The practical portion of an inspection examination is the one candidates most often underestimate. It is not a knowledge test with props. It is a measurement test with a document, and the skills it measures are physical, procedural, and time-constrained in a way that reading cannot prepare you for. People who have inspected welds for years still lose points here because their normal working pace assumes they can go back and check something later.
This guide covers what the hands-on part actually measures, the gauges you will pick up, how to read each one without second-guessing yourself, how to work from a specification you have never seen before, and how to manage the clock. It describes general inspection practice and equipment. It does not restate any published acceptance criteria, and any numbers that decide accept or reject on a real job come from the governing code or contract documents, never from a study guide.
GaugePass is a study aid and is not affiliated with or endorsed by the American Welding Society. The practical portion of the AWS Certified Welding Inspector (CWI) exam is defined by the certifying body, and their current published documentation is the authoritative source for what is supplied, what is expected, and how it is scored.
What the practical part is really measuring
Strip away the format and the practical portion is asking three questions about you. Can you physically measure a weld feature accurately and repeatably with the correct tool? Can you take a written acceptance requirement you have never seen before and apply it literally to what you measured? Can you record a disposition clearly enough that someone else could audit your decision without asking you anything?
Notice what is not being tested. It is not testing whether you can spot a defect from across a shop, and it is not testing whether you know a code from memory. Experience gives you speed, but experience also creates the biggest trap in this part, which is answering from the way your employer does things rather than from the specification you were handed on the day.
The examination environment adds a fourth constraint that nobody trains for: you are working on unfamiliar samples, with equipment that may not be the exact model you own, under a clock, with no colleague to sanity-check a reading. Preparation that ignores those conditions produces a candidate who knows how to inspect and still runs out of time.
The gauges you will actually pick up
The working set is small. Fillet weld gauges give you leg size and let you check for convexity or concavity against a nominal profile. A bridge cam gauge is the multi-tool of visual inspection, covering reinforcement height, undercut depth, fillet leg, angle of preparation, and misalignment depending on how you set it. A hi-lo gauge handles internal alignment on pipe. A basic rule, a magnifier, and adequate light round it out.
Fillet gauges come in two common styles and it is worth being fluent with both. One style is a stack of individual leaves, each cut for a nominal size, which you place against the joint until you find the one that fits. The other is a pair of gauges with sliding or fixed profiles, one for concave welds and one for convex, which read the size directly. They give the same answer when used correctly and they fail differently when used carelessly.
Your own tools are worth buying early even if the examination supplies equipment. Familiarity removes hesitation, and hesitation is what costs you the clock. Practise on real welds at work rather than on clean samples, because the whole skill is seating a gauge correctly on a surface that is not flat, not clean, and not the shape the diagram assumed.
- Fillet weld gauges: leg size and profile against a nominal shape
- Bridge cam gauge: reinforcement, undercut, fillet leg, preparation angle, misalignment
- Hi-lo gauge: internal alignment across a pipe joint
- Rule, magnifier, and a light source you control
Reading a fillet weld gauge without second-guessing
A fillet gauge only tells the truth if it is seated on both legs of the joint. The most common error is letting the gauge rock or sit on a bead crown rather than on the base metal, which reports a size the weld does not have. Set the gauge flat against one member, bring it into contact with the other, and confirm both faces are touching metal before you read anything.
For a convex fillet, the size is governed by the leg length, and the gauge is finding the largest isosceles right triangle that fits inside the weld profile. For a concave fillet, the throat is the limiting dimension rather than the leg, which is why concave and convex gauges are shaped differently and are not interchangeable. Using the wrong one produces a confidently wrong number.
Measure at more than one point and record the governing one. A fillet is rarely uniform along its length, so a single reading at a convenient spot tells you about that spot. The number that matters is the one at the worst location, and finding that location quickly is a scanning skill worth practising deliberately: run the gauge along the weld rather than placing it once.
The bridge cam gauge: one tool, several measurements
The bridge cam gauge does several jobs because it combines a bridge that spans the weld with a rotating cam and a pointer. Set on the base metal either side of a groove weld, the cam is rotated until it just touches the weld crown, and the pointer reads reinforcement height. The same geometry, applied to a depression instead of a crown, reads undercut depth. Set differently, the tool reads fillet leg size, preparation angle, and misalignment across a joint.
Each of those uses has its own seating requirement and its own way of going wrong. For reinforcement, the bridge feet must both be on sound base metal rather than on spatter or on the toe of the weld. For undercut, the pointer must be in the deepest part of the groove, which is not always where it first drops in. For misalignment, the tool measures the offset between two surfaces and tells you nothing about which one is wrong.
Slow down on the setup and speed up on the read. Candidates lose time by taking a reading, doubting it, and repeating the whole measurement. A deliberate two-second seating check followed by one confident read is faster and more accurate than three hurried attempts, and it is the habit worth building in practice sessions.
- Reinforcement height on groove welds, with both feet on sound base metal
- Undercut depth, with the pointer at the deepest point of the groove
- Fillet leg size as an alternative to a dedicated fillet gauge
- Angle of preparation on a prepared joint
- Misalignment or offset between two surfaces across a joint
Working from a specification you have never seen
The specification handed to you in a practical examination is written for that examination. It is not a code you can study in advance, and that is the design intent: the ability being measured is applying an unfamiliar written requirement literally and quickly. Your years of knowing what is normally acceptable are actively unhelpful here, because the document in your hand may set a different requirement and it governs.
Read the acceptance section before you touch a sample. Work out which characteristics are addressed, what units they are expressed in, and whether any requirement depends on material thickness, weld type, or location. Then measure against that list rather than measuring everything you can think of. Candidates who inspect first and read second end up measuring characteristics the document does not care about.
When you write a disposition, cite the provision you applied. On the day it demonstrates that your decision came from the document. On the job it is the difference between a finding that survives a challenge and an opinion that does not. Getting into that habit during practice makes it automatic under pressure.
- Read the supplied acceptance requirements before handling any sample
- List the characteristics the document actually addresses, in the units it uses
- Note any requirement that varies with thickness, joint type, or weld type
- Inspect each sample against that list rather than against habit
- Measure at the worst location, not the most convenient one
- Record the measurement, the provision applied, and the disposition together
Weld replicas and what they simulate
Practical examinations commonly use plastic replicas alongside or instead of steel samples. Replicas reproduce weld surface geometry faithfully, which is what visual examination reads anyway, and they are consistent between candidates and locations in a way that a batch of welded coupons could never be. Treat them exactly as you would treat metal, because the measurement task is identical.
The differences that matter are practical rather than conceptual. Replicas can be lighter and can flex slightly under gauge pressure, so seat gauges without leaning on them. They do not have the surface texture, mill scale, or oxide that helps you judge features on real steel, so lighting matters more than usual. Angle a light across the surface rather than straight at it, because raking light is what makes a shallow surface condition visible.
The habit to build is scanning before measuring. Look over the whole sample under good light and identify every candidate condition before you pick up a gauge. Working sample by sample in one pass is far faster than discovering a second condition after you have already written up the first and moved on.
Managing the clock
Time is the constraint that decides most outcomes in this part. The failure pattern is consistent: candidates spend too long on the first few items, become aware of the clock, and then rush the remainder. Unanswered items score nothing regardless of how carefully you handled the earlier ones, so pacing is worth as much as accuracy.
Divide the available time by the number of items before you start, and set a rough per-item budget. When an item runs over budget, record your best current answer, mark it, and move on. Returning to a marked item with five minutes left is a much better position than never reaching the last three items at all.
Practise this deliberately. Set a timer, work a batch of samples end to end, and note where the time went. Most candidates find the loss is not in measuring but in re-reading the specification repeatedly, which is solved by digesting the acceptance requirements properly at the start rather than dipping back into them for every sample.
Practice that actually transfers
The practice that transfers has three ingredients: real gauge handling, an unfamiliar written acceptance requirement, and a clock. Drop any one and you are training something easier than the thing you will be asked to do. Reading about gauges builds none of it. Measuring welds without a document builds half of it. Measuring welds against a document you wrote yourself builds most of it, but you already know what it says.
Manufacture unfamiliarity where you can. Have a colleague write acceptance criteria you have not seen, or take criteria from an unrelated area of work and apply them to samples you have on hand. The point is to force the read-then-apply loop rather than the recall loop, since recall is exactly the instinct that fails you.
If you cannot get hardware, an interactive trainer that puts a gauge scale against a weld profile and asks for a reading still builds the interpretation half of the skill, and it builds it in short sessions you can do on a phone during a break. Combine it with whatever physical measurement you can get at work, and add the clock as soon as your readings are consistent.