Why train welders, and what training must achieve

In 1919, Lorn Campbell Jr., the officer in charge of welding instruction for the United States Army’s Ordnance Department, opened his manual with a complaint that would be familiar to any fabrication manager today. The demand for welders had always been greater than the supply, he wrote, and when the government came to form its wartime Army and Navy it found so few people familiar with the process that it set up its own schools to train them, something industry had been thinking of doing for years. A century later, fabricators still say the same thing, and many still do what the Army did: train their own.

This chapter sets out the choices a fabricator has when it needs welders, defines what training must achieve in terms that can be observed and measured, and puts a number on the cost of poor welding, which is the main thing training is meant to reduce. It introduces the fabricator and the four trainees whose training the rest of the book follows.

Three ways to get welders

A business that needs more welding capacity has three broad options. It can hire qualified welders, if it can find them; it can train its own, from apprentices or from people in other roles; or it can subcontract the welding to another business. Each has its own cost, risk and lead time.

Figure 1Three ways to meet a need for welders

	Lead time	Cost	Main risk
Hire qualified welders	weeks to months, if available	market wage plus recruitment	none available; skills not as claimed
Train your own	months to years	training cost plus lower output while learning	trainees leave once trained; slow results
Subcontract	days to weeks	subcontractor's margin	quality and delivery outside your control

Hiring is fastest when welders are available, but in a tight market a business may wait months, pay a premium, and still find that a welder’s claimed skills do not match the work. Subcontracting buys capacity without commitment but hands quality and delivery to someone else. Training is slow and costs money before it returns any, but it builds capacity the business controls, in the processes and standards it actually uses, and it is the only option that adds welders to the market rather than competing for the existing ones. It also lets the business shape its welders around its own procedures and quality expectations from the first day. Most fabricators that take welding seriously train some of their own.

What training must achieve

“Train a welder” is too vague to plan, cost or measure. Training needs a target that can be observed: what the trained person must be able to do, under what conditions, to what standard. READOS sets out the general method in Building workforce capability in manufacturing; for a welder, the target typically covers:

• the processes, such as flux-cored, solid-wire or manual metal arc welding, and the equipment for each;
• the joints and positions, such as fillet and butt welds, flat, horizontal, vertical and overhead;
• the materials and thicknesses, such as carbon steel plate from 6 to 20 mm;
• the quality standard, meaning welds that meet the acceptance criteria of the work, judged by visual inspection and tests;
• the productivity, meaning welding at a rate the business can afford, not just eventually producing a good weld;
• the knowledge, such as reading a welding procedure, setting the machine to it, recognising defects in one’s own work, and knowing why the rules exist;
• the safety, working safely and protecting others, as Chapter 9 describes.

A target written this way can be tested. A trainee either can or cannot make a vertical fillet weld in 10 mm plate that passes visual inspection and a break test, at a travel speed within the procedure’s range, without supervision.

The cost of poor welding

Training is an investment, and the return comes mainly from welds that do not have to be repaired. A weld that fails inspection must be ground out and rewelded, often taking longer than the original weld, and a defect that escapes the factory can cost far more in the field. The cost of poor welding is therefore a good place to start in deciding how much training is worth.

Figure 2Annual cost of weld repairs

repair cost = metres welded × repair rate × cost to repair a metre

Metres of weld a year

about 6,000 m

Cost to repair a metre

about A$60

1. at a 3% repair rate: 6,000 × 0.03 × 60 = A$10,800 a year
2. at an 8% repair rate: 6,000 × 0.08 × 60 = A$28,800 a year
3. difference between the two rates: about A$18,000 a year

The repair rate is the fraction of weld length that needs repair, from inspection records. It depends heavily on the welders’ skill, and a few poorly trained welders can raise the whole shop’s rate. Every percentage point of repair rate is worth about A$3,600 a year to this fabricator, before counting the cost of a defect that reaches a customer.

Figure 3The fabricator's annual repair cost at different repair rates

2% repair rateA$7,200 a year, excluding GST

3% repair rateA$10,800 a year, excluding GST

5% repair rateA$18,000 a year, excluding GST

8% repair rateA$28,800 a year, excluding GST

Productivity counts too

A welder who makes good welds slowly costs money in a different way. Arc Welding by the Numbers shows in its chapter on arc time and operator factor how welding time per job is built from arc time and everything around it. A trainee who takes twice as long as an experienced welder to complete a frame doubles the labour cost of that frame. Training must therefore aim at speed as well as quality, and the learning curve in Chapters 2 and 3 describes how both improve with practice.

The running case

This is an illustration. The fabricator in this book has twenty-two staff, eight of them welders, and makes portal frames, platforms, stairs and handrails for commercial and industrial buildings. It welds about 6,000 m of weld a year, mainly fillet welds in carbon steel from 6 to 16 mm, with flux-cored and solid-wire processes in the shop and manual metal arc on site. Its work is mostly to structural steel welding standards, which require qualified procedures and qualified welders.

Two of its experienced welders are due to retire within three years, and it has failed to hire replacements for a year. It has therefore taken on four trainees:

• Ana, a first-year apprentice straight from school;
• Ben, a second-year apprentice transferred from another employer;
• Chris, a boilermaker’s labourer with the business for four years, who has tacked and ground but not welded production work;
• Dee, a career changer from automotive repair, with some light-gauge welding experience.

The workshop manager’s target for each is the same: to weld production fillet welds in the flat, horizontal and vertical-up positions with the shop’s flux-cored process, to the shop’s qualified procedures, and to pass the welder qualification test for that work, within a year if possible. The apprentices will continue to a full trade qualification beyond that.

Measuring the starting point

Before training starts, the manager records where each trainee begins: a short practical test on a few fillet coupons, a check of eyesight and hand steadiness where relevant, and a conversation about what they already know. The starting point is not a pass or fail; it is the first point on each trainee’s learning curve, against which progress is measured. Without it, the business cannot tell whether a trainee who welds well after three months has learned quickly or simply arrived with more skill than expected, and it cannot plan the practice each trainee needs.

Training is not a single event

Campbell was clear that welding could not be learned by watching others work. It was learned by actual practice at the torch, with thought and determination. That is the central fact of welder training, and it shapes everything in this book. A course of lectures, a demonstration, or a week beside an experienced welder starts the process; skill comes from hundreds of hours of practice with feedback, spread over months. A training plan is therefore a plan for practice: how much, of what, in what order, with what feedback and what tests along the way.

Defining “ready”

A trainee is ready for production work when they can meet the target reliably, not once. The workshop manager defines “ready” for each trainee by three tests: a reject rate on practice coupons below a set level over the last 20 coupons (Chapter 3), a travel speed within the procedure’s range, and a pass in the formal qualification test (Chapter 7). Until all three are met, the trainee welds production work only under supervision, with every weld inspected. The definition matters because without it, “ready” becomes a matter of opinion, and opinions differ most when the work is short of welders. Writing the definition down before training starts, and showing it to the trainees, also tells them exactly what they are working towards.

Choosing trainees

Who to train matters as much as how. Welding needs steady hands, good eyesight with correction where needed, patience, attention to detail and the willingness to practise the same movement hundreds of times. It also needs someone who will stay in the work long enough to repay the training. Previous experience helps in some ways and hinders in others: a person who has tacked and ground in a fabrication shop knows the work and its hazards, while a person who has welded light-gauge material in another trade may need to unlearn habits before learning the shop’s methods. The fabricator chose its four trainees for different reasons, the apprentices for the long term, Chris for his knowledge of the shop, and Dee for her experience and maturity, and expected each to learn at a different rate. Chapter 3 shows how to measure those rates rather than guess them.

Keeping the people you train

The risk most often raised against training is that trained welders leave for better-paid work elsewhere, taking the investment with them. The risk is real, but it is not an argument against training so much as an argument for training well and making the business worth staying in. A welder who has been trained thoroughly, given a clear path to qualification and a trade, paid fairly as their skill grows, and given interesting work has fewer reasons to leave. The numbers in Chapter 10 include a simple allowance for trainees who leave, so that the decision to train is made with that risk counted rather than ignored.

The informal alternative

Many fabricators train informally: the new person is put beside an experienced welder, picks up what they can, and is gradually given more work. Informal training has real strengths, because it is practical and immediate. Its weaknesses are that it varies with whoever gives it, passes on bad habits as readily as good ones, has no defined end point, and produces no record. Campbell complained in 1919 that the new welding classes each taught in an entirely different manner from the others, and argued for a standard method. The same argument applies inside a business: four trainees taught four ways by four different welders will reach four different standards.

A failure: the habit that was passed on

This is an illustration. Some years earlier, the fabricator had trained a welder informally, beside a long-serving welder who made vertical fillet welds downhill, travelling from top to bottom. The long-serving welder’s downhill welds on thin material had served well for years. The new welder learned the same technique and used it on 12 mm stair stringers, where the qualified procedure called for vertical-up welding.

A visual inspection and a break test of a sample showed lack of fusion at the root of several welds: downhill welding on the thicker plate had run the weld metal ahead of the arc, over metal that had not been melted. Twelve metres of weld were ground out and rewelded, at a cost of about 12 × 60 = A$720, plus a day’s delay. The new welder was retrained on vertical-up welding, and the manager concluded that informal training would no longer be the only training in the shop.

Exercise: what a repair point is worth

The fabricator expects its four trainees, once welding production work, to weld about 1,500 m of the shop’s 6,000 m a year between them. If their repair rate is 6% in their first production year and the experienced welders’ rate is 2.5%, estimate the extra repair cost the trainees add, at A$60 a metre. Then estimate what it would be worth to bring their rate down to 4% with extra supervised practice.

At 6%, the trainees’ 1,500 m would need about 1,500 × 0.06 = 90 m of repair; at the experienced welders’ 2.5%, about 37.5 m. The extra 52.5 m costs about 52.5 × 60 ≈ A$3,150 a year.

At 4%, the trainees would need about 60 m of repair, 30 m fewer than at 6%, saving about 30 × 60 = A$1,800 a year. If the extra supervised practice to reach that rate costs less than about A$1,800, it pays back within a year on repair costs alone, before counting the lower risk of a defect reaching a customer. Chapter 5 shows how to cost that practice.