Lapping floor joists over girder

[Dutchman]

I will give you an engineers answer to this question. The load on the beam is the same whether the joist is continuous or split. The stress and the deflection in the continuous beam are lower at the mid-spans than if the joists are split. This practice is common in commercial and bridge construction. Its just not that practical in residential construction.

As usual, the specifics of each job are not the same, but according to my PE, the load is less if they are not continuous. Keep in mind, this is in reference to a continuous span or simple span ridge. I'm not sure how it works out in the case of a floor joist with load bearing walls on either end.

His note back to me after my request to chime in:

Bob,
The statics of the problem dictate that a two span continuous beam of equal spans, carrying a uniform load imparts a 25% higher reaction at the center column than if the continuous beam were actuall cut into two simple span beams. The beam diagrams are attached.

[dlhunter]

I just can't get my head around that.

[Dutchman]

I just can't get my head around that.

I'm not an engineer either. I trust him though, he just designed the airport expansion in our town, this type of stuff is like building a simple wall to us.

[mike maines]

Think about it this way: with two separate beams, the center support carries exactly half the load from each side. With a continuous beam, the center support acts as a cantilever, and supports a bit more of the load from each side.

There is not more load on the beams, it's the reaction load in the column (and the stresses inside the beam) we're talking about. If you notice, the end columns see less reaction load than the center column.

It's intuitive that a continuous beam is stronger or stiffer than two individual beams, right? That intuition has some math behind it.

[Dick Hackbarth]

I think we have gone over this problem in the past. I seem to remember a long drawn out discussion with a guy by the name of Robert Riversong about this problem a couple years ago. Here are some of my basic thoughts on the matter, as I read through the thread. To set a few ground rules for the discussion; we have two equal spans as the OP says, two exterior found. walls and a beam down the center. The math gets kinda crazy pretty quickly if the span lengths are not very close to equal; to the extreme condition where on a short second span you could get uplift at its exterior reaction. We have the same DL & LL on each span, but we might have one span with full LL, while the one on the other side of the beam has very little LL. You will see that in some instances doing the continuous jsts. over the beam is o.k. from the jst. standpoint, but detrimental from the beam and pier/column standpoint. My comments relate to various posts, and that post # starts each comment.

#2. There are some obvious handling advantages to one long 2x jst. as long as you can get them. Normally, you guys think in terms of camber up when setting jsts., and you don’t want any defects in the middle half of a typical single span jst. near the bottom edge, that’s the tension edge, and knots, cracks, checking, notches, and the like are a big problem. With the two span cont. (continuous) jst. the stresses in the jst. are the same over the beam, as they would have been in the simple span jsts., that’s why you were told it’s o.k. to do this, most of the time. But, remember that now the top edge of the jst. are in tension, over the beam, so now avoid knots, checks, etc. in this region, at the top. Now you need a long 2x and it better be good on both edges. Lumber grading comes into play. With I jsts. or parallel chord fl. trusses consistency and length are not an issue, but the I jsts. need web stiffener blocks over the beam. And, they all likely need roll-over blocking in the jst. spacing over the beam. Another advantage of the cont. jst. is that it is stiffer w.r.t. deflection, so the deflection at center span will be less with this jst. arrangement.

#16. Mighty good point. Follow the plans unless you want to assume the designers responsibilities and liabilities.

#17. The builder is usually the lowest bidder also, if you were casting any aspersions. But, I don’t generally love Archs. either.

#24, 25 & 31. Duchman is right and interprets what his engr. told him correctly. And, his attachment is even better and is exactly what I’d have drawn for him. Give or take a few inches or lbs. of extra 2x, the total load hasn’t changed enough for me to worry about Olllee’s concern of more weight. As the attachment shows, the two simple span jsts. load the beam with (2)(wl/2) and this is intuitively obvious to all of you, a simple beam, uniformly loaded. Whereas, the continuous jsts. cause a beam loading of (10wl/8), not so intuitively obvious any more, damn math and engineering mechanics. But, we also see that the two ext. reactions are (3wl/8), so the total load is accounted for, (10wl/8) + 2(3wl/8) = (2wl). This is the static equilibrium condition for that jst. arrangement. MikeM @ #34 calls this the cantilever affect. An engineer would say that for these continuous jsts. the negative moment (-M) over the beam draws or pulls load or shear toward the -M region or the beam.

Try this experiment: take a yard stick and support it 6" from each end, this represent a 2' long simple beam, with (wl/2) reaction at each end; imagine some uniform load on the yard stick and we can see the deflection and bending with both being max. at center of this simple span; obviously the load can’t be so stiff as to span of its own accord; now, on one end only (representing the end at the center line beam), press down on the yard stick, you can actually lift the other end off its support (the ext. wall support) or cause uplift at that reaction, now we have (wl) reaction on the beam, not just (5wl/8); your downward pressure is akin to the negative moment caused by the adjacent, continuous, span; never mind the exact magnitude of this -M at the moment, that’s the messy engineering part, and all the messier of the spans aren’t equal, or the loads uniform. But, what we have shown is that the application (existence) of the -M will draw load/shear/reaction to the region of the -M. When w & l are the same on each span these -M’s will be identical and balanced about the axis of the beam. If w or l change the -M’s won’t be balanced and this can lead to uplift. That’s why the IRC talks about a min. back span for cantilever situations, so as never to cause uplift.

The reason this continuous jst. arrangement works for you guys is that we rarely see the max. total load and we have factors of safety on the jsts., beams and columns. So, the beam or columns are not very likely going to see their design loads or stresses. Whereas, on Dutchman’s ridge beam you have much larger contributing area, longer beam spans, more likelihood of full roof snow load or drifting, consequently much larger beam loads, reactions and column loads. And, the columns are likely 2 stories high so loading and stability are critical on them. An engineer would not be wise to ignore this condition, nor should you.

#26. The continuous jst. is stiffer w.r.t deflection and vibration or bounciness, because of the adjacent continuous span coupled to it. However, it needs to be the same strength in bending in either case, since the -M = wl^2/8 over the beam, in the continuous case, and the simple jst. +M at center span is also wl^2/8. And, these bending moments lead to the max. fiber stress in the jst., thus with the same M, the bending stresses will be the same. Stiffness and strength are related through the engineering mechanics of the problem, but they are different things. But then, the shear stress in the continuous jst. is now a function of a higher shear (5wl/8) near the beam, and this could be a problem. Concentrated loads near the ends of jsts. make this shear problem worse. The real killer is the guy who designs and builds a fairly light fl. system, and then moves a baby grand piano onto it. Those three leg loads bear on only a couple jsts.

You may get away with the continuous jsts. most of the time, but the admonition to follow the plans is a good one from your own liability standpoint. At least ask the Arch. or Engr. if you can do it.

[dlhunter]

Well it took me three times reading it but I think you've explained it to me. Thanks

[MarkMc]

The real killer is the guy who designs and builds a fairly light fl. system, and then moves a baby grand piano onto it. Those three leg loads bear on only a couple jsts.

Got that one once on a remod. The piano was to be moved from adjacent room which had a 10" flexicore on a ~22' span; new room as existing was 2x6@24 w/ 9'4" span, 5/8 cdx +carpet. That was fun.

[jimAKAblue]

Think about it this way: with two separate beams, the center support carries exactly half the load from each side. With a continuous beam, the center support acts as a cantilever, and supports a bit more of the load from each side.

There is not more load on the beams, it's the reaction load in the column (and the stresses inside the beam) we're talking about. If you notice, the end columns see less reaction load than the center column.

It's intuitive that a continuous beam is stronger or stiffer than two individual beams, right? That intuition has some math behind it.

I like this explanation best: you explained it in one paragraph with no formulas!

We have a winner!

[Dutchman]

Great discussion guys.

Now, let's keep these fundamentals in our heads when we're building things.

This, actually, is one of those threads that I find very productive and enjoyable.

[mfsmithbuilder]

I like this explanation best: you explained it in one paragraph with no formulas!

jim.... i don't have your current email address

good discussion... thanks to all

[jimAKAblue]

jim.... i don't have your current email address

good discussion... thanks to all

Mike! Good to hear from you again! I'll shoot you an email tomorrow. I almost shot you one last month: just curious how biz has been.

[Stovepipe]

removed comments

[DaveProfitt]

Dick's explanation was very thorough, though it might leave the average GC and his framer scratching their heads and wondering what the plain English version might be.

Well, I'm not sure there really is one, but from 10 yrs on Dicks side of the table and 27 on the other side, here's my translation.

Don't lap your floor joists over a girder unless you really want your floor to squeak under the wall plate and follow the framing plan if there is one. If it doesn't look right - question it. Everybody makes mistakes, some more often or with greater ramifications that others.

If you are framing with dimension lumber with no engineered plan, keep in mind the building codes are written to yield a structure "suitable for use as a habitable dwelling". There's nothing in there about squeaks, bouncy floors and unhappy homeowners.

On residential construction framing with dimension lumber, most of the time it won't matter if you run continuous length joists and in most cases it will give you a superior floor over using multiple shorts. There are a few exceptions to that, notably involving large dim lumber and heavy loads on some spans (like the piano) and light loads in adjacent spans. But, most applications will benefit from long length lumber.

Framing with I-joists can be a different matter. If the I-joist layout is prepared accurately, it may specify the use of shorter pieces in lieu of longer lengths. As I-joists are sold by the LF, normally with no premium for long lengths, it can be tempting to ignore the layout and run them continuous. You can get the best of both worlds by ordering the long length joists, then cutting then about half the depth over the specified break point before installing your blocking (only over the bearing wall, by the way).

The ruler is a good example of what the effects of, say running a 36' joist over a 16', 14', and 6' span. the uplift on the 6' span can cause some headaches, and most joist design programs will fail that application. Cutting the joist partially thru (from the top down), takes the uplift out of the short span.

The engineering theory behind the continuous beam span was developed for steel components and though it applies equally to most wood-based applications, there are exceptions.

Steel is assumed to be relatively consistent in composition (it doesn't have knots, splits and barked edges). Engineered lumber FOR THE MOST PART does not have these issues either, but dimension lumber does. Beware of the I-joist that has been stored outside on the yard for extended periods. If it doesn't look new, it won't perform like it either, as weather takes it's toll on the integrity of the wood fibers and the bond of the adhesives. Also anywhere the flange has been dinked by a careless forklift driver or over-pressure banding has been compromised. if you can't cut that section out, don't use the joist.

Also, the steel design theory is based on one of 2 different support conditions, welded to perpendicular support or a pivot connection under the bottom of the beam.

Neither translates exactly to a wood frame construction situation. The welded connection locks both the top and bottom flanges (where all the stress is) in a way that nailing a piece of wood cannot approach. The pivot is pretty close but by engineering theory it's assumed to be in the center of the support structure (wall, post or pier). In reality, just as the example using the yardstick indicated, the joist pivots on the nearest edge of the support. If you're bearing on a triple girder, the other side of that face will see uplift. That lap is an under-the-wall squeak in the making.

As long as you have enough bearing surface on your beam or girder (normally assumed to be 1 1/2" on dimension lumber, and determined by load on engineered lumber) to support a butt joint connection, you will have a quieter floor and if the member sizing is correct, strong enough to do the rest of the job effectively.

A few other notes:
- If your engineered framing plan calls for a engineered lumber or steel post - don't run the top plate over the post. The span member needs to bear directly on the post.

-The loading in a structure starts at the top and works it way down thru the structure. Framing starts from the bottom and works it's way up. Make sure you have addressed all the load points from the roof down thru the rest of the structure when you lay that first floor system.

-Always solid block between the joists when you cross a load bearing wall or girder. If you are using dimension lumber, rip a half inch off the blocking depth before installing - and see below.

-If you're not framing your floors with engineered lumber - ask yourself why.