More Cutting, Drilling and Notching

[Energystar]

R802.7 Cutting, drilling and notching.

In my area current economics have caused many builders to now opt for stick-framing over trusses. This section of the IRC has been discussed at length by this forum, yet no one has come up with a complete explanation of its apparent contradictions. There are two separate but related issues that have been discussed: 1) The depth of notch limit and 2) Toe support.

The diagram below makes no sense to me except for the part about the rafter cantilever. Why would the code limit the depth of the notch on a rafter to D/4 as measured in the figure below? I have seen rafters framed off of subfloors where the notch was nearly full depth. Jack rafters for overlays are full depth. If we ignore the tail, and the toe of the birdsmouth is solidly supported, what’s the problem?



Hip and Valley Rafters

The reason I insist on beating this horse is that if one follows the code precisely using the smallest bearing allowed (1.5”) for rafters and then uses the next size up (2” nominal larger) as required for the hip, it is impossible (using standard techniques) to make the notch on the hip less than D/4 regardless of the pitch. See the figure below. Interestingly, it is also impossible to support the toe unless using 2x6 walls or using a 12/12 pitch for 2x4 walls.

One must assume the code is not banning the common practice of stick-framing. Are hip and valley rafters, therefore exempt? If one uses engineered material like an LSL or LVL, it is possible to engineer your way out of these issues, but with conventional lumber it is much more difficult. There are many videos online that describe techniques of fabricating hip and valley rafters. All seem to go out of their way to avoid discussing these issues while all appear to violate the D/4 limit as shown above.





If we read R802.7.1 carefully, it appears to refer to the overhang cantilever in figure R802.7.1.1 as an exception, “. . . except that cantilevered portions of rafters shall be permitted in accordance with Section R802.7.1.1.”

R802.7.1 also refers to another section of the code in Floors, R502.8.1, where the following figure shows the allowable notching of floor joists. Again, they use the D/4 limit. It doesn’t take an engineer to see that on the left side of this figure the notch will definitely weaken the joist. The deeper the notch, the greater the impact.



The general statement it makes regarding structural members is, “Notches at the end of the member shall not exceed one-fourth the depth of the member.” I would suggest that, as it pertains to rafters, this figure would be analogous to an unsupported toe of the rafter as shown below with the D/4 dimension changed to reflect the impact to the weakening of the rafter.

This is a figure comparing 2x10’s notched to the maximum depth of D/4. The figure on the right is code compliant. Is the figure on the left? Many would argue that neither are “proper” or at least ideal and should be avoided. The question is, is this what the ICC is attempting to say? The way we measure the depth of the notch should be related to its impact on the structural integrity of the rafter.

 

[e hilton]

In your second diagram, none of the remaining dimensions agree with the code diagram. Your dimensions are vertical, the code dimension is perpendicular to the top surface of the lumber.

 

[steveray]

It's all moot as you can't use conventional lumber anymore and meet energy code unless you are using spray foam....

 

[e hilton]

It's all moot as you can't use conventional lumber anymore and meet energy code unless you are using spray foam....

Educate me … what is unconventional lumber? LVL? Metal framing? If you’re thinking trusses ... they are built from old style 2x material.

 

[Energystar]

e hilton,

The second diagram has depth dimensions on the hips shown in the lower figures.
Conventional, commodity, dimensional--whatever term you want to use. Lumber that is sawn from trees and not formed into chips or veneer and bonded using resins.

 

[Mark K]

The prescriptive code provision are only applicable for limited conditions. When those limitations are exceeded you need to engage the services of an engineer. This is one of the reasons why every building department should have an engineer on staff or on contract.

The detail showing a bottom notch of a floor joist should raise red flags. since this may result in horizontal splits in the floor joist.

 

[e hilton]

e hilton,

The second diagram has depth dimensions on the hips shown in the lower figures.
Conventional, commodity, dimensional--whatever term you want to use. Lumber that is sawn from trees and not formed into chips or veneer and bonded using resins.

Ok, i see the dimensions for the 2x8’s. I’m not disagreeing with your original point, i just wanted to see correct illustrations.

 

[Energystar]

I am genuinely curious about what you all think about this interpretation. If any of you have connections with the ICC, (I do not), it would be cool to run it past them.

Here is a related thread:
https://www.thebuildingcodeforum.com/forum/threads/rafter-bearing.31030/

 

[wwhitney]

The second diagram has depth dimensions on the hips shown in the lower figures.

On first read-through, it wasn't clear to me what your diagrams were showing, so in case it helps others:

For rafters of a fixed size at a given pitch, bearing directly on the top plate, using the minimum bearing length will put the top of the rafters as high as possible. The top of a hip/valley rafter has to be in the plane determined by the top of the common rafters on each side. [Well, less the effect of the thickness of the hip/valley rafter, unless you bevel the top.] Then the deeper the hip/valley rafter, the bigger the necessary seat cut and resulting notch depth. The diagrams illustrate this for 2x6 common rafters, 2x8 hip rafters, and 6/12 and 12/12 pitches.

It seems to me most of this effect is from using a deeper hip rafter than the common rafters. For example, just for the common rafters, if you bump one rafter up from 2x6 to 2x8, the depth of the notch in the 2x8 obviously needs to 1.75" greater in the 2x6s, if the tops are to remain in plane. Which is most of the allowable D/4 notch depth of the 2x8.

Is using a deeper hip rafter required by R802 somewhere?

Cheers, Wayne

 

[steveray]

Educate me … what is unconventional lumber? LVL? Metal framing? If you’re thinking trusses ... they are built from old style 2x material.

Non-engineered products....2x

 

[Energystar]

Thank you Wayne. You summed it up perfectly. Sorry to all. I didn't realize I botched my explanation that bad.

As with a ridge board, the hip or valley rafter must cover the end of the jack rafters coming into it. Therefore, you must increase the depth of the hip member.

 

[e hilton]

Non-engineered products....2x

Right. That’s how i read it. So going back to steveray’s comment, he is saying you can’t meet energy code with 2x lumber. How is common 2x different from engineered material, from an energy point.

 

[wwhitney]

I didn't realize I botched my explanation that bad.

Maybe it was just me. : - )

As with a ridge board, the hip or valley rafter must cover the end of the jack rafters coming into it. Therefore, you must increase the depth of the hip member.

Ah yes, R802.4.3. Whose first sentence says "Hip and valley rafters shall be not less than 2 inches (51 mm) nominal in thickness and not less in depth than the cut end of the rafter." Which confused me at first reading in isolation, as I took "the rafter" to mean the only rafter mentioned in the sentence so far, the hip or valley rafter. But based on the similar section for ridges, it means a jack rafter bearing on the hip or valley.

Related questions:

Is there a requirement in R802 that I missed that says a rafter bearing on a wall must be supported at the heel?

Also, take a rafter with no overhang that bears on a wall with a plumb cut at the exterior face of the wall framing. Are we really to interpret R802.7.1.1 as saying the notch depth is based on the height (in the frame of reference of the rafter) of the point of the rafter where plumb and level cuts meet? That would imply that it's better to have less bearing on the top of the wall, the minimum 1.5", rather than full bearing on the wall with the heel (inside end of level cut) at the interior face of the wall framing. Is that idea supported by engineering?

Cheers, Wayne

 

[Energystar]

There is a much simpler argument to make here that demonstrates that figure R802.7.1.1, which has been the source of much confusion, is in error. It seems to be referring to the structural support necessary to support a cantilevered overhang. It limits the minimum thickness of the tail over the notch to 3-1/2” and the length of the overhang to 24”. This much is understandable. Now, what does this have to do with the depth of the notch? The strength of the tail is solely determined by its length and depth over the notch. One could easily make the depth greater than 3-1/2” with a notch depth greater than D/4 by using deeper rafters like 2x12s.

This leads one to believe that the D/4 dimension in the figure must be referring to rafters in general. However, this also doesn’t seem possible. One could imagine the extreme case where the wall could be quite thick, say 2x8. The rafter could be only 2x4 with no tail. The notch in this case could be the full depth D with the entire level cut of the rafter supported by the thick wall.

 

[Beniah Naylor]

Right. That’s how i read it. So going back to steveray’s comment, he is saying you can’t meet energy code with 2x lumber. How is common 2x different from engineered material, from an energy point.

Thread drift, sorry.

e hilton, it has to do with the thickness of dimensional lumber not allowing enough inches of insulation to get the required R-value at the edges of the attic where they land on the top plate unless you change your techniques to give yourself more room. Not impossible, but does require you to totally rethink how you frame.

 

[wwhitney]

This leads one to believe that the D/4 dimension in the figure must be referring to rafters in general.

Also because the text in the D/4 notation references IRC 802.7.1, which is the general limit for "cuts notches and holes" in rafters. So it seems that Figure 802.7.1.1 is doing double duty illustrating the limits for 802.7.1.1 while also illustrating an example of applying the notching limits in 802.7.1.

If that 802.7.1 text in the illustration were absent, we could reasonably interpret the D/4 limit to apply only to the case of reduced depth rafter tails covered by 802.7.1.1; i.e. the rafter tails would have to be at least 3/4 of the depth of the rafter, for whatever reason. Or perhaps we are to take the D/4 limit to be determined using the portion of the rafter inbound of the outer surface of the exterior wall framing; i.e. the level cut could be much longer, and the rafter tail thinner down to the minimum of 3-1/2", but that extra length of the level cut would have to be outbound of the exterior face of the wall.

Both of those seem like unnecessary limitations, but are still physically possible.

However, this also doesn’t seem possible. One could imagine the extreme case where the wall could be quite thick, say 2x8. The rafter could be only 2x4 with no tail. The notch in this case could be the full depth D with the entire level cut of the rafter supported by the thick wall.

Well, while that configuration would be possible, interpretations of 802.7.1 offered so far would seem to rule that out, without necessarily ruling out 2x4 rafters. It depends on how we are to measure the notch depth as far as the impact of the plumb cut on the rafter, which is not specified. If we get to ignore the plumb cut, then for a 2x4, we can make the level cut just 1.5" in length to get the 1.5" bearing, and put the point of the end of the rafter at the framing exterior wall plane. As long as the slope is low enough, that 1.5" length will give a notch depth of 7/8" or less, so 2x4s are not precluded. But if the slope is 12:12, the 1.5" level cut length would mean a 1.5" / sqrt(2) notch depth, which is bigger than 7/8", so 2x4 rafters would be ruled out by this interpretation of this rule.

Cheers, Wayne

 

[wwhitney]

Still wondering:

Is there any explicit guidance in the IRC that says the heel of the rafter has to have bearing on the exterior wall or other support? Or are we just getting implied constraints along those lines from the notch depth limit, depending on how we interpret we are to measure notch depth?

Also, R802.7.1 sends us to R502.8.1, which allows a D/4 depth bottom notch at the end of a joist. It seems to me that the structurally analogous allowance for rafters would be a D/4 limit for unsupported length at the heel. In that in both cases we have the structural bearing starting at a location D/4 from the bottom edge of the member, with the resulting possible issue of development of a longitudinal crack at that point.

Lastly, has the 2024 IRC made any changes to R802.7.1? I don't know where to find the draft version of it, although I expect it is final or close to it.

Cheers, Wayne

 

[steveray]

Lastly, has the 2024 IRC made any changes to R802.7.1? I don't know where to find the draft version of it, although I expect it is final or close to it.

Long way from final thanks to the energy code...

Is there any explicit guidance in the IRC that says the heel of the rafter has to have bearing on the exterior wall or other support?

Not really...But I am more willing to accept a notch than a rafter that wants to slide off the roof....

 

[wwhitney]

Not really...But I am more willing to accept a notch than a rafter that wants to slide off the roof....

Sorry, what failure mode is attributable to the heel being unsupported, with support starting maybe 1.5" from the heel? Assuming a conventionally framed roof with all the prescriptive details present, like the ceiling joists forming a tension tie between the bottoms of opposing rafters.

Cheers, Wayne

 

[ICE]

Is there any explicit guidance in the IRC that says the heel of the rafter has to have bearing on the exterior wall or other support?

Would this be that?

2020 CRC:
R802.6 Bearing. The ends of each rafter or ceiling joist shall have not less than 11/2 inches of bearing on wood or metal and not less than 3 inches on masonry or concrete. The bearing on masonry or concrete shall be direct, or a sill plate of 2-inch minimum nominal thickness shall be provided under the rafter or ceiling joist. The sill plate shall provide a minimum nominal bearing area of 48 square inches.

 

[wwhitney]

Would this be that?

That section describes the required amount of bearing, but I don't see any language that specifies that for a rafter the bearing area needs to include the heel.

Cheers, Wayne

 

[Energystar]

Sorry, what failure mode is attributable to the heel being unsupported, with support starting maybe 1.5" from the heel? Assuming a conventionally framed roof with all the prescriptive details present, like the ceiling joists forming a tension tie between the bottoms of opposing rafters.

In the conventional sense, I am assuming that they are concerned about nailability (if that's a word). This is an important connection between to diaphragms.

 

[Glenn]

It's all moot as you can't use conventional lumber anymore and meet energy code unless you are using spray foam....

Not all roofs are over conditioned structures. Garages, sheds, porch roofs all still need conventional framing methods that make sense. I believe these notch limits make sense and I will reply to the OP separately.

 

[wwhitney]

I believe these notch limits make sense and I will reply to the OP separately.

My main question at the moment is about the implication from Figure 802.7.1.1 that the level cut at the end of a rafter counts as a notch in all cases, something that is not otherwise spelled out in the IRC text to my knowledge. Seems to me that at least to the extent the level cut is supported, it should not count as a notch for the purposes of applying section 802.7.1.

Cheers, Wayne