Skip to content
Arc & Bead

Beginner welding mistakes, and how to fix them

Most bad first welds come from the same ten causes. Two of them look exactly like a faulty machine, which is why people replace welders that were fine.

Ryder M.Published Updated How we pick →
A person welding a steel joint in a bright workshop

The ten, in order of how often they cause problems

  1. Welding on dirty metal. Rust, paint, oil and mill scale cause porosity. Grind to bright steel — not wire-brushed, ground.
  2. Bad fit-up. A gap you weld across becomes a hole. Cut accurately, clamp everything, and remember a joint that moves produces a weld that is wrong before you start.
  3. Wrong polarity. Flux-core runs DCEN, solid-wire MIG runs DCEP. Getting it backwards produces heavy spatter and no penetration — and looks exactly like a faulty machine.
  4. A helmet you cannot see through. You cannot steer a puddle you cannot see. This is the most underrated cause of bad welds there is.
  5. Inconsistent stickout. The wire between tip and work changes the current reaching the arc, so drifting mid-weld changes your heat without you touching a dial.
  6. Pushing when you should drag. If there is slag, you drag. Pushing flux-core drives slag into the joint ahead of the arc.
  7. Work clamp on the bench, not the work. A poor return path causes unstable arcs and arcing at unexpected places.
  8. Guessing at settings. Your machine has a chart inside the wire door and your wire manufacturer publishes procedure tables. Start from those.
  9. Changing two things at once. Change one variable, run a bead, look. Two changes teach you nothing.
  10. Welding at the machine's limit. A 120-volt machine at maximum on quarter-inch plate is working at its ceiling. Bevel the joint, or accept multiple passes.

The two that look like machine faults

Wrong polarity is the big one. Self-shielded flux-core runs electrode negative; solid-wire gas MIG runs electrode positive. On most gas-capable machines you swap two leads inside the wire compartment.

The symptoms are dramatic — heavy spatter, spitting unstable arc, a bead sitting on top without fusing — and they arrive the moment you change process. If everything suddenly went wrong after a change, check polarity before anything else.

A worn contact tip is the other. It causes a wandering arc and inconsistent feed, it is the cheapest part on the machine, and it is the most commonly overlooked cause of a welder that "suddenly stopped working properly".

Why preparation matters more than settings

Beginners tend to treat cleaning as optional and settings as the answer. It is the other way round.

Contaminated metal produces porosity — small holes through the bead that weaken it — and no setting fixes it. With gas MIG in particular, porosity looks like a shielding problem, so people chase the gas flow for an hour instead of spending thirty seconds with a flap disc.

Grind to bright metal. It takes seconds and it eliminates an entire category of problem.

Why fit-up matters more than technique

A gap is the enemy, and the thinner the material the more true that becomes.

On thick plate a small gap fills with filler metal. On sheet, a gap is a hole you are trying to weld across, and welding across a gap is the main cause of burn-through — MIG welding sheet metal covers the techniques.

Cut accurately, fit tightly, clamp firmly. Preparation taking longer than welding is the correct ratio, not a sign you are being slow.

The learning loop that actually works

Set from the chart. Run a bead on scrap of the same thickness. Chip, brush, look at it. Change one thing. Run another bead. Compare.

That loop is the whole of learning to weld, and it is why machines with continuous controls matter so much more than headline amperage — you cannot run a feedback loop with four fixed settings.

It is also why you should practice on flat plate with no joint at all before attempting a joint. A joint adds fit-up, position and heat-sink variables to a problem you have not solved yet.

Three mistakes that are about safety, not quality

Welding in a closed garage. Fume accumulates with nowhere to go, and the symptoms often appear hours later. Open two things so air can cross-flow, and put a fan where it pulls air away from your face — flux-core welding safety.

Welding galvanized steel without grinding it back. Zinc oxide fume causes a flu-like illness that starts hours later. Grind the coating off the weld area, ventilate, wear a respirator.

Letting someone watch without protection. The arc does not distinguish between the welder and a bystander, and arc eye arrives six to twelve hours later. A cheap passive hood in the corner solves it.

The mistake that is not on the list

Buying the wrong machine. It happens, and it is genuinely expensive — but it happens far less often than the ten above, and most people who conclude their machine is wrong have actually hit one of these.

Before replacing a welder, check polarity, contact tip, metal cleanliness and circuit. In that order.

What we would buy

1. Fixes mistakes 1 and 2

DeWalt DWE402 4-1/2 in angle grinder

Preparation and fit-up are the two biggest causes, and both are grinder work.

The two most common causes of a bad weld are welding on dirty metal and welding across a gap, and a grinder addresses both: it takes material to bright steel and it cuts an accurate edge.

It also bevels, which is how a 120-volt machine reaches thicknesses it otherwise cannot.

What works

  • Every weld needs prep and cleanup, so this earns its place from your first day regardless of process
  • Cut-off, grinding and wire-wheel discs all fit the same tool
  • No air, no gas, no dedicated circuit

What does not

  • Slow and physical on anything thicker than about 1/4 in
  • Cannot follow a curve accurately — that is the thing a plasma cutter buys you
  • Makes far more mess and noise than plasma cutting

2. Fixes mistake 4

Antra AH6-660

Four sensors and a shade-4 light state — you cannot steer a puddle you cannot see.

A large share of beginners whose welds look bad are welding nearly blind and correcting by feel.

Antra publishes four arc sensors, a 3.86 by 2.50 inch view, a 4/5 to 9 and 9 to 13 shade range and ANSI Z87.1 compliance. The shade-4 light state lets you position the gun accurately before striking, which is where most welds go wrong.

What works

  • Four sensors and a shade-4 light state at a price where two sensors is the norm
  • ANSI Z87.1 rated, which is the line between eye protection and a novelty
  • Cheap enough that starting here and upgrading later costs less than agonizing over it

What does not

  • Headgear is the weak point — it is the part owners most often report replacing
  • Optical class is not published as 1/1/1/1, and over a long session you can feel the difference
  • Switching speed is adequate for MIG and stick but is not a low-amperage TIG helmet

3. Fixes mistake 8

Lincoln Electric Innershield NR-211-MP flux-cored wire

Published procedure tables mean a documented starting point, not guesswork.

Trying to find settings by trial and error is slow and dispiriting. Lincoln publishes full procedure tables for this wire, so there is a stated starting voltage and wire feed speed for every thickness.

AWS E71T-11, all-position, 0.030 and 0.035 inch, run DCEN.

What works

  • Lincoln publishes full procedure tables for it, so there is a documented starting setting for every thickness
  • All-position rated — it will run a vertical fillet, not just a flat bead
  • Available in 1 lb, 5 lb and 10 lb spools, so a small machine does not need a big spool

What does not

  • Noticeably more expensive per pound than unbranded E71T-GS wire
  • Like every self-shielded wire it produces slag and more spatter than a gas process

Questions people actually ask

Why does my weld sit on top of the metal without fusing?

Usually not enough heat, or the wrong polarity. Check polarity first — flux-core runs DCEN and solid-wire MIG runs DCEP, and getting it backwards produces exactly this along with heavy spatter.

If polarity is correct, increase wire feed speed to raise the current, and make sure the metal is ground clean rather than brushed.

Why is my weld full of small holes?

Porosity, and the cause is almost always contamination or lost shielding. Rust, paint, oil and mill scale all produce it, as does a draft blowing the gas shield away on MIG.

Grind to bright metal first. Do not increase gas flow — excessive flow becomes turbulent and draws air in, making it worse.

How do I stop burning through?

Reduce heat into the part. Lower the wire feed speed, travel faster, switch to a finer wire, and stitch weld in short bursts rather than running a continuous bead.

On thin material, tight fit-up matters more than any setting, because welding across a gap is the main cause.

My machine welded fine yesterday and badly today. What changed?

Check the contact tip first — it wears, and a worn tip produces a wandering arc and inconsistent feed. It is the cheapest part on the machine and the most commonly missed cause.

Then check the work clamp is attached directly to the workpiece, and that your wire has not absorbed moisture.

How long should it take to learn to weld?

Most people produce an acceptable flux-core or MIG bead within an afternoon and a consistent one within a few sessions. TIG takes considerably longer.

The people who struggle longest are usually changing several variables at once, which makes it impossible to learn what any of them does.

Sources

Read next

Back to Find Your Level, or start over at Find Your Level.

Ryder M.

Ryder M. is the hands-on owner behind Magoo Media Group. He writes Arc & Bead's buying guides and picks every machine on the site. What every recommendation is based on — and what it is not based on — is set out on How We Pick.