Inspection Tools for Welding Inspectors: A Practical Guide

A fabricator’s QA lead handed the inspector a fillet weld that looked textbook — smooth bead, even reinforcement, no undercut a naked eye could find. Two weeks later the client’s radiograph came back rejected, and the cause was embarrassingly simple: nobody had ever put a gauge on the weld. A welding inspector’s toolbox is not a luxury; it is the difference between an opinion and an acceptance criterion, and the right tool used correctly is what turns “it looks fine” into a defensible, code-compliant pass or fail.

Why the Right Inspection Tools Change the Outcome of a Weld

The difference between a weld that passes on paper and one that passes in service often comes down to one thing: measurement. A bead can look textbook under shop lighting, but undersized legs, undercut, or root misalignment can go unnoticed until a radiograph or a failed inspection catches them. The inspector isn’t there to out-weld the welder. They’re there to measure what the welder cannot see and compare those measurements against a written acceptance criterion.

What “Inspection Tools for a Welding Inspector” Means in Practice

For a welding inspector, “inspection tools” aren’t just a shopping list. They’re any device that converts a subjective judgment into a number, shape, or image that can be verified against a requirement. That requirement comes from the drawing, the welding procedure specification (WPS), and the governing code, whether that’s AWS D1.1 for structural steel, ASME B31.3 for process piping, ASME Section VIII for pressure vessels, or ISO 17637 and ISO 5817 for international work.

The tool doesn’t make the accept or reject call. The code does. All it does is show exactly where the weld sits relative to that limit. A recorded reading, such as a 7 mm leg length against an 8 mm requirement, can be audited or defended. A statement of “it looks fine” cannot.

Three Families of Tools, One Decision

No single tool can cover an entire weld. The most effective inspections rely on three families working together to answer different questions:

• Visual and dimensional hand gauges. Fillet weld gauge sets, bridge cam (Cambridge) gauges, hi-lo gauges, undercut gauges, taper and feeler gauges, protractors, dial calipers, and steel rules answer how big, how deep, or how far out of position a feature is.
• Optical and lighting aids. Flashlights and headlamps at a raking angle, inspection mirrors for root-side welds, borescopes, and 10x loupes answer is it there, and is it surface-breaking?
• NDT and destructive testing equipment. Ultrasonic flaw detectors and thickness gauges, radiographic equipment, magnetic particle yokes and consumables, penetrant kits, eddy current probes, bend coupons, macro sections, and hardness testers answer what is buried inside, and how deep. Even if a subcontractor performs these tests, the inspector must be able to interpret their results.

A gauge may confirm a fillet meets size requirements, raking light may find a toe crack, and ultrasonic testing may find lack of fusion none of the others could reach. Together, they form one defensible decision.

The Stakes: What Certification Actually Asks of You

This isn’t theoretical. The AWS Certified Welding Inspector program requires demonstrated, hands-on competence. Per the CWI examination guide, Part A is a 150-question fundamentals exam requiring a minimum score of 72% on each part, and Part B is a hands-on practical where candidates must use a kit that includes fillet weld gauges, a V-WAC gauge, a dial caliper, a ruler, a C4.1 surface roughness gauge, a protractor, a hi-lo gauge, a magnifying glass, a mirror, and a flashlight. These are the same items referenced within the Part B basic requirements.

Furthermore, AWS B5.1:2025 defines the duties for Associate Welding Inspector (AWI), Welding Inspector (WI), and Senior Welding Inspector (SWI). For all three levels, it requires inspectors to read Jaeger Number 2 letters or smaller at a minimum distance of 12 inches in at least one eye, with or without correction, as well as pass testing for color perception and contrast differentiation of shades of gray. According to B5.1:2025 Table 8.1, Visual Examination is weighted at a minimum of 10% and NDE Fundamentals at a minimum of 8% of the fundamentals exam content.

Who This Guide Is For, and What You’ll Gain

Whether you’re a fabrication shop QC technician, a welder working toward inspection, a certification student, or a career-changer about to sign off on someone else’s work, the instruments will look familiar across pressure vessel, pipeline, and structural shops. Acceptance criteria may differ, but the habits of measuring against them do not. If your work branches into more specialized fields, our guides to aerospace welding standards and careers can be useful references alongside this one.

The goal of this guide is simple: know which tool answers which question, use it without fooling yourself, and build a kit that fits both your budget and the code you work to. Often, a $40 fillet gauge used correctly is the cheapest insurance for a weld worth exponentially more.

Inspection vs. Examination: How the Toolkit Maps to Acceptance Criteria

Inspection is a quality function performed by or for the owner, by the owner’s Inspector or an agency on the owner’s behalf. Examination is a quality control function performed by the manufacturer, fabricator or erector while producing the work. ASME B31.3 draws that line in paragraph 340.1, and it carries authority: a shop examination record is the fabricator’s evidence of conformance, an inspection report is the owner’s basis for accepting the work. The ASME BPVC certification and accreditation guide shows how this intersects with jurisdictional and National Board oversight.

Independence and personnel qualification

Independence is written into the code, not left to judgement. ASME B31.3 paragraph 342.2 requires in-process examination to be performed by personnel other than those doing the production work: you cannot examine your own weld while you are welding it. So an inspector walks a pipe spool before the root pass, checking fit-up, root gap and alignment, and that record survives to handover. The paper trail is worth far less when the same person runs the arc and signs the examination.

Qualification is written in as well: paragraph 342.1 requires NDE personnel to be qualified and certified per a procedure described in ASME BPVC Section V, Article 1, T-120(e) or (f). ASME Section V never applies on its own. It is a referenced code, invoked by B31.3, Section VIII, Section I and others, supplying what those codes lean on: written NDE procedures (Article 1, T-150), calibration records available to the Inspector, and personnel certification rules (T-190, sitting on T-120). Certification under Section V is employer-based: the Certificate Holder writes the practice, leaning on a recognised baseline such as SNT-TC-1A, ANSI/ASNT CP-189 or ACCP, in the edition the referencing code accepts. So the useful answer names the employer practice, the standard behind it and the certificate numbers.

The same split shows up outside

Inspection vs. Examination: How the Toolkit Maps to Acceptance Criteria

Inspection is a quality function performed by or for the owner, by the owner’s Inspector or an agency on the owner’s behalf. Examination is a quality control function performed by the manufacturer, fabricator or erector while producing the work. ASME B31.3 draws that line in paragraph 340.1, and it carries authority: a shop examination record is the fabricator’s evidence of conformance, an inspection report is the owner’s basis for accepting the work. The ASME BPVC certification and accreditation guide shows how this intersects with jurisdictional and National Board oversight.

Independence and personnel qualification

Independence is written into the code, not left to judgement. ASME B31.3 paragraph 342.2 requires in-process examination to be performed by personnel other than those doing the production work: you cannot examine your own weld while you are welding it. So an inspector walks a pipe spool before the root pass, checking fit-up, root gap and alignment, and that record survives to handover. The paper trail is worth far less when the same person runs the arc and signs the examination.

Qualification is written in as well: paragraph 342.1 requires NDE personnel to be qualified and certified per a procedure described in ASME BPVC Section V, Article 1, T-120(e) or (f). ASME Section V never applies on its own. It is a referenced code, invoked by B31.3, Section VIII, Section I and others, supplying what those codes lean on: written NDE procedures (Article 1, T-150), calibration records available to the Inspector, and personnel certification rules (T-190, sitting on T-120). Certification under Section V is employer-based: the Certificate Holder writes the practice, leaning on a recognised baseline such as SNT-TC-1A, ANSI/ASNT CP-189 or ACCP, in the edition the referencing code accepts. So the useful answer names the employer practice, the standard behind it and the certificate numbers.

The same split shows up outside ASME. The CSWIP welding inspector scheme divides duties the same way, and in aerospace documented qualification is the baseline for employment rather than an extra, as our guide to aerospace welding schools shows.

Tools follow criteria, never the other way around

You cannot decide what to measure until the governing document defines the acceptance criterion. The gauge does not set the limit; the code, drawing or contract does, and only then does the gauge tell you whether you are inside it. Choose the tool first and you end up defending a measurement nobody asked for.

Governing system	What it governs	Where the acceptance numbers live
AWS D1.1	Structural steel: buildings, bridges, general fabrication	Undercut limits, fillet leg and convexity caps, groove reinforcement maximums
ASME B31.3	Chemical and petrochemical process piping	Root reinforcement, internal misalignment (high-low), radiography acceptance
ASME Section VIII	Pressure vessels and their welds	Radiographic acceptance, heat treatment, coupon and retest rules
ISO system	International work: EN 1090, offshore, shipbuilding	ISO 17637 visual testing, ISO 5817 levels B, C and D, ISO 17635 acceptance levels

Same job, different numbers. A bridge cam gauge does not care which document you are working to, but a reading only means something once you can name the clause it came from.

The ISO system differs in form but not logic. ISO 17637 sets the rules for visual testing, while ISO 5817 defines three quality levels for imperfections, B, C and D, with B the most stringent. A contract, drawing or enquiry document should state which level applies, and ISO 5817 says it should be prescribed before production begins, ideally at the enquiry or order stage. ISO 17635 then ties those quality levels to acceptance levels for the NDT methods, so “quality level B” becomes a radiography or ultrasonic acceptance limit.

Two ISO 5817 limits are worth knowing early: with macro-examination, only imperfections detectable at a maximum of tenfold magnification may be considered, and systematic imperfections are permitted only in quality level D.

That detail matters because the numbering tells you where a flaw sits, and location decides which tool can see it. ISO 6520-1 groups imperfections into families: lack of fusion types in the 1.x series, cracks in the 2.x, lack of penetration in the 3.x, porosity and holes in the 4.x, inclusions in the 5.x. A 1.x flaw buried inside the joint is invisible to a fillet gauge, a loupe or a flashlight, passing visual inspection without a murmur, and only volumetric or planar NDT finds it: radiography, or ultrasonic testing where the orientation suits the flaw.

The Core Dimensional Gauges: Fillet Weld Gauge, Bridge Cam Gauge, and Hi-Lo Gauge

Three gauges carry most of a welding inspector’s workload. The fillet weld gauge answers “is this fillet the size it should be?” The bridge cam gauge answers “is this weld the right shape, and how deep is the undercut?” The hi-lo gauge answers “is this pipe joint aligned well enough to weld?”

Certification treats them seriously. Part B of the AWS Certified Welding Inspector exam is performance based, and candidates must know how to use and apply each measuring and inspection instrument in the examination kit, including how to properly zero the instrument where relevant. US and Canada candidates use the Customary Unit kit; internationally it is the Metric Unit set.

The Fillet Weld Gauge: Size, Throat, and Convexity

A fillet weld gauge is a flat plate with etched or stamped right-angle profiles for standard leg sizes, plus features for throat thickness and convexity or concavity. The classic design cuts a different radius slot for each leg length: slide the weld into the slot that fits and check whether the plate edge clears the weld face.

New inspectors get caught by a fillet that looks generous on one side and thin on the other. The spec calls for a 10 mm fillet. The gauge seats easily on the long leg, but the short leg has only about 8 mm of metal. From a few feet away it looks fine. It is undersize, and the gauge catches it in seconds.

Three techniques make the reading trustworthy:

• Seat it flush on both legs. A gauge that rocks, touches one corner, or bridges spatter is telling you about the spatter.
• Read the concave side for leg size. On a convex fillet, seat the gauge on the convex face and slide it across until the profile matches the leg dimension; check convexity from the opposite side.
• Check both legs every time. A 10 mm leg with a 6 mm leg fails the same way a 6 mm fillet fails.

AWS D1.1 permits a reduction in fillet weld size of up to 2 mm for a specified size L of 5 mm or less, 2.5 mm for L = 6 mm, and 3 mm for L of 8 mm or greater. Note the direction: it is a tolerance on the specified size, and the code restricts where reduced fillets may be used, so read the full clause instead of treating it as blanket permission.

The Bridge Cam Gauge: The Most Versatile Tool in the Box

The bridge cam gauge (also sold as a Cambridge gauge or weld gauge) is a sliding caliper-style instrument with a fixed jaw and a moving jaw driven by a cam lever. It measures undercut depth, groove weld reinforcement height, root opening, misalignment, bevel angle, and fillet leg size. Own one quality hand tool and this is usually it.

Picture a groove weld with a smooth face on a drawing that calls for a flush weld. The welder ran a slightly proud cap and filleted the edges, and visual inspection cannot quantify how proud. The bridge cam slides its blade across the reinforcement and the scale reads 3.2 mm. AWS D1.1 says groove weld reinforcement shall not exceed 1/8 in. [3 mm] unless a flush weld is specified, so that reading is a reject with a number attached.

Here is the part that trips people up. The bridge cam has a scale, but the reading is only right if the instrument is zeroed. Close the jaws fully, confirm the scale reads zero, and adjust if it does not. An un-zeroed cam gives a wrong answer that looks entirely plausible, and plausible is not good enough when you are the last check before the weld ships. Check the fillet gauge set the same way, against a known straight edge or square.

The Hi-Lo Gauge: Internal Misalignment in Pipe

The hi-lo gauge is the pipe inspector’s tool. It hooks over the edge of one pipe wall and reaches across the joint to compare the internal surfaces of two pipe ends, measuring the offset between them. It can also read internal root reinforcement, the extra metal left inside the bore by the root pass.

Consider a 12 in. schedule 40 stainless pipe with no visible problem. The hi-lo gauge shows 2 mm of high-low at the joint. That offset means the root pass cannot fuse consistently across the joint, so you get a high-low defect right where the misalignment is, and potentially inadequate penetration that no amount of grinding will fix. Catching 2 mm before the root pass takes minutes; catching it after the pipe is in service is a much larger conversation. Used as a fit-up check before welding, the hi-lo gauge is one of the cheapest ways to prevent a nonconformance.

The Numbers These Gauges Verify

Under AWS D1.1, the limits these gauges help you confirm are:

Acceptance criterion	AWS D1.1 limit
Undercut, material less than 25 mm thick	0.8 mm (1/32 in.) maximum
Undercut, material 25 mm thick or greater	2 mm (3/32 in.) maximum
Undercut, primary members, weld transverse to tensile stress	0.25 mm (1/10 in.) maximum
Fillet weld convexity, weld 5/16 in. [8 mm] and under	1/16 in. [1.6 mm] maximum
Fillet weld convexity, larger welds	1/8 in. [3 mm] maximum
Groove weld reinforcement	1/8 in. [3 mm] maximum unless flush weld specified
Fillet weld size undersize allowance	L ≤ 5 mm: 2 mm; L = 6 mm: 2.5 mm; L ≥ 8 mm: 3 mm

Undercut is measured from the weld toe into the base metal, and the depth gauge gives a straight reading instead of a guess along a curved surface. The convexity limits are where the fillet gauge earns its place; reinforcement above 1/8 in. [3 mm] on a non-flush weld is where the bridge cam earns its.

One porosity rule is worth memorizing. For all groove welds other than CJP butt joints transverse to computed tensile stress, and for fillet welds, AWS D1.1 limits the sum of visible porosity 0.8 mm in diameter or greater to 10 mm in any linear 25 mm of weld. You are not counting holes. You measure the longest dimension of each visible pore, add them up, and compare the total with the length of the inspection zone.

Two habits tie these gauges together. Zero the instrument before you measure, every time. And when a reading looks wrong, check your seating before you reject the weld. A proud cap that reads 3.2 mm on a bridge cam may be a cam that was not zeroed, not a nonconforming weld.

Undercut Depth, Root Gap and Bevel Angle: The Secondary Gauge Set

The primary gauges say whether a finished weld is acceptable. This set covers the questions that come first: undercut depth, root gap, bevel angle and alignment, most of which are set before the arc is struck.

Start with joint preparation. ISO 17637 clause 6.2 requires checking that the shape and dimensions of the weld preparation meet the welding procedure specification, that fusion faces are clean, and that parts are correctly fixed relative to each other. Checked means measured, not assumed.

The Undercut Gauge

An undercut gauge is a small rigid scale measuring the depth of the groove eaten into the base metal at the weld toe. Typical resolution is 0.01 in. (0.25 mm). AWS D1.1 limits undercut to 0.8 mm (roughly 1/32 in.) for material under 25 mm (1 in.) thick, 2 mm for material 25 mm and over, and 0.25 mm (0.010 in.) maximum in primary members where the weld is transverse to the tensile stress.

Measure from the weld toe into the base metal, not along the curved profile of the weld. Slide the gauge in flat against the base metal until the scale edge touches the toe, then read where the marked edge meets the groove surface. Seating the gauge against the weld reinforcement makes you read the profile angle and under-report the depth. Check both sides; undercut on one side only still counts.

Root Opening with Taper and Feeler Gauges

Root opening is the gap between the two edges being joined, read with a taper gauge that slides into the gap and shows width on a graduated scale, or with feeler gauges used blade by blade. The taper gauge is faster; feeler gauges reveal a non-uniform gap.

The WPS specifies a root opening or root face dimension, and the joint must fall inside that range before welding starts. Too wide and you get excessive root reinforcement, high-low misalignment and lack of fusion. Too narrow and the root pass cannot penetrate, giving incomplete penetration or a cold lap under the cap.

Bevel Angle and Surface Roughness

A protractor measures the bevel angle of a prepared edge; a bevel gauge or weld gauge does it faster once you know its graduations. Angle controls pass count, filler metal and arc penetration into the sidewall. A shallow bevel on a thick joint is a common cause of lack of fusion.

Angle is only half of it. AWS B1.11M:2015 recommends the AWS C4.1 surface roughness guide for evaluating weld edge preparation surfaces. AWS D1.1 limits manual and semiautomatic processes to no rougher than Sample 3 with no gouges deeper than 1.5 mm, mechanized and automatic processes other than SAW to no rougher than Sample 4 with no gouges, and SAW edges to no rougher than Sample 3 with no gouges.

C4.1 is a pocket-sized card of etched sample surfaces held against the edge for visual comparison, not a measuring tool. A gouged, stepped or burned edge is a rejected edge preparation.

Dial Calipers, Steel Rules and Squares

Dial calipers handle plate thickness, root face dimension, fillet leg checks on larger welds and reinforcement spot checks. Close the jaws lightly so the dial shows a stable value, not a force-deflected one, keep the jaws perpendicular to the surface, and read straight down the scale to avoid parallax. Rock it until the reading stops changing. Steel rules and machinist squares cover length, flatness and squareness; out-of-flat plate makes every later dimension suspect.

Misalignment: Three Different Problems

• Angular misalignment is measured with a protractor or weld gauge comparing the two plates on either side of the joint.
• Offset misalignment is checked with a straightedge laid across the joint: if the two plate faces are not in line you will see light under the straightedge or a measurable step.
• High-low is internal. Only the hi-lo gauge, hooked over the pipe wall, can measure the step between the inside of one pipe and the inside of the other.

AWS D1.1 limits angular and offset misalignment in butt joints to 1/8 in. (3 mm) for plates over 3/8 in. (10 mm) thick, with proportionally more allowance for thinner material.

None of these tools are expensive, but a weld that fails a radiograph because an undercut was measured with the wrong zero is. The Air Gas welding supply guide reminds us that consumables are bought constantly while gauges are bought once, and gauges are what turn inspection into a record an auditor accepts.

AWS B1.11M notes that gauges are supplemental to the main method of visual examination: the gauge does not replace the trained eye, it replaces the argument. The AWS CWI Part B exam kit lists a C4.1 gauge and protractor alongside fillet gauges and a hi-lo gauge, because joint preparation is inside the inspector’s scope.

Lighting and Optical Aids: The Most Underrated Tools in the Box

For many inspectors, gauges take priority while lighting is treated as an afterthought. The reality is the opposite. A gauge can only measure what can already be seen, and without the right light, even a surface-breaking defect can go completely undetected.

Raking Light: Exposing What Direct Light Hides

The single most effective visual technique is raking light. This means holding your light source at a shallow angle, roughly 10 to 15 degrees to the weld surface. At this angle, any surface-breaking discontinuity interrupts the beam and casts a shadow, pulling defects like tight toe cracks into full view that would be lost under direct lighting.

Movement is key. After cleaning the weld, sweep the light across it from end to end and watch for shadows that shift with the beam. A shadow that moves with the light is relief from a physical discontinuity. A shadow that stays in place is far more likely to be dirt, spatter, or a grinding mark.

What ISO 17637 Requires

Proper lighting isn’t up to personal preference. ISO 17637:2016 requires white light illuminance of at least 350 lx at the weld surface, and recommends 500 lx. It specifically states that tinted goggles increase this minimum, which is why dark, auto-darkening welding filters are not suitable for close visual inspection.

The standard also sets viewing geometry. The examiner’s eye must be within 600 mm of the surface and at an angle of no less than 30 degrees. An additional light source may be used to increase contrast and relief. If a visual inspection is inconclusive, it must be supplemented by other NDT methods for surface inspection. Where direct line of sight is impossible, remote inspection with mirrors, borescopes, fibre optic cables, or cameras shall be considered.

Mirrors, Borescopes and Loupes

Access is just as much of a limitation as light is. A simple dental-style inspection mirror can allow a view of root sides, box sections, behind flanges, or pipe spools with no room to position your head. When using one, apply the same raking light technique to the reflected image, keeping in mind that left and right are reversed.

For areas with no line of sight at all, a rigid or USB borescope is the answer. Its greatest advantage is being able to place a light source directly beside the surface and rake it remotely. However, resolution is limited, so while a borescope can find cracks and obvious defects, dimensional measurements will still require a gauge.

For fine detail, a 10x loupe is invaluable for finding porosity, toe cracks, and crater cracks. Always rake light in from the side when using one, as reflecting your light source directly into the lens will wash out the very shadows you’re looking for.

Flashlights, Headlamps and Real-World Obstacles

The tool itself matters less than how it’s used, however there are practical considerations for your kit. A handheld flashlight allows for precise control of a 10-degree angle, whilst a headlamp frees both hands for tasks such as holding a mirror or gauge. For most inspectors, the most practical pairing is both.

There are also environmental hurdles to watch out for. On reflective materials, such as those covered in aluminum MIG welding guides, contrast can be worse despite bright light. Direct sunlight, LED colour shifts, tinted safety glasses, dirty lenses, and night shift fatigue can all degrade visibility. Many of these can be countered by changing angles, shielding from glare, or stepping away when contrast sensitivity begins to drop.

Eyesight and Qualified Inspection

Just as lighting must meet requirements, so must the inspector. AWS B5.1:2025 requires testing of colour perception, contrast differentiation, and shades of grey for this exact reason. These requirements are put into practice for certification, as candidates taking the CWI Part B exam are supplied with a magnifying glass, mirror, and flashlight as part of their standard kit. This is outlined within the AWS CWI Part B body of knowledge and exam materials and further detailed in the CWI Examination User Guide.

NDT Equipment the Welding Inspector Has to Understand (Even When They Do Not Run It)

Most welding inspectors never operate a flaw detector or expose a radiograph. A certified NDE technician does that work, under a fixed procedure with a calibrated instrument. The inspector’s job is different, not lesser: read the result, judge whether the technique suited the material, geometry and flaw, then decide what the indication means under the acceptance criteria.

Terminology varies by shop. AWS B1.11M notes that AWS A3.0M/A3.0 uses “nondestructive examination (NDE)” as the standard term, though NDI and NDT remain everyday language. Here NDE is the formal term, NDT the common one.

Who Runs the Test and What the Inspector Must Do

A shop welding pressure-critical piping for a client whose spec calls for radiography on 10 percent of butt welds shows the split clearly. The inspector is not the radiographer. When the film returns with what looks like a porosity cluster, the inspector must know whether it is real, whether it sits inside the wall, and whether the answer changes the joint’s disposition.

Comparing the Five Methods

Method	Materials	Access	Best for	Record	Depth?
PT (liquid penetrant)	Any nonporous metal	Clean, accessible surface	Surface-breaking cracks, some porosity	Report plus photos	No
MT (magnetic particle)	Ferromagnetic only, not austenitic or aluminium	Accessible surface	Surface and near-surface cracks	Report plus photos	Indicative only
UT (ultrasonic)	Most metals	One side, contact probe	Planar flaws: cracks, lack of fusion	Instrument data plus report	Yes
RT (radiographic)	Most metals, best in thick sections	Both sides of joint	Volumetric flaws: porosity, slag	Film or digital image	No
ET (eddy current)	Conductive surfaces	Probe access or immersion	Surface and shallow subsurface cracks	Data trace and report	Sometimes

Three Questions That Pick the Method

What is the part made of? Magnetic particle testing needs a ferromagnetic material to magnetise, so it is out on austenitic stainless and aluminium. Our guide to aluminum flux core welding covers the process, but the inspection lesson is the same: aluminium needs different techniques, including penetrant testing and eddy current. Liquid penetrant is the flexible fallback, working on any nonporous material such as carbon steel, stainless, aluminium or nickel alloys.

Where is the flaw? Surface or buried. PT and MT are surface and near-surface methods; UT and RT reach inside. A subsurface crack, or one hidden under deposited weld metal, is not found by a visual or penetrant check.

What shape is the flaw? This is what separates UT from RT. Planar flaws with a reflective surface, such as cracks and lack of fusion, are strongly oriented to the ultrasonic beam, so UT is sensitive and also gives depth of indication. Volumetric flaws such as porosity and slag scatter the beam in every direction and are best seen by radiation, because a radiograph images the total metal the beam passed through.

Recognising the Equipment

The inspector should be able to tell a flaw detector from a thickness gauge. A thickness gauge times the echo off the back wall and is easy to misread: too close to the weld edge and the echo returns from the plate edge, or over a reinforcement it reads high. A flaw detector shows the location and length of a reflector along the sound path, which is why it is the instrument used for crack sizing.

For radiography, the imaging rule matters. Less metal in the beam path means more exposure, so porosity and slag image dark. A tungsten inclusion is denser than the surrounding steel, blocks radiation and images light. That single rule explains most of what an inspector sees on a film, and why an apparently light spot is usually the serious one.

Magnetic particle equipment is simple, but the consumables are not optional. The yoke needs good contact and clean, uncoated metal. Dry powder is suspended in a carrier over a magnetized part. Wet fluorescent penetrant is brighter and higher contrast, and needs UV-A illumination and darkened conditions. Consumable type and magnetization method both belong in the written procedure.

Penetrant testing is a four-part kit: cleaner, penetrant, developer and the report. The critical failure mode is a coating. Paint, primer or a heavy oxide film seals the opening of a crack, so the penetrant never enters and the crack never shows up. Surface preparation with an appropriate metal cleaner for welding prep is not optional when penetrant testing is specified. Eddy current testing needs conductive surfaces and a probe; it finds surface and shallow subsurface flaws, but meaningful depth requires a reference standard and a skilled operator.

Inspecting the Inspection: Procedure, Calibration, Certification

Under ASME Section V, each NDE method and technique needs a written procedure, and the demonstration that the procedure is capable must be witnessed by the Authorized Inspector under Article 1, T-150. The National Board and ASME certification guide also notes that examination equipment must be calibrated, with the calibration recorded and made available to the Inspector. If the inspector cannot produce that record for the instrument used on his job, the result is not usable.

Personnel qualification differs by method. Radiographic and ultrasonic examiners must be qualified and certified to the employer’s written practice or through a central certification program. Magnetic particle, liquid penetrant and visual examiners may be certified by the Certificate Holder based on demonstrated proficiency plus an annual eye examination. Certification records must contain the items required by Article 1, T-190.

AWS offers separate CWI endorsements for dry powder yoke magnetic particle testing and for Type II Method C penetrant testing, both listed on the AWS Certified Welding Inspector program page. And AWS QC1 makes the responsibility split explicit: a CAWI may perform inspections only under the active supervision of an SCWI or CWI, who maintains responsibility for determining final acceptance.

Destructive and Supplementary Testing Tools

Dimensional gauges can confirm leg size or throat, but they cannot prove a weld is strong. A fillet or groove weld can measure to spec and still conceal slag, cold lap, or a crack within its cross-section. When the contract demands proof or concerns remain after visual checks, inspectors turn to destructive and supplementary testing tools to verify fusion, ductility, and material identity.

Bend Test Coupons: Proving Fusion and Ductility

Bend tests are one of the most direct ways to prove that the weld metal and heat-affected zone can deform without cracking. A coupon is machined from the weld, cut across the joint, and bent over a mandrel to a specified included angle. Under AWS D1.1/D1.1M, acceptance is strict: the bent surface must show no cracks in the weld metal or heat-affected zone. Cracks that start outside the HAZ are assessed against the base metal provisions instead.

Orientation is just as important. A face bend places tension on the weld face, while a root bend places tension on the root, making it far harsher for root fusion. For full penetration groove welds, the reinforcement at the bend face must be ground flush before testing. Without this step, the extra cap metal can act as reinforcement and mask a lack of fusion. These tests are a staple of procedure and welder qualification, and are mandatory when the code or drawing calls for them.

Macro-Examination and Macro Etch

Macro examination is as close to an autopsy as a weld gets. An inspector cuts a cross-section, grinds and polishes it, then etches it with a reagent to contrast the deposited metal from the base metal. At 5x to 10x magnification, the fusion line becomes visible, revealing imperfections such as lack of fusion, trapped slag, gas porosity, or cold shuts.

However, its scope is limited. Per ISO 5817, if macro-examination is used to detect imperfections, only those detectable at a maximum of tenfold magnification may be considered. Micro lack of fusion and microcracks fall outside of this assessment and require other methods. Even with these limits, macro examination is routine for qualifying processes, particularly in industries such as aerospace welding, where confirming full fusion is non-negotiable.

Hardness Testing as a Nondestructive Proxy

Hardness testing cannot directly measure strength or locate buried defects. Instead, it gauges a weld’s resistance to indentation using methods such as Brinell, Rockwell, or Vickers. Portable, rebound-type (Leeb) instruments allow these readings to be taken directly on the finished joint in the field.

Its value lies in context. Codes use hardness as a proxy for hydrogen cracking susceptibility or to verify a material responded correctly to heat treatment. For that reason, results must be compared against a specified limit, not judged by the number alone. A hard weld can still contain an unfused root, so it will never replace volumetric or destructive proof testing.

The Supplementary Toolbox

Not every job requires cutting coupons. Many rely on smaller, targeted checks to confirm material identity, wall integrity, or coating coverage.

Tool	What it verifies	Typical trigger
PMI / XRF alloy analyzer	Base metal and filler match the specified alloy	Unknown heat numbers, mixed deliveries, or stainless contamination risk
Spark test	Rough alloy family sort	Fast triage on incoming stock before a full PMI scan
Coating (DFT) thickness gauge	Dry film thickness of paint or coating	Pre-coat and post-coat inspection over welds
Ultrasonic thickness gauge	Remaining wall thickness	Erosion, corrosion, or replacement of pipe sections
Hardness and corrosion test coupons	Production-level strength or corrosion behavior	When the code mandates production coupon testing

Spark testing is strictly a triage tool, not acceptance-grade verification. PMI is what closes the alloy question when a grade is specified. Similarly, an ultrasonic thickness gauge is a wall gauge, not a flaw detector, and a coating thickness reading over a weld bead will include the reinforcement beneath it.

Pre-Weld and In-Process Inspection Tools

The cheapest defect to fix is the one that never gets made. A root gap 10 mm over the procedure limit, mill scale on a fusion face, the wrong heat number, an unverified preheat: each makes a weld that looks fine and still gets rejected. Once the arc is struck, the options narrow to record, reject or repair. Before that, the fix is free.

The Document Set Is Itself an Inspection Tool

Before a gauge leaves the bag, the primary instrument is the document set. Five items belong on the table at every pre-weld check:

• The WPS and its supporting PQR, the exact procedure the welder will work to.
• The welder’s qualification for that process, position and thickness range, with a current continuity record.
• The drawing and governing code, so acceptance criteria are set upfront.
• The inspection plan and its hold points and witness points.
• Mill certificates and heat numbers for the material in the joint.

AWS B1.11M:2015 sorts visual examination into three categories, prior to welding, during welding and after welding, and states that welding should not begin before the visual acceptance criteria are clearly defined. It also notes that the examiner should be familiar with the principles and methods of visual examination, and that certification may be required by the welding standard or the law. “The drawing looked fine” is no defence when the criteria were never written down.

Physical Pre-Weld Checks

ISO 17637:2016 clause 6.2 sets the physical baseline: clean fusion faces and adjacent surfaces, required surface treatment carried out, parts correctly fixed relative to each other.

Pre-Weld Check	Tool	What You Are Proving
Joint fit-up	Fit-up gauge, straightedge, square	Parts fixed, misalignment within limits
Root gap	Feeler gauge, hi-lo gauge	Opening within procedure, high-low corrected
Bevel prep	Bevel gauge, AWS C4.1 roughness guide	Angle, root face and land as specified
Surface condition	Visual, solvent cleaner, abrasive	No mill scale, oil, paint or oxide
Material identity	Marking check, heat trace	Plate matches drawing and certificate

Heat numbers on the plate and the mill certificate must reconcile, and for stainless or aluminium the marking can be part of the acceptance criterion, so wiping a stencil off is a real nonconformance. Cleaning aluminium properly before you weld is an inspection topic, not housekeeping.

In-Process: Heat, Sequence and Access

During welding the inspector monitors the procedure: preheat verified before the first arc, interpass temperature under the WPS maximum, and fit-up rechecked before the root pass, once filler metal lands the joint is fixed. Temperature crayons and temp sticks are cheap, but they measure one point and can be fooled by a hot interpass region nearby. Thermocouples give a number and a record; an infrared thermometer gives speed at the cost of accuracy on shiny, scaled or emissivity-mismatched surfaces.

Two ISO 17637 clauses define the in-process obligation. Clause 6.3 requires each run or layer to be cleaned before it is covered, visible imperfections reported so remedial action comes before more weld metal is deposited, and the depth and shape of any gouging to match the WPS. So carbon arc gouging for weld removal is legitimate only when the excavator knows the WPS limits for gouge depth and shape. Clause 6.1 adds that visual testing before, during or after welding must happen while access remains: once a pass is buried under three more layers, nobody can see it.

When You Find a Defect Mid-Weld

The hard part is deciding. Spot a defect while the arc is hot and the sequence is fixed: stop, document, report to the supervisor before more metal is deposited. Do not talk the welder into grinding out an unsanctioned repair, fix it yourself, or let the next pass bury the evidence. Record the location with a weld map reference or a photograph, describe the condition, state the criterion it fails, and let the engineer decide the disposition. Cleaning a defect away without a record is indistinguishable from hiding one.

Safety deserves the same discipline. Arc radiation and welding fume do not pause because an inspector carries a flashlight. Correct eye protection, respiratory protection and fume extraction matter on an inspection walk as much as on the welding floor.

Buying a Kit, Calibrating It, and Keeping Records That Hold Up

Building Your Kit in Tiers

A welding inspector does not need to buy every measuring device on the market to be effective. The smartest approach is to build your kit by necessity, not novelty. Start small with the items that will be used on almost every inspection, and expand only when your code, job scope, or qualifications call for it.

The essential starter set should be the backbone of any kit. At a minimum, this should include a fillet weld gauge set, bridge cam gauge, hi-lo gauge, undercut gauge, taper/feeler gauge, protractor, steel rule, dial caliper, and a machinist square. These are the tools that verify the most common acceptance criteria on the shop floor.

The optical tier is where visibility becomes measurable. A reliable flashlight or headlamp, inspection mirrors in multiple sizes, a 10x loupe, a borescope, and a phone camera with a clip light will allow you to inspect tight joints and root sides without having to contort yourself into an unsafe position.

The documentation tier is just as physical as any gauge. Keep WPS/PQR forms, a weld map, inspection report forms, a gauge calibration log, and nonconformance and repair tracking forms in a dedicated folder. Without these, a correct reading is still not a defensible record.

The advanced tier is reserved for inspectors with a defined need for them. Items such as a portable hardness tester, UT thickness gauge, coating thickness gauge, temperature measurement devices, and a PMI instrument are invaluable, but they are not required for every assignment. When sourcing some of these specialty items, it can be worth looking into reputable sources such as Airgas Welding Supplies to compare options for your specific scope of work.

When purchasing, consider metric versus US customary sets. The AWS CWI Part B examination is supplied in Customary Unit (inches, degrees-Fahrenheit) for exams delivered in the US and Canada, and in Metric Unit (kilograms, millimeters-Celsius) for exams delivered internationally. Owning both sets can be costly, so only invest in both if your work crosses between the two systems regularly.

For gauges themselves, stick to brand-neutral practicality. What matters is stamped graduations, flat seating surfaces, rigidity, and no glare. Marketing claims will not make an unseated gauge read correctly. As far as high-end digital or optical measuring devices, such as laser distance tools or digital profilometers, they can add convenience, but they will never replace an understanding of the acceptance criterion for the code that you are working to.

Calibration: Trust, But Verify

Every measuring device in your kit has a breaking point, and trust must be earned through verification. There is a key distinction to be made here. For hand gauges, calibration is the process of comparing them against a traceable standard to confirm their graduations are accurate. For NDE instruments, calibration is far more in-depth, and as stated within ASME Section V, examination equipment must be calibrated and that calibration must be recorded and available to the Inspector. This is also one of the knowledge areas listed within AWS B5.1:2025 for Quality Assurance and Welding Coordination.

A few rules should be non-negotiable. A gauge that has been dropped onto a concrete floor is no longer trustworthy until it has been verified against a known standard. Calibration should be performed before use, at periodic intervals based off of your employer’s program, and immediately after any damage. Finally, know the difference between calibration and verification: verification is a quick check to confirm a gauge is still fit for use, whereas calibration is the documented comparison to a traceable standard with a record to prove it.

Records That Hold Up to an Audit

A pass or fail means nothing without a record. For visual testing, ISO 17637:2016 clause 7 is the gold standard for minimum test record information. Your reports should include the component manufacturer name, testing body name, identity of the object tested, material, type of joint, material thickness, welding process, acceptance criteria, imperfections exceeding the acceptance criteria and their location, extent of testing, test devices used, the result of testing, and the name of the tester and the date of test.

Additionally, welds tested and approved should be suitably marked or identified. If a permanent visual record is required, photographs or accurate sketches (or both) must be made, with any imperfections clearly indicated.

At the end of the day, the most expensive tool in your kit is an unverified one, and the most valuable one is an accurate, traceable record. Buy for the code that you inspect to, verify before you measure, and record every decision. That is what turns an opinion into a defensible, code-compliant result.

Common Mistakes, Troubleshooting, and the Conclusion: A Working Routine

It only takes one oversight to put an inspector’s credibility on the line. Most nonconformances come from a short list of habits that can be corrected with discipline and a working knowledge of the code.

Ten Mistakes That Keep Repeating

1. Not seating the gauge. A fillet gauge must sit on both legs and the base metal. Resting on a single leg will always give a false reading.
2. Reading the wrong scale. Many combination gauges have both inch and metric graduations. Double check which scale applies before recording any number.
3. Skipping a zero check. A bridge cam gauge must be zeroed against a flat surface and verified every time it’s picked up. An un-zeroed gauge is not an instrument.
4. Inspecting under flat light. Hairline cracks are easily missed head on. Rake a flashlight at a shallow angle of 10 to 15 degrees to cast the shadow needed to reveal them.
5. Inspecting before cleaning or too early. Visual inspection may begin once completed welds have cooled to ambient temperature under AWS B1.11M, but all slag must be removed. Spatter is only acceptable where NDT other than visual is not required and nothing else is specified.
6. Only checking one leg. For fillet welds, both legs must be measured. An undersize on either leg is still undersize.
7. Using a damaged gauge. A drop onto concrete can bend an edge or throw a zero off. If there’s any doubt of damage, take it out of service.
8. Relying on “looks good”. An opinion will never hold up on a report. Always record the number, location, date, and name.
9. Inspecting over coatings. For PT and MT, coatings must be removed prior to inspection. If a hold point exists before coating, inspection must happen at that hold point.
10. Running MT on non-ferrous metals. Magnetic particle testing only works on ferromagnetic metals. Aluminium requires alternate methods, such as penetrant testing, as outlined in our aluminum MIG welding guide.

Troubleshooting When Numbers Disagree

• One leg reads short. Before rejecting, verify fit-up, ensure the gauge is seated on both legs, and confirm which leg is called out on the drawing. Distortion is often the culprit.
• Reinforcement is off. Check the joint type against the drawing. A joint specified as flush allows no reinforcement, meaning a proud weld is a nonconformance, not a negotiable measurement.
• High-low after the root pass. The issue is fit-up, not deposition. Realign, then gouge or cut out the root as required. Shops often use air arc gouging and carbon arc gouging to remove a bad root before re-fitting.
• UT disagrees with visual. Never average the two. A crack parallel to the ultrasound beam can reflect very little energy and appear clean, so a visible indication must be escalated for qualified evaluation.

Safety, Ethics, and the Report You Owe

Inspection puts you close to live welding. Control exposure to weld fume and arc radiation with the correct PPE and respiratory protection for the process and material. Professionally, the obligation is simple: report what you find, with no adjustments. AWS B5.1:2025 places establishing and enforcing visual requirements with the employer, and AWS QC1 states that a certification makes no determination of an individual’s capabilities in a specific work environment or under actual working conditions. The certificate opens the door. Your record keeps it open.

Conclusion: A Working Routine

The hardware is the cheapest part of inspection. Knowing which criterion governs the joint, applying the correct tool properly, and documenting exactly what was measured is what carries weight. For those looking to build that discipline, the AWS Certified Welding Inspector pathway is a solid start, but consistency in the field is what makes it effective.

Frequently Asked Questions

What inspection tools does a welding inspector actually need to get started?

The minimum working kit is a fillet weld gauge set, a bridge cam (Cambridge) gauge, a hi-lo gauge, an undercut gauge, a taper or feeler gauge, a protractor, a steel rule and a dial caliper, plus a flashlight, inspection mirror and 10x loupe. AWS’s own CWI Part B practical examination kit is built from essentially that list. Add a copy of the governing code’s acceptance criteria and you have enough to run visual inspection on most structural and pipe work.

Is a digital weld gauge better than a manual fillet gauge?

Not for the measurement itself. Manual fillet and bridge cam gauges remain the reference tools because they are fast, rugged, cheap and have no battery to die on a job site. Digital tools help with recording, reproducibility and measuring dimensions a gauge cannot reach, but they do not change the acceptance criterion, and an inspector who cannot read the criterion cannot use any tool.

How often should welding inspection gauges be calibrated?

There is no single mandated interval for a hand gauge, but a sensible practice is a visual and functional check before each shift or inspection, a periodic verification against a master or certified standard on a defined schedule, and re-verification after any impact, drop or suspected damage. ASME Section V requires calibration of examination equipment with the calibration recorded and available to the Inspector, which is the standard most audits will hold you to.

What is the difference between a bridge cam gauge and a fillet weld gauge?

A fillet weld gauge set checks leg length and the convexity or concavity of a fillet profile. A bridge cam gauge is the general-purpose instrument: it measures undercut depth, groove weld reinforcement height, root opening, misalignment, bevel angle and fillet leg size. Most working inspectors carry both because fillet gauges are quick and bridge cam gauges are comprehensive.

Can a welding inspector perform MT, PT or UT themselves, or do they need separate certification?

The welding inspector normally interprets results rather than producing them. Under ASME Section V, radiography and ultrasonic examiners must be qualified and certified to the employer’s written practice or a central certification program, while magnetic particle, liquid penetrant and visual examiners may be certified by the Certificate Holder on proficiency plus an annual eye examination. AWS also offers separate CWI endorsements for dry powder yoke MT and Type II Method C PT.

What lighting and access does the code require for visual inspection?

ISO 17637 requires at least 350 lx of white light at the surface, with 500 lx recommended, and access sufficient to place the eye within 600 mm of the surface at an angle of not less than 30 degrees. Tinted safety glasses increase the minimum illuminance required, so plan your lighting for the glasses you actually wear.

How do I measure weld undercut depth without a bridge cam gauge?

You can use a dedicated undercut gauge, which reads finer increments (typically down to 0.01 in. or 0.25 mm), or use a depth micrometer with a sharp point placed against the toe. Just remember to measure from the weld toe into the base metal along the deepest part of the groove, and to compare against the limit for the specific material thickness and loading case, which in AWS D1.1 varies with thickness and with whether the weld is transverse to tensile stress.

When is a hi-lo gauge mandatory rather than optional?

Any time the code or specification calls out root misalignment, which is common in pipe and tubular work where offset or high-low drives inadequate penetration and high stress concentration. The hi-lo gauge hooks over the pipe wall and measures the offset between the inside surfaces at the joint, so it should be checked during fit-up, before the root pass is welded, not after.

What are the most common measurement errors new welding inspectors make?

Three dominate: seating the gauge on the edge of the plate or on spatter so the reading is off by the thickness of spatter, failing to zero the bridge cam gauge before use, and checking only one leg of a fillet weld. Reading a combined metric/imperial scale in the wrong units and measuring before the weld has cooled or after the surface has been coated are close behind.

Where should an inspector look first if a weld looks good but the client’s NDE rejects it?

Start with orientation. A crack running parallel to the radiographic beam can show as a faint line or not appear at all, while a crack perpendicular to the beam images well; that mismatch between what you saw and what was reported is often a technique and beam-angle issue rather than a fabrication defect. Then check the weld’s location against the NDE symbol on the drawing and the technique identifier on the report before condemning the weld.