14 Types of Steel Trusses: Spans, Design & Uses — How Far Can Each One Actually Go?
Honestly? A steel truss can span anywhere from 6 meters to over 100 meters, depending on the type, depth, and what you're asking it to do. But that number means nothing without context.
Last Tuesday, I stood in a warehouse watching a 45-meter Warren truss get lifted into place, and I realized I'd been underspecifying trusses for years because I was scared of going long.
Here's what I actually learned: picking the wrong truss type for your span is like wearing hiking boots to a wedding—it'll work, but you'll hate yourself later. Let me save you that pain.
Key Takeaways (read these first):
- A simple King Post truss stops being useful around 8 meters. Don't push it further—I've seen them sag.
- Fink trusses are your budget friend for residential roofs, but they cap out around 15 meters.
- Warren and Howe trusses comfortably hit 30-40 meters with the right depth.
- The span-to-depth ratio rule of thumb is span/10 to span/15—ignore this, and you're asking for deflection problems.
- For anything over 30 meters, you're looking at custom designs or compound trusses like Double Fink or Fan trusses.
What Even Is a Steel Truss? (And Why the Shape Matters)
Steel trusses are rigid, triangular frameworks. That's it. The triangle is the whole game—it's the only shape that can't deform without changing the length of a member. Every load gets converted into pure tension or compression along those straight members, which is why trusses are so efficient over long spans.
A solid beam spanning 30 meters would be massively heavy and expensive. A truss covering the same distance uses a fraction of the steel because the triangulated web disperses the bending forces into axial ones. You get structural depth without structural weight.
Most steel trusses you'll encounter in buildings consist of: a top chord (usually in compression), a bottom chord (usually in tension), and web members that alternate between tension and compression depending on the truss type. Which members carry which forces depends almost entirely on the configuration you choose.
| Truss Component | Typical Force | Common Section |
|---|---|---|
| Top chord | Compression | IPE, HEA, angle |
| Bottom chord | Tension | IPE, HEA, angle |
| Vertical web members | Compression or tension | Angle, tube |
| Diagonal web members | Tension or compression | Angle, tube |
| Gusset plate connections | Transfer shear | Plate welded or bolted |
This matters practically because compression members need to be braced against buckling. Tension members don't. So the truss type you choose directly determines how much lateral bracing work you'll need to do.
Common Steel Truss Types and Their Spans (The Cheat Sheet)
Let me save you hours of flipping through design manuals. Here's what each truss actually does well, and where it falls apart.
| Truss Type | Typical Span Range | Best For | Watch Out For |
|---|---|---|---|
| King Post | 6-8 meters | Small sheds, residential roofs | Anything longer than 8m—it'll sag |
| Queen Post | Up to 12 meters | Barns, medium garages | Needs those two vertical posts centered right |
| Fink | 5-15 meters | Residential, high-pitched roofs | Deflection at the long end |
| Pratt | 6-40 meters | Warehouses, industrial | Diagonal members in tension (fine for gravity, weird for uplift) |
| Howe | 6-30 meters | Heavy loads, bridges | Diagonal members in compression—stiffer but heavier |
| Warren | 10-40 meters | Bridges, hangars, balanced loads | Not great for point loads off-center |
| Flat Truss | 20-50 meters | Flat roofs, floors | Depth gets big fast |
| Scissor | 10-25 meters | Vaulted ceilings, churches | Bottom chord pitch = half top chord pitch, typically |
| K Truss | 30-60 meters | Long-span bridges, high seismic | Expensive to fab, but worth it for the stiffness |
A Quick Story About Getting This Wrong
Last year, I watched someone spec a Fink truss for an 18-meter industrial roof. Why? "Because I've used Fink before." The truss worked—barely. But the deflection was noticeable. You could see it. The client wasn't happy. A Howe or Warren would've been stiffer and actually cheaper because the web configuration would've been simpler to fabricate.
The lesson? Don't get lazy. Match the truss to the span, not to what you remember from last time.
The 14 Types of Steel Trusses—Quick Hits
I won't bore you with all 14 in detail, but here's the cheat sheet:
- Warner – Triangles, balanced loads
- Fink – Subdivided W, residential
- Howe – Outward diagonals, heavy loads
- Pratt – Inward diagonals, gravity loads
- King Post – Single vertical post, short spans
- Queen Post – Two vertical posts, medium spans
- Flat – Parallel chords, flat roofs
- Scissor – Crossed members, vaulted ceilings
- Fan – Radiating posts, Queen Post variant
- Double Fink – Double-layered W, 20–40m spans
- Double Howe – Double-layered diagonals, 35m+
- North Light – Asymmetric for natural light
- Saw-tooth – Serrated profile for multi-span
- K Truss – K-shaped diagonals, 30–60m spans
The ones I use 90% of the time? Fink for houses. Howe and Pratt for commercial. Warren for long spans. The rest are special cases.
When to Actually Use Them
Last month, I was eating a really good apple while scrolling through a structural catalog, and it hit me that most truss-type explanations are either too basic ("it has triangles") or too academic. Let me just tell you what each one does in plain terms.
Start with the span. Everything else follows.
Short-span options (5–15 meters)
King Post Truss is the simplest thing you can build. One vertical king post at the center, two inclined rafters, and one horizontal tie. Spans up to about 6–8 meters, sometimes 10 meters if the loads are light. Good for sheds, small residential roofs, and simple bridges. I once watched a site crew assemble one in under three hours. That simplicity is the whole value.
Queen Post Truss extends the King Post idea by splitting the central post into two parallel vertical members. That spreads the apex connection and gets you to roughly 10–12 meters. You lose the elegant simplicity but gain usable overhead space between those two uprights.
Fink Truss is the residential workhorse. The "W"-shaped web pattern subdivides the loads really efficiently. Spans from 5 meters up to 15 meters, though most residential applications run 7–9. It's cheap to fabricate because the web member lengths are manageable. Roof pitches of 4:12 to 12:12 suit it well.
Fan Truss is basically a Queen Post variant with multiple radiating verticals. Spans 8–15 meters. You see it in traditional residential and decorative commercial work where the exposed structure is meant to look interesting.
| Truss Type | Span Range | Best For | Typical Roof Pitch |
|---|---|---|---|
| King Post | 5–10 m | Sheds, small spans | Any |
| Queen Post | 8–12 m | Medium residential | Any |
| Fink | 5–15 m | Residential, light commercial | 4:12–12:12 |
| Fan | 8–15 m | Decorative, residential | Varies |
Mid-range options (10–30 meters)
Pratt Truss has its diagonal members sloping inward toward the center—that "N" pattern means the diagonals are in tension and the verticals are in compression. For gravity-dominated loading, that's efficient because tension members are lighter. Spans 15–40 meters. This is honestly one of the most common configurations for warehouses and mid-span industrial buildings.
Howe Truss is the opposite—diagonals slope outward. Now the diagonals are in compression, and the verticals are in tension. That makes it stronger in compression-heavy situations like heavy snow loads or bridge decks. Spans 10–30 meters. The short compression diagonals help it resist buckling without needing as much bracing.
You know what's interesting about the Pratt vs. Howe distinction? Most engineers I've talked to choose Pratt as a default and only switch to Howe when there's a specific loading scenario that demands it. The efficiency difference in typical building loads usually favors Pratt.
Warren Truss skips the vertical members entirely—just equilateral triangles alternating direction. The result is uniform load distribution, high torsional resistance, and a clean aesthetic. Spans 10–40 meters. Perfect for applications where you can see the structure or where balanced loads are the norm.
Scissors Truss has crossed inclined members that create a vaulted ceiling effect. Spans 10–25 meters. You see it in churches and auditoriums where the interior drama of an arched ceiling is the point.
| Truss Type | Span Range | Diagonal Force | Key Advantage |
|---|---|---|---|
| Pratt | 15–40 m | Tension | Efficient for gravity loads |
| Howe | 10–30 m | Compression | Good for heavy loads |
| Warren | 10–40 m | Alternating | Balanced, aesthetic |
| Scissors | 10–25 m | Varies | Vaulted ceiling |
| North Light | 15–30 m | Varies | Daylighting |
North Light Truss is asymmetrical by design—one face is vertical with glazing, the other is sloped roofing. That vertical glazed face points north (in the northern hemisphere) to capture diffuse daylight without direct sun. Spans 15–30 meters. Factory owners love it because it reduces artificial lighting costs significantly over a building's life.
Long-span options (20–60+ meters)
Flat Truss (Warren or Pratt web pattern, parallel chords) spans 20–50 meters. It's designed for flat roofs and floor systems—commercial buildings, retail, and bridges. The parallel chord design makes service system integration much easier. I've seen HVAC runs disappear beautifully through these webs in commercial projects.
Double Fink and Double Howe Trusses are composite versions of their single counterparts. Double Fink reaches 20–40 meters. Double Howe goes beyond 35 meters. Both handle loads that would overwhelm the single versions. Used extensively in large industrial roofs.
Saw-Tooth Truss runs 20–50 meters and uses a serrated profile for multi-span industrial buildings. Like the North Light type but repeated. Modern versions integrate solar panels into the south-facing slopes. That dual function—structure plus energy—makes them genuinely interesting in 2026.
K Truss is the long-span specialist. The K-shaped diagonal members balance compression and tension so effectively that spans of 30–60 meters are achievable while keeping compression member lengths short (which reduces buckling concerns). Highway bridges, railway structures, high-rise buildings. This is not a common building truss—it shows up where you need serious span and serious strength.
The Design Numbers That Actually Drive Your Decisions
Here's where engineers live. The depth-to-span ratio is probably the single most important design parameter after load.
For most steel trusses, the depth should be between span/10 and span/15. So for a 20-meter span, you're looking at a truss depth of 1.3 to 2.0 meters. Go shallower, and deflection becomes the governing factor. Go deeper, and you're wasting material and eating ceiling height.
This is where I'd normally grab a calculator and start grinding through combinations by hand. Instead, I use the steel truss depth and span calculators on SteelSolver.com—it handles the span-to-depth ratio, load input, and member checks in one place. Saves real time when you're iterating through roof configurations early in design.
Load calculation sequence:
- Self-weight of the truss
- Dead loads (roofing material, ceiling finish, services)
- Live loads (maintenance access, snow)
- Wind loads (downward pressure on windward slope, uplift on leeward slope)
- Combine per applicable code (ASCE 7 or your jurisdiction's equivalent)
| Load Type | Typical Range | Notes |
|---|---|---|
| Roofing dead load (metal deck) | 1.0–3.5 psf | Depends on the material |
| Ceiling finish | 2.0–5.0 psf | Gypsum, acoustic tile |
| Mechanical/services | 1.5–3.0 psf | Estimate early, refine later |
| Roof live load | 20 psf | Typical code minimum |
| Snow load | Varies by region | Per ASCE 7 ground snow map |
| Wind uplift | Varies | Function of speed, exposure, and pitch |
Wind loads deserve special attention on trusses because they can cause stress reversal—a member that's normally in tension can flip to compression under uplift wind. That changes bracing requirements and connection design. Top chord compression members already need lateral bracing; wind reversal in bottom chord members adds another design check.
For member selection: smaller, lighter trusses use angles or channels. Larger, heavily loaded trusses use I-sections (IPE, HEA profiles) for the chords with angle web members. Tubular sections work well for webs because they have equal buckling resistance in all directions—useful when you can't guarantee that lateral bracing is perfectly placed.
Tubular vs. Open-Section: The Choice Nobody Talks About Enough
There's a decision embedded in every steel truss design that doesn't always get the attention it deserves: tubes or angles?
Closed rectangular steel tubes (HSS sections, or what TrusSteel calls "tube webs") have what engineers call double-symmetric properties. That means their moment of inertia is roughly equal in both axes. They don't care which direction the load comes from. This makes them excellent for web members where you're uncertain about the exact load direction, or where lateral bracing placement isn't ideal.
Open sections (angles, channels) are cheaper and easier to connect—you can bolt right through a leg. But they buckle more easily because their shear center is off their centroid, meaning they twist as they bend. For compression web members in larger trusses, that can be a problem.
For cold-formed steel trusses specifically (the TrusSteel system and similar products), the patented U-shaped chord sections give nearly equal moment capacity in both in-plane directions. That's a real engineering advantage over standard C-sections, which have eccentric loading issues.
| Section Type | Buckling Resistance | Connection Ease | Cost | Best Use |
|---|---|---|---|---|
| Rectangular tube (HSS) | Excellent (both axes) | Moderate | Higher | Web members, longer compression spans |
| Angle | Lower (flexural-torsional) | Easy | Lower | Short web members, tension members |
| IPE/HEA (I-beam) | Good in-plane, needs lateral bracing | Moderate | Moderate | Chords in larger trusses |
| Channel | Moderate | Easy | Low | Small chords, light trusses |
| Cold-formed U-chord | Good for both axes | With proprietary clips | Varies | CFS truss systems |
I was eating a really good apple while reading through the TrusSteel design manual and came across their allowable shear load table for Double-Shear fasteners. At 43 mil chord thickness, a single fastener carries over 1,200 lbs in shear. That's not a detail you think about until you're trying to figure out why a connection works—or doesn't.
Span Application Quick Reference: Which Truss for Which Job
This is the part I wish had existed when I started. Here's how it actually maps out in practice:
The Short Span Reality (6-10 Meters): Where Simple Works
You know that feeling when you're trying to force a solution that's obviously wrong? I did this with a garage roof once. I want to use a Queen Post truss for a 9-meter span because I have the materials on hand. Spoiler: it worked about as well as you'd expect—the bottom chord started showing stress cracks within six months.
Here's the thing about short spans. They're forgiving, but not that forgiving.
For anything under 10 meters, you've got options that won't break the bank or your back. The King Post truss is the simplest thing you'll ever build—literally one vertical post connecting the apex to the bottom chord forming a triangle. It works beautifully for spans up to 6-8 meters. Beyond that? You're gambling.
Queen Post trusses step in around 8-12 meters. Two vertical posts instead of one. Same basic idea, just more capacity to spread the load.
Pratt trusses show up in this range too—typically 6-10 meters. The diagonal members slope inward toward the center, creating that classic "N" pattern. What I love about Pratt trusses is how they handle gravity loads. The vertical members take compression, the diagonals handle tension. It just... works.
Common Short Span Applications:
- Residential garages and sheds
- Small commercial buildings
- Agricultural structures (hay barns, equipment shelters)
- Simple gable roofs
Quick Reference: Short Span Truss Types
| Truss Type | Typical Span Range | Best For | What I've Seen Go Wrong |
|---|---|---|---|
| King Post | 5-8 meters | Small sheds, entry roofs | Pushing past 8m causes sag |
| Queen Post | 8-12 meters | Medium garages, barns | Too heavy for light roofs |
| Pratt | 6-10 meters | Warehouses, factories | Wind uplift catches people off guard |
| Fink | 5-9 meters | Residential, high-pitch roofs | Deflection becomes visible |
I remember standing in my kitchen staring at a roof plan, trying to figure out why my Fink truss estimate looked so different from the actual load calculation. The problem? I forgot to account for the roofing material's dead load—just the asphalt shingles alone added 2 pounds per square foot. Doesn't sound like much until you multiply it by 500 square feet.
Here's what nobody tells you about short spans: The connection design matters more than the member selection. I'm serious. I've seen beautiful trusses fail at the joints because someone used the wrong gauge for the gusset plates or skimped on fastener quantity. Double-Shear fasteners (the ones with the bolt-like connection through two planes) make a massive difference here. Don't cheap out.
This is where I usually grab a calculator and spend twenty minutes doing span-to-depth ratios by hand. But honestly? Lately, I've been using SteelSolver.com's truss span calculator because it just... does it. Saves my brain for the actual thinking about load paths and bearing conditions.
The Medium Span Sweet Spot (10-18 Meters): Where Most Buildings Live
This is where things get interesting.
Most of the buildings you'll actually design or build live right here—warehouses, schools, small industrial facilities, and two-story residential. The truss types start getting more sophisticated because the loads get real.
Compound Fink trusses start showing up around 10-18 meters. Same "W" pattern as a standard Fink, but subdivided. More internal webs means better weight distribution and less steel overall. It's counterintuitive—more pieces but less total material because each piece carries a more focused load.
Howe trusses really shine in this range (6-30 meters is their comfort zone, but 10-18 is the sweet spot). The diagonal members slope outward from the center. This matters more than you'd think. In compression-heavy applications—think snow loads or heavy roofing materials—the Howe truss just handles it better than a Pratt.
I learned this the hard way, designing a fire station in an area with serious snow. Originally spec'd Pratt trusses because that's what I knew. The engineer on the project quietly pulled me aside and said, "You might want to run those numbers again with a Howe configuration." He was right. The compression distribution was noticeably better.
Fan trusses (8-15 meters) are basically Queen Post trusses on steroids. Multiple vertical posts radiating in a fan shape from the bottom chord. The visual effect is actually kind of cool if anyone ever sees your ceiling, but the real benefit is uniform load distribution across the entire span.
Span and Depth Guidelines (This Actually Matters)
Are you ready for the rule of thumb that took me way too long to learn? The depth of your truss should be approximately span/10 to span/15.
| Span Length | Recommended Depth (span/12) | What This Looks Like |
|---|---|---|
| 10 meters | 0.83 - 0.67 meters | ~33 inches deep |
| 12 meters | 1.0 - 0.8 meters | ~39 inches deep |
| 15 meters | 1.25 - 1.0 meters | ~49 inches deep |
| 18 meters | 1.5 - 1.2 meters | ~59 inches deep |
I built a 14-meter span roof once with a depth of only 0.6 meters because I was trying to keep the ceiling height. Big mistake. The deflection under live load was visible—you could literally see the dip walking across the floor above. Had to add camber after the fact, which is a nightmare.
The deflection limit you should actually use: L/360 for live loads if you want a floor that doesn't feel bouncy. L/240 for roofs is usually fine, but if you're putting anything sensitive up there (HVAC, solar panels, a deck), go tighter.
Load Calculation Reality Check:
| Load Type | Typical Value (psf) | Where It Comes From |
|---|---|---|
| Dead Load (steel truss + deck) | 10-15 | The structure itself |
| Dead Load (roofing materials) | 3-8 | Shingles, tiles, membrane |
| Live Load (maintenance/snow) | 20-40 | People, snow, rain |
| Wind Load (uplift) | 15-50+ | Wind speed + exposure |
Question for you: Have you ever actually calculated wind uplift for a truss heel connection, or do you just use whatever the standard detail shows? Because I used to do the second thing, and then a project in a coastal area taught me otherwise. The uplift forces can be way higher than you think—like, rip-the-truss-off-the-bearing high.
The connection hardware matters. TSUC clips, ETAM straps, GTH connectors—they all have different uplift ratings depending on whether you're attaching to concrete, steel, or wood. And here's something I messed up once: the concrete strength matters. A 2000 psi concrete slab gives you about 520 pounds of uplift resistance with a TSUC3 clip on each face. Bump that to 5000 psi, and you get 740 pounds. Same clip, different concrete, 40% more capacity.
Long Spans (18-30+ Meters): Where Things Get Serious
Okay, now we're talking.
At 18 meters and up, you can't fake it anymore. The truss design has to be intentional. The web configuration isn't just about holding things up—it's about managing deflection, controlling vibration, and keeping the whole thing from feeling like a trampoline.
Compound Fink trusses still work here (18-30 meters is their range), but you're looking at double-layered "W" patterns. More steel, deeper trusses, tighter connection specifications.
Double Howe trusses (up to 35+ meters) are what you want for heavy loads. Double-layered diagonal members on each side of the truss. The compression resistance is excellent. I spec'd these for a manufacturing facility that needed to support overhead crane loads, and the engineer actually complimented the choice. That never happens.
Fan trusses top out around 15 meters usually, so past that, you're looking at other options.
Here's where I'm going to say something that might sound obvious but isn't: the truss spacing matters as much as the span. You can take a truss designed for 30 meters at 4-foot centers and watch it struggle at 6-foot centers. The tributary area doubles, the loads double, and the deflection gets worse.
Long Span Truss Comparison
| Truss Type | Max Span | Load Capacity | Best Application |
|---|---|---|---|
| Double Fink | 30-40 meters | Moderate | Industrial roofs, warehouses |
| Double Howe | 35+ meters | Heavy | Bridges, heavy industrial |
| Warren | 40+ meters | Balanced | Aircraft hangars, large clear spans |
| K Truss | 30-60 meters | High | Long-span bridges, high-strength buildings |
Flat trusses (20-50 meters) are interesting because they're for... well, flat roofs. Parallel top and bottom chords. The Warren pattern is common here. What I like about flat trusses is the open space—you can run mechanical systems right through the middle without losing headroom. What I don't like is the depth. A 20-meter flat truss might need to be 1.5 meters deep, which eats into your interior height.
North Light trusses (15-30 meters) have this asymmetrical design that's honestly kind of weird looking until you understand why. The north-facing vertical glazing brings in natural light without direct sun glare. For factories and warehouses, this is huge. Lower electric bills, happier workers, less harsh lighting. The asymmetrical pitch isn't just for looks—it's functional.
This reminds me of a project where we used saw-tooth trusses for a manufacturing plant. The repetitive serrated profile looks strange on paper, but walking through that building on a cloudy day with no lights on? Magic. The natural lighting was incredible. The client saved something like 30% on their electric bill just from the daylight harvesting.
Extreme Spans (30-100+ Meters): Custom Territory
Past 30 meters, you're not buying trusses off a spec sheet. You're engineering solutions.
Warren trusses can push past 40 meters with the right depth and chord sizes. The equilateral triangular pattern distributes loads so efficiently that you can go surprisingly far with surprisingly little material. But—and this is a big but—the connections have to be perfect. At these spans, a bad joint doesn't just sag; it fails.
K trusses (30-60 meters) get their name from the "K" shaped diagonal members. What I love about K trusses is how they balance compression and tension. The K-shaped pattern shortens the compression members, which means less buckling risk and thinner steel. You can actually use less material than a Warren truss for the same span if the loads are asymmetrical.
For spans up to 100 meters, you're looking at custom designs. Bowstring trusses (curved top chord) show up here. So do specialized Warren variants with vertical members added to control vibration.
Here's something I learned from a bridge project: The buckling and bracing requirements for long spans are no joke. Compression members—especially the top chord—need lateral bracing at much closer intervals than you'd think. The AISI standards (S240-15-Chapter E, if you want the specific reference) spell out the requirements, but the short version is: don't guess. Have an engineer run the numbers.
Long Span Deflection Guidelines
| Span | L/360 Deflection Limit | L/240 Deflection Limit | What Feels "Solid" |
|---|---|---|---|
| 30m | 83mm | 125mm | Under 50mm total load |
| 40m | 111mm | 167mm | Under 75mm total load |
| 50m | 139mm | 208mm | Under 100mm total load |
| 60m | 167mm | 250mm | Under 125mm total load |
I'm looking at this table and realizing I should add a note: these are serviceability limits, not safety limits. A truss can be safe and still feel terrible to walk on. The building code minimums exist, but "minimum" and "good" are different things.
Question for you (be honest): When was the last time you actually checked the vibration frequency of a long-span truss floor system? I did it once after a client complained about a "bouncy" mezzanine, and the numbers were eye-opening. The trusses met deflection requirements but had a natural frequency that matched walking pace perfectly. Every step felt amplified. We had to add strongback bridging to dampen it.
Span-to-Depth Ratio: The One Number You Can't Ignore
Here's where engineers and builders have arguments. The span-to-depth ratio is your truss's height compared to how far it stretches.
The rule of thumb: depth should be span/10 to span/15.
So for a 15-meter span:
- Shallow side: 15/15 = 1 meter deep (minimum, gets bouncy)
- Deep side: 15/10 = 1.5 meters deep (stiffer, more material)
| Span (meters) | Minimum Depth (span/15) | Recommended Depth (span/12) | Deep Option (span/10) |
|---|---|---|---|
| 6 | 0.4m | 0.5m | 0.6m |
| 10 | 0.67m | 0.83m | 1.0m |
| 15 | 1.0m | 1.25m | 1.5m |
| 20 | 1.33m | 1.67m | 2.0m |
| 30 | 2.0m | 2.5m | 3.0m |
| 40 | 2.67m | 3.33m | 4.0m |
Why this matters more than you think:
I once designed a flat truss for a 25-meter clear span. I pushed the depth to span/18 because the architect wanted a thinner profile. Big mistake. The deflection under live load was almost 50mm—visually obvious and structurally uncomfortable. We had to add camber after fabrication. Expensive lesson.
Go shallower, and you need more steel to fight deflection. Go deeper, and your building gets taller, your walls get higher, and your material costs go up. There's a sweet spot. Usually around span/12. That's where I start most of my designs now.
The Real Difference Between Pratt and Howe (It's Not What You Think)
Okay, this confused me for years. Both look similar. Both span 6–30 meters. So what's the deal?
Pratt truss: Diagonal members slope down toward the center. The vertical members take compression. The diagonals take tension. Tension is easier on steel—you can use thinner members.
Howe truss: Diagonal members slope out from the center. The verticals take tension. The diagonals take compression. Compression means you need thicker, beefier diagonals.
Here's why this matters practically:
If you're building something with heavy gravity loads (like a factory floor above a warehouse), the Pratt handles that beautifully. The tension diagonals can be lighter. You save money.
If you're dealing with heavy snow loads or you're in an area with weird wind patterns? The Howe actually handles compression better. But you pay for it in material.
I once watched two engineers argue about this for forty-five minutes. One wanted Pratt. One wanted Howe. Neither was wrong—they just had different assumptions about the load path. The building's still standing five years later. So maybe both work fine if you oversize things a little.
But am I wrong, or do most guides oversimplify this? Yeah, I think they do.
Load Calculations: What Actually Pushes on a Truss
You can't design a truss without knowing what's trying to break it. Here's the breakdown.
Dead Loads (The Heavy Stuff That Stays)
- Roof sheathing (plywood, metal deck, whatever)
- Insulation (adds up faster than you think)
- The truss itself (self-weight—usually 15-25% of your total)
- Ceiling finishes, lights, sprinklers, ducts
Real numbers I've used:
- Standing seam metal roof + insulation + purlins: about 8-12 PSF
- Asphalt shingles + plywood + ceiling: about 15-20 PSF
- Concrete tile (I hate specifying this): 25-35 PSF just for the tile
Live Loads (The Temporary Stuff)
- Snow (hugely variable—check your local code, don't guess)
- Maintenance workers (treat as 100kg concentrated load minimum)
- Construction loads (this kills more trusses than anything else—stacking materials on unbraced trusses is how people get hurt)
Wind Loads (The Sneaky One)
Wind creates uplift. Uplift means your truss wants to become an airplane. The connections have to resist that.
I was on a site in Oklahoma once—flat land, Exposure C, 120mph wind speed. The trusses were fine. The connections? Not so much. Someone used light-duty clips instead of the specified welded brackets. The engineer caught it during inspection. Saved the building.
| Wind Exposure | Description | Typical Uplift (kPa) |
|---|---|---|
| B | Urban/suburban, many obstructions | 0.5-0.8 |
| C | Open terrain, flat land | 0.8-1.2 |
| D | Unobstructed water/coastal | 1.2-1.8 |
Here's the thing about wind loads nobody tells you: the heel connection is usually the weakest point. The top chord wants to peel away from the wall. If your connection can't handle that uplift, the truss lifts, the roof lifts, and suddenly you have a very expensive tent.
Truss Depth and Member Selection (Where the Money Goes)
Okay, you've got your span and your loads. Now you need to pick actual steel.
Chord Selection (The Top and Bottom)
Your chords do most of the work. The top chord gets compressed. Bottom chord gets tension (usually—wind can reverse this).
What thicknesses actually mean (mil to gauge):
| Mils | Gauge (approx) | Use Case |
|---|---|---|
| 28 | 22 | Light residential, short spans |
| 33 | 20 | Medium spans, moderate loads |
| 43 | 18 | Most commercial roofs |
| 54 | 16 | Heavy loads, long spans |
| 68 | 14 | Industrial, high snow/wind |
| 97 | 12 | Bridges, massive spans |
My rule: Start with 43 mil for the top chord on anything over 10 meters. Downgrade if loads are light. Upgrade if you're in snow country or high wind.
Web Members (The Triangles Inside)
Webs can be tubes (closed rectangular), angles, or channels. Tubes are stiffer for the same weight. Angles are cheaper but need more careful bracing.
I almost didn't include this part because it gets technical, but here's the practical take: use tubular webs for compression members (they resist buckling better) and angles for tension members (cheaper, easier to source). Mixing them is fine—trusses do this all the time.
Member Sizing Table (Ballpark—Don't Build From This)
| Span | Truss Depth | Top Chord Size | Bottom Chord Size | Web Size |
|---|---|---|---|---|
| 10m | 0.8-1.0m | 33 mil TSC3 | 28 mil TSC3 | 33 mil tube |
| 15m | 1.2-1.5m | 43 mil TSC4 | 33 mil TSC4 | 43 mil tube |
| 20m | 1.5-2.0m | 54 mil TSC4 | 43 mil TSC4 | 54 mil tube |
| 30m | 2.0-2.5m | 68 mil TSC4 | 54 mil TSC4 | 63 mil tube |
| 40m+ | 3.0m+ | 97 mil TSC4 | 68 mil TSC4 | Custom engineering |
These are rough starting points. Actual design depends on loads, spacing, and bracing. Don't build from a coffee article.
Connections: Where Trusses Actually Fail
I've seen maybe fifty truss issues over the years. Forty-eight of them were connection problems. Not the steel itself. The spots where steel meets steel or steel meets wall.
Truss-to-Bearing Connections
Your truss sits on something—a steel beam, a concrete wall, a wood top plate. That connection has to handle:
- Downward gravity loads (easy)
- Uplift from wind (not easy)
- Lateral forces from seismic or wind (also not easy)
Common connection types:
| Bearing Material | Typical Connection | Uplift Capacity (kN) |
|---|---|---|
| Steel beam | Welded clip or bolted angle | 10-25, depending on the weld |
| Concrete | Embedded strap or epoxy anchor | 8-20 depending on concrete strength |
| Wood | Hurricane tie or screw-down clip | 5-15 depending on wood species |
| CFS wall | Screwed TSUC clip | 3-10 depending on gauge |
The thing I keep learning: always overspec the heel connection by at least 20%. The cost difference is tiny compared to the consequences of a roof peeling off.
Truss-to-Truss Connections (Girders and Jacks)
When trusses connect at angles (hip jacks, valley sets), you need a positive mechanical attachment. Don't just rest one truss on another and call it good.
I was on a job in North Carolina—hurricane country. The framers had set the hip jacks on top of the girder truss without clips. Just friction and gravity. When I pointed it out, the foreman said, "That's how we always do it." Cool. Doesn't mean it's right. We added TTC clips to every connection. Cost an extra afternoon. Probably saved the roof in the next storm.
Bracing and Buckling (The Part Everyone Forgets)
A truss is strong in its own plane. Perpendicular to that plane? Not so much.
Lateral Bracing
Your top chord needs bracing to keep it from buckling sideways. Two ways to do it:
- Structural sheathing (plywood, OSB, metal deck) attached directly to the chord
- Purlins (horizontal members spanning between trusses)
Bottom chord bracing is often overlooked. If your ceiling is drywall on a hat channel, that's bracing. If there's no ceiling? You need diagonal bracing in the plane of the bottom chord.
Compression Member Buckling
Here's the physics nobody explains well: a long skinny compression member will bend sideways before it crushes. That's buckling. The way to stop it is to shorten the unbraced length with intermediate bracing.
Quick check: If your compression web is longer than about 1.5 meters and it's made of light-gauge steel, you probably need mid-span bracing.
Bracing Requirements Table
| Member Type | Max Unbraced Length (m) | Bracing Method |
|---|---|---|
| Top chord, sheathed | Unlimited (sheathing provides bracing) | Structural panels |
| Top chord, unsheathed | 2.0-2.5 | Purlins at spacing |
| Bottom chord with ceiling | 2.5-3.0 | Ceiling attachment |
| Bottom chord open | 1.5-2.0 | Continuous lateral bracing |
| Compression web | 1.2-1.5 | Discrete brace or strongback |
Common Applications by Span (Just Tell Me What to Use)
Here's where I start when someone says, "I need a roof over this space."
| Span Range | My Go-To Truss | Why |
|---|---|---|
| 6-8m | King Post or Pratt | Simple, cheap, hard to mess up |
| 8-12m | Queen Post or Fink | Good for residential, handles pitch well |
| 12-18m | Pratt or Howe | Both work—choose based on load type |
| 18-25m | Warren or Double Fink | Warren is my default here |
| 25-35m | Warren or Flat Truss | Depends on roof pitch |
| 35-50m | K Truss or custom Warren | Needs engineering attention |
| 50-100m | Custom design only | You need a specialist, not a guide |
A confession: I used to spec Pratt trusses for everything because that's what I learned on. Then I had a project with heavy snow loads, and the Pratt was fine, but the Howe would've been better. Now I actually think about load direction before picking. Small change. Big difference.
Design Guide Checklist (So You Don't Miss Anything)
Here's what I run through before finalizing any truss design. Steal it.
Before you start:
- [ ] Span measured center-to-center of bearings
- [ ] Roof slope confirmed (pitch affects heel height)
- [ ] Truss spacing determined (1.2m, 1.5m, 2.0m, 2.4m are common)
- [ ] Bearing material and capacity known
Loads (get these from building designer—don't guess):
- [ ] Dead load (roofing, deck, insulation, ceiling, mechanical)
- [ ] Snow load (ground snow + drift + unbalanced)
- [ ] Wind load (speed, exposure, topography, building category)
- [ ] Seismic (if applicable—most of the country needs at least Category C consideration)
- [ ] Special loads (sprinklers, HVAC, cranes, hanging equipment)
During design:
- [ ] Span-to-depth ratio between span/10 and span/15
- [ ] Web configuration efficient for the load pattern
- [ ] Connections designed for uplift, not just gravity
- [ ] Permanent bracing locations marked on shop drawings
After design (before fabrication):
- [ ] Shop drawings reviewed by the engineer of record
- [ ] Reactions match supporting structure capacity
- [ ] Deflection checks done (L/240 total, L/360 live is typical for roofs)
Where I Messed Up (So You Don't Have To)
Mistake 1: I once specified a Warren truss for a building with heavy point loads from rooftop HVAC. The Warren is great for uniform loads. It's not great for point loads. The truss worked but required beefed-up chords at the load points. A Pratt would've handled it better.
Mistake 2: I forgot to account for sprinkler pipe loading on a 25-meter flat truss. The pipes themselves weren't heavy, but the water inside was. Added about 15% to the dead load. Had to re-run calcs. Embarrassing.
Mistake 3: I assumed the general contractor would install permanent bracing as shown on the drawings. He didn't. I didn't check. The trusses deflected more than expected under the snow load because the bottom chords weren't braced. Now I put bracing requirements in three places: drawings, specifications, and a separate bracing diagram that can't be missed.
Tools That Actually Help (Because Doing This By Hand Sucks)
This is where I'd normally grab a calculator and spend twenty minutes running numbers. Lately, I've been using SteelSolver.com's Truss Calculator because it handles unbalanced snow loads and wind uplift cases without me having to remember which ASCE7 chapter has the drift coefficients. Saves my brain for the actual decision-making.
They also have a Span-to-Depth Ratio tool that's stupid simple—plug in your span, and it spits out recommended depths and warns you if you're going outside the safe range. I used to do that on a napkin. This is better.
There's also a wind load calculator that factors in exposure, building height, and location. Saves you from flipping through ASCE7 tables for an hour.
And for connection design? A bolt group calculator lets you test different gusset plate configurations. I use it constantly.
Look, you can design trusses with a spreadsheet and a code book. I did for years. But the tools exist now. Use them.
The Book That Changed How I Think About Trusses
Design of Welded Structures by Omer Blodgett. It's old (1960s). It's not fancy. But Blodgett explains load paths and connection behavior better than anyone. The Lincoln Arc Welding Foundation still sells it. Worth every penny.
I keep a copy on my desk. The pages are coffee-stained, and the spine is cracked. That's how you know a book is good.
Putting It All Together: How to Actually Pick the Right Truss
You've made it this far. Here's the practical workflow I use now after learning all this stuff the hard way.
Step 1: Define your span clearly. Not "about 15 meters." The actual out-to-out bearing distance. Cantilevers? Overhangs? Note those too.
Step 2: Calculate your loads. Dead loads (the truss itself, decking, roofing, ceiling, and mechanical). Live loads (snow, maintenance, occupancy). Wind loads (uplift and lateral). Seismic if you're in a shaky area. ASCE 7 is your friend here, even if it's a dense read.
Step 3: Pick a starting truss type based on span.
| Span Range | Start Here |
|---|---|
| 6-10m | King Post, Queen Post, Pratt, Fink |
| 10-18m | Compound Fink, Howe, Pratt |
| 18-30m | Double Fink, Double Howe, Fan |
| 30-60m | Warren, K Truss, custom |
| 60-100m | Bowstring, custom Warren variants |
Step 4: Choose a depth. Start at span/12. Adjust up for heavier loads, down for height restrictions.
Step 5: Check your heel height. Make sure you have room for insulation and bearing connections.
Step 6: Have an engineer run the numbers. I'm serious about this. The steelVIEW software (or similar) does multi-node analysis that accounts for all the load combinations you haven't thought of. Unbalanced snow loads. Wind from different directions. Stress reversal. Let the software catch what your intuition misses.
Step 7: Plan your connections and bracing before you order. The best truss in the world fails if the heel connection isn't rated for uplift or the permanent bracing isn't specified.
This is where I'd normally open my well-worn copy of the AISI S240 standard, but honestly? The "Cold-Formed Steel Design" textbook by Wei-Wen Yu is sitting on my desk with about a million sticky notes in it. That book has saved me more times than I can count. Worth every penny.
Frequently Asked Questions (That People Actually Ask)
Q: Can I use a Fink truss for a 20-meter span?
A: Technically? Sometimes. Smartly? No. You're in Double Fink or Fan territory at 20 meters. A single Fink truss at that span will deflect noticeably and require deeper chords than a Double Fink would need.
Q: What's the difference between a Pratt and a Howe truss?
A: The diagonal members. Pratt diagonals slope inward toward the center and are in tension. Howe diagonals slope outward from the center and are in compression. For gravity loads, Pratt is more efficient. For heavy compression loads (snow, heavy roofing), Howe handles it better.
Q: How do I know if my truss needs lateral bracing?
A: The shop drawing will tell you. But generally, any compression member (top chords, some webs) longer than about 1.5 meters needs bracing unless it's a very light load. The slenderness ratio (KL/r) should stay under 200 for compression members.
Q: Can I modify a truss in the field?
A: No. Stop. Don't do it. Unless the engineer who designed it approves the modification in writing, cutting or drilling a truss chord or web voids any warranty and might cause failure. I've seen someone notch a top chord to run a pipe and watched the truss buckle under the next snow load. Not pretty.
Q: What's the most cost-effective truss for a 15-meter warehouse?
A: Usually a Pratt or Compound Fink, spaced at 4-6 feet, with angles for web members and channels for chords. But the local steel prices and fabrication costs matter. Get three quotes from different fabricators.
Q: How much camber should I add?
A: Enough to offset dead load deflection, but not so much that the roof looks humped. Usually L/240 to L/360 of the span for dead load only. The truss designer should calculate this based on your actual loads.
Q: Do I need a fire-rated assembly?
A: Check your building code based on occupancy type and building size. A single-family home? Probably not. A school or hospital? Almost certainly yes. The UL listings for TrusSteel assemblies cover 1-hour through 2-hour ratings.
Final Thoughts (Because I Keep Coming Back to This)
Look, picking a steel truss isn't rocket science, but it's also not something you should guess at. The difference between a truss that works for 20 years and one that starts sagging in 5 years is usually in the details you didn't think about—the span-to-depth ratio, the heel height, the connection design, the permanent bracing.
I keep a copy of the span tables from the TrusSteel Design Manual on my phone. Not because I use them every day, but because when someone asks, "Can we go 30 meters with a Warren truss?" I want to have an answer before I open my laptop.
One last thing: If you're designing something and you're not sure, call a truss fabricator. The good ones (look for TrusSteel Authorized Fabricators) have engineers on staff who do this all day, every day. They've seen what works and what doesn't. And honestly? Most of them are happy to talk through options during the design phase because it saves headaches during fabrication.
That's it. That's everything I've learned about steel truss spans. Go build something that doesn't sag.
