When and how to allow notches in wood bending members? 2

[AaronMcD]

When notches are not permitted, do you permit them with reinforcement? Does it make a difference if the beam is not highly stressed?

I have a condition where I have 1 3/4"x 9 1/4" LVL joists at 12" o.c. These are sized for deflection. Bending utilization is only 35%. The interior designer went and specced sliding door tracks in the joist space that would have required 2" notches near the middle 3rd. Now they changed the track spec, but it still requires 1/2" notches.

According to NDS, "8.4.1.1 The tension side of structural composite bending members shall not be notched, except at ends of members for bearing over a support, and notch depth shall not exceed 1/10 the depth of the member."

Would you take this to mean "no notches of any kind permitted"? What about a 1/32" "notch"? 1/4"? 1/2"? At what point does a slight scrape off the wood become a notch? Is a notch permitted with reinforcement?

What about a dado parallel with a beam? I have another beam where they will have to do this. This seems reasonable to me, as splitting isn't much of a concern when most of the member retains it's full depth.

http://www.hellodwell.design

 

[JoshPlumSE]

Would you take this to mean "no notches of any kind permitted"? What about a 1/32" "notch"? 1/4"? 1/2"? At what point does a slight scrape off the wood become a notch? Is a notch permitted with reinforcement?

I would take that to mean that the capacity of the member should be determined based on the un-notched sections for any notching that doesn't comply with the NDS restrictions.

If you meet deflection limitations and capacity limitations using the un-notched cross section then you're okay. At least that's my take.

 

[AaronMcD]

Well this is in the NDS so I assumed it is written for engineers that know how to calculate capacity of a reduced section, but it seems there has to be some "out" to design around notches. Can notches split upwards from the notch corner on laminated veneer members? I read sometimes LVL are all oriented longitudinally, in which case splits should run fairly horizontal. Sometimes a few veneers are run vertically, in which case it seems they would arrest mild splitting.

http://www.hellodwell.design

 

[KootK]

Even designing based off of a reduced section is probably not entirely kosher given that:

1) The issue is really about stress concentrations and fracture mechanics rather than raw, available cross section. When some fracture occurs, it may or may not occur parallel to the member.

2) In going with a reduced effective section, you also compromise the original lumber grading of the member. That would have been based upon things like how close to the member edge knots are located etc. Parameters which your notch has now altered.

3) While I'm still not sure that it's code compliant, screws can be used to reinforce notches and, depending on the situation, I might do this: [link link]https://mtcsolutions.com/wp-content/uploads/2019/04/Reinforcing_Wood_with_Full_Thread_ASSY-screws.pdf[/url]. It's better suited to timber rather than a bunch of repetitive joist though.

4) In practice, I do exactly what JP suggested along with direction than I don't want to see any knots or other imperfections in the outer 1/3 of the member depth on the tension side. Obviously, I cannot justify that completely in a theoretical sense. But, then, one must be practical in our field and I feel that it's impractical to doggedly enforce the no notch thing without exception.

 

[KootK]

I would expect the situation with LVL to be better than it is with lumber for exactly the reasons that you're thinking. That said, I know of know formal recognition of that in any code or standard.

 

[XR250]

Consider making the notch curved instead of angular to reduce or eliminate the stress concentrations.

 

[dhengr]

AaronMcD:
Use a 1” dia. by 2” long, or larger, straight bit in a router to cut the notch. That way you will end up with nice rounded reentrant corners where the stress raisers would normally be. Make a guide frame/template to run the router base against to accomplish a nice clean notch. If you get tear out at the exit end of the notch, do half the length of the notch from each side of the joist. Don’t let them do the cut-out with a hatchet or recip saw, where they tend to get over-cuts and square corners at the reentrant corners. Check that the bending stresses and deflection with the reduced depth are still o.k. As KootK mentioned above the big issues are the stress raiser at the reentrant corners and the fact that any splits may run with the grain (kinda unknown, uncertain direction) reducing the member strength (depth) even further. Tell the interior designer to forget about the notches and cut .5” off the bot. of the doors, you think that the structure is more important than the fact that they didn’t bother to study the plans before they ordered the doors.

 

[JoshPlumSE]

4) 1) The issue is really about stress concentrations and fracture mechanics rather than raw, available cross section. When some fracture occurs, it may or may not occur parallel to the member.

2) In going with a reduced effective section, you also compromise the original lumber grading of the member. That would have been based upon things like how close to the member edge knots are located etc. Parameters which your notch has now altered.

3) While I'm still not sure that it's code compliant, screws can be used to reinforce notches and, depending on the situation, I might do this: [link link]https://mtcsolutions.com/wp-content/uploads/2019/0...[/url]. It's better suited to timber rather than a bunch of repetitive joist though.

4) In practice, I do exactly what JP suggested along with direction than I don't want to see any knots or other imperfections in the outer 1/3 of the member depth on the tension side. Obviously, I cannot justify that completely in a theoretical sense. But, then, one must be practical in our field and I feel that it's impractical to doggedly enforce the no notch thing without exception.

Good point about compromising the original lumber grading rules.... especially related to the increased fracture risk at the notch. Shouldn't be as much of an issue with engineered lumber like LVLs. But, for regular dimensional lumber that is an excellent point.

 

[KootK]

I'm not in a position to speak with any real authority on this but my gut instinct for how to handle the fracture mechanics aspect is different from the other gentlemen. A radius makes sense to me for an isotropic, ductile metal. For something that is highly isotropic like wood, however, I lean towards something like this.

 

[Settingsun]

A drill hole at the re-entrant corner is good practice, ie a small radius then a gentle slope as shown by KootK. The eurocode requires 1:10 or flatter to disregard the stress concentration.

 

[AaronMcD]

@steveh49
Good to know there is some precedent for this even if it isn't the US code. 1:10 isn't that far for the tiny notches I'm being asked for.

http://www.hellodwell.design

 

[Agent666]

The Australian timber standard has an interaction equation for notched beam design, whereby you're checking the interaction of moment and shear stresses at the notch and comparing it to an acceptable design stress. This formulation includes an additional reduction factor on the allowable stress that is based on the geometry of the notch depth to the notch transition length. This additional reduction factor also varies if your notch depth is greater than 10% of the original member depth. See screen captures below:-

If you need further info on what some of the other factors account for let me know and I can dig through standard and try find them (I'm not intimately familiar with the standard but have used certain aspects from time to time where our NZ stanndard is lacking).

 

[AaronMcD]

Looks like the first term is bending stress. Do you know why the 6 in the second term?
Also wondering if there is a difference in "maximum" and "nominal maximum" with regards to bending and shear respectively. If you can track down those k coefficients it would be appreciated!
Also, what is f'sj? Is wood design allowable stress with properties modified by stuff like wet service, duration of load, size and orientation, etc. like in the US?

http://www.hellodwell.design

 

[Agent666]

Not sure on the 6 in the shear term. Perhaps it's some factor for how much shear stress there is contributing at the extreme tension fibre due to V* (since max shear stress occurs at mid depth in rectangular member).

Don't think there is any difference between 'maximum' and 'nominal maximum', adding nominal seems redundant in the context.

k1 is a loading duration term
k4 has to do with the moisture content
k6 is to do with temperature effects
f'sj is a specific shear stress used for joint design as far as I can tell. It's specific to each type of wood available in the local market, so it kind of means the procedure cannot be easily applied elsewhere.

 

[Settingsun]

The Australian code (AS 1720) gives low capacity. As you noticed, the first term is bending stress, while the second term is 4*max shear stress. The f'sj capacity term depends on species and moisture condition but is in the order of 10% of the bending strength at un-notched sections. For your little notch, the g40 term is 2.8 (for 1:4 slope), so you're down to about 28% of the un-notched bending strength even if the shear force is zero. It's tough to make this stack up except right at the support. Though you're sized for deflection so you have a chance.

Here's a relevant paper with test results.

https://na.eventscloud.com/file_upl..._Notched_Timber_Sections_JCU_Justin_Dewey.pdf

 

[AaronMcD]

So...
~10% for vertical cut
~18% for 1:2
~28% for 1:4

I have a hard time seeing this trend up to anywhere near 100% at 1:10

I guess I'll stick with a reinforcing requirement. I'm sure they'll do it, this house seems to have never ending construction administration.

http://www.hellodwell.design

 

[Settingsun]

But As1720 is known to underestimate the strength. And you only need 35%.