Deflection performance in floor trusses compared to other joist types
alexdorf
| Posted in General Questions on
Looking for any input on this. In planning a build, we’re looking to achieve a high deflection rating, such as L/720 for super stiff floors. The preference is to go with a floor truss.
All else being equal (depth, span, load, etc.), how do floor trusses perform in deflection when compared to other types of joist materials (e.g. I-joist, Common lumber, LVL/PSL, etc.)
thank you!
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Search and download construction details
Replies
I typically design 16”-20” floors trusses at 24” o.c. with L/480 for most floors and 1 1/8” T&G Advantech floor sheathing. For heavy tile and stone floors, we bump it to L/600 or L/720 as required. I try to keep clear spans between 20’-24’ before I need to increase deflection or truss height. I work really close with all truss designers, and compare and review their layout, concentrated loads and details for each and every single truss to make sure we have no problems on the jobsites. All my Builders and framers love that!
I never use dimensional lumber for floor framing because I hate creep, consistent with that style of framing, nor I-joists because I like to design chases in the trusses for the metal trunk and branch system, which makes it easier for HVAC contractors to install it.
LSLs, LVLs, PSLs, Glulams and truss girders are used as headers or support beams making sure they don’t interfere with the duct layout.
You need to make sure that you protect the floor frame in wet areas really well, since you can lose deflection when structural members get wet.
Thank you for the detailed response!
When bumping up to the L/600 - L/720 - what's your floor assembly look like then? My spans will also be kept between 20-24'.
Love that you're using the LVLs as headers.
Thank you!
Different floor manufacturers, suppliers or installers have their own guidelines and preferences to install heavy flooring. I've had to coordinate details showing waterproof membranes on top of the subfloor, latex-modified mortar or epoxy, and then floor stone. Other requirements can have cement backerboard, especially in wet areas. If your are installing it over rigid foam, you may want to use gypcrete.
As precaution, install a calked joint anytime tile or stone flooring meets a wall, a tub or any obstruction. Use elastomeric membranes where you find two different types of substrate material, concrete cracks or control joints.
Most soft wood species are about the same stiffness. Some are slightly stiffer but not enough to matter in most cases.
So what you end up with is geometric differences. When it comes to floor joists, most of the work is done by the top and bottom of the joist, the further these are apart the stiffer the floor is. The middle section is mostly there to keep the top and bottom apart, this is why you can get away with 3/8" OSB web for I-joists.
The is simply no substitute for joist height. Doesn't matter how the joist is constructed, most of the stiffness comes from how tall it is, this is where the rough rule of 1" height for each foot of span.
How wide a joist is also helps but nowhere near as much as height. You can definitely get much more span out of wider joists, this is why you can usually span much longer with I-joists with a 2x4 flange.
One thing to watch is for longer spans (over 20') the deflection criteria no longer works for bounce. What you care about is the natural frequency of the floor. Lot of engineered lumber manufacturers have tools to calculate this to help you get better floors.
When you have loner spans, one way to get better floors without needing to go tall is to span with steel I beams and hang floor joists off the steel. This does make running plumbing and ducting a bit more challenging as the steel beam is in the way but doable with good planning. This way it possible to clear span 30' with only 2x12s (or 11 7/8 I-joist).
Thank you, very informative. There are some steel I-beams I've come across that have adequate pre-bored holes for one of the floors we're working on.
"The is simply no substitute for joist height. Doesn't matter how the joist is constructed, most of the stiffness comes from how tall it is"
That was my (non-engineer) feeling as well.
I wonder how much gluing a good subfloor helps (acting like a wide continuous top flange).
You'd have better results with a glued ceiling (using subfloor materials) as that plane would be in tension under stress. That would be a pretty odd structure though :-) I did some experiments building a variety of torsion box assemblies and was pretty amazed with how long you could free span (used them for clear span work surfaces) these structures. My best effort was using 1/2" oak veneer ply for the top surface, and 5.2mm Revolution ply for the bottom surface with 1 5/8" stringers separating them. https://www.garagejournal.com/forum/threads/woodys-works-garage.263351/page-37#post-6392685
That assembly had about 1/16" deflection over 6ft with a 240 lb point load (me) at center...but only weighed 40 lbs. Applying the same philosophy to floors brings up some interesting options :-)
Floor trusses should be code honestly as they make subsequent mechanical work so much simpler later. My commercial building used them at the 2nd floor and our extensive retrofit was super simple when it came to routing all the new plumbing, electrical and HVAC.
DennisWood,
"Floor trusses should be code honestly as they make subsequent mechanical work so much simpler later."
Amen!
🙌
I really like a 7 ply t&g plywood subfloor glued to the joists and fastened with screws. I also add a bit of construction adhesive in the tongue of the plywood for more airtightness and to eliminate possible squeaking. Also some 2x6 vertical stiff back tying the web trusses together makes for a very solid floor system.
I've often wondered if a row of full depth blocking at mid-span of the joists, with a continuous strap on the under side, would act as a kind of beam perpendicular to the joists and help to stiffen up the entire assembly, by engaging multiple joists when a point load is applied to one of them. "In theory" it should. But would it provide a noticeable difference and be worth the extra bother?
Anybody have experience with that?
Blocking has two purposes: to share the load on one joist with the adjacent joists, and to prevent the joists from rolling under load. What you are describing is a version of load-sharing. It is somewhat effective, but also tends to result in squeaks, and it's not as effective as you might think--the tension force acting linearly on the bottoms of beams is pretty large and it's very difficult to get blocking or straps tight enough and fastened securely enough to maintain structural integrity over time.
One of the biggest differences between engineered joists (I-joist, open web or steel) vs. sawn lumber joists is that engineered joists are more prone to harmonic vibration because of their uniformity; sawn joists have enough variation in density that they are less prone to vibration. There are formulas for predicting vibration and I've heard that they are part of Canadian building codes but I don't have experience with them.