Weld It or Replace It: When the Repair Costs More
A decision framework for repair versus replacement — the parts that are always worth rebuilding, the ones that never are, and the safety-critical categories where the answer is not yours to make.
The short version
- Wear parts are almost always worth rebuilding. They fail by losing material, and putting material back is cheap.
- Fatigue cracks are the opposite. A fatigue crack means the design or the loading is wrong, and welding it up restarts the same clock.
- Count preparation and finishing, not arc time. Access, cleaning, machining, and coating usually dwarf the welding itself.
- Some categories are not a judgement call — lifting equipment, structural members, and pressure vessels have rules that override economics.
- Repair beats replacement most decisively when the part is obsolete, custom, or on a long lead time.
The instinct in a well-equipped shop is that anything can be welded, and it usually can. That is a different question from whether it should be. A good repair decision weighs the total cost of getting the part back into service against the cost of replacing it, and factors in whether the repaired part will actually last — which depends far more on why it failed than on how skilled the welder is.
Disclosure: Welding Gear is reader-supported. Links to Amazon and eBay are affiliate links, and we may earn a commission on qualifying purchases at no extra cost to you. We do not accept payment for placement, and we do not publish prices, ratings, or reviews we have not verified.
Start with the failure mode
This is the question that decides most cases, and it comes before any cost arithmetic.
| Why it failed | Repair outlook | Reasoning |
|---|---|---|
| Abrasive or impact wear | Repair, usually easily | The part is structurally sound and has simply lost material. Buildup and hardfacing restore it, often better than new. |
| Overload — a single event beyond design | Repair is often reasonable | The part was fine until something unusual happened. Fix it and address the cause; check for distortion beyond the obvious break. |
| Fatigue cracking | Repair with caution, and only with a design change | A fatigue crack means the part is loaded beyond what its geometry tolerates over cycles. Welding the crack shut restarts the same countdown unless you add material, change the geometry, or reduce the stress concentration. |
| Corrosion | Depends on extent | Localised loss can be built up. General thinning across a section means the whole part is compromised, and patching the hole you can see ignores the metal around it. |
| Manufacturing defect or bad prior repair | Often worth doing properly | Cut out the defect entirely rather than welding over it. A repair on top of a bad repair fails at the interface. |
| Wrong material for the job | Replace, or remake in the right material | No amount of welding fixes a specification error. |
If you find a crack that started at a stress concentration — a sharp corner, a weld toe, a drilled hole, a section change — and grew progressively, welding it closed without changing anything else means it will return in roughly the same service time, at the same place. Gusset it, radius the corner, grind the weld toe smooth, or reduce the load. Repairing the symptom is the classic false economy in fabrication repair.
Count the real cost
People compare the price of a new part against an estimate of welding time, and that comparison is wrong nearly every time because welding is the small part.
- Access and removal. Getting the part off the machine, and back on, is frequently the largest labour item.
- Cleaning and preparation. Degreasing, removing paint and coatings, gouging out cracks, grinding to sound metal. Repair welding onto contaminated material is how you get a repair that fails immediately.
- Preheat and heat control. Unknown steels, heavy sections, and castings need it, and it takes time and gas.
- Machining and finishing. Restoring a bore, a mating face, or a bearing surface after welding may need equipment you do not have.
- Coating. Repainting or re-galvanising after a repair on a coated part.
- Downtime. Often the dominant cost, and it can point either way — a repair today beats a replacement part next month, and vice versa.
Run those honestly and a surprising number of small, common, in-stock parts turn out to be cheaper to replace, while large, custom, or obsolete parts are overwhelmingly worth repairing.
Repair almost always wins here
- Obsolete or unavailable parts. If it cannot be bought, the comparison is against fabricating from scratch, and repair wins easily.
- Long lead times. Downtime cost frequently exceeds the entire repair.
- Wear components on heavy equipment. Buckets, teeth, augers, hammers, mixer paddles — designed to be rebuilt, and hardfacing often outlasts the original surface.
- Large weldments. The material and fabrication cost of a big structure dwarfs the cost of repairing a local failure.
- Cast components on older machinery. Difficult to weld, but often impossible to buy.
Replace instead when
- The part is cheap, standard, and in stock. Shop time is worth more than the component.
- Corrosion is general rather than local. Patching one hole in a section that is uniformly thin buys weeks, not years.
- Restoring dimensional accuracy is beyond your equipment. A repair that leaves the part out of tolerance has not repaired it.
- The material cannot be welded reliably with what you have — certain aluminium alloys, some castings, heat-treated components whose properties would be destroyed.
- The part has already been repaired several times. Each cycle adds heat-affected material and residual stress.
The category where economics do not apply
Some components are not a shop judgement call regardless of how good the repair would be. Lifting equipment — hooks, slings, lifting eyes, crane and hoist components. Structural members in buildings and bridges. Pressure vessels and pressurised piping. Suspension, steering, and braking components on vehicles. Roll-over protection structures and safety cages. Anything with certification, a rating plate, or a design code attached to it.
These are governed by codes and qualified procedures, and repairs typically require qualified welders working to an approved procedure with inspection. Welding a lifting eye back on because it looked simple is how people get killed, and it will also void any insurance position you thought you had. If a component falls into these categories, the correct question is who is qualified to assess it, not whether you can weld it.
Gouging and preparation consumables
Where repairs are won or lost$$Repair welding is preparation work with some welding at the end. Air-arc gouging carbons, gouging-specific plasma consumables, and a good stock of cutting and grinding discs are what let you remove a crack completely and get to sound metal rather than welding over the problem.
The discipline is to keep going until the defect is gone. A crack that has been gouged most of the way out is still a crack, and it will propagate straight back through the repair.
Strengths
- Removes defects completely rather than burying them
- Air-arc gouging is fast on heavy sections
- Prep quality determines repair life more than welding does
Trade-offs
- Air-arc gouging is loud, dirty and needs a large compressor
- Consumables are an ongoing cost
- Requires eye and hearing protection beyond normal welding kit
Buildup and hardfacing consumables
Rebuilding wear parts$$For the category where repair wins most decisively. A tough machinable buildup consumable restores dimension, and a hardfacing alloy matched to the wear mechanism goes on top. Together they return a wear component to service with a surface that frequently outlasts the original.
Match the hardfacing alloy to how the part actually wears — abrasion, impact, or metal-to-metal. Putting a brittle high-carbide alloy on an impact part produces a surface that spalls in the first hour.
Strengths
- The clearest economic case in all of repair welding
- Rebuilt surfaces often outlast the original
- Buildup layer is machinable back to dimension
Trade-offs
- Alloy must match the wear mechanism
- Layer limits apply — check the datasheet
- Does not fix cracked or fatigued base components
Measuring and inspection gauges
Prove the repair before it ships$$Fillet gauges, straight edges, calipers, and a dye penetrant kit let you confirm that a repair is dimensionally correct and free of surface-breaking cracks before the part goes back into service. On repair work this matters more than on new fabrication, because you are welding onto material with unknown history.
Penetrant testing in particular is cheap, needs no equipment beyond the aerosols, and reliably finds the surface cracks that a repair is most likely to produce.
Strengths
- Confirms the repair is sound before it is installed
- Penetrant kits find surface cracks with no equipment
- Cheap relative to a repeat failure in service
Trade-offs
- Surface methods only — will not find internal defects
- Requires clean, dressed surfaces to be meaningful
- Not a substitute for qualified inspection on coded work
The framework, in four questions
Ask them in order. One: is this component in a safety-critical, coded, or certified category? If yes, stop and get it assessed properly. Two: why did it fail, and will the repair address that cause or just the symptom? Three: what is the full cost of repair including removal, preparation, finishing, and downtime, against the full cost of replacement including lead time? Four: can you restore it to the dimensional accuracy and material condition the job requires with what you have?
Four clear answers make the decision obvious. What causes bad outcomes is skipping straight to question three, which is how a fatigue crack gets welded shut on a part that will crack again in the same place, and how a lifting eye ends up repaired by someone who was never asked whether they should.
Frequently asked questions
Can I weld a crack in a lifting eye or a hook?
Do not treat this as a shop decision. Lifting equipment is rated, certified, and governed by regulations that generally require assessment and repair by qualified personnel to an approved procedure, followed by inspection and recertification. A failure here endangers people directly. The correct action is to take it out of service and have it assessed.
Why do repaired cracks keep coming back in the same place?
Because the crack was a symptom of fatigue loading at a stress concentration, and welding it closed does not change the geometry or the load. To fix it properly you have to reduce the stress concentration — radius the corner, grind the weld toe smooth, add a gusset or doubler — or reduce the loading itself.
Is it cheaper to weld or replace a worn bucket tooth?
Rebuilding wins clearly for wear components whose base structure is sound. They fail by losing material rather than breaking, and buildup plus hardfacing restores dimension with a surface that often outlasts the original. The economics only reverse if the underlying part is cracked or has worn past its load-carrying section.
How do I know if corrosion damage is repairable?
Measure the surrounding metal, not just the visible hole. Localised pitting in otherwise sound material can be built up. General thinning across a section means the whole area is compromised and patching one spot buys very little time. Ultrasonic thickness measurement is the practical way to tell the difference.
How many times can a part be repaired?
There is no fixed number, but each cycle adds heat-affected material, residual stress, and accumulated weld metal of uncertain history. A component that has been repaired several times in the same area should be viewed sceptically, and on anything load-bearing it should be replaced rather than repaired again.