NCCER Boilermaker common mistakes
NCCER Boilermaker · Understand the traps and how to avoid them.
Entering a boiler drum, mud drum, or fired heater shell to do tube or baffle work without first testing the atmosphere and posting a confined-space entry permit, and working alone with no attendant at the opening.
Candidates picture 'confined space' as a tank or sewer and do not recognize a drum with two 18-inch manways as permit-required; the space looks like a normal work area once the cover is off, and the job is treated as routine maintenance.
What to do instead
Before any vessel entry: classify the space, test oxygen (19.5–23.5%), flammables and toxics with a calibrated meter, force-ventilate, post the permit, and station a trained attendant with a retrieval system at the opening for the entire entry.
Mounting a grinding or cutting wheel on an angle grinder without checking that the wheel's maximum RPM rating meets or exceeds the tool's no-load RPM, and running it with the guard removed or the wrong-size flanges.
Wheels are grabbed by arbor size alone and often look interchangeable; guards are removed because they block access in tight spots, and the RPM stamp on the blotter is small and easy to skip.
What to do instead
Match wheel RPM rating to the grinder's nameplate RPM every time, ring-test or inspect new wheels, keep the guard and correct flanges installed, and stand clear of the wheel plane for the first 30 seconds of run-up.
Shutting down an oxyfuel torch in the wrong order - turning off the oxygen first, or laying a lit torch down and letting the tip contact the plate - and operating without flashback arrestors at the regulators.
Under time pressure the torch is set down between cuts rather than extinguished, and the shutdown sequence is memorized as a single blur so the fuel gas is left flowing into a hot tip.
What to do instead
Light with a friction lighter, never a match; on shutdown close the fuel gas valve first, then oxygen, then the torch valves, then the cylinder valves, and bleed the lines. Keep arrestors at both regulators.
Running a long arc and travelling too fast on a vertical-up or overhead SMAW pass, leaving visible undercut at the toes and slag trapped between passes that is only found later on a bend or RT test.
The welder is watching the puddle crown rather than the arc length and the wet-in at the edges, and raises the arc to see through smoke or in poor light.
What to do instead
Hold arc length near one electrode diameter, pause at the toes to let the puddle wet in, keep the electrode angle within about 10° of perpendicular to the joint, and chip and brush each pass fully before the next. If light is the problem, fix the light - not the arc length.
Using nominal pipe size as if it were actual outside diameter when laying out a fit-up or computing a take-out, and mixing dimensions taken from different reference points (top of steel vs. centreline vs. face of flange) on the same spool.
Nominal sizes match actual OD only for 8-inch and larger; smaller sizes are counterintuitive, and drawings often give elevations to different datums that are not labelled prominently.
What to do instead
Work from an OD table for every size, mark a single datum on the drawing and carry all dimensions from it, and total the chain of dimensions for each spool rather than trusting individual segment checks.
Cutting layout lines exactly on the scribe with no kerf allowance and no weld-shrinkage allowance, so a shell course, transition piece, or nozzle cutout comes out undersized and out of square after welding.
The layout looks correct on the plate because the line was followed precisely, and shrinkage is invisible until the weld cools - candidates treat layout as pure geometry.
What to do instead
Offset the cut line by half the kerf for the process being used, add the specified shrinkage/contraction allowance to developed lengths, and verify square, diagonals and circumference on a trial fit before final welding.
Sizing slings and shackles by the load's weight alone while ignoring sling angle - for example assuming a 2-leg bridle at 30° from horizontal carries half the load per leg.
The weight is known and the angle is eyeballed, and the multiplier (tension rises sharply as the angle flattens) is not applied because the sling tag only lists a vertical capacity.
What to do instead
Compute leg tension from the vertical share divided by the sine of the angle from horizontal, apply the hitch factor (choker, basket, vertical), and de-rate accordingly; flatten nothing below the manufacturer's stated minimum angle without engineering approval.
Reading the crane or hoist load chart at the radius of the pick-up point instead of the radius at the point where the load will be after booming out, and omitting the weight of the rigging, spreader beam and tag lines from the load.
The chart is checked once at the start of the lift and the boom is then extended over the vessel or roof line while the operator watches the load, not the radius indicator.
What to do instead
Determine the maximum radius over the full travel path before the lift, add rigging and attachments to the load weight, confirm outrigger or mat configuration matches the chart being used, and re-check the indicator at the farthest point.
Pressure-testing a vessel or boiler with the wrong test pressure, with air left in the system, with a single gauge, or by bringing pressure up quickly with a shop compressor through an unregulated line.
The code multiple and temperature correction are half-remembered, and venting high points is skipped because the top taps are hard to reach on a tall drum.
What to do instead
Use 1.5× the maximum allowable working pressure (or the code-specified value), correct for metal and water temperature, vent all high points to remove air, use two calibrated gauges - one at the low point and one at the top - and pressurize in controlled increments with a relief valve set below the test pressure.
Over-rolling tube ends in a tubesheet - turning the roller past the point of tube-sheet contact to 'make sure it seals' - which thins the tube wall, produces a visible rolled lip, and causes leaks on the hydro test.
The expansion is judged by feel and by how far the roller is pushed in rather than by a controlled torque or measured wall reduction, and a weeping tube tempts the operator to roll harder.
What to do instead
Expand to the tube and tube-sheet manufacturer's specified wall reduction (typically in the 5–8% range), use a torque-controlled or pre-set roller, roll a sequence pattern across the sheet rather than randomly, and verify with a tube gauge before hydro.