Welder Training by the Numbers
Chapter 2 of 10 · 10 min read
How welding skill is learned
Follow a trainee from conscious effort to automatic skill, count the repetitions an hour of practice gives, and see why feedback, spacing and variety decide how fast welding skill grows.
Welding is a manual skill of an unusual kind. The welder must hold a torch or gun at a steady distance and angle, move it at a steady speed, weave it in a pattern, feed filler, all while watching a small, bright, moving pool of molten metal through a dark filter and adjusting every movement to what the pool is doing. None of that can be learned from a book or a lecture. Campbell put it simply in 1919: watching other welders will not teach anyone to weld, though it may help a beginner; the skill comes from practice at the torch, thought and determination.
This chapter describes how such a skill develops, from conscious effort to automatic performance, counts how many repetitions an hour of practice provides, and explains why feedback, spacing and variety decide how fast a trainee improves. It prepares for the learning curve in Chapter 3, which puts numbers on the improvement.
What the welder controls
Every welding process gives the welder a handful of variables to control by hand and eye. For the fabricator’s flux-cored process, the main ones are:
- contact tip to work distance, the stickout, which affects the current and the arc;
- travel speed, which sets the size of the weld and the heat put into it;
- torch angle, both along and across the joint, which directs the arc and the shielding gas;
- weave, the side-to-side movement used on some welds to spread the pool across the joint;
- position of the arc in the pool, at the leading edge where it melts the parent metal, not on top of the pool where it buries unfused metal.
A good weld needs all of them right at once, continuously, for the length of the weld. Each varies with position, plate thickness and joint type. The skill lies in holding them steady and in reading the pool well enough to adjust them when something changes.
The 1919 description of torch skill
The Davis-Bournonville course of 1919 described the gas welder’s version of the same skill precisely. The welder had to learn to hold the torch so that the tip of the white-hot cone stayed about an eighth to three-sixteenths of an inch, roughly 3 to 5 mm, above the pool, and to do it unconsciously, while giving the torch a movement across the joint that distributed the heat. The course explained that a semicircular weaving movement was better than a plain zigzag because the flame dwelt longer at the edges of the joint, where heat was lost by conduction and fusion was hardest to achieve; a straight zigzag passed over the edges too quickly and left them cold, causing laps and cold shuts. It also warned against melting the filler rod in the flame and letting it drip onto the joint, which gave the same faults.
The word that matters in that description is “unconsciously”. A beginner can hold the right distance and make the right movement only by concentrating on them, and then has no attention left for watching the pool. A trained welder does them automatically, and watches the pool.
Three stages of skill
Psychologists studying motor skills, notably Paul Fitts and Michael Posner, described learning a skill as passing through three broad stages, and the description fits welding well.
- Cognitivethe trainee thinks about each step: distance, angle, speed; welds are slow and uneven
- Associativethe steps are linked; the trainee starts to read the pool and correct; errors fall quickly
- Autonomousdistance, angle and speed are automatic; attention goes to the pool and the joint
In the cognitive stage, the trainee is working out what to do and doing it deliberately, step by step. Progress is fast, because every early mistake is large and easily corrected, but the welds are slow and inconsistent. In the associative stage, the movements are becoming linked and the trainee begins to notice the pool’s behaviour and adjust; errors fall but the welder still needs to concentrate. In the autonomous stage, the movements take care of themselves, and the welder’s attention is free for reading the joint, which is what allows a skilled welder to cope with poor fit-up, a change of position or a gust of wind. Each new position or process starts a partial return to the earlier stages.
Why lectures come second
Campbell observed that students could take in only so much from lectures, however quickly the instructor moved through the material, and that his school had spent considerable time working out the right ratio between lecture subjects and shop work. His conclusion was that shop practice was the most important part of the course. Theory matters, because a welder who knows why a weld cracks or why a gas must be shut off in a certain order can solve problems that a trained hand alone cannot. But in the cognitive stage the trainee’s attention is already full, and theory that cannot yet be connected to practice is mostly forgotten. Chapter 4 uses this to plan the balance of a course.
Where watching helps
Campbell allowed that observation may at times assist the beginner, and a good demonstration does. A trainee who has watched an instructor set up a coupon, set the machine, take up a stance and run a weld has a picture of the whole task before trying it, which shortens the cognitive stage. The demonstration works best when it is short, shows one thing at a time, and is followed at once by the trainee trying the same thing. Watching a skilled welder work for a day, with no chance to try, adds little, because the trainee cannot see what the welder is controlling: the important movements are small, and the important information is in the pool, behind the welder’s helmet.
The instructor’s eye
The instructor sees what the trainee cannot: the trainee’s stance and the angle of the gun from the side, the speed of travel against the joint, and the drift of the arc away from the leading edge of the pool as concentration fades. Watching each trainee weld for part of every session, through a suitable filter, and giving one or two specific corrections at a time, is one of the most effective things an instructor can do. Video recordings through a welding camera, now inexpensive, let trainee and instructor watch the pool together after the weld, which turns the fastest and hardest feedback into something that can be discussed.
Confidence and frustration
Learning a manual skill is frustrating. Progress comes in bursts, separated by plateaus where nothing seems to improve, and a bad day can undo a trainee’s confidence. Trainees who can see their own progress, in the records of reject rate and speed that Chapter 3 describes, are more likely to persist through a plateau than those who have only a feeling that they are not getting anywhere.
Practice as repetition
Skill grows with repetition, so it helps to count repetitions. A “run” here is one practice weld of a set length on a coupon.
- Arc time per run
- 150 ÷ 250 = 0.6 min
- Arc time fraction in practice
- about 20%
- arc minutes per hour = 60 × 0.20 = 12 min
- runs per hour = 12 ÷ 0.6 = 20
- 400 hours of practice ≈ 8,000 runs
Twenty runs an hour sounds a lot, but each run is only 36 seconds of arc, and the rest of the hour is spent cutting and tacking coupons, cleaning and inspecting welds, changing wire and resting eyes and hands. The arc time fraction in practice is usually lower than in production, because inspection and feedback take time, and it should be: practice without inspection is repetition of whatever the trainee is already doing, right or wrong.
Feedback
A trainee improves only by finding out what went wrong and changing it. Welding gives feedback at three levels, at different speeds.
| When it arrives | What it shows | Who gives it | |
|---|---|---|---|
| The pool, while welding | instantly | wetting, size, arc position | the trainee, once they can read it |
| The finished bead, cleaned | within minutes | profile, undercut, overlap, spatter, size | the trainee and instructor |
| Break, bend or macro test | within an hour | fusion, penetration, internal defects | the instructor, with the trainee |
The pool is the most immediate feedback, but the beginner cannot yet read it, so early in training the instructor’s job is to make the slower feedback fast: inspect each run with the trainee as soon as it is cleaned, break a coupon every few runs to show the fusion inside, and connect what the trainee sees on the outside to what was happening in the pool. Campbell’s school did exactly this, breaking welds and examining the cross-section to show students that metal added where the parent metal had not been fused was not a weld at all. Testing Welded Joints by the Numbers describes the bend and break tests in its chapter on bend tests and ductility.
Spacing and fatigue
Practice is more effective spread out than crammed together. Several sessions of two or three hours, separated by production work or rest, generally build skill more durably than a single long session, because concentration and fine control fall with fatigue, and because skill consolidates between sessions. Welding is tiring: the posture, the heat, the concentration and the dark filter all take their toll, and runs at the end of a long session are often worse than runs in the middle. READOS explains the general case for spaced practice in The forgetting curve and why training fades.
Variety
Practising one weld in one position builds that weld; practising a mixture builds the ability to adapt. Once a trainee can make an acceptable flat fillet, mixing flat, horizontal and vertical practice, and varying plate thickness, builds a more robust skill than perfecting the flat fillet first and only then moving on, even though the mixture feels slower and harder at the time. The fabricator’s plan in Chapter 4 introduces the vertical position early and returns to each position repeatedly.
The simplified start
Campbell described his school’s method as deliberately simple. Certain principles were instilled in the beginner, and minor exceptions were set aside to avoid confusion, to be learned later. For the beginner at gas welding, his manual recommended a method of adding filler that needed little torch movement and could be repeated until enough metal had been added, letting the student make a sound, reinforced weld before mastering the more elaborate techniques. The principle is sound for any process: give the beginner one reliable method that works, and add refinements once the basic movements are automatic.
The running case: the first fortnight
This is an illustration. In their first fortnight, each trainee spent about 20 hours on fillet practice in the flat position, about 400 runs each. Their progress differed as the manager had expected. Dee, with some welding experience, made acceptable-looking flat fillets within a few hours but had to unlearn a habit of holding the gun too far from the work. Chris, who knew the shop and the equipment, was steady and careful, slow but improving fast. Ana and Ben started from almost nothing and spent the first days on striking and holding an arc, but by the end of the fortnight both were making continuous beads. The manager recorded each trainee’s reject rate on break-tested coupons and their average time per run, which became the first points on the learning curves in Chapter 3.
A failure: flat only
This is an illustration. In an earlier attempt at training, the fabricator had kept a trainee on flat fillets for two months, until they were excellent, before introducing vertical welding. The trainee’s vertical welds were poor for weeks, and the trainee became discouraged, because the movements perfected in the flat position were partly wrong for vertical-up welding, where the pool must be supported against gravity and the weave and speed are different.
The manager’s conclusion, built into the new plan, was to introduce each position early, once the trainee could hold a steady arc, and to mix positions in practice from then on. Progress in the flat position was slightly slower, but the trainees reached acceptable welds in all three positions sooner, and none of them came to see the vertical position as a wall.
Exercise: runs and arc time
A trainee practises for 3 hours a day, four days a week, on 200 mm runs at a travel speed of about 200 mm/min, with an arc time fraction of 25%. How many runs does the trainee make in a week, and how many hours of arc time? How many weeks to reach 5,000 runs?
Each run takes 200 ÷ 200 = 1 minute of arc. In an hour, the arc time is 60 × 0.25 = 15 minutes, so the trainee makes about 15 runs an hour. In a week of 12 hours, that is about 180 runs and 12 × 0.25 = 3 hours of arc time.
To reach 5,000 runs takes about 5,000 ÷ 180 ≈ 28 weeks, more than half a year at this schedule. The arithmetic shows why welding skill takes months, and why the hours of practice, not the weeks on the calendar, are the measure of a training plan.
