What's new
What's new

# OT: Anyone ever run across a simple table of steel roof truss loads?

#### JasonPAtkins

##### Hot Rolled
I have a question. It's probably better suited to WeldingWeb, but in general I think PM answers have more informed math, lol.

About once a year, someone asks my fab students to do a roof. Sometimes it's over a simple school building, sometimes a simple one-story house.

In the past, I've enlisted the help of a volunteer engineer from the States to help me come up with proper truss sizing and spacing.

The roofs we're talking about are custom welded metal trusses (involving some combination of angle and i-beam), then set in place and purlins (usually rectangular tube) welded across them. I'm carefully overseeing the student welding, as obviously something that could fall and kill someone needs to be done right. Lots of work we do here is seat-of-the-pants engineering, but we like to stay away from that for roofs.

Unfortunately, that engineer isn't available to help me anymore, and I have a new project, which is the biggest span we've been asked to do.

This leads me to my question. For manufactured wood trusses like we're used to in the States, the manufacturers obviously provide loading info, and spacing requirements from one truss to the next. Has anyone ever seen such a table for steel trusses? What I was doing with the engineer on previous projects was essentially reinventing the wheel every time, and I have to imagine that after a hundred years of building buildings out of metal, someone has already run these numbers and published them somewhere.

The information I'm looking for would be like: For a standard double cantilever truss where the top and bottom cords are a pair of 2x1/4" angle, with 1.5"x3/16" crosses, you can support X psf of load if spaced every 2', Y psf of load if spaced every 3', Z pfs of load if spaced every 4', etc.

It would be great to have a chart to be able to look that information up, which would let me design a simple roof, but the design be backed by solid engineering. That's the general case.

Specifically for this case, if anyone knows of an engineering firm that would be interested in donating time to help build a training center in West Africa, put me in touch with them, please! In this case, I came late to the party, after the spans and spacing of the reinforced concrete columns had already been formed, so the spacing is (to my uneducated eye) wider than I would have liked, requiring a really heavy truss. That, combined with the spotty supply of strong steel here, adds to a fun little predicament.

#### DDoug

##### Diamond
The bar joist catalogs contain the roof loading PSF.

However they are flat.
Also, they don't give out the material sizes.

IF you can use the flat bar joist's, go to the eng-tips.com forum, someone uploaded a chart
"Delong Truss" showing the loading tables, but also the construction details.
BUT only for 18" tall and up.

#### Garwood

##### Diamond
I've designed and built my own steel trusses, but I'd probably want to limit my liability if I were asked to do the same for others.

One way I've seen it done is to just use an I-beam for the span with a light framework on top to form the roof pitch. You certainly add some weight this way, but you don't need an engineer for the deflection calcs anymore.

How long is the span?

#### JasonPAtkins

##### Hot Rolled
I've designed and built my own steel trusses, but I'd probably want to limit my liability if I were asked to do the same for others.

One way I've seen it done is to just use an I-beam for the span with a light framework on top to form the roof pitch. You certainly add some weight this way, but you don't need an engineer for the deflection calcs anymore.

How long is the span?

The clear span is 33'. The columns, which I unfortunately wasn't part of the project when they were being poured, are spaced every 11.5' (10m span, columns every 3.5m)

The two ends of the building have block wall where a truss would go, and then there are four sets of columns in between (17m long by the 10m wide).

So the entire 1850 sq ft roof will be supported by four trusses, plus walls on each end. The good news is that roofs here don't include any decking (what's the word, like we use OSB under shingles?) - they just screw the steel roof panels directly to the purlins. So, the roof isn't as heavy as a roof that size would be in the States. Still, it's a lot of weight on six supports. The other fun part is that it's a second story roof on an open atrium, so the floor is 24' away from the roof trusses for most of that area.

We don't have heavy i beam available here either, unless scavenged from the scrap yard. Otherwise, hiring a crane to set heavy beams like you're describing would be an interesting idea. The i beams we get (I think they'd be called mini beams in the US) are IPN100, sometimes IPN120, so it'd be like a narrow flange 5" i beam.

#### Garwood

##### Diamond
The crane bay in my shop is 34' wide. The roof is kind of a hodge podge of heavy I-beams, wood trusses and one truss made from 3x4 .120 tubing that also supports a side roof.

I lifted the 1500 lb I-beams 26' in the air with my 3000 lb forklift and an 11' extension. Lift beam up and balance on extension then pick extension and beam, tilt all the way back and up she goes. Crane would have been nice, but I made due.

If you don't have beams it's a moot point though.

If you PM me your email I can send you a DXF or something of the 34' truss I built with tubing. It has a 3/12 pitch and didn't deflect a measurable amount with the roof on (3/4" plywood, 2x6's and steel).

#### Rudd

##### Stainless
When I was in school (long long ago in a galaxy far away) we learned a graphical method to solve trusses - could be done on a drawing board. Better for us mathmatically challenged Architects to be.
Of course,that assumed pin connections at truss members, and I don't think included forces on any particular connection other than the reactions.
I see there is quite a bit out there on the 'net explaining this method.

For just "gut figgering" purposes (how much space do I allow for structure) on beams and joists, we usually take number of feet,divide by two,call that inches and add 1 to that number, this gives a rough depth.

#### D Bronson

##### Aluminum
Jason:

The go-to people for steel joists in US is the Steel Joist Institute. They have lots of free downloads on designing roofs using steel joists. However, most joists in this country are ordered form a manufacturer - not fabricated by the user. Go to the publications page:

Steel Joist Institute – Everything for steel joists

Also, Vulcraft has a lot of information on their website:

Home

They are a major manufacturer of the joists.

Pay special attention to the end support conditions and joist cross-bracing. Lots of good diagrams in the SJI documents.

Regards,

DB

#### john.k

##### Diamond
Dont forget your wind loadings,if you get violent thunderstorms ,or hurricanes.....Mezzanines are also problematic.....loadings can be huge ,so consider additional direct post supports of the framework.

#### Street

##### Hot Rolled
Unfortunately there is a lot of variables so software has been made to cover this, you still need to input the loadings such as wind and others, these are still done by hand using the Appropriate code in the australian standards. AS1170 covers windloads.
The software is available from SPACE GASS Home Page it is not cheap to buy and is used extensively by structural engineers here as its much quicker to do complex analysis.

Wooden frames can be fudged in spacegass by inputing the material properties of wood in the input data page, however connecting joints made of steel still need to be done by hand from the output reactions you get from spacegass. The joints are the most important part as joints can weaken members as you drill holes in the members making the effective area smaller.

They design for deflection limits and check for strength. Deflection most of the time is the first thing to look at.

It is safer to get a engineer to design it as there are many traps, they are trained in it and can get insurance etc to cover risk.

#### MattiJ

##### Titanium
It is safer to get a engineer to design it as there are many traps, they are trained in it and can get insurance etc to cover risk.
I'm guessing you haven't seen the "building code" in third world countries. Insurance? Engineer? Certified Welder? Structural steel? maybe on their largest skyscrapers if that.
OP is residing(doing volunteer work?) in Guinea-Bissau that has 800 USD GDP per capita vs western world with 60000 USD per capita.

#### EmanuelGoldstein

##### Diamond
While waiting for a \$150,000/yr structural engineer to come along and give you a hand, you could get your feet wet

there's some other truss analysis programs there also. And I could swear my HP-41 had some modules for that. Those are still available used.

Or if you want to get carried away

Tochnog Finite Element Analysis download | SourceForge.net

not the same as knowing what you are doing but better than trial and error

#### Street

##### Hot Rolled
I suppose he could ask for a design to be done for free as a donation from them, but it also needs fabrication, assembly drawings as well as the design aspect.

He could get one design done and duplicate that how ever many times he likes, he would have to be all over the quality aspect of it like using good welders, trained people, good welding wire, good materials, proper assembly, making sure it meets the drawing and proper erection and take on the risk of anything going wrong due to a failure in any part of the process.

If he is doing one offs for end use commercial purposes its really a business and to do good quick cost effective work they use spacegass here and a engineer.

Its up to him how he goes about it, how he manages the risk. so just pointing him in a direction where he should get a good result, as good as what is done here if that is what he wants.
I suppose they would be all over him if anything went wrong with any of his one off buildings, just as much as here no matter their earning GDP figure.

##### Plastic
When I was working in Ghana, the metal we got was soft (weak) to the point we brought a container load for repairs from the states. This was on a farming operation. I’d hate to be the one designing anything using that junk we were getting.

#### thermite

##### Diamond
When I was working in Ghana, the metal we got was soft (weak) to the point we brought a container load for repairs from the states. This was on a farming operation. I’d hate to be the one designing anything using that junk we were getting.

Yah could use the "do one, show me" approach, Construct ONE truss, clear personnel safely, artificially load it, measure deflection, eg: TEST the sumbich.

Doubt you have snow loading, so wind, fail in side-loading - and seriously, here, do more than just "write it on the wall", PREVENTING the daring-ignorant from trying to use it for convenient HOISTING, 'coz you can bet your living ass more than one fool will WANT to do - are your major worries.

My late Da' could actually DO all the Math, not even need paper, but.. as Post Engineer, Watertown Arsenal, Korean War, had to get a guard tower up YESTERDAY, didn't miss it by much more than a week.

First Army tried to climb up his ass over lack of planning and approvals, he just ordered up a bulldozer with winch, cleared the tower, ordered the hook attached to the top of the structure.

Words to the effect:
Now, what we are going to demonstrate right here, right NOW, is to prove that the widely utilized standard government schedule re-purposed pre-fab forest service fire tower meets its already tested and published specs and was bolted-together properly.

When the D6 starts hauling you can watch the sumbich collapse before it BEGINS to tilt the multiple tons of brute force, but blitz FAST concrete slab it is anchored to. That's a lot more than Metro Boston, Mass, wind loads have ever been or are ever likely to be.

Not all CE Majors had been Regimental Construction Sergeant Majors, NACA inspectors, CE flood control gurus before that. This one had, so that was the end of discussion.

Pragmatism.

Some days common sense is all you get to work with, just concentrate on the "sense" more than the "common" and cover yer ass EARLY ON, while change is still cheap, not too-late and already hurting.

First "truss" ever seen on this Earth wasn't built by a computer nor from a table. LAST one built before the sun goes dark may not be, either.

First and last are more likely to be sub-triangulated Bamboo or similar local materials than steel, BTW.

Amazing how so many third-world countries were smarter before Europeans came back to haunt them. Most any tribe in Africa, Polynesia, or Asia could put a sun and rain cover that large into place and not harm a soul.

Exclusive patent on brains having expired when Eve et the apple, don't be TOO surprised some of your students can actually "do the math", even if first-go at it needs a team effort, not just one lone contributor.

Harness that. Hands-on pride in a job well-done pays VERY well in better futures, lasts a long time, can never be taken away by any mortal once earned.

Going forward, it will be better for all-hands, too.

#### DDoug

##### Diamond
Yup, after being in/around/over trusses, bar joist, and general steelwork, your "Mark One Eyeball" get's a calibration of sorts.

Designed by eye as it were.

You can "see" the loading, and what the FEA print out would look like, with all the shades of red and orange.

#### JasonPAtkins

##### Hot Rolled
Thank you all for your useful replies.

Failing a real engineer offering to help, I think the plan is going to have to be the guestimate and check method.

I need to sit down and calculate the total weight of the roof panels and purlins, and then can add that to the published truss weight to make sure it's within the safety margin - that will let me choose the materials to use and the sizing, and then from there we can build one, prop up the ends, and artificially load it to measure deflection. One of the docs I read said they were going for deflection of less than 1/360th of the span. For this 33' span, that'd mean 1.1" of deflection. Sounds like a reasonable goal? Then, once one has been made and proven to deflect by no more than that 1.1", we can move on and make the others.

The ASD tables for standard K joists say that a 33' span on a 18K7 joist can carry 145# per linear foot.

So that's an 18" tall bar joist. I found the chart below, and am wondering if the K7 in the joist name corresponds to a chord size 7 in the chart below?

If that were the case, I'd need to find 1-3/4 x 1-3/4 x 3/16 angle for the top chords, which we can get here (most of the time).

33' roof trusses using 1-3/4 x 1-3/4 x 3/16 angle - seem like I'm in the right ball park? Make one, set it up, load it, measure deflection, and then if I'm ok against the weight of the roof plus some live load, move forward?

(BTW, I totally understand the "you're not an engineer, leave it to professionals, there's too much liability" sentiment. In my context here though, if I don't build the roof, someone with way less mathematical skill and interest in safety is going to do it. At least if I do it I'll know the welds are done properly.)

#### DDoug

##### Diamond
Thank you all for your useful replies.

Failing a real engineer offering to help, I think the plan is going to have to be the guestimate and check method.

I need to sit down and calculate the total weight of the roof panels and purlins, and then can add that to the published truss weight to make sure it's within the safety margin - that will let me choose the materials to use and the sizing, and then from there we can build one, prop up the ends, and artificially load it to measure deflection. One of the docs I read said they were going for deflection of less than 1/360th of the span. For this 33' span, that'd mean 1.1" of deflection. Sounds like a reasonable goal? Then, once one has been made and proven to deflect by no more than that 1.1", we can move on and make the others.

The ASD tables for standard K joists say that a 33' span on a 18K7 joist can carry 145# per linear foot.

So that's an 18" tall bar joist. I found the chart below, and am wondering if the K7 in the joist name corresponds to a chord size 7 in the chart below?

View attachment 276790

If that were the case, I'd need to find 1-3/4 x 1-3/4 x 3/16 angle for the top chords, which we can get here (most of the time).

33' roof trusses using 1-3/4 x 1-3/4 x 3/16 angle - seem like I'm in the right ball park? Make one, set it up, load it, measure deflection, and then if I'm ok against the weight of the roof plus some live load, move forward?

(BTW, I totally understand the "you're not an engineer, leave it to professionals, there's too much liability" sentiment. In my context here though, if I don't build the roof, someone with way less mathematical skill and interest in safety is going to do it. At least if I do it I'll know the welds are done properly.)

Yup, and YUP.
You don't need to zig-bend the center round rod.
While I went to great lengths to make a bender and actually make them, simple cut pieces
work fine, although slightly more welding.
Another compromise is making simple "Vees" out of the rod, thereby making less double welds.

#### JasonPAtkins

##### Hot Rolled
Yup, and YUP.
You don't need to zig-bend the center round rod.
While I went to great lengths to make a bender and actually make them, simple cut pieces
work fine, although slightly more welding.
Another compromise is making simple "Vees" out of the rod, thereby making less double welds.

Thanks - I did do a set of them (12" bar joists) a few years ago that used bent (rebar, round rod isn't easy here). I have a bending jig, that's no problem. But, that table, if I'm reading it correctly, would want thicker round bar than we can get. So, I think we'll go for using angle with pinched ends like you often see, since we can get that and bend it easily.

#### DDoug

##### Diamond
Thanks - I did do a set of them (12" bar joists) a few years ago that used bent (rebar, round rod isn't easy here). I have a bending jig, that's no problem. But, that table, if I'm reading it correctly, would want thicker round bar than we can get. So, I think we'll go for using angle with pinched ends like you often see, since we can get that and bend it easily.
I didn't like that chart, Mine called out the web rods separately.

Can you get square rod (barstock) ? Would save that flattening op.
If you don't flatten all to the same dim, the flange angles weave in/out funny.

I can see it now:
bar joist manufacturing at DuckDuckGo

Last edited:

Replies
4
Views
371
Replies
32
Views
3K
Replies
52
Views
3K
Replies
30
Views
804
Replies
3
Views
518