This question may be properly addressed to Mr. Levin, but I'm sure others can answer as well.

I have been asked to review a couple of timber frames by a joinery company. In regards to a previous semi-frame related job, I purchased the "Timber Frame Joinery and Design Workbook" (Red Book), and "Timber Frame and Joinery Design Volume 2" (Green Book). In the Red Book, Mr. Levin presents an analysis of a model house frame using an FEA model and designs a brace for a 3,500 lb compressive load. His approach makes perfect sense to me and I agree with the logic.

The frame I am considering for discussion consists of three bents with 10'-0" x 14'-0" bays on a 42'-0"L x 20'-0"W 1-1/2 story house and contains a bearing Ridge. The NYS ground snow loads for this area are 85 PSF, resulting in an approximately 59 PSF projected roof snow load and approximately a 12 PSF dead load. The walls and roof are sheathed using SIP's. Combined with the second floor diaphragm, wind loads and tension in the braces are not a concern. The ridge columns are 8x10 D-F-L No. 1 and the ridge beams are 8x12 D-F-L No. 1. The knee braces are 4x6 D-F-L- No. 1 and have 1-3/4"W x 3-1/2"L tenons with single 1" diameter white oak pegs, and not housed.

My FEA model is producing compressive axial forces in the braces anywhere from 8,000 lb for single braces up to 14,000 lb for pair braces. A cursory check of load and span ratios shows the 8,000 lb load on a single brace is much in line with the result Mr. Levin got for his brace. However, the larger compressive loads appear to be the action of adjacent braces transferring gravity induced compressive forces across the column to the opposite brace. Checks of these braces as designed shows they are only adequate for approximately 5,000 lb if I provide a 1/2" housing and use one half the tenon area as suggested by Mr. Levin. At the larger loads, this brace will not work no matter the housing, and braces with larger horizontal bearing area must be employed.

What disturbs me is that the joinery company informs me that in some 400 designs, an engineer has never requested a housing on this size brace of over one inch. This raises the question of why I am seeing these loads and how is it others apparently don't or ignore them? I can take a simple beam and put two braces on it and support the ends of the beam in the Y-direction, and the brace bottoms in the X & Y- directions and the results are much the same. All supports and connections are modeled as simple.

Can anyone shed some light on what I should be expecting here? Is the common theory that the beam bearing will take the load prior to the brace failing? Any help would be greatly appreciated.

BWW

It is my understanding that many engineers assign no gravity load to the braces, or at least no more than they can take (load goes to stiffness as you alluded to in the last paragraph).

If I am wrong I would love to be corrected.


Gabel

Hi Gabel,

Thank you very much for your response. I had surmised load is assumed to go to stiffness based on my search through the archives here. While I didn't see any posts that stated this, it seemed that is was alluded to.

I don't have any reason to doubt the numbers I'm seeing since all member properties are included in the model. But the old saying is you don't reinvent something that has been shown through history to work.

Hopefully somebody else will take the opportunity to set me straight.

Thanks again.

Braces can do bad things to a frame, in theory. The provide apparent stiffness at low forces, but cause havoc on column bending (and bearing forces, as you've seen) at high loads. It's usually a trad off between bigger columns and smaller beams, in my experience.

Will the model survive without the braces?

In my models, braces get full loads, since they will actully see nearly full loads (any load added after the tenon makes contact with the housing). I try to have the framer avoid them when possible.

Remember that (1) very few structures are made with lumber which is of minimum size and grade strength and (2) very few have actually withstood full environmental loads. The change of 85psf ground load is just 2% per year. How many houses by the TF have been in service for more than 50 years? Of those, how many were designed with minimally sized members. Now that your answer is close to, or exactly zero, can you really extrapolate success from such a small sample?

As engineers, we design to save lives, not necessarily to minimize cost or promulagate tradition. If youre analysis is good, then the forces are real. I have worked on both architectual, consumer-ish, and space flight hardware, and have had the opportunity to model strucutres which I subsequently broke or had stress monitors on in a test environment - the forces you are seeing are real.

Hi Jordan,

Thank you very much for your input. As I said, I have no reason to doubt what I am seeing is accurate. The sheathing at perimeter walls attached to frame beams will also lend support and reduce loads on these braces, but this is not the case at the interior bent I am concerned about.

My feeling is that the knee braces are being viewed as more architectural than structural. As you know, this is simply not the case. I have increased the size of the beams such that they will not fail even at brace failure as a bit of added insurance.

The chances of failure as you mention are slim at a 2% recurrence interval, however, who wants to gamble on even those odds?

I gather you try to steer framers away from knee braces all together based on results similar to mine? Based on your experience of a trade off between beam and column size, I may experiment with this a bit and see if the reactions on the braces are reduced. I concur with your statement about wreaking havoc on column bending. The deflected shape results foretold the situation.

Thanks again for increasing my awareness.

I just have to wade in and ask how the old ontario barns stand? Most of our barns have 3x4 pine braces with 3x2 tenons and 1" shoulders. And much bigger buildings.. 30 or more wide queen frames and 40 or more wide aisled frames. The posts tend to be 10x10.
The frame you are talking about(although I can't understand your 10x14 bay measure with three bents{do you mean 20x14?}) are smaller and better braced than our barns...

Hi Mark,

To clarify the structure, three bents - two eave and one ridge. The spacing between bents is 10'-0". There are three 14 foot bays on each bent adding to the 42'-0".

As I said, history shows knee braces work, or at least don't seem to suffer catastrophic failures. Roof covering and pitch have a lot to do with the loads seen. I'm dealing with a shingled roof. Shingled roofs don't shed snow. Metal roofs readily shed snow. Not knowing what kind of roof you have, I couldn't even comment.

I believe that in reality some load does go to stiffness, but as Jordan stated, the codes establish design criteria which minimizes the likelihood of failure in the interest of both life and property protection. The building codes generally establish loads based on a 2% probably model of the load ever being exceeded. The NDS establishes wood design properties based on the percentage of members whose strength falls below the design strength. Thus, it is reasonable that structures which don't necessarily pass a design check still do not fail. But when a PE starts accepting designs that don't meet established design criteria, it's a slippery slope if one wants to keep their license. I know I do.

Thanks for your reply.

BWW:
I have come up against the same results, and this is my approach and rationalisation: I have never seen a brace buckle or a post snap where a brace pushes on it. In cases where I have examined leaning barns, the only failure observed at compression braces is crushing of the bearing surfaces of the brace housing. Before I model a frame, I draw up a typical brace end connection and calculate the allowable force in the brace based on bearing areas of the joint as discussed in TFJ&DWB. (This has always been less than the allowable axial strength of the brace per NDS.) I then model the frame with braces as compression only members. My first pass always comes up with brace forces too high. I then reduce the cross sectional area of the brace members and repeat until the model brace forces are less than the joint can handle. When I need a tension brace to make things work, I let a steel rod into the top edge of the brace and run it through the post and beam.
Happy modeling,
Chris

tension braces.........???????? lets get to it. how do you guys design something. wood is not isotropic, live and let live. bye

Hi Christopher,

Brace buckling is a fairly common phenomenon in timber framed buildings especially in respect of roof (wind) braces. These tend to be made a bit thinner than wall braces and are less well secured (lap top, mortice & tenon bottom). Wind braces are frequently lost over time making buckling problems worse for surviving braces.

Regards

Ken Hume P.Eng.