Hi, I was looking through Historic American Timber Joinery at these types of joints.

I was reading and Sobon says "the full depth of the joist may be notched into the girder, or , more commonly, the joist is reduced in depth at the joint so as not to cut too much out of the girder."

Was this due to square rule issues and different heights of joists? I would not think it was because of girder strength (I-beam). I have long wondered why the removal of wood at the bottom of the joist reduced shear stress. Anyone have thoughts on this? And to quote NH, "Any and all other comments welcome".

Mo, nice illustrations. What software do you use for that?
Reducing height/depth of joist is mainly to leave more continuous, uncut, long fibers in the girder, thus preserving more of the girder's strength.

Careful angled or arced cutting of the joist end from its bottom upwards aids in keeping the shear forces within tolerable limits.
"Tolerable" means you've calculated loads. So you've traded some weakening of the joists to preserve some more strength of the girder.
Grain slope, swirls and knots in the reduced joist end are even more important to observe and account for. I've seen a row of housed, reduced-end joists in a girder where c. 20% of the joists were split/failing from their ends. Main cause was "bad" grain, i.e. carpenter's bad judgement.
Steve

Waccabuc, were the joist cut with a 45 degree angle or via adze?

I see the adzed cuts better because there is no starting point for the check to start at. The 45's have a corner for the check to start. Worse than the 45 is a square cut for the reduction. The best is the fully housed, but then you start to compromise the girder, if it is not deep enough. And I like the visual effect. That is all by theory, I have seen square reduction with no check or failure. Most failures I see are due to excessive loads and shifting of the building creating the excessive loads.

I try to avoid knots within the adzed section, they are a pain, much nicer with clear wood. Also drop them in upside down, adze, flip up right, no layout.

Tim

As mentioned the curved cut is done to remove the "point of fracture" from the floor joists.
In order to support a floor load the beam has to be sized correctly for the intended load. That is at the mid point of the beam where the bending will be the most. The ends of the beam have to be sized to prevent the beam from failing in shear. That is as if the beam was sheared vertically.... Solving for shear is usually quite easy as wood is strong perpendicular to the grain...
Modifying a straight beam with curve cut ends doesn't weaken the beam as the middle is thick where it needs to be for strength, and thick enough at the ends.
And as mentioned not reducing the tie or girder it's set into, to full depth of the joist makes it stronger as less of the tie or girder has to be removed...

Hi Mo & Jim,

Ideally connecting joists should enter and bare on the supporting beam on the neutral axis of that beam i.e. centre line. Medieval joists were turned flat to achieve this (e.g. 4 x 6 joists into an 8 x 8 beam) plus also it meant that a slightly curved tree could be used to make 2 flat joists when the tree (puncheon) was halved providing they were laid flat. This can result in a springly floor and later on when better quality straight timber was employed then the joists were fitted to the beam "on edge" but the joist tenons still sat on the beam centre line. It is rare to see floor joists profiled as illustrated above. I have seen one house in Farnham where the carpenter hedged his bets by setting each alternate joists on edge then flat.

Regards

Ken Hume

Hi guys,
A word of caution on notching: The National Design Specification (NDS) for Wood Construction only allows a notch (with straight, curved, or angled transition) equal to 1/4 the total height of the beam being notched.

Andrea

Hello all, thanks for the info. So if joists need to carry a heavier load, and their plumb height required is larger than half the height of the plumb height of their carrying beam cut a housing no deeper than the neutral axis of the carrying beam( Ken, I assume this is to not go into tension wood?).

Therefore, if the carrying beam is 9 1/2" in height you would have to cut for housing no more than 4 3/4" from top. Which would force you to use no more than a 5 15/16" joist in height.

(4 3/4")(1.25)= 5 15/16"

Should this be a good guideline for joist and carrying beam sizing when determining loads? Or should we just oversize?

Hello all, thanks for the info. So if joists need to carry a heavier load, and their plumb height required is larger than half the height of the plumb height of their carrying beam cut a housing no deeper than the neutral axis of the carrying beam( Ken, I assume this is to not go into tension wood?).

Therefore, if the carrying beam is 9 1/2" in height you would have to cut for housing no more than 4 3/4" from top. Which would force you to use no more than a 5 15/16" joist in height.

(4 3/4")(1.25)= 5 15/16"

Should this be a good guideline for joist and carrying beam sizing when determining loads? Or should we just oversize?

Steve, the software is Cadwork. Sorry about the double post. Im still shaking from the fishing trip.

This was washed up in South Carolina yesterday. Seriously.

Now there's a face only a mother could love. What blows my mind is the size of some of the fossil sharks teeth... so much for the good old days!

Scroll down this page to the green area. You can download a free copy of the NDS commentary. Sections 3.2, 3.4, 4.4.3 deal with notches. Technically you can't notch in the middle third of span of a beam by the NDS.
http://www.awc.org/Publications/download.html

Code allows holes up to 1/3 depth anywhere as long as they are not closer than 2" to either edge.

Also googling reentrant joist notch or similar will probably yield some more info. That's the fifty cent name of the sharp cornered notch's point of fracture, or where concrete slabs always crack. A smooth transition helps reduce the shear concentration from that one point.