Roof valley connections

[Don P]

Will the bearing area support the load?

[Algorythm]

Hi Don,

Yeah, that's what I was hoping to do here. The roof valley would rest on top of the post (3"x10" bearing surface). The steel bracket would maintain the roof valley in place and give it a larger foot print over the bearing surface of the post. Let me know if you think something is wrong or if I am missing something ???.

Here is a picture of bearing surface for the roof valley for info:



 

Also had a message for Beenthere: "Regards your bracket, offset the bolt holes so they don't align with the same grain in the wood."

Thanks a lot, I will make sure to correct this.

[Don P]

To pile on to beenthere's comment, the hole should be at least 7 diameters down from the end of the notch.

The steel should cover the entire contact area between post and valley, just reiterating that you should develop full bearing no matter how you do it.

What I was driving at with my last comment was that you should quantify the load on the bearing surface and compare it to the allowable stress*
This is a pic showing the tributary area supported by the valley (both shade colors). Half of each jack rafter is bearing on the ridge, half on the valley. Then in yellow shading is the halfway point between your new mid valley support and the post, so that is the area to quantify load on the valley heel bearing. It will take some judgment to fudge the load of sliding consolidated snow laying in the valley on the worst day, that is the design day.





Using this as a teaching moment for others following this, notice the valley is a beam with the load uniformly increasing towards the lower end. If you visualize a purlin rather than a rafter arrangement that "kite" of tributary area flips end for end, the load uniformly increases towards the top.

*The post in parallel grain end bearing is probably not the control, the valley heel bearing on something between parallel and perp to grain is the member to check for crushing. The Hankinson formula gives the interpolation between perp to grain and parallel to grain design values in compression to use. You run into this with rafter and truss heels as well.

Sorry for the aside, back to the discussion in progress.

[Algorythm]

Alright I will be working to improve the design considering placement of holes and full bearing of the valley.

Also, I looked online for a Hankinson formula calculator. I ended up on the American Wood Concil website. They had a calculator but I am not certain of all the values I need to put into this ???. I played with values to get a feeling of how it affected the capacity. What surprised me most are the variations of the fastener's diameter but I am still not sure how to interpret all of this and what to compute for my design :-\. Here is a picture to the calculator:



 


Don, I was also wondering how the tie beams placed below the Valley would affect the tributary area. They would likely help support some of the extra weight and also prevent flexion up to a certain point.

[Don P]

Let me sketch a little more and show it in a different view. The shaded areas are the entire tributary area supported by the valley. Now look down the valley the inside corner of the yellow area is halfway between the supporting tie and the post, that is the area of concern right now, the load the post and valley heel are carrying. The green area is loading the tie, the blue is carried by the peak. Can you see how the load is being divided by halves, halfway up each jack rafter, halfway between truss and first jack, halfway between points of support? Just remember snow is going to slide down into the bottom of the valley and lay there, this will substantially increase the load beyond ground snow load values.




Hankinson... first see if we need him. Quantify the load in the yellow area and the area of the valley heel bearing area. Multiply the bearing area by the allowable compression perp to grain. If that allowable is higher than load then you check without going further.

The actual grain angle is heading towards compression parallel to grain which is stronger than perp to grain, we are somewhere in between the two. That is what Hankinson does, it adjusts the allowable compression for the angle to grain. Hmm, it would take longer to figure out how to write the equation here legibly than to google it, here we go;
https://en.wikipedia.org/wiki/Hankinson%27s_equation
I'll start trying to put that into a calc for the toolbox, that might take a bit.

Now, what is the steel doing? You aren't using it to increase bearing area, the heel controls there either way. You are really using it to hold the valley in place, against what? Uplift?, drift? I'm just exploring that. Would a strap running in the valley topside and down the post outside do the same thing but hidden? Just thinking not directing.

[Don P]

Hankinson calc is ready for a test drive;
http://forestryforum.com/members/donp/hankinson.htm

[Algorythm]

Hi Don (and everyone following the post),

I was away from home in the past few days and just got back. I have carefully read everything you said and got back to work today. I am not ready yet for Hankinson but I will be gathering and putting together in a spread sheet the values I need to calculate the roof load. Once I got things sorted out, we can try the formula.

As for the metal bracket, I first needed to rework my model a bit in order to answer your question. The bearing of the roof valley was suboptimal. By cutting the bottom of the valley by 1" inch I was able to improve the bearing surface by a significant margin (approximately 28% without the steel bracket and 35% with the steel bracket. The first image below shows the modification on the roof valley (cutting 1" at the bottom). The second image shows the bearing surfaces of the roof valley before and after modification and with or without the steel bracket.



 


 

Without the modification, I have approximately 10% gain on the bearing surface value by adding the steel bracket. After modification I have approximately 20% gain on the bearing surface value by adding the steel bracket. I believe that the steel bracket could be interesting but it has downsides and requires more work. It needs to be aesthetic and carefully calculated. What do you think?

[Don P]

The calc is pretty simple so I'll walk through it real quick with what I think I remember. If the roof pitches are 45° (12/12 pitches) then the valley would be around a 35° angle from horizontal. The calc is asking for the angle between load direction, vertical in this case, and direction of grain. I plugged in 55° and #2 EWP which has an allowable compression parallel to grain of 400 psi and perpendicular to grain of 350 psi. The resulting allowable compression would then be 365 psi. Multiply that by your bearing area, I'll try 13.35 square inches=4872 lbs allowable load. I'm coming up with roughly 35 square feet bearing on that area so 4872/35=~140 psf total load allowable on the heel (that was not a check of the valley size itself only of the compressive capacity of the heel).

[Algorythm]

Alright Don I think I got it. I tried calculating the dead and live load of the roof for the 35m² surface. I approximated at 910 lbs the dead load. For the live load, I considered that the snow was very wet and icy (55lbs/ft²) and that there was 1 foot all over the 35m² area. The sum of dead+live load gave me 2835lbs for that 35m² section of the roof so approximately 81 psf total load. If there was 2 feet of snow it goes up to 126 psf total load. I am not expecting to have that much snow on the roof with a metal sheeting and a 45deg roof pitch but better be safe than sorry.

[Don P]

I was using 35 sf, I think you just typo'd but check.
Moving up the valley, I'm getting around 50sf in the green section center point loading the tie which has about ~11' span. The sketch now has the valley just touching a corner there... next detail to think about, that might be the bracket ???.

[Algorythm]

With 50 ft²:

4872 lbs / 50 ft² = 97,5 psf. Withstands a bit more than a foot of wet snow including the dead weight.

I think at that point I might need to improve something else and adjust some values to make up for uncertainty. I don't want to monitor the roof on a daily basis. I see a few options I could explore ???.

1. Using the bracket

16.7 inch² x 365 psi = 6095,5 lbs allowable
6095,5 lbs / 50 ft² = 121,91 psf.

2. Improving support of the tie beams under the valleys by adding braces, spline joinery for the tie beam to central post joints. Also going to see if I can improve bearing area of the roof valleys over those tie beams. I might need to adjust the central post.

3. Reviewing my roof weight calculations to improve accuracy of the results.

I will be making adjustment on the model shortly to improve its strength. Here is an image of the improved version of the bracket considering recommendations I had. I feel like It is slowly getting better.



 

[Algorythm]

Hello all,

I have been trying a lot of different things in order to support the valley's. Here is what I came up with (see pictures below or skp file attached). I still think that it is not perfect but I believe I am making progress. I decided to change the staircase a bit and use the staircase post to support one of the valley. As for the other, I went for the spline joinery at the center post. As for the other end, I did not want a post in the middle of the dinning room so I used a shouldered mortise&tenon with a bigger brace to support the tie beam. Also posted an exploded view of the junction of the valley with the king post.