If you're ordering a steel shed or a temporary fabrication building, the weld spec is where the deal either holds or fails. I've been a quality compliance manager in structural steel for over a decade—I review every fabrication before it hits the site. Roughly 200-some projects annually. In 2023, I rejected 12% of first deliveries for weld quality issues alone. That cost us, give or take, $18,000 in rework on just one job.

Here's the thing: most of those rejections were preventable. Not by better welding—by better specification on the front end. This checklist is for anyone who buys industrial structural fabrication. It's five checkpoints to cover before the steel gets cut. Not comprehensive, but these are the ones I see missed most often.

5 minutes of specification verification beats 5 days of rework.

Checkpoint 1: Specify the Welding Process (and Stick to It)

The first question I ask: what welding process are we using? The answer should be specific. Not just 'welding.' That's like saying 'woodworking'—it tells me nothing about quality.

For structural steel in a typical metal office warehouse building or steel shed, the options are usually:

  • FCAW (Flux-Cored Arc Welding) – common for field work, high deposition rate. But it's sensitive to wind and operator technique.
  • SMAW (Shielded Metal Arc Welding) – old-school stick welding. Reliable but slower.
  • GMAW (Gas Metal Arc Welding) – fast, but needs clean conditions.

In our Q1 2024 audit, we had a batch of fifty I-section steel columns where the vendor had switched from FCAW to GMAW mid-run—no notification. The GMAW joints failed tensile tests at a 22% rate. They redid all fifty at their cost, but we lost six days on the schedule.

I'm not a structural engineer, so I can't give you the perfect process for every alloy. But I can tell you from a quality perspective: get it in writing. Specify the process by name and subcategory.

Checkpoint 2: Define the Effective Throat Thickness (Not Just Fillet Size)

Most buyers focus on the visible leg size of a fillet weld. But the carrying capacity comes from the throat—the shortest distance from the root to the face. This is one of those outsider blindspots I see constantly.

The question everyone asks is 'how big is the weld?' The question they should ask is 'what's the effective throat?'

For a fillet weld with equal legs, the throat is leg size × 0.707. So a 1/4-inch fillet gives you about 0.177-inch throat. That's straightforward in theory. In practice, I've measured welds that looked 1/4 inch on the outside but had only 0.12 inch of actual throat because the weld was convex—more visual metal, less functional metal.

Take this with a grain of salt: I'd argue most specification disputes I've mediated come down not to whether the weld looks good, but whether the throat meets the design requirement. Specify a minimum throat measurement. Not a leg size. A throat.

Checkpoint 3: Require a WPS (Welding Procedure Specification) Before Any Steel Is Cut

This is the one I'm most stubborn about. A WPS is a document that says exactly how a weld will be made: voltage, amperage, travel speed, preheat, interpass temperature. It's not optional—it's table stakes for quality work.

Under industry practice, a qualified WPS should reference standards like AWS D1.1. I know that's the governing standard for structural steel, though I'm not a code expert—the legal compliance side gets beyond my pay grade. But I do know that a project without a WPS is a project waiting to fail.

What a solid WPS includes for a typical steel shed:

  • Process type and filler metal classification
  • Joint design (single V, double V, fillet, etc.)
  • Preheat and interpass temperature range
  • Electrical characteristics (voltage, amperage range, polarity)
  • Travel speed range

I still kick myself for not demanding a WPS on a project back in 2019. It was a temporary steel building for an equipment storage yard. The welds looked fine—until we had a 40-knot wind during a storm, and one column-to-beam connection failed. The investigation found the vendor had used undersized filler metal. The WPS would have caught it before the column ever hit the shop floor.

Checkpoint 4: Inspect for Discontinuities Before the Paint Goes On

Once the paint is on, you can't see the weld quality. And I've seen more than a few fabrications where paint was applied thick enough to hide porosity, undercut, and even incomplete fusion.

Honestly, I'm not sure why some fabricators rush to paint. My best guess is that it makes the steel look finished and hides imperfections temporarily. But paint inspection should happen after weld inspection, not before.

Part of me wants to blame the schedule pressure. Another part knows that a fabricated steel shed project with a 6-week lead time often has 4 weeks of buffer and gets squeezed into 2 weeks of work. I've been there. But in my experience, visual inspection of the raw weld—before any primer—catches 80% of critical defects.

What to look for:

  • Porosity: Small holes in the weld face. More than 3 per inch in any 6-inch segment is a reject per AWS D1.1.
  • Undercut: A groove at the weld toe. More than 1/32 inch deep on a critical joint is a problem.
  • Overlap: Weld metal that rolls over the base metal without fusing. This is often hidden by paint.

I ran a blind test with our site inspection team: same fabricated I-section steel with a painted weld versus an unpainted weld. The paint hid visible porosity in 3 out of 8 samples. That's a 37% miss rate on visual inspection—just from a coat of primer.

Checkpoint 5: Confirm the Test Coupon Program (Even If It Costs Extra)

I have mixed feelings about destructive testing on production welds. On one hand, it's expensive—you're destroying perfectly good welded plates. On the other hand, it's the only way to confirm the weld meets tensile and bend strength requirements.

For a metal office warehouse building or a steel shed fabrication, I'd argue a test coupon program is justified for any job over about 200 linear feet of weld. That's around $500-$800 extra in most shops. But it's an insurance policy.

Not ideal if you're on a tight budget. But consider this: in 2022, we had a job where coupons showed the weld had 60% of expected tensile strength. The root cause was an incorrect shielding gas mixture—something no visual inspection would catch. The vendor swapped the gas and re-qualified. Total cost: about $600 for the coupon test. The rework if we'd waited for failure? Estimated $4,200 plus a three-week delay.

Final Note: Common Mistakes That Happen Anyway

Even with a perfect checklist, problems happen. Here are the three I see most often:

  1. Assuming the shop floor follows the WPS. They don't always. Ask for process control records on the first run. Verify before production scales.
  2. Skipping the pre-production meeting. A 30-minute call with the fabrication manager and the welder lead can catch 90% of specification mismatches. I've never had a call like that go wrong. I've had plenty of projects without one go sideways.
  3. Trusting paint to 'seal' a marginal weld. It doesn't. Paint adds corrosion resistance. It does not add structural strength. A defective weld under paint is still a defective weld—it's just harder to find.

That's the list. Not complex. But skipping any one of these checkpoints has cost me and my clients real money, real schedule time, and real frustration. If you're specifying an industrial structural fabrication job, run these five checks before the steel gets cut. It's the cheapest insurance you'll buy.