Bad Weld Job: How to Spot, Fix, and Prevent Bad Welds Like a Pro

A contractor picked up a used dump trailer for $4,200 because the price was right and the frame ‘looked solid’ — until a hairline crack opened at a crossmember weld and rust started bleeding from a joint that had clearly been welded over, not repaired. That single bad weld job turned a bargain into a tow-away hazard, and it is exactly the kind of failure this guide teaches you to catch before money, time, or safety is on the line.

The Story Behind Every Bad Weld Job (and Why This Guide Matters)

A hobbyist spent a Saturday afternoon welding a new jack mount onto a garden trailer. The beads were wide, evenly rippled, and shiny enough to photograph. He ground one spot smooth, wiped off the soot, and called it done. Two weeks later a crack traced the toe of that pretty weld, and the mount pulled away from the frame with a lazy pop under load. Nothing about the setup screamed danger from five feet away. The weld simply never bonded with the base metal the way the joint required — and nobody, including the welder, knew how to check for that.

That gap between “looks finished” and “actually sound” is where bad weld jobs live. It is also why people type the phrase bad weld job into a search bar in the first place.

Who This Guide Is Written For

After years of reading questions and shop complaints around this topic, the audience almost always sorts into three groups:

• The customer with a bad feeling. You paid a shop or a handyman for repairs, and the result looks off — or it already cracked. You need to know whether you overpaid for sloppy work and what to demand next.
• The DIYer or apprentice checking their own bead. You suspect your weld is wrong but you do not have a clear standard to hold it against. You want honest, non-judgmental guidance on what is acceptable and what is not.
• The shop owner fighting rework. Every rejected weld is scrap metal, burned consumables, lost labor, and a delivery date slipping. You need root causes and a prevention system, not another lecture about trying harder.

All three are really asking the same question: what makes a weld bad, and how do you know?

What “Bad Weld Job” Actually Means in Trade Language

In the trades, quality is not a matter of taste. Welding quality is governed by written acceptance criteria published in standards such as ISO 5817 (quality levels for fusion-welded joints), AWS D1.1 (the structural welding code used across most American steel construction), and ASME Section IX (the qualification standard for welding procedure and welder performance used on boilers and pressure work).

The vocabulary matters too. The American Welding Society distinguishes a discontinuity — any interruption in the uniform structure of a weld — from a defect, which is a discontinuity large enough to make the weld unserviceable under the applicable code. In other words, a tiny porosity pit may be perfectly legal on one job and an automatic reject on another. The code decides, not the welder’s eye. AWS publishes a practical guide to exactly this kind of evaluation in AWS B1.11, Guide for Visual Inspection of Welds, and acceptance categories are laid out plainly in resources like PCC’s Weld Quality Standards reference.

This is not academic hair-splitting. AWS inspection literature is blunt about the stakes: substandard quality and unreliable welds cost lives, and it points to the 2010 San Bruno, California gas pipeline explosion as the kind of consequence a single failed joint can produce. Whether the job is a trailer frame, a stair rail, or a gas line, the same principle holds: someone defined “acceptable” before the arc was struck.

What This Guide Will Walk You Through

This article moves from recognition to prevention in ten practical steps:

1. Define what counts as a bad weld job and when a flaw becomes rejectable.
2. Catalogue the defects you actually see, with causes and danger levels.
3. Trace those defects back to root causes — parameters, prep, procedure, and people.
4. Run a hands-on visual inspection routine you can use today.
5. Understand the internal flaws visual inspection cannot catch, and the NDT methods that find them.
6. Price the damage in reject rates, rework, warranty exposure, and risk.
7. Repair the right way — and avoid the classic “fixes” that make things worse.
8. Build a prevention system around acceptance criteria, written WPS documents, and qualified welders.
9. Close with a buyer’s checklist for comparing shops and quotes.

Expect specific numbers, code references, and step-by-step checks rather than vague advice — starting with the visual red flags, such as undercut along a weld toe, that separate an acceptable bead from one headed for the grinder. Along the way we will lean on inspection training literature from the American Welding Society, because recognizing a bad weld is a trainable skill, not a matter of intuition. If you are still setting up your machine and guessing at parameters, our MIG welder setup guide is the place to start before you read another line of defect theory.

What Exactly Counts as a Bad Weld Job?

Before you can judge any weld, you need the same vocabulary an inspector uses. Most arguments about whether a weld is “bad” come down to two people using the same word to mean different things. So let’s draw the lines clearly.

Discontinuity vs. Defect vs. Rejectable Defect

A discontinuity is simply any interruption in the normal structure or properties of a weld. A defect is a discontinuity that fails to meet the applicable acceptance criteria. And a rejectable defect is one the governing code or specification says must be repaired or the part rejected outright.

This distinction matters more than any bead picture on the internet. As Kobelco’s weld imperfections handbook puts it, a surface irregularity or discontinuity is not a rejectable defect unless it exceeds the permissible range of the governing specification. In other words, no weld is “perfect.” Every weld has some interruption somewhere. What makes it bad is whether it crosses the limit written in the standard that applies to your job. You can review the Kobelco handbook yourself to see how they classify imperfections by type and severity.

Surface Irregularity vs. Internal Discontinuity

The second split is location. Surface irregularities include undercut, overlap, excessive reinforcement, spatter, arc strikes, underfill, and visible porosity. You can see them with your eyes, a flashlight, and a mirror. Internal discontinuities include lack of fusion, incomplete joint penetration, slag inclusions, and subsurface porosity. These hide inside the weld and root, and no amount of staring at the cap will reveal them.

That is the trap that catches DIYers and customers alike. A cap can look like stacked dimes and still hide a cold root with zero penetration. If you are still building your visual skills, our guide to undercut causes and acceptance limits walks through one of the most common surface defects in detail.

Workmanship Acceptance vs. Code Acceptance

Workmanship acceptance asks: does this look like a competent welder produced it? Code acceptance asks: does it meet the numeric limits in the governing standard, such as ISO 5817, AWS D1.1, or ASME IX? These are not the same question. A weld can look rough and still pass. A weld can look beautiful and still fail.

ISO 5817, for example, grades welded joints at three quality levels (B, C, and D), with different allowable limits at each. ASME Section IX is different again: it focuses on whether the welding procedure and the welder are qualified for the job, not on cosmetic appearance. AWS D1.1 is the common reference for structural steelwork in the United States.

How Acceptance Criteria Actually Work

Inspection standards sort discontinuities into rough severity bands:

• Not acceptable at any size: Cracks. All major codes, including ISO 5817, AWS D1.1 and ASME IX, classify cracks of any type as unacceptable and require immediate repair. Hot cracks, cold cracks, crater cracks, toe cracks, and underbead cracks are automatic rejects, full stop.
• Acceptable within limits: Things like slight undercut, minor porosity, and small underfill. Under structural visual acceptance criteria, maximum undercut commonly allowed is about 1/32 in. (roughly 0.8 mm) in depth. Beyond that limit, it is a rejectable defect.
• Usually acceptable: Porosity is a cavity-type discontinuity caused by gas trapped during solidification. Because of its spherical shape it is considered the least detrimental common discontinuity. But “least detrimental” is not “free.” Visual porosity is still generally unacceptable under most codes.

For a deeper look at how inspectors apply these rules, the AWS B1.11M/B1.11 guide to visual weld inspection is the standard reference, and Michigan Technological University publishes a useful webinar on discontinuities versus defects that explains the severity bands in plain language.

Why Home-Shop “Looks Good to Me” Standards Fail

This is exactly how welded-over trailer frames and re-welded bucket arms end up in service. If you are fabricating for anything structural, or if you are hiring someone who is, acceptance criteria must be agreed on before the arc is struck, not after the invoice arrives. When the work involves machinery that needs reliable wire feeding and settings from the start, even something as basic as setting up a MIG welder correctly is part of preventing the defects you will later have to argue about.

The short version: a bad weld job is not a weld with a flaw. It is a weld whose flaws exceed the limits that were supposed to apply to it, or a weld where nobody ever defined those limits at all.

The Most Common Bad Welds: A Defect Field Guide

Before you can fix a bad weld job, you have to name it. Calling something “ugly” tells the next person nothing; calling it “undercut at the toe of a fillet weld” says where to look, what caused it, and whether it can be accepted. Match what you see on the metal to the entries below and you already know half the repair.

One framing point: in trade language, most of these are discontinuities, not automatic failures. A discontinuity becomes a defect only when it exceeds the acceptance criteria of the governing code. Cracks are the exception, and the American Welding Society treats them as the most serious class of weld defect, period.

Surface Discontinuities You Can See

Porosity. Look for pinholes, wormholes, or a cratered, sponge-like surface. Gas gets trapped in the weld pool as it solidifies, usually from inadequate shielding gas, wind, oil, rust, or moisture. Surface porosity can sometimes be ground out and rewelded; clustered or internal porosity usually means the whole section comes out.

Undercut. A groove melted into the base metal at the weld toe or root and left unfilled, showing as a sharp notch along the bead edge. It matters because it concentrates stress and gives cracking a starting point. Excessive amperage, too long an arc, or whipping the electrode sideways cuts it in. Our deep dive on welding defects: undercut causes, acceptance limits, and repair covers the fix step by step.

Overlap (cold lap, roll over). The mirror image of undercut: weld metal spills past the weld toe without fusing into the plate. Most codes reject it because the sharp end condition it leaves behaves like a crack starter. Low heat input and travel speed too fast for the deposition rate are the usual culprits.

Underfill. The weld sits below the required reinforcement level, leaving the joint thin. It looks innocent and quietly removes load-bearing metal, usually from too small a weld, too large a joint, or a gap that was never budgeted for.

Excessive reinforcement. The opposite problem: a proud, humped bead. Codes cap reinforcement height because a steep weld-to-base transition creates a notch on the far side of the bead. Overbuilding a joint rarely adds strength and often adds a stress riser.

Burn-through and melt-through. A hole burned clean through the plate, common in stick welding thin metal where heat management is unforgiving. Either too much current, or a travel speed that stalled over the puddle.

Arc strikes and spatter. Stray arc marks outside the weld, each one a tiny, rapidly quenched, hardened spot with microcrack risk, so codes require them ground smooth and blended. Spatter is mostly cosmetic but forces extra grinding and can hide real surface flaws.

Distortion. Not metallurgical, but real: uneven shrinkage pulls frames, plates, and tubular assemblies out of alignment. If the finished assembly no longer fits the jig, the weld job failed whether or not the beads are pretty.

Fusion Discontinuities

Lack of fusion. The weld metal never bonded to the base metal or a previous pass. It is frequently invisible from the surface, which is exactly what makes it dangerous. Insufficient heat, wrong electrode angle, or a weave that never wetted the sidewall are the usual causes.

Incomplete joint penetration (IJP). The weld did not reach the full depth of the joint; in a groove weld you may see the root never bridge. Poor joint design, a gap that is too tight, or conservatively low amperage all play a part.

Slag inclusion. Non-metallic flux trapped inside the weld, common in flux core and multi-pass stick work where interpass cleaning was skipped. It reduces the effective cross section and can hide unfused areas behind it.

Cracks: The Automatic Reject

Cracks are the most serious defect in welding, and every major code treats them as unacceptable regardless of size or location. You will meet them in several flavors:

• Hot cracks form while the metal is still solidifying, often in restrained joints or sensitive chemistries.
• Cold cracks in steel develop after solidification, below roughly 200 °C, driven by diffusible hydrogen, a brittle microstructure, and restraint stress. That is why hydrogen-controlled consumables like E7018 matter and why preheat is not optional on thick, restrained joints, and why front-end settings count whether you are running 3/32 E7018 or 5/32 E7018.
• Crater cracks star out at the stop of a bead when the arc is pulled away abruptly and the crater solidifies unconfined.
• Toe cracks start at the weld toe, where undercut and stress concentration meet.
• Underbead and transverse cracks run beneath the weld and across it, and they are almost always internal, so visual inspection alone will not catch them.

Every entry shares a pattern: the visible symptom sits on the surface, but the real cause sits in joint prep, parameters, consumables, or procedure. That is the thread the next section picks up.

Why Bad Welds Happen: The Real Root Causes

Ask a shop full of welders what caused a rejected bead and the usual answer is that someone’s hand was off that day. It is the most satisfying explanation and often the least accurate. Porosity, undercut, cold lap and lack of fusion are rarely random. They are visible symptoms of a decision made minutes or weeks before the arc was struck: a dirty plate, a wrong gas bottle, parameters guessed from memory, a joint no torch could actually reach.

Industry analysis backs this up. The Fabricator’s breakdown of weld defect root causes narrows nearly every shop’s reject pile down to two buckets: poor instruction and workmanship on one side, and poor weld design or material selection on the other. That framing tells you where to look. Half the battle is training and process discipline. The other half is decided by the engineer, the purchasing department and the person who cut the fit-up long before anyone turned on a machine.

Contamination and Sloppy Joint Prep

The most common upstream failure is a joint that was never properly cleaned. Mill scale, rust, oil, paint, cutting-fluid residue and moisture all break down in the arc and dump gas and oxides into the puddle. Kobelco’s technical defect guide lists rust, oil, paint and moisture on the joint faces, moisture in coatings or flux, insufficient shielding gas, strong wind, and excessive amperage or arc length as the main drivers of porosity. Most of those are preparation problems, not skill problems. A grinder, a wire brush and five minutes of attention cost far less than a cut-out and re-weld.

Parameter Guesswork: “What I Always Use”

Wrong amperage, voltage, travel speed and contact-tip-to-work distance will defeat good technique every time. An AWS Inspection Trends feature on weld quality describes shops where welders set their machine from habit, running “what I always use” instead of the qualified ranges written into the welding procedure specification. Too much heat on thin material produces burn-through and undercut, too little produces lack of fusion, and a stickout that has crept an inch past spec quietly changes the entire arc. Getting the numbers right is learnable, which is why setup guides like How to Set Up a MIG Welder: Complete Step-by-Step Guide and 3/32 E7018 Amperage: Settings, Polarity & How to Run It matter more than most welders admit.

Shielding Gas: Turbulence, Starvation and Wind

Gas is the invisible variable that wrecks welds fastest. Flow too high pulls air in through turbulence; flow too low lets the puddle go hungry and porosity follows. Practical guidance looks like this:

Drafts matter just as much as dial settings. A fan, an open bay door or a breeze can strip shielding gas off a puddle in seconds, and open-root and vertical-up work are especially vulnerable. Flux Core Wire Welding Tips: Pro Techniques for Cleaner Stronger Welds covers the same gas-and-draft discipline for FCAW, where operators often assume the flux makes them immune. It does not.

Moisture, Consumables and Bad Procedure Paperwork

Wet electrodes, damp flux, expired low-hydrogen rods stored in an open box and contaminated wire all feed hydrogen into the weld, which shows up later as underbead cracking or porosity. So does an out-of-date or nonexistent WPS. When there is no written procedure, welders fill the vacuum with memory and guesswork, and every shift inherits a slightly different process. A quick look at Pipe Welding Procedures: The Complete Guide to WPS, Positions, Passes & Inspection shows how much a real procedure settles before anyone strikes an arc.

Design, Access and Overwelding

Some defects are designed in. A joint that cannot be reached with a torch, a gap too wide to bridge cleanly, a fillet specified thicker than the plate can absorb, or a detail that traps slag will defeat a capable welder. Overwelding is the sneaky version: adding a fat bead to “be sure” increases heat input, distorts the assembly, amplifies residual stress and sometimes makes the joint weaker, not stronger. Corrosion-prone and pressure details make this worse, as the rules in Gas Line Welding: Codes, Safety, and Purging Rules You Must Follow make clear.

Human Factors: Training Gaps and Lost Focus

Finally, people. Welders who were never taught how to read a drawing, interpret a spec or self-inspect their own work will repeat the same defect indefinitely. Fatigue, distractions and personal problems behind the hood do the rest, and a moment of inattention at a start-stop or crater is exactly where cracks begin. AWS has argued directly that welders need training in inspection, because a welder who cannot spot a defect cannot prevent one. Prevention is a system of clean joints, written parameters, controlled environment and aware people, not a lucky pair of hands.

How to Spot a Bad Weld Job: A Practical Visual Inspection Guide

You do not need a lab or a $12,000 ultrasonic rig to catch most bad welds. Visual examination is still the single most-used inspection method in the industry, and when it is done properly it detects the overwhelming majority of rejectable work before anything else gets touched. The catch is that “looking at it” and “inspecting it” are two different things. Here is the routine professional visual inspectors actually use, adapted so you can run it in a garage, on a job site, or on something you just bought.

The 7-Step Visual Inspection Routine

1. Clean the weld first. Wire-brush away spatter, slag, smoke film, and grinding dust. You cannot judge a toe line through a crust of flux. If the joint has been painted or undercoated, do not chip at it blindly; coatings can hide a crack completely.
2. Get the light right. Aimed light beats ambient light. Sweep a flashlight or work lamp across the weld from several directions so shadows expose ripples, craters, and undercut. AWS-referenced practice calls for roughly 1000 lux at the weld surface for final visual inspection — bright, even, and free of glare.
3. Get close and get low. A welder’s visual inspection handbook instructs that the inspector’s eyes be within 24 inches of the surface and at an angle of at least 30 degrees to it. Standing back and glancing down at a weld from three feet will hide underfill, lack of fusion at the toe, and incomplete crater fill.
4. Inspect the correct zone. The examination area is not just the shiny bead. AWS guidance puts the zone at the completed weld face plus 1/2 in. on both sides of the weld, for its entire length. That half-inch margin is where cracking, arc strikes, and undercut show up.
5. Gauge the smallest-looking section. Do not measure the prettiest inch. Find the narrowest, thinnest, or most washed-out area and measure there, since a chain fails at its weakest link. A cheap fillet gauge or a weld comparator tells you more than your eye alone.
6. Check the trouble spots deliberately. Spend most of your time on the toes, the crater at every stop, every start-and-restart, and the root where two members actually join. On tubular or closed sections, get a small mirror behind the joint to confirm the root actually fused rather than just capping over.
7. Compare and document. Symmetric joints should look symmetric. If the left gusset has a smooth, even bead and the right one has a lumpy, wandering one, that difference is information. Photograph every finding with a scale reference before anyone starts grinding.

Top Visual Red Flags

• Hairline cracks at the toe, in the crater, or across the bead — automatic reject, always.
• Undercut — a notch cut into the base metal beside the weld that acts as a stress raiser. See Welding Defects Undercut: Causes, Acceptance Limits, and How to Fix It for limits and fixes.
• Rust bleeding from a joint, uneven beads, visible gaps, leftover slag, misaligned parts, or components that move when they should not — all classic warning signs of failing welds on trucks and trailers.
• Porosity clusters, overlap/cold lap, underfill, excessive globs of reinforcement, and arc strikes outside the weld zone.

Remember that appearance cuts both ways: a glassy, perfectly rippled bead can still hide lack of penetration. The visual check tells you what is on the surface; anything load-critical and hidden needs NDT, which the next section covers.

Run this routine on every joint, every time, and photograph what you find. Ten focused minutes of visual examination is far cheaper than a structural failure, a warranty claim, or the argument you will have when someone insists it “looked fine when it left the shop.”

Bad Welds You Cannot See: Internal Defects and NDT Explained

Visual inspection is always the first step, and it catches plenty: undercut, overlap, craters, and surface porosity. But it has a physical limit. A weld cap hides what lies beneath. Three of the most dangerous discontinuities in welding—lack of fusion, incomplete joint penetration, and internal porosity—live where no flashlight, magnifier, or trained eye can reach. That is why AWS B1.11, the guide for visual inspection of welds, treats inspection as a surface examination, and why codes pair it with something more.

That something is nondestructive testing (NDT): methods that look inside a finished weld without cutting it apart. Here is what each one actually does in shop terms.

What Each NDT Method Finds Best

• Penetrant Testing (PT): A colored or fluorescent liquid is sprayed on a clean surface, wiped off, and then developed. Dye trapped in a crack bleeds back out and lights up. PT finds surface-breaking defects only—cracks, pinholes, and hairline lack of fusion at the toe. It works on any nonporous material, costs almost nothing, needs no power, and sees nothing below the surface.
• Magnetic Particle Testing (MT): The part is magnetized and fine iron particles are dusted on. Where a crack interrupts the magnetic field, particles pile into a visible line. MT is fast and field-friendly, excellent for toe cracks and underbead cracking. But as AWS notes, it is limited to ferromagnetic materials, so it won’t work on aluminum, austenitic stainless, or other non-ferrous alloys.
• Ultrasonic Testing (UT): A transducer sends high-frequency sound into the weld and reads the echoes. Discontinuities that reflect the beam appear as signals on a screen. UT is one of the primary ways to catch lack of fusion and incomplete joint penetration—the very defects a visual check misses. It needs couplant gel, trained interpretation, and has a small near-surface dead zone.
• Phased Array UT (PAUT): Instead of a single fixed beam, a phased array probe steers dozens of beams electronically and scans a volume in a single pass. The big advantages are a permanent digital record and fewer false calls than conventional UT. The tradeoff is higher equipment cost and specialized training.
• Radiography (RT): X-ray or gamma rays project a literal image of the weld’s interior, much like a medical film. RT is the classic answer for porosity, slag inclusions, and lack of fusion. But it is directional: the source must be aligned through the defect, and beam direction determines what you can and cannot see.
• Eddy Current Testing (ET): A coil induces circulating currents in the metal; flaws disturb them. ET is fast, automated, and excellent for surface and near-surface checks on conductive material, especially tubing. Depth is limited, and the material must be electrically conductive.

Why Shops Are Moving From RT to PAUT

Radiography requires strong radiation sources and specially licensed operators. That means cordoning exclusion zones, clearing adjacent work, notifying neighbors, and pausing production—real disruptions on a busy floor. Those constraints are why many fabricators are exploring alternatives; AWS Inspection Trends notes growing interest in phased array UT as a safer, more flexible option that produces a recordable digital result without exclusion zones.

The methods do not actually disagree much on quality. A Florida DOT study of bridge welds reported remarkably similar rejection rates across methods: 8.7% for phased array UT (8 of 92), 7.4% for conventional UT (4 of 54), and 9.3% for radiography (10 of 108). The choice comes down to geometry, access, material, cost, and safety—not which is stricter.

When a Visual Check Is Enough—and When NDT Is Mandatory

For non-code, light-duty work—a decorative bracket, garden cart, or non-structural part—a careful visual inspection with sensible acceptance limits is usually proportionate. The moment a weld is code-controlled, visual alone is not enough. NDT is typically required or strongly expected when the work involves:

• Structural and bridge work governed by AWS D1.1 or DOT specifications.
• Pressure-retaining welds under ASME or ISO 5817 acceptance levels; our Pipe Welding Procedures: The Complete Guide to WPS, Positions, Passes & Inspection shows how NDT scope is built into a procedure.
• Gas and hazardous service lines where a leak is a public safety event—see Gas Line Welding: Codes, Safety, and Purging Rules You Must Follow.
• Lifting, overhead, and load-bearing equipment, including trailer frames, cranes, hoists, and vehicle chassis.
• Rail, pressure vessel, and boiler work, including specialized joining such as Railroad Thermite Welding: How Aluminothermic Rail Joining Works.

One final habit worth building: agree on the NDT method, acceptance criteria, and extent of examination before the first arc strike. Chosen after the fact, NDT becomes an argument. Chosen up front, it becomes a standard.

The Real Cost of a Bad Weld Job: Reject Rates, Rework and Risk

Rejection is not an abstract concept on a shop floor. Every scrapped part, every gouged-out groove and every angry phone call has a number attached to it, and those numbers add up faster than most owners expect. Understanding what a bad weld job actually costs, in dollars, downtime and danger, is the fastest way to justify spending an extra ten minutes on joint prep and settings before the arc is struck.

What a “Normal” Reject Rate Actually Looks Like

Start with the benchmarks, because you cannot improve what you refuse to measure. Among industry practitioners, a weld rejection rate of roughly 1–2% is considered expected for a competent operation, while a 5% reject rate is generally considered too high. Many clients go further and write 2% or below into their purchase orders as a condition of doing business, according to discussion in AWS Inspection Trends.

Those numbers sound abstract until you see the outlier. An AWS Inspection Trends question-and-answer described a fabricator running flux cored arc welding who was rejecting around 20% of fillet welds and 10% of groove welds. One in five fillet welds went to the rework pile. At that rate, a welder spends more time fixing than producing, the schedule slips, and the profitability of the entire job evaporates. It is a pattern that shows up again and again in FCAW work, which is why parameter discipline and flux core wire welding tips matter so much in production environments.

The Per-Welder Price Tag of Rework

An American Welding Society shipbuilding study put hard numbers on the opportunity: an estimated $3,244 per welder per year could be saved by reducing rework, scrap and rejects, forming part of a larger $17,044 total per-welder opportunity when productivity losses are included. Multiply that across a crew of twenty welders and you are looking at a six-figure annual sum that never reaches the bank account.

The visible costs are only the tip. Poor weld quality drives what accountants call non-value-added costs: rework labor, warranty repairs, replacements, delay costs, liquidated damages, penalties, and lost business from a tarnished reputation, as itemized in AWS Inspection Trends editorial coverage on welder inspection training. A delayed shipment can trigger contractual penalties that dwarf the value of the part itself, and none of that shows up on a welding rod invoice.

Warranty, Liability and the Costs Nobody Quotes Up Front

Then comes the risk side. A cracked trailer crossmember, a fatigued equipment frame, a leaking pressure joint or a failed structural connection does not simply fail an inspection; it creates exposure. Under OSHA’s welding, cutting and brazing standard, employers are responsible for protecting workers from cuts, burns, fires and fumes, and a defective weld that fails during service can turn a paperwork issue into an injury investigation. Reputations in this trade travel by word of mouth, and one high-profile failure can erase years of good work. Proper protection and hazard control, covered in our guide to MIG welding safety precautions, is the first line of defense against exactly these outcomes.

Code-controlled work raises the stakes further. On pipe welding procedures governed by ASME or equivalent standards, a rejected weld rarely means a quick grind-out. It means documentation, NDT re-testing, possible hydrotesting, and a paper trail that follows the shop on future bids.

The Customer-Side Cost: Paying Twice

From the buyer’s seat, a bad weld job means paying for the same work twice. First there is the original invoice, then the diagnostic visit, the dismantling, the repair, the re-inspection and the lost use of the equipment. Add the dispute over who absorbs it, the second opinion from an independent inspector, and the delay while the item sits out of service. Even when the welding shop eats the cost, the customer has already lost time and confidence. That is why acceptance should be judged with gauges and code criteria, not a casual look at a pretty bead, and why defects such as undercut deserve real attention rather than a shrug.

The bottom line: a bad weld job is never just a ugly bead. It is a measurable drain on shop profitability and a genuine safety liability, which makes the inspection habits in the next section worth mastering.

How to Fix a Bad Weld Job the Right Way (and the Wrong Ways)

Finding a bad weld is the easy half. What you do in the next ten minutes decides whether the joint comes back stronger than before or comes apart again six months later, this time with a load on it. Repair is where most shops and DIYers get sloppy, because a quick bead over a crack looks fine and passes a casual glance. It is also the fastest way to turn a fixable problem into a hidden one.

The Seven-Step Repair Workflow That Actually Works

1. Stop. Halt production or service on the part. Do not tack, do not “touch it up,” and do not keep welding past a defect hoping the next pass buries it.
2. Find the full extent. Clean the area and determine how far the crack or lack of fusion actually runs, including its orientation. AWS guidance is explicit: locate the crack ends before you strike an arc, because cracks commonly run further than the visible hairline suggests.
3. Verify with NDT. Confirm you have reached the true end of the discontinuity using dye penetrant or magnetic particle testing. Grind a half-inch past the visible tip, then re-check. If you cannot confirm the end, you are not ready to weld.
4. Remove metal completely. Grind, gouge, or arc-air the defective metal out down to sound base metal. For cracks, that means full removal along the entire length — never a surface scratch followed by a cover pass.
5. Fix the root cause. Correct the parameter that caused the defect before restarting: dial in amperage with a reference like our 3/32 E7018 amperage chart, fix gas flow, clean the joint, or change position. Repair the same mistake and you simply recreate it.
6. Re-weld to a qualified procedure. Use the correct filler for the base metal, follow the WPS requirements for preheat, interpass temperature, pass sequence, and travel technique.
7. Re-inspect and restore protection. Visual first, then the same NDT method that found the flaw, then reapply coating, primer, or galvanizing where the heat affected zone was stripped.

What the Standards Say About Specific Defects

• Cracks (hot, cold, crater, toe): Always rejectable. Locate extent and orientation, grind out completely, confirm with NDT, then re-weld under the proper procedure including preheat when required. There is no acceptable crack, and surface inspection alone won’t catch every one.
• Undercut: Repair generally means light grinding to remove the notch, then additional welding to replace the base metal you just removed. Grinding alone leaves a thinner section, which is why an undercut fix that is only abrasive work is a second defect in disguise.
• Overlap or cold lap: Grind down excessive reinforcement, then add weld metal if the finished size falls below the required dimension. Avoid “washing” the weld face — dragging the arc flat across the surface smears the toe and masks lack of fusion instead of correcting it.

The Classic Wrong Fixes

• Welding over a crack. The single most dangerous repair. The crack stays live underneath, and the cover bead hides it from visual inspection.
• Just adding a bead. If the root cause was wet shielding gas, a dirty joint, or wrong stickout, the next pass inherits the same problem.
• Over-reinforcing with a giant glob. A bloated, out-of-profile weld concentrates heat, traps slag, and looks reassuring while performing worse.
• Wrong filler. Matching by color or availability rather than base metal and service conditions produces a metallurgical mismatch.
• Skipping preheat. On thicker or high-carbon steel, cold welding invites hydrogen cracking that may not appear for 24 to 48 hours. AWS quality reporting keeps circling back to procedure discipline as the fix.
• Grinding a structural crack “smooth.” Hiding a defect is not repairing it, and it is a documentation violation on code-controlled work.
• Bolting on a gusset as a fix. An incorrectly placed reinforcement plate moves the stress concentration somewhere new and creates a fresh failure point. Reinforcement must follow an approved design, never an on-the-spot guess.

The honest summary: a real repair restores the joint to its original designed strength, follows a written procedure, and passes the same inspection the first weld should have passed. Anything less is a bad weld job with a second chance at failing.

How to Prevent Bad Weld Jobs Before They Start

Every bad weld job in this guide shares one thing: someone could have stopped it before the arc was struck. Prevention is cheaper and less painful than repair, and high-performing shops do it with simple habits applied consistently.

Define Acceptance Criteria Before Anyone Strikes an Arc

The cheapest prevention step is the most overlooked. AWS guidance states clearly that welding should not be initiated before the visual acceptance criteria are clearly defined. Decide what “acceptable” means before welding starts, not after someone argues about a finished bead.

Acceptance criteria answer the questions that spark disputes: how much undercut is tolerable at the toe, whether porosity is allowed on a capped surface, the minimum fillet leg size, the acceptable reinforcement height. Whether you are working to AWS D1.1, ISO 5817, or a customer drawing, write those answers where both welder and inspector can see them.

Write and Actually Follow a Real WPS

A written procedure specification is prevention built into paper. It states base metal and filler classification, thickness range, joint design, preheat and interpass temperature, polarity, current and voltage ranges, travel speed, shielding gas and flow, and qualified positions. Follow it and the parameters are right every time, whether the operator is a 20-year veteran or a second-year apprentice running their first uphill root. For a full walkthrough, see our pipe welding procedures and WPS structure guide.

Qualify the Welder, Not Just the Paperwork

A WPS is only as good as the hands executing it. A welder performance qualification coupon test proves the welder can produce sound welds in the positions, processes, and materials they will encounter. Qualification is not gatekeeping; it is data. Shops that track first-pass yield by welder find targeted training on specific defects beats blanket retraining or swapping in a “better welder” for repairs.

Control the Consumables, Gas, and Environment

Contamination causes a huge share of field defects, almost all preventable:

• Clean and dry consumables. Store rod ovens hot and use low-hydrogen electrodes within the manufacturer’s holding time. Damp flux is a porosity generator.
• Verify shielding gas. Check the mix, cylinder date, regulator setting, and hose for leaks. Low flow starves the puddle; high flow pulls in air.
• Kill the drafts. Wind screens, enclosure, or moving indoors. A surprising amount of porosity is just airflow.
• Clean the joint. Mill scale, oil, paint, rust, and moisture at joint prep are preventable with zero technical difficulty.

Wrong machine setup is another recurring offender, and a proper pre-shift check avoids it. Our MIG welder setup guide covers the baseline dial-in most shops skip.

Mock Up, Coupon Test, and Inspect in Process

Before production runs, run a mock-up. A coupon weld on the actual material, thickness, position, and joint geometry exposes parameter and access problems while they are still cheap. Then inspect during the job: catching undercut on pass one costs seconds, while catching it at final inspection costs disassembly.

Track Defects and Chase Root Causes

Run a simple defect tracking chart: defect type, location, process, welder, cause, corrective action. A chart alone changes nothing. Industry root-cause analysis finds the same pattern: defects come from workmanship and instruction gaps combined with design or material choices, not from “careless welders.”

Assigning a better welder to redo a failed joint is an expensive short-term fix that eats margin. Ask why: wrong gas, wrong preheat, rushed joint prep, bad access? Fix the cause once and the defect stops recurring.

Pre-Weld, During-Weld, and Post-Weld Checklists

Two checklist items deserve emphasis. First, prevention includes the welder: burn and eye-injury prevention starts with proper gear, as covered in our complete welding PPE guide. Second, a quick refresher on electrode selection and storage prevents a whole family of hydrogen and porosity defects.

Prevention is a loop: define the standard, qualify the people, control the inputs, inspect in process, track defects, fix causes. Run it and bad weld jobs stop being surprises and start being statistics you can drive down.

Bad Weld Job Checklist: Buying Work, Comparing Options and Next Steps

Knowing the warning signs only helps if you act on them. Whether you are hiring a shop, buying a previously welded trailer, or deciding whether to attempt a repair yourself, a little due diligence up front can save thousands later. This final section gives you a clear comparison of your options, a pre-acceptance checklist, and the questions to ask before you greenlight any work.

DIY Garage Welds vs a Certified Professional

The urge to save money by grabbing a machine is real, and hobby work is very possible with the right resources, such as How to Set Up a MIG Welder: Complete Step-by-Step Guide, Is Stick Welding Hard? An Honest Beginner’s Guide, or the MIG Tack Weld Guide. But certification, or simply owning a machine, does not guarantee a repair is correct. Structural work demands an understanding of the load path, steel type, procedure, heat input and coating restoration.

A qualified shop works to strict parameters, the kind covered for specialty consumables in 3/32 E7018 Amperage: Settings, Polarity & How to Run It and 5/32 E7018 Amperage: Settings, Polarity & Technique, and handles projects ranging from Pipe Welding Procedures down to delicate Stick Welding Sheet Metal. Garage-style bend, break and macro-etch tests are useful for practice coupons, but they should never be used to accept roll cages, trailer hitches, vehicle frames, lifting points, pressure vessels or code-controlled work.

Cheap Quote vs a Qualified Shop

The cheapest bid almost always skips the most paperwork. Behind a low price can sit undocumented repairs, an unqualified welder, or no inspection scope at all, so compare what is actually included before you pay a deposit. Safety is also non-negotiable for whoever does the work: shops that value protection in line with Welding PPE: The Complete Guide, What Does a Welder Wear? and MIG Welding Safety Precautions are a far better investment than the shop that cuts corners.

As AWS Inspection Trends puts it, nondestructive testing is an adjunct to the visual inspection program, not a substitute for it. A flawless-looking bead means nothing without proof.

Appearance-Only Acceptance vs Gauge-and-Code Acceptance

A clean toe with zero undercut can still hide lack of penetration. That is why code-based acceptance, such as the limits set in AWS B1.11 and interpreted through PCC’s Weld Quality Standards, always outranks “looks good to me.” These standards prevent the exact hazards that plague bargain jobs, from trailers to high-risk work such as Gas Line Welding or Railroad Thermite Welding. Undocumented repairs remain unknowns regardless of how attractive the bead looks, so buyers should ask what failed, who repaired it, what procedure was followed, and request the records.

Pre-Acceptance Checklist for Buyers

Save this list before paying for any critical weld work. Check off each item and request proof where it applies:

• Documentation: A documented WPS and the qualified welder’s WPQ for the process, position and base metal.
• Scope: Stated acceptance criteria (AWS, ISO 5817 or ASME IX) for the joint in question.
• Verification: An agreed NDT scope (PT, MT, UT, RT or PAUT) with pass/fail noted. Never “as needed”.
• Repair History: Records of what was removed, preheat, filler and any gouging. Never accept “just welded over it”.
• Finish: Coating restoration (primer, paint or galvanizing) to prevent post-repair corrosion.
• Accountability: A hold point for inspection, photographs and written sign-off before return to service.

Questions to Ask and Next Steps

When hiring, get straight answers to these five questions. Ask for evidence, not opinions, and do not be afraid to get a second bid:

1. The Failure: “What failed, and what was found (visual or NDT) before repair?”
2. The Process: “Which qualified procedure (WPS/WPQ) will you use, and what are the preheat, interpass and heat input limits?”
3. The Proof: “Will inspection be performed, and can I receive the results and photos?”
4. The Safeguards: “Who is responsible for coating restoration and post-repair checks?”
5. The Risk: “If it fails inspection, will it be repaired to spec at no charge, or will replacement be offered?”

Buyers should obtain these answers before funds change hands; welders can use them as a pre-weld gate. Hobbyists stretching into new gear may find MIG TIG Combo Welders, Flux Core Wire Welding Tips, Spot Welding Copper or How Does a Spot Welder Work? useful, but those processes belong in non-critical applications only.

Final Takeaway

A bad weld job is defined not by how ugly it is, but by how unknown it is. Appearance can lie; a traceable paper trail paired with the correct inspection cannot. As ESAB’s Welding Defects Guide and The Fabricator’s breakdown of root causes reinforce, prevention always costs less than rework, liability or injury. Keep reading with our undercut breakdown, or refine your craft through precision processes like TIG Welding Sheet Metal. Ultimately, do not ask whether it looks good. Ask whether it is provably sound. That single question separates a bargain from a bad weld job.

Frequently Asked Questions

How do I know if I got a bad weld job?

Look for cracks, undercut grooves at the weld toe, cold lap or overlap, clusters of porosity, visible gaps, undersized or lumpy beads, slag left in the joint, and distortion of the parts. Any crack, or a weld on a load-bearing part that looks or feels wrong, should be professionally evaluated before the item is used again.

Which weld defects are the most dangerous?

Cracks are the most serious because they can propagate under load and cause sudden failure, and codes generally require them to be repaired immediately. Lack of fusion and incomplete joint penetration are also high-risk because they reduce strength while remaining invisible on the surface.

Is porosity always a reason to reject a weld?

Not always. Acceptance depends on the governing code and the amount and location of the porosity; minor scattered porosity may be acceptable in some applications while visual or elongated porosity in a sealed or structural joint is typically rejected. Always check the acceptance criteria written in the drawing, specification or standard.

Can a weld that looks good still be bad?

Yes. A smooth, uniform bead can still hide incomplete penetration, lack of fusion, slag inclusions or internal porosity, none of which are visible from the outside. That is why critical welds require nondestructive testing such as ultrasonic or radiographic inspection in addition to visual checks.

How much does a bad weld job cost?

Costs include removal and rewelding, grinding and filler consumption, lost arc time, inspection, delays, coating restoration and possible liability or warranty exposure. Industry surveys have estimated thousands of dollars per welder per year in avoidable rework, scrap and reject costs, so prevention is far cheaper than repair.

What is a normal weld reject rate?

Shop practices vary, but practitioners generally treat 1-2% as expected and 5% or higher as a problem that demands corrective action. Some shops have reported FCAW reject rates as high as 20% on fillet welds, which signals process and training issues rather than bad luck.

Should I just weld over a crack to fix it?

Never. Welding over a crack conceals the defect without stopping it from growing; the crack must be fully located, ground or gouged out to sound metal, confirmed clear with NDT, and then re-welded under a qualified procedure with correct preheat and filler metal.

When does a weld repair need a professional inspector?

Bring in a qualified inspector whenever the weld is structural, load-bearing, pressure-related, involved in lifting or towing, mounted on a vehicle frame or suspension, governed by a code or engineer’s requirements, or failing repeatedly after repair. If failure could injure someone or damage property, professional evaluation is not optional.

What is the single best way to avoid bad welds?

Define acceptance criteria before welding starts, work to a written welding procedure specification, keep the joint and consumables clean and dry, and inspect in-process rather than only at the end. Catching a problem during the first pass is dramatically cheaper than discovering it after the assembly is finished.