I believe the consensus around here is that wood moment connections don't work that great due to fastener tolerances and slip. The demand is pretty low here though. I might be ok with it if the joint was braced and some glue was used and more fasteners.

Wood moment splices like this are unreliable due to fastener slip. Will it fall down? Not likely. Will it deflect wildly with age? Almost guaranteed.

I agree with jayrod, long term deflection will probably be an issue.

Can you sandwich the overlapping section between some steel plates and clamp it together with threaded rod?

Honestly, what's the application too. Is this a one off supporting nothing important? Or are you doing this a series of times to build 25 foot long floor joists? Also, 145mm deep lumber spanning 7m doesn't check out in my head. that's an L/d=48. That's far too slender for the length.

So it's a 3x6 spanning 23 feet, I agree with jayrod about what is this supporting. Barely it's own dead weight right?

Come on, stop speculating. It is just to demonstrate the question. Actually I spliced a rafter that is angled at 45 degrees and has a design moment of around 2 kNm. I'll post a screenshot it when I get to my PC later. But thanks for the answers.

This is copy and paste of EC-5 Clause 5.4.2 Frame structures , Item 9


− The splice connection has a load-carrying capacity which corresponds to at least the
combination of applied force and moment, provided that the timber members are not subject
to bending stresses which are greater than 0,3 times the member bending strength, and the
assembly would be stable if all such connections acted as pins.

My suggestion would be ,

- The depth 145 mm seems not suitable for span 7m ( choose around for rafters around L/24 )..

- Prefer in line splice

- You may select the splice location where moment is less critical ( say 0.25 L )

- Consider dowel fasteners , Flitch plates

My opinion...

Not to know is bad;
not to wish to know is worse.

NIGERIAN PROVERB

It is not even 7 m. The strut makes it a two-span beams with around 3 m spans.

Ugh, somehow I figured that was 12", yea that does not work.

Quote (nivoo-boss)

It is not even 7 m. The strut makes it a two-span beams with around 3 m spans

You are scaring me. I hope you have a competant supervisor who reviews your work.

I'll present my scary rafter soon.

So it is supported in the middle to be a two-span beam? I have no clue what's going on.

Quote (nivoo_boss)

It is not even 7 m. The strut splice(?) makes it a two-span beams with around 3 m spans.

You realize that if you splice two beams together then it becomes one beam, right? Or maybe you're calling them "span beams" and there are two of them. "two span-beams"... Even then, the splice will make it a single span-beam.

This is going great.

Alright, I'll explain. I can't believe how far off-topic this has gone.

First, a screenshot from my Tekla model for the rafter. The splice is highlighted red, the strut blue:


Next, a bending moment diagram of the rafter. Look at the diagram - the strut sort of makes the rafter a two-span beam, doesn't it? (I know the connection slips a bit, and the strut has some axial stiffness, but still, it supports the rafter near the mid-span)


Next, the strut with compression:


And lastly, a picture from the local engineer's handbook about this sort of rafter:


It is really common around here (Estonia) but not so much in the US it seems. The thing with these rafters is, that the horizontal supports must be rigid enough to take the horizontal reactions, otherwise the strut becomes a tie and the rafter is basically a simply supported beam.

Sorry, English is not my native language, perhaps you would call this "strut" some other name.

We would generally call those collar ties. And I understand how you think that this may change our opinion. To me, it doesn't. I've seen too many of these rafter roofs with bowing rafters just due to the rafter deflection and creep without a splice. I wouldn't want to automatically introduce a source of flexibility into an already dubious system.

Around here we can get engineered lumber in 7m lengths, I don't see why you wouldn't just go that route.

Because here I cannot get over 6 m.

I was about to respond to the original question right about when you sent the post above, which seems like an entirely different problem. In any case, I'll respond to this new information:

1. I think it's somewhat accurate to consider the rafter as a continuous beam with the strut acting as a support, but only if the bottom of the rafters are restrained from spreading outward. Otherwise, the strut is doing the opposite. Instead of providing support, it is instead imposing a huge point load on the rafters. The tendency in this case will be for the rafters to snap at the strut location. It looks like you're modeling it with the bottom of the rafters restrained. You really need to be sure that that's the case. It may be so if the rafters are connected into the floor.
2. Regarding the original question about the splice, if the structural model shown above is accurate (considering the outward thrust issue I noted), I would try to put the splice at a location where the moment is zero (inflection point). I would still try to avoid a splice though, if possible. Can you get a longer piece if you use a deeper rafter? That might be less expensive to construct. Finally, if the splice is the only option, I would not offset the rafters to make the splice. I would rather butt them together and use wood or metal plates on either side.
3. Without getting into the numbers, the rafters seem a little small if I'm understanding the overall dimensions.

I remember seeing an older post on here about a somewhat similar situation with not long enough roof rafters. It was say a 22' span, but the lumber only came in 20 ft lengths (or something like that). The general consensus was to get two full length pieces and lap them. That way the lap distance was nearly the entire member. I don't see why you couldn't do that in this situation. That's probably the only way I'd feel comfortable with a wood "splice".

Quote (nivoo boss)

Perhaps you can help. I have this timber beam that has to be spliced since 7 m long elements are not available. I want to make a 1,3 m overlap, that has 2 bolts at each end of the splice so the bolt groups are spaced 1 m apart. Is this thing viable? I need to design the bolted connection and so I need the shear force acting on the bolts. Is shear due to moment here moment at the bolt location divided by the lever arm (1 m in this case)? Have you used this kind of splice before?

I really don't like the concept of this overlapping connection, but first, the statics. The shear diagram for bolts indicates a different shear value for the two bolt groups. That cannot be correct. The shear for each pair of bolts must be the same for a uniform load.

If the load per meter is w, the reactions are 3.5w at each end. Each member carries w for a length of (7-1.3)/2 = 2.85m and w/2 for the central 1.3m. W = 7w for the beam without splice. Reactions are 3.5w each end.

Each spliced member has a reaction of 3.5w at one end and a moment of (w/2)*(2.85^2+4.15^2)/2 = 6.34w at the splice, so with each bolt group separated by 1m, the shear is 6.34w in each group. In your case, w = 4.42/7 = 0.631kN/m so bolt shear = 4.00kN for each bolt group.

Split rings or shear plates could be used to reduce fastener slip, but eccentricity puts significant torsion in each member between bolt groups which needs to be considered.

I don't like it, but would not rule it out without further calculations. The depth of beam should likely be in the order of 300 or 400mm for deflection control.

We have discussed similarish ideas in the office, though in our case it was in the context of lapping two pieces of timber to get an outrigger roof to work
Something to consider is whether two bolts actually achieves the performance you expect - I'll elaborate below

Your scenario has two pieces of overlapping timber with two bolts at the front, and two bolts at the back
This is idealised as a pin-pin support so resolving the moment between front and back as vertical reactions/bolt shears

Imagine that you now remove the back pair of bolts - what happens? Is it theoretically unstable?
Well, the two bolts at the front have some ability to form a moment couple, so it's 'stable' (though with a low capacity)
This highlights the concern that I have with such details

The idealised pin-pin support is not quite correct as you actually have a moment connection at the front and back of the splice
When you try and rotate the cantilever portion of the rafter about the front bolts, they act to try and restrain that rotation as a cantilever/fixed connection - not the pin-pin vertical shear couple you've idealised
This will cause locally high shears either parallel or perpendicular to the grain (depending if you stack your bolts vertically or horizontally)
As these shears increase, the timber will locally crush around the bolts, allowing some rotation and forming some magnitude of pin-pin vertical shear couple as you've expected
However, if too much crushing is required to fully mobilise the pin-pin model then you will instead cause local failures in your timber around the front bolts

We tried to research this and couldn't find any literature on the topic to understand how severe this effect is in practice
We then resorted to building some dummy models relevant to our situation and weren't convinced that the required rotation could be safely sustained
We then tried to build a representative model on the squat rack in my garage...we managed to break a piece of 2x4 but I can't claim that we actually understood anything useful from that (besides, timber go bang loudly...)

I would be more concerned about the axial load (buckling). It is definitely unstable around the weak axis, maybe even around strong axis. Oh, and I hope you have a sufficient tie at +4.000 level.
Regarding material length - specify it as a glulam beam or finger jointed timber beam which will get you over 6 m.

I agree with 0.125pLL. It would be much better design to use a full length member.

I managed to get finger-jointed 7 m C24 elements, so thankfully I will not be using this splice.