32x48 timberframe plans.. In need of some review

[Ernest T. Bass]

Hi all! This is kind of a lot to ask of you guys, and we may get an engineer's help to finalize our plans, but I'm still in the design stage and would like to make sure I'm on the right track..

The plan in question is a timber frame for a 32'x48' cordwood home with a gambrel roof. Here's the tentative first floor plan:

Note the measurements at the top, indicating the bent spacing. That one space says 13', but it's been modified so none of the bents are distanced more than 12' apart. The big weird thing below the kitchen is a wood cook stove. The wall behind it is only 3'-4' tall, and their may also be a wood stove facing the living room, back-to-back with the cook stove.

Here's the loft:

I'm mostly posting for advice regarding the timber frame, but comments or critiques about the floor plan are also not only welcome but appreciated!

Here's what the cross-section of one of the "bents" might look like. I say "bent", because we won't exactly be framing in typical "tip up the bent" fashion; more like one post and beam at a time. Also, not all of the joints will be complex works of art... Anything beyond a dovetailed loft joist and we will probably settle for metal plates and bolts to save time.

The rafters are standard 2x12's on 24'' centers. The main thing I'm concerned about is the offset 2nd floor post. However, I've seen this done in books many times, and it helps to shape the gambrel roof. I'm planning on 3''x6'' knee braces on all of the interior posts like the one pictured (I drew only one out of laziness), so that will help transfer the weight of the roof to the 1st floor posts. Also, the 8x8 cross beam is only notched 2'', so it should retain most of its shear strength. It seems strong enough to me, but the calculations for the shear+whatever strength the knee brace adds are beyond me.. The wood will most likely be spruce.

Moving right along, here are the 1st floor posts.


Here are the beams and 3''x6'' loft joists.

Sorry about that first beam looking so goofy.. A SketchUp pro I am not! The beams connecting the interior posts are 8''x8''s.. The one center beam supporting the ends of the loft joists is 6''x8''. I have no knee braces pictured but am planning on them... The cordwood infill will brace the exterior framework, but I'm feeling that a house this big should probably have braces on the interior posts as well..

And finally..

With the rafters and some sort of sheathing I'm thinking I won't need to put knee braces on the upper posts and beams. Would that be okay?

Sorry for the length of this post and the complexity of the questions.. Any advice would be greatly appreciated! I can also give more info if I failed to be descriptive enough.. Our snow load is in the 70 psf range.. I'm probably forgetting some important info.. That's the problem with late-night posting!

Thanks a bunch,
Andrew

[Whitlock]

Not that I'm a expert but I think that summer beam needs to be a 8x12 instead of a 6x8. The frame should competly support itself the cord wood is just for filling. I would say you need knee braces in both derections on the upper posts and beams. If you are worred about the posts under the 2nd floor off set you could gun stock the top of them or use bigger posts. You need a engineer to look at this I think he is going to make you go bigger on a lot of your beams

Nice job on the design it is a step in the right direction

You can also get some answers and ideas at this web site might even be a engineer for timber framing that will give advise on the forum[noidea'

http://tfguild.org/

http://www.tfguild.org/forums/ubbthreads.php

[Ernest T. Bass]

Thanks for the info; I'll post my load calculations for the beam sizes later, but I just wanted to ask why you think knee braces are needed on the 2nd floor post and beams? It looks like the rafters would brace the posts in one direction if they were attached well, and the diaphragm bracing of the roof sheathing and/or metal roofing would brace the frame in the other direction.

[Ernest T. Bass]

Okay, check my math regarding the summer beam in question...

(I'll repost the picture for easy reference.)


It spans less than 12', but I'll round it up.. It supports 3.5' worth of loft. 12' x 3.5' x 47psf = 1974lbs total.

The beam is under a two point load.. At 12' with an allowable deflection of 1/360, my table shows an 8'' deep spruce timber at 386 lbs per inch. 1947lbs/386 = a beam width of at least 5.11''.

On the other end of the house, where the loft spans the full width, there is an extra post under the summer beam (next to the bedroom door) to aid with the extra load.

Whilst I'm at it, I'll present the calcs for the other beams as well...

The main 8x10 girders running both sides the length of the house..
They support 6' of roof and span no more than 12'.   6' x 12' x 92psf = 6624lbs. Because of the steep pitch of the roof, I can multiply that by a reducing factor of .45, which equals 2981lbs. The beam also supports 4' of loft, which is 2256 lbs. (In reality, the live load of the loft will be very minimal next to the eves, with this roof design.)

5237 lbs total. A 10'' deep timber under a uniform load and spanning 12' is capable of supporting 1028 lbs an inch, meaning a 5x10 is really all that is needed, but there will be a porch roof added on as well.


The 8x8 girders simply support 6.8'x12'  worth of loft.. 7.5' x 12' x 47psf = 3835lbs. At 386lbs an inch for a 1/360 deflection, the beam is undersize, but by allowing a bit more deflection an 8x8 is fine. I think that allowing 47lbs for a loft load is generous, so I'm not too worried about that...

Anything incorrect with my logic? I could very easily be mistaken in my thinking here...

[Don_P]

I'm not sure I'm understanding correctly. Can you highlight the beams in question with various colors and the text in your posts with the same colors?

I'm assuming "Eastern Spruce", that would be Black, Red ,or White Spruce?
If so, in #2 grade it has an Fb of 575 psi, E=1.0, Fv=135 psi.
Plugged into a two point load equation here;
http://www.windyhilllogworks.com/Calcs/2ptbeam.html

You asked about the upper posts on the floor offset from the posts below. They are close to the posts below. If you count on the braces check out the posts for combined compression with a side load. I'd be tempted to sum the upper post's point load and the uniform floor load together, move the load point to the right a proportionate distance and checking the beam as a single point loaded beam. Would you like a calc like above for that equation?

I would agree, if the roof diaphragm is providing the bracing the upper posts do not need kneebraces. If it's board sheathing then its a different matter.

I'll check in when I freeze again, busted a lift arm on Fergie

[Ernest T. Bass]

I'm not sure I'm understanding correctly. Can you highlight the beams in question with various colors and the text in your posts with the same colors?

I gave that a try.. Does it help?

I'm assuming "Eastern Spruce", that would be Black, Red ,or White Spruce?

You know, I'm not even sure.. I don't know my wood that well, and it'll be coming from a private party. Does it really make that much difference?

You asked about the upper posts on the floor offset from the posts below. They are close to the posts below. If you count on the braces check out the posts for combined compression with a side load. I'd be tempted to sum the upper post's point load and the uniform floor load together, move the load point to the right a proportionate distance and checking the beam as a single point loaded beam. Would you like a calc like above for that equation?

If it isn't too much trouble! You're the expert..

I would agree, if the roof diaphragm is providing the bracing the upper posts do not need kneebraces. If it's board sheathing then its a different matter.

What about a heavy metal roofing? Is that considered an adequate form of diaphragm bracing? I've noticed very large pole barn roofs that have no other form of bracing, but perhaps that's just bad practice. We don't have to worry about hurricanes or earthquakes though..

I'll check in when I freeze again, busted a lift arm on Fergie

Hmm.. Bummer. Sounds like a relaxing way to spend a Sunday. Thanks for the help!

[Whitlock]

With timber framing the beams do a lot more that support a load the summer beam will tie the building together so it is in tension also not just compression.



Good luck,W

[Don_P]

Awesome! Thanks   
I put 1000 lbs on each of the points in this calc;
http://www.windyhilllogworks.com/Calcs/2ptbeam.html
At full dimension 6x8 it failed, at full 8x8 it passed. If there is joinery we're not out of the woods though.
I'm pretty sure these are your spruces, the range between species looks to be 25 psi in #2, not huge.
I'm trusting your assessment of the loads on the far end of the building.



I think it is reasonable to call the loading here uniformly distibuted; http://www.windyhilllogworks.com/Calcs/beamcalc.htm
I failed slightly in bending, again without joinery. If there are sheds then the slope reduction is unconservative, I don't take those but do bump the Fb 15% for snow typically... I didn't here since you had already told a whopper  .


At my design values I needed 1/3 more section modulus to pass, a 9x9 made it.

I am using current #2 design values which is pretty typical, I'm leery of calling it better unless you are trained to grade or have some really clear stock.

Yes the metal roofing is considered a diaphragm when attached according to spec.

You'll need to recalc the remaining section modulus after joinery.

I'll post it if I'm successful on the other calc, please go ahead and figure the heaviest area's post and floor loads.

The tension in this direction is minimal, you have resolved the infamous thrust issue with the purlins in the other direction. These are really bending members rather than tension or compression members. Nice building Whitlock.

[Whitlock]

It's not mine Don just showing that knee braces were used in this aplaction.

[Ernest T. Bass]

Awesome! Thanks   
I put 1000 lbs on each of the points in this calc;

Wow, isn't 1000 lbs pretty high? The 3x6 joists are on 4' centers. They span 7' from the blue girder to the red summer beam. 3.5' x 4' x 47lbs = 658 lbs per point. (And I think considering 47 psf total load for a loft is pretty conservative..)

I think it is reasonable to call the loading here uniformly distibuted; http://www.windyhilllogworks.com/Calcs/beamcalc.htm
I failed slightly in bending, again without joinery. If there are sheds then the slope reduction is unconservative, I don't take those but do bump the Fb 15% for snow typically... I didn't here since you had already told a whopper  .

I agree about the porch roof somewhat negating the slope reduction factor; I hadn't thought of that, but I'm surprised that your calc is showing such different results than the tables I'm using, which are from several reliable timber framing books.. I'll have to come back to that one..

At my design values I needed 1/3 more section modulus to pass, a 9x9 made it.

Yeah, those should probably be beefed up a bit. However, most of the spans are closer to 10' than 12', and the beam will not be notched at all, except for the knee braces underneath, which will actually help support it.

I'll post it if I'm successful on the other calc, please go ahead and figure the heaviest area's post and floor loads.

Do you mean the weight of one of those 2nd floor posts? Using the same process as before I came up with 11788 lbs.

Thank you very much! Cool building there, Whitlock. Looks very stout.. Another example of the offset posts.

[Don_P]

Awesome! Thanks   
I put 1000 lbs on each of the points in this calc;

Wow, isn't 1000 lbs pretty high? The 3x6 joists are on 4' centers. They span 7' from the blue girder to the red summer beam. 3.5' x 4' x 47lbs = 658 lbs per point. (And I think considering 47 psf total load for a loft is pretty conservative..)

I divided the 1974 lbs between the two joists, ~1000 lbs each. Yes you can figure the moments more accurately.

I think it is reasonable to call the loading here uniformly distibuted; http://www.windyhilllogworks.com/Calcs/beamcalc.htm
I failed slightly in bending, again without joinery. If there are sheds then the slope reduction is unconservative, I don't take those but do bump the Fb 15% for snow typically... I didn't here since you had already told a whopper  .

I agree about the porch roof somewhat negating the slope reduction factor; I hadn't thought of that, but I'm surprised that your calc is showing such different results than the tables I'm using, which are from several reliable timber framing books.. I'll have to come back to that one..

What design values are you using, or which books are the tables from? The upper roof will dump onto the steep roof. That will crash down to the shed roof transition and pile there. Next snow it will do it again and the pile will grow. I have that situation on my barn.

At my design values I needed 1/3 more section modulus to pass, a 9x9 made it.

Yeah, those should probably be beefed up a bit. However, most of the spans are closer to 10' than 12', and the beam will not be notched at all, except for the knee braces underneath, which will actually help support it.

The weak link is the one that fails, it is not typical to count on the braces in the load calcs. We can massage the data but in the end this is an attempt to model reality and apply a margin of safety. We aren't going to change reality, its mighty easy to unwittingly eat into our safeties. I stay very conservative and let the engineer earn his money shaving it down based on his experience.

I'll post it if I'm successful on the other calc, please go ahead and figure the heaviest area's post and floor loads.

Do you mean the weight of one of those 2nd floor posts? Using the same process as before I came up with 11788 lbs.

Thank you very much! Cool building there, Whitlock. Looks very stout.. Another example of the offset posts.

Thanks for the post load, I had it up and running but crashed when I tried to save it. I'll try again tonight.

[Ernest T. Bass]

I see what you're saying about erring on the side of caution... I'm just trying to keep the timbers at a manageable size, since there will only be a few guys working without heavy equipment.

I've been using tables from 'The Craft of Modular Post and Beam' by James Mitchel, 'Building the Timber Frame House' by Tedd Benson and James Gruber, and 'A Timber Framer's Workshop' by Steve Chappell.

I have a table for basic stresses from the Forest Products Laboratory that lists dry spruce at 1600 psi in fiber bending, 120 psi in shear and 1.2 million psi for modulus of elasticity. I'm not arguing, just trying to figure out what's safe without going too overboard.

[Don_P]

I think there is a link to the NDS Supplement on one of the calcs I posted. These are the design values recognized by code and all engineers working in wood.  Benson's engineer and head of operations, Ben Brungraber, was part of a roundtable after giving presentations along with the head engineer from AF&PA who publishes the NDS. A short article by him preceeds the NDS design values in the Guild's "TF Joinery & Design Workbook". I can see how the values were derived by Mitchell and Benson, they speak to an evolving understanding at the time. Don't use them, they are above Select Structural grade (clear) for a sawn heavy timber. Bensons description of working through the design process is excellent, just use current accepted design values. I suspect the FPL table is specifying those values for uniform loading of "dimensional" stock with several adjustment factors included. The FPL among others does testing of small clear samples (2x2's) and those numbers are the tables in chapter 4 of the Wood Handbook, that is their role in the process. The numbers there are nearly useless to the average person for anything more than comparison between species. The grading agencies using the methods described in ASTM D245 and working with the AF&PA then derive the allowable working stresses for each grade and publish them. They have backchecked these design values by doing "In Grade" testing. Breaking loads of graded lumber.

Notice the difference in design values in the Supplement between dimensional lumber and heavy timber. Heavy timber has not been tested in large scale. The shots I've seen from the lab when they do are mighty impressive. In one series the technician is diving behind the equipment as half a flying 8x12 ricochets into the ballistic glass. It is harder to grade boxed heart heavy timber and there are fewer alternate load paths if a timber fails. If a 2x joist fails, there are others close by. If a timber fails a large load gets thrown to its few neighbors. These are all reasons for caution.

I've been through the training to grade lumber. In lumber I can roll the board and visualize with confidence just about all the defects. Think about this. Branches grow from the heart, they grow radially from heart to bark. Many are pruned by one thing or another at some point in the growth of the tree and are then covered by new growth. Knots are one of the big factors in grading. In a piece of dimensional lumber if it doesn't include the heart and if I identify the grain orientation as I approach it, I can tell you everything about the knot I'm looking at. I know how much sound wood surrounds it and what it's grain orientation is like. In a boxed heart timber there can be more lurking under the surface that I have very few, or no, clues about. Most stress is carried on the surfaces of a beam (an I beam tells you how little is needed in the middle) so there's a grain of salt. Your eye is still capable of saying a whole lot about the load carrying abilities of a timber.

It is not the knot so much that weakens a timber, sure there is essentially missing wood, but it is the slope of grain surrounding it that plays a larger role. If the grain swirls substantially around a knot and then we saw across that grain in making the board, the short grain will break relatively easily. So here's the other end of the spectrum; a log is not manufactured, a saw has not cut across any of the grain. Even if it swirls around a knot it is a continuous strap of fiber. Unsawn round logs carry the highest design values, eastern spruce;Fb 1400, E 1.2, Fv 135. Saw up to 3/10 of the diameter off the top and Fb drops to 1150 in a #1. Saw any more and you just made a heavy timber.

Notice the difference between #2 eastern spruce beams and stringers and #1, big difference. I don't assign anything above a #2 unless it is a very sweet timber. I hope this explains some of the reasons behind the values and which to use. This is the biggest variable and requires the best judgement.  We can do elegant calculations and in the end what I'm hoping is that the timber you choose has an ultimate strength at least double what the numbers say, (remember the one that broke above).

I don't consider you to be arguing, its part of developing a feel for how the design values relate to the stick in front of you. What I'm doing I hope is confirming where I can and directing you to the resources your engineer will be using. The supplement and those calcs are from the NDS. Good judgement allows you to adjust the variables and I don't mind if we disagree on those details. For me on residential work, a card carrying grader is going to mark on the timber the grade and thus the design values.

[Don_P]

Seems to be working;
http://www.windyhilllogworks.com/Calcs/AnyPtLd.htm

Make sure the left and right sides sum up to the total span. I'd check the post at correct location and then drift it right a bit while watching moment, deflection and shear.

[Whitlock]

Following Steve Chappell's book and you will have as good of a idea of what to do and how to do it

Cool web site Don

[Don_P]

waay off topic but this got me thinking. On our higher elevation peaks around here we have some red spruce. The tree by all rights shouldn't be growing here. These are remnant colonies from a colder time left stranded by natural climate change. Quite a few are now dying, acid rain got some and the southern pine beetle got some others in the Nat'l Forest that I got to saw up. Nice trees, the really good ones are making music as the tops of guitars. If you read "Clapton's Guitar" those spruces have a walk on part. The local guitar maker got those, I sawed the other ones  .

[Ernest T. Bass]

Thanks for taking the time to explain all that to me! I tried to download the NDS Supplement, but it failed to open for me.. I'll try and find it somewhere else.

I would prefer to work with round timber as much as possible; I feel that nature knows how to make a strong chunk of wood regardless of the knots in it, and sawing into the grain severely weakens it like you describe.. It's just a lot more work!

I'm surprised that the experienced framers who wrote my books would use inferior tables.. You would think they'd be aware of the stuff you're talking about.

[Don_P]

Basically, you can publish anything. The numbers I've posted are the correct ones to use for design. I have checked the '82 NDS, the closest I could come to publication era, the numbers were the same. I've pm'd you a little more.

It is also not customary to take Mitchell's roof slope snow load reductions, ASCE-7 does allow reductions with caveats. Use good judgement there, I take full ground load. I've been called when a client had 3' of consolidated snow stuck on his 12/12 metal roof... never would have believed it. Happily the engineer had specced rafters on 12" centers, I quit laughing at him about then  .

These are from the '01 NDS Supplement. There were adjustments to the shear values that have changed with the '05 version so use the new values I posted earlier. I have a pdf of the '05 Supplement I could email. PM me your email if you want to try that.
http://www.windyhilllogworks.com/Calcs/Fblist.htm

I feel that nature knows how to make a strong chunk of wood regardless of the knots in it, and sawing into the grain severely weakens it like you describe..

The FPL has been trying to work on that. We have an overstock of small diameter roundwood nationally. It needs thinning but the timber has alot of weak juvenile wood. If you can saw just 2 faces and leave the others round the resulting timber is about 50% stronger than if it were squared.

[Ernest T. Bass]

Seems to be working;
http://www.windyhilllogworks.com/Calcs/AnyPtLd.htm

Make sure the left and right sides sum up to the total span. I'd check the post at correct location and then drift it right a bit while watching moment, deflection and shear.

How exactly did you work this calc? I tried the point load @ 10'' from the support and an 88'' 8x8 failed. I imagine figuring out exactly how much additional support the brace adds would be complex and would depend on the size and type of joint, but if the figures came close without the brace it would be pretty safe, right?

[Whitlock]

Even though the brace will support some weight that is not it's job. Don't minimumly engineer this building Over build it Over build it Over build it.

[Don_P]

Sorry to be slow, been doing a big takeoff, hard to concentrate with bare feet in the winter  

Whitlock nailed it, I was just being verbose, but its writ so my .02

Braces are normally not factored into the beam sizing, here's my take on it. The post will shrink some in service, this will move its bearing surface away from the brace's end a little. The brace will shrink a bit in width, the 45 degree angle you cut will become more acute, pointier.

This brace was cut by a computer controlled machine in green oak. ALL the tolerances were within a very few thousandths when assembled. It has now dried for about 3 years. Yes you should house them but it is good to see what is hidden in many housings. Dry brace stock cuts that scenario at least in half.



If asked to do work the point will move over till it contacts the post and then the point will have to drive into the post until it crushes a wide enough bearing surface to support the load. "Load goes to stiffness", where is the load? The ends of the beam are supported on the vertical grain of the post top (note to self, check beam end on post for crushing), the posts' length really won't change appreciably so there are our stiff supports. The load will definitely be entirely upon them initially. If the beam deflects enough, near it's end, it will begin to bear on the brace to some degree. At the point where all this "hooks up" have we bent the beam enough to begin weakening it? Your shears are bad near the loaded end, if that begins to shear the beam will seperate into horizontal planes and bend like the pages of a phone book. This may happen before the brace takes the load, you are beyond me there. Braces are ideally just that, they resist instant racking loads like wind (and they don't do real a good job of that).

A post with a loaded brace is considered to be also side loaded as opposed to just axially loaded. Instead of the load being applied to the top of the post and travelling straight down its vertical axis, the post has a vertical load as well as a bending load from the side. Take a thin stick and put it between thumb and forefinger and squeeze. That is an axial load. Now push in on the side of the stick while squeezing the ends, combined loading. It should have been much easier to buckle the stick in the second experiment under the same axial load. There is a complex interrelationship equation to account for combined loading but it is modelling this simple fact. The post can do a certain amount of work, the more axial load it takes, the less bending load it can handle and vice versa.

I've tried to make a calc out of the modern combined loading equations and haven't succeeded yet. I might scan them in just so you can get a laugh, complex doesn't begin  . This WWII era math for the interrelationship is cruder but better than nothing and might help get a feel for the concept;
http://www.windyhilllogworks.com/Calcs/44axbend.htm

The point loaded beam calc is only figuring for a point load. We have a point load from the post and a uniform floor load on the beam. This affects the location of the maximum bending moment... the effective bearing point of our point load, that's why I said to drift right a little, out into the span. I've been lazy and thrown the entire load onto the post so we just have a point load. The actual maximum bending moment is going to be a short distance to the right of the post. As I've played with it a bit this is more critical than I thought if you are trying for the minimum beam size. There is a balance between shear and bending failure as the post moves that looks pretty fine. We need to read more on figuring the true point of maximum bending, it will be where shear passes through zero so approaching it from either of those directions should get us there. Please remember you're talking to a carpenter, I like to err big and brutally basic. This is all above my pay grade, do get it checked if you are pushing around the edges or if any of this isn't adding up for you. A dairy barn up the road collapsed yesterday, another old TF gable barn is twisting and listing, it won't take much more snow and is a deadfall trap waiting to be thumped. You sure don't want that call  .

Lifting big beams is just a matter of outsmarting them. Cathedrals were built without modern cranes.

[Ernest T. Bass]

Okay, I've got to ask one more question before I can let this thread die..

First of all, I'm rethinking the frame a little bit. By adding two posts to the middle of the living room for 2nd floor support, I'd like to scoot the two inner rows of post out to sit directly under the 2nd floor posts (no offset worries).

My question is about knee braces and if they are necessary, considering the cordwood infill that will completely brace the perimeter of the building. My thought was that they wouldn't hurt, but I'm having trouble designing the frame (in a straightforward way), to incorporate large beams that intersect over interior posts for brace attachment.

Conventional houses are only braced by their exterior sheathing though, so could the knee braces be safely omitted for simplicity's sake?

If the answer is no, than what is the minimum size for a floor joist to accept a let-in knee brace from below? It's a ridiculously simple question that I haven't found an answer to.. I know that the braces themselves can be as small as a 3x6, but would notching into a 3x6 joist weaken it too much?

A thousand thank-you's.

[Don_P]

I'm enjoying the conversation myself.
The cordwood infill could probably provide some level of in plane lateral resistance, I don't know what it would be. Out of that wall plane the floor or ceiling diaphragm can be bracing the building. As long as the whole is braced I don't feel the bracing has to be timber knee braces, they aren't much anyway.

As far as how much can you remove from a joist, the correct way would be to check the bending moment at the location of the mortise and then figure the joist dimension there as the mortised section. Check it as a simple beam. Often the sizes of timbers revolves around the size of timber remaining after joinery.

[Ernest T. Bass]

I don't know how you could possibly determine exactly what amount of bracing the infill will provide, especially since we will be using a cob mortar. You wouldn't think a massive 18'' thick wall would have any problems, though. In order for a wall to rack, the infill would have to totally smush..

I'm just a tad nervous because of the size and height of the place.. It'll pick up a lot of wind.

[Don_P]

Can you not integrate some big diagonal X's of something across the frame in the cordwood as you build? If not timber in compression, Hi tensile steel in guying tension x's, think about the end of a fence?