised considerably. What I found partially
agrees with you about there being more tension up higher,
but that is what the collar ties are there for- to handle
tension. Corollary-you wouldn't want your basement posts
near the walls just because there was less of a load on
them-you want them where the load IS.
A secondary benefit to collar ties is keeping the rafters
from sagging or developing independent movement in other
ways. Having them higher, again, works better for this
purpose.
Joists often do not help as much as you might imagine,
depending on how they are secured to the plates. On that bedroom job,
the ceiling joists each had just a single
toenail into the plate-they were notched to sit on the plate
as low as possible with no consideration for outward push.
A big key here is how the collar ties are attached. I screw
them with deck screws-if the rafters are really old and
won't take a screw very well you can use bolts (yes I know it SHOULD
be reroofed, another story)
Blackie, let us know what the company said. Good luck with the
problem.
Mad Dog
xxxxxxxxxxxxxxxxxxxxxxxx
*
You attach collar ties with deck screws? I wouldn't have thought that would be appropriate.
Rich Beckman
*Deck screws are a bad idea. They have small shank diameters, and aren't made for structural applications. Stick with nails..........
*How about a steel tension rod on the middle rafter pair? That pair would have to be doubled or trippled or more depending on the load and the walls below it would have to be strong enough. This would make the ridge effectivly have only half of the span and the 5x15 may be more than enough. The rod could be an interesting design element. A tensioned chain would be an alternative. Of course an engineering evaluation would be required (paid for by whoever screwed up????????). As long as the rafters are stiff enough to not over-deflect under the roof load the lower the tension member, the better.
*Ron, what you and others keep forgetting; and something I keep saying. Is that screws do not do the work. They squeeze the wood together, create a compressive stress, and friction holds one piece of wood to the other. Put a little glue in there and now the load is transferred over the entire area the boards overlap. Look at the compressive stress region as a cone with the peak at the underside of the head of the screw. The area of compression gets larger and larger until it hits the other face of the board. The math can be done and the optimal spacing can be found. It is possible to get all of the overlapped wood to be transferring loads.A nail draws wood together in compression for about a second. Ring shanks "may" cause it to stay under compression a little longer. The area transferring the load via shear is now the exact area of cross section of each of the nails totalled up.This is why four cheap screws and some of that stupid glue can do the work of ten nails!I actually wanted to build test pieces and have them tested in our mechanical lab, but got fired nefore I could. I've still got contacts there so it may still happen.-Rob
*I'm not sure I read all the posts because the browser is screwing up on my Mac but I'll weigh in (not much weight I suppose.I am consulting for an exact job like this. Though a lot smaller.The roof is 2x6 10 feet span (on their axis). Roof is 8 1/2 on 12. House is 15 feet wide in the gable direction, 32 feet in the ridge direction. The ridge is a 1x10!!!! No ties anywhere. This was a new roof built over the old one!!The old roof is 4x6 rafters spanning 9 or so feet. 4 on 12 pitch (originally). The old skip-sheathing was removed, but the old ceilings are (archeologically) L&P, beaded wainscoting, 2 layers of drywall!! There are literally 6 tons of dead weight spreading the walls apart. The walls deflected enough to topple the chimney!! About six inches in the center of the wall.What am I recommending for my project? Looks like a new double LVL 1 3/4 x 9 1/2 with a post in the middle. Collar ties only for hdidng the beam and not for structural reasoins. I would like to clear up a few things said above.1 - Nothing you can do will change the amount of "load" the rafters, plates or walls will see. All you can do is change the amount of deflection.2 - Collar tie placement depends on what you want them to do. If you want them to stop the walls from blowing out the lower the better, on the top plates is best**Scientific explanation follows**You may want to skip this.The technical term for needing collar ties is bending moment. It is expressed in ft-lb ( the number of pounds times the number of feet from the pivot point). There is a tremendous amount of moment at the ridge (if your walls can easily deflect outward). In structural engineering the connection from rafter to ridge and rafter to top plateand sole plate to subfloor is a "pin," free to rotate at will. Thus if the top of the wall will move sideways, the rafter will go along for the ride. The collar tie is another pinned connection accross the rafter. Due to it's installation it is in tension (in a gable roof). It is in place specifically to not stretch and allow the rafter to pivot about the ridge. By not stretching it wants to bend the rafter ( if the walls easily move). If the tie is near the ridge it is under much more tension to make the same counter moment to the rafters. More tension in the tie means more likely to stretch, more likely to be loose ( see diatribe above about nails vs screws), and more likely to deflect the rafter. Perhaps even to the point of breaking the rafter. So it is a complex mathematical problem to solve, but the boneheaded, easiest best answer is to set the ties on the top plates.Real quick – assume 1000 ft-lb moment at the ridge. If you put ties 1/3 of the way down (say 3 ft in a ten foot span) they need to generate 333 pounds of tension. If you put them at the bottom (10 feet in a 10 foot span) the amount of tension they need to generate is 100 pounds. The 333 pounds is a point load on the rafter, ants to bend it. The 100 pound load cannot bend the rafter, it is holding the rafter in compression which may bend it, but this is how the rafter is supposed to be loaded. The collar tie at the bottom adds no additional load.OK – so you don’t want to set the ties (wood cable or otherwise) on the top plates for various reasons – or say you don’t even want ties at all. Then you have to stop the relative deflections of the ridge beam (this was your original question) OR the walls.For the ridge – Easily, additional gluelams can be sistered on the sides of the existing. This is a simple no-brainer that requires a few temporary jack posts to hold the ridge straight (or preload it up slightly). You can tear it all out and add steel (pain in the ass) or add permanent posts at the center of the span. To halve the span you decrease the deflections of each half to a small percentage of the current deflection of the whole span. Specifically, if you halve the span of the existing beam, the deflections will be much less than half of the current delflection. The problem with this is that the beam is carried by the transverse roof rafters. If they are not heavily deflected yet they will only deflect slightly more due to the added dead weight of the additional gluelams or steel beam. You will not be adding much more load to this roof.For the walls – something must be done to stop the tops of the walls from moving outward. Ties or cables across the top plates is the easiest way to do it. The other way is to go to the outside of the wall and stiffen it horizontally so that cannot deflect, and allow the rafters to move outward, and allow the ridge to deflect. How would you do this? You need to have good sized soffits to do anything substantial. You can add (horizontally) plywood box beams, open web wood joists (floor joists), I-joists, steel bar joists, hybrid wood/steel joists or more gluelams. The other classic way to do this (as in roman empire) is to add buttresses. This is how they built soaring churches without using 8’ deep ridge beams. Your building is not large enough to warrant the buttresses.This is an easy job for an engineer and yours is apparently sandbagging. He may be doing a very thorough job because the stakes are high (large new house, banks, buyers and sellers attorney, etc.) I don’t blame him with our current litigious society his heirs will be liable for your roof after a tornado tears it of and your brother-in-law fixes it with scraps left over from other jobs.-Rob
*OK, I understood everything in his explanation ... up to "you may want to skip this."Why would you want collar ties high besides practical needs such as headroom?Frankly, i think some flying buttresses would enhance this project. It already has the roofline of Notre Dame. Is this seriously going to be someone's home?
*Rob, I can't go with ya on this one. Drywall screws are made without standards, have smaller shank diameters, and are generally made from the cheapest possible material. They simply aren't made for structural use. Do you really want to count on friction to hold your house together for the next 50 years ?
*Has anyone suggested running a cable betrween the walls at ceiling height? I've done this with large 40'+ spans in old buildings, barns, lodges, etc, with good sucess. However, it may be best to suggest thsi to the engineers first.
*Rob, I'm having a little trouble agreeing with your explanation of the screw and glue method of fastening the collar ties to rafters. In my neck of the woods, fasteners are considered to be either shear or tension; in calculating shear load for fasteners, any contribution from friction is considered nil and is not used to bolster the shear limit. By adding glue what really is happening is the glue is now being put into shear (and the wood fibers). The screws are adding some amount of shear strength albeit small if the the shanks & quantity are small. I guess it's the friction thing that is giving me some trouble.....Sam
*
Ron:
I did say deck screws, not drywall screws, as you converted
in your reply to Rob. These are large diameter screws. I
am open on this but still believe them to be better in this
application; more for holding power and keeping the collar
ties from gradually loosening rather than for sheer
strength, though that is certainly important. I guess I
felt like if it got down to a situation of sheer strength,
these fasteners might fail, but other things would have gone
first. This would have to be some heavy duty snow or wind,
and in the case of the job I did, the roof pitch was about
18 in 12, I think, and therefore snowload is not going to be
much of an issue.
Also, if you think about it, deck screws should compare with
framing nails driven through a gun. Which is probably what
a lot of guys would use in this situation.
One more thing: collar ties do function to combat ridge sag.
I tried to scan back to refer directly to the post above
stating they don't, but couldn't find. Sorry to be vague.
The best thing to eliminate the outward push of the walls is
indeed, the joists, provided they are well secured. Take
them out, next best thing are collar ties, properly attached
and loca
*
This type of connection is commonly used in structural steel work: high tensile strength bolts are torqued high enough that the friction between to the two steel plates, rather than the shear across the bolts themselves, are used to resist the forces put on the connection. Given the unstable nature of wood in comparison to steel, I'd have thought it would be a tough job to engineer this type of connection in wood.
*Ron - re-evaluate that last question of yours. Everything that is held together is held together with friction.Deck screws vs drywall screws I don't have information on. I think I'll do this test just for fun.-Rob
*Barry, Can't say too much about structural steel work but when it comes to airplane design, shear joints utilize fastener capability, pattern and hole preparation to determine shear load and fatigue resistance. Fatigue resistance isn't something that most residential construction is going to need to account for but shear load of course is. As noted in my previous post, any contribution from friction is considered nil. I wasn't aware of this sort of joint being used with steel structure; interesting.... Sam
*Their was a snippet in FHB or elsewhere about a guy who was seriously injured when the ceiling joist he was sitting on gave way. The hangers had been fastened with 4 drywall screws each. Oops.
*My understanding of why drywall screws are no good in sheer situations (BTW, what is an example of a non-shear situation?) is that they are drop forged. Poured into a mold, instead of coming from a long wire, like nails or wood screws. Is this inaccurate? - jb
*All intellectual theorey and logic aside, with the way wood moves, cups, dries, and otherwise changes over time (is it fair to expect a house to last 200 years?) I would have a hard time relying on vertical friction for much. I believe I under stand how mechanical fasteners, certain joints, and gravity hold a building together. I would be very leery of lookin' someone in the eye and sayin' "yup, friction". - jb
*Blacky,After looking at the picture you posted, let me make sure I understand this. Because the glulam was undersized, it sagged in the middle causing the rafters to push the walls out of plumb. Do I have that right? Am I even close?JonC
*
Blacky...what was your final solution?
*Blacky. Having worked extensivley with large rafter framed roof systems, I have been through the mill with architects and others about what works and what does'nt. It sounds like you wher shooting for A structural ridge beam and that did'nt happen. On many log homes, you will see A collar tie bearing dirrectly on the wall topplates, this holds the downward force of the roof system and prevents the walls from "shitting out". I secure my collar ties (ussually 14" to 16" round logs or solid 12x12 timbers ) with large metal plates fabricated by any good welding shop. My plates are wide enough to lap over the top plate, inside and out, three inches, rest on the top plate, ussually 51/2", and rise above the top plate the thickness of the collar tie, provided your heal heighth will allow. If your heal heighth is'nt enough , you can drop your collar tie down in A pocket in the wall, or notch your collar tie top. The span of 25 ft. is'nt that great of A span, so top plate to top plate collar ties would work, the only real hassle of not having pre planed collar ties from plate to plate is getting them in if you are already sheeted. Good luck.
*
Well finally as of today I got some additional lumber I need to place two more beams running paralell with the ridge at half span so the beam manufacture did all the final engineering on the job is suppling the beams and a crane to slide them through a window eeeeek but wheres theres a will theres a way so it looks as if all I got to eat is the labor and the loss of time but all the dilema has really impressed the homeowners on how I acted to correct the problem and they are going to give us enough work to keep us out of the snow and the additional beams will look hot especially with the large window surrounded by the beams so now all thats left is gettin them uggggh in the window with the crane roughly 1200#s each thanks for all the input it was greatly appreciated I especially liked the flying butress idea but it didn't really go with the scheme of the project the boss lady said I'll post a picture of the job in the end
*Sre you saying they are run as purlins at the halfway point of the rafter span. Or are they sistered to the ridge beam?-Rob
*the solution (or as far as the engineneers deem is two 5x23 mid span ridge beams roughly 2/5 of the way down roof from the ridge
*Hoorah! Uh, where do you buy 5x23's?
*well progress finally (and a bunch of side jobs later) the two 5x25 x 32' beams were placed today well not quite in place but in the general area now they just have to go from 8' to 15' up tommorrow gawd they look huge on the high side they will show 29" (after a filler strip is installed) again thanks for all the input and a pic will follow after there finished
*
Amazing that the original ridge was to be 5x24 and now your fix involved TWO 5x25's. Thanks for keeping us posted.
*Blacky, for a guy that really abuses the "run-on sentence" it looks like you did a good job of taking care of your customers. Glad you got it handled.
*
I've got a undersized gluelam spanning 32 feet 5x15 when alas it should have been a 5x24 eeek well two different engineers have give me two ideas one with steel and another with microlams
its a 12/12 pitch 16"oc 2x10 25'rafterspan(wall to wall) 30#snowloading with plaster seems no engineer wants to use collarbeams is this feasible I was thinking 5 like mid rafter heigth please some opinions
*Do you mean collar ties (horizontal conectors of opposite rafters) ? Your frrter span, if I understand you correctly is actually something less than 12' - the horizontal (level) span from wall pate to center below ridge. How are your rafters connected to the ridge ? Is it cathedral ceiling or a conventional attic assembly ? What is your roofing system ? What type of wall system do you have? These need answers before anyone can reply.Assuming conventional attic and stud framing:You have lot of accumulated weigth in your ridge's engineered lumber and your rafters' dimensional lumber. Collar ties will stop the bearing weight of your roof structure from spreading your walls and they will stop your rafters from spreading at level they are connected (if you faster them well) but they won't stop your rafters from sagging under the load nor will they stop the ridge from sagging gable to gable. Collar ties recycle the load back to the rafters.Are your rafters butted into the ridge or sitting on top of ridge? If you combine colar ties with plywood ridge gussets you can turn the system into a truss. Suggesting you back up the system up with steel and more glue-lam seems over doing it structurally, cost wise and weight wise. Diagonal bracing on either end, starting at about the 10' point on the ridge and conected down to your top plates would split the ridge's load. You can shorten the ridge brace point to reduce room interference and you can you use posts on the gable end walls to lift the height of the lower connection point.
*Blacky,
Joseph FuscoView Image
*ok 2x6 16"o.c. wall framing 40 yr. shingles no ceiling joists all open to supposed bearing ridge assembly no collar ties at all as of yet all rafters are nailed together with solid blocking to each other and solid blocking to gluelam also toenailed to gluelam
*here is a lousy picture but it might help clarify things a bit there is a large window at the right side that would prohibit placement of any diagonal bracing also if at all possibile I am trying to keep the view to the window as clear as possible
*Blacky - Collar ties don't reduce the amount of load the ridge beam is carrying. If you've had engineers look at the house, I'd go with one of their recommendations.
*But collar ties are an alternative scheme to a ridge beam, right? (I'm asking innocently here, not rhetorically.) With collar ties the rafters brace each other against tipping (hopefully) and bear the roof weight fully on the side wall top plates -- if done properly here, the ridge could be degraded to a mere nailer. Or the ridge could bear some of the load down the gable walls, the plates taking the rest on the side walls -- if the structure below is set up for this arrangement. The side walls as designed and built already have to hold up the 50% of the roof that the ridge doesn't, maybe they can't take more.The problem I saw was that joists halfway down aren't collar ties. Ties have to be farther down -- their purpose is to prevent the plates from spreading (a simple truss). Collar ties, it's true, have nothing directly to do with bearing roof loads, but by keeping the roof from blowing apart the walls they do enable it to bear its weight on the side walls. Halfway down would bow the rafters possibly to breaking. The rule of thumb I recall is to put them in the bottom 1/3. Undersized rafters will bow with either the ridge or the collar ties.Collar ties can be added to every other rafter in pairs or otherwise done decoratively if they won't reduce headroom too much. If they would, how about a scissors truss? Unfortunately none of these solutions is going to give the cathedral ceiling you are likely after.I've wondered about hybrids like this. Probably your engineers know best, but they may not be sensitive enough to considerations of cost. I don't suppose you could field "laminate" another gluelam to the one you have to get the appropriate strength? You saved some money getting a small one right?
*From the looks of this picture I hope you didnt finnish framing. Who decided on that size beam? Was it just guessed at? How did you figure out it was to small. You may be better off ripping it out. and replacing it with a proper one. If you have a good relationship with your local supplier he may be able to take the old one back, or know someone who would buy it.- Rick Tuk
*well unfortunately it was some sorta snafu at the manufacturer money was no object while this was bid out or at least it was totally up to the cost of the right beam (which I thought I had recieved) and yes its framed roffing siding soffit etc (I only found this problem while installing a shower unit and found out the walls were pushed out of plumb 5/8" over 8')'tis a good thing it was before my board hanger came so unless I get any other good ideas the steel is going up this week eeek oh well progress must be made get in get out and next time have the manufacturer sign off (and I'm still not done with them)
*Andrew: Hate to disagree, but collar ties are more effective in the upper 1/3 of the rafter, at least as far as space consideration vs. structural strength go in protecting the outward push of the sidewalls. I know it seems backwards, but if the collar ties are securely screwed or well nailed, the top 1/3 area works out best.Dogx
*Andrew,That was my thought on the collar beams however it seems tough to find a engineer that will collaborate with me but today it took a turn for the better the beam manufacturer finally is going to come and look at the beam why there coming to look is beyond me I guess to confirm my suspisions? so hopefully they will agree to do a lil more indepth structural analysis of the whole dilema sooo I'll just have to be patient a 'lil bit more but Hey slow and steady wins the race right?
*Good luck. If the glulam mfgr is this interested maybe THEY will help engineer a solution. They certainly are the best qualified to mickey mouse the undersize glulam!
*The higher they go, the more tension they get -- just like the farther you put a nut into the nutcracker, the more force on the nut. Higher is advantageous for headroom, but harder on the rafters. It can become impossible to get enough fasteners in there to resist the load. The ideal "collar tie" for the rafters is probably the ceiling joist.
*
ok finally beams up and in place. It took a crane and 10 guys to get em in they were cut 1/4"shy of the distance from exterior OSB to the oppsite stud face tight fit for such a large chunk of wood. I wanted to thank everyone again for there input. when its completed I'm gonna send a pic to the magazine perhaps they will print it:)