Welding Pad Eyes: Design, Welding, and Safety Guide

A fabrication shop welds a pair of small steel lugs onto a two-ton steel platform, slings it with a forklift, and lifts. Halfway up, one lug tears out of the base plate and the load swings into a storage rack. Pad eyes look like the simplest part on any job—two holes, a few inches of weld—but they are the single point of failure between a lifted load and the floor, and the weld is usually the part that fails, not the steel.

What Are Welding Pad Eyes and Why Do They Matter?

A pad eye is a plate or bar with a hole through it, welded permanently onto a structure so it can be lifted, tilted, rigged, or tied off. A shackle pin passes through the hole, a sling or chain connects to that shackle, and a crane takes the load from there. While fabricators may call them lifting lugs, weld lugs, tie-off lugs, or just “the ear”, they all serve the same purpose.

The biggest distinction from other rigging hardware is permanence. Shackles, eyebolts, or hoist rings come from a manufacturer with a stamped rating. A pad eye is made by a welder on the spot, has no catalog number, and becomes part of the structure forever. That makes it just as responsible for a suspended load, without the convenience of a ready-made rating.

Two Types to Know

Not every pad eye is meant for a crane hook. Mixing the two is one of the most common and costly mistakes.

• Structural or lifting lugs: These are load-bearing. They’re used to lift skids, tanks, machinery bases, weldments, ship sections, or other heavy assemblies. They must be sized for the load, welded to a procedure, and treated as a structural component.
• Tie-off and guide lugs: These are much smaller, often used for tag lines, temporary alignment, or securing covers. They’re rarely designed for full lifts, and the danger comes from their reputation as “temporary”. Far too often, they’re reused years later as a lifting point without ever being checked.

You’ll see both across shipbuilding, construction, heavy transport, and nearly any fabrication shop where steel needs to be moved.

The Two Obligations of a Pad Eye

There is one rule to carry through this guide: a pad eye inherits both a load-rating obligation and a weld-quality obligation.

1. Know the rating. Someone must determine what the lug can hold, and anyone rigging to it must know that limit before tension is applied.
2. Know the joint is the load path. The weld attaching the lug to the structure is just as much a part of that path as the lug plate itself. In most field failures, the eye doesn’t bend or tear. The weld or the base metal right beside it tears out, as the joint proves to be the weak link.

Standards, Responsibility, and Approval

Welding a lifting lug onto a structure is standard practice, but it is not improvised work. Several frameworks dictate that responsibility:

• OSHA: Rigging hardware and lifting devices must be rated for their load and inspected before use under 29 CFR 1926.251 for construction and 29 CFR 1910.179 for general industry.
• ASME BTH-1: The go-to engineering reference for the design of below-the-hook lifting devices, including lifting lugs. See ASME BTH-1 for its scope.
• ASME B30.26: Governs the rigging hardware that connects to the lug, such as shackles and other sling hardware. See ASME B30.26.
• AWS D1.1: Governs the weld itself, including procedure qualification, welder qualification, preheat, inspection, and acceptance criteria. Check the AWS standards listing for the applicable edition.

What This Guide Covers

This guide follows a pad eye from concept to retirement: geometry and hole sizing, weld design, preparation and welding, applicable codes and responsibility, common failures and prevention, inspection and maintenance, alternatives to welding, and a practical shop workflow.

For steel structures, the AWS D1.1 and ASME BTH-1 path above applies. For aluminum, allowances differ greatly, so consult our Aluminum MIG Welding: The Complete Practical Guide before adapting any design.

Above all, keep this in mind: a pad eye looks like the simplest piece on the job, but it’s the single point of failure between a lifted load and the floor. Treat it accordingly.

How to Design a Pad Eye: Geometry, Hole Size, and Load Rating

Designing a pad eye looks like shop trivia: cut a plate, punch a hole, weld it on. The reality is that three separate load paths have to survive the same lift, and the weakest one sets the rating for the whole thing. Get the geometry right and the welding follows. Get it wrong and no amount of good arc time saves you.

Plate Material and Thickness

Most lifting lugs are cut from ASTM A36 plate, which has a specified minimum yield strength of 36 ksi and a tensile strength of 58–80 ksi. A36 is the default for a reason: it is well characterized, ductile, and forgiving to weld. For heavier lifts, fabricators step up to higher-strength plate, but the stronger the base metal, the more carefully the filler metal and procedure have to be matched to it. Our guide to Alloy Steel Welding Electrodes: Selection, Use & Safety covers that relationship between filler strength and base metal.

Thickness drives three things at once: net section through the hole, bearing on the shackle pin, and how much weld throat you can hang around the perimeter. There is no single magic number, but shop practice generally treats 3/4-inch plate as a sensible floor for structural lifting lugs, with smaller tie-off or guide lugs allowed to run lighter. Adding thickness is the cheapest way to gain capacity, because net section scales directly with it.

Hole Size, Edge Distance, and Proportions

Size the hole around the shackle, not the other way round. The pin should pass through freely with a little clearance, and the shackle bow needs to clear the base plate completely. If the bow rubs or seats on the plate edge, the shackle picks up bending instead of pure tension, and your load rating is fiction.

Edge distance is where small lugs usually fail. A widely used rule of thumb is that the center of the hole should sit at least 1.5 to 2 times the hole diameter from the nearest edge. That mirrors the thinking behind AWS D1.1’s minimum edge distance provisions for bolts, and for the same reason: too little material in front of the hole turns a tension load into a tearing load. In width, leave the same margin on both sides of the hole; in length, give yourself enough room beyond the hole for a real weld group, typically a lug roughly twice as long as it is wide.

The Load Path: Net Section, Bearing, and the Weld Group

Run three checks, because three different things can break:

• Net section in tension. The plate tears across the narrowest slice, which is the width minus the hole. The gross plate section is irrelevant here; the net section through the hole controls tensile capacity.
• Bearing on the pin. The pin crushes or elongates the hole. With a reasonably thick lug this rarely governs, but it still needs checking.
• The weld group. The fillets around the base of the lug resist shear and the prying action that appears whenever the load is not perfectly aligned. Welds on one side only create a lever arm, so a full fillet all around is the standard answer.

Design Factor, Service Class, and Proof Testing

A working load limit is only meaningful next to a design factor. ASME BTH-1 organizes below-the-hook design around design categories and service classes: Design Category A covers general applications, Category B covers severe service, and Service Classes 0 through 4 describe how many load cycles the device will see. BTH-1 calls for a minimum design factor of 3 for Category A and 5 for Category B, which is why experienced riggers treat 3:1 as the floor for any hand-designed lug and go higher for lift points that will see repeated or shock loading.

Worked Example: A 1-Inch Hole Lug in 3/4-Inch A36 Plate

Suppose the lug is 3 inches wide, 2.5 inches long, drilled for a 1-inch hole, welded all around with a 3/8-inch fillet (about 8 inches of total weld length). Illustrative numbers only:

Check	Math	At 3:1
Net-section tension	(3.0 − 1.0) × 0.75 = 1.5 in²; × 58 ksi = 87 kips	≈ 29 kips
Gross-section yield	3.0 × 0.75 = 2.25 in²; × 36 ksi = 81 kips	≈ 27 kips
Fillet weld group	throat 0.707 × 0.375 = 0.265 in; × 8 in = 2.12 in²; × 21 ksi ≈ 44 kips	≈ 15 kips

The weld governs at roughly 15,000 lb per lug: the weld, not the plate, is usually the weak link. Bumping the fillet to 1/2 inch narrows the gap but does not eliminate it. Pin bearing is checked separately and is not critical at these proportions.

Then apply the sling angle. Tension in each sling leg rises as the angle flattens away from vertical, and in the worst case a sling set 90° from vertical can double the tension you would measure on a straight vertical lift. A lug rated at 15 kips vertically can be over capacity on a long, flat sling, which is why the lift plan has to be finished before the geometry is finalized.

And remember what AWS D1.1 implies: a rated lug is a designed lug. If nobody ran the numbers, signed them, and tagged the lug, the honest rated capacity is zero.

Preparing the Joint: Fit-Up, Chamfers, and Base Metal Prep

By the time the arc is struck, most of a pad eye’s success or failure is already decided. Technique matters, rod choice matters, but nothing turns a bad joint into a good one faster than careful preparation. Lifting lug failures usually trace back to the five minutes before welding: the wrong location, a painted surface, a sloppy fit-up, or a cold plate on a January morning.

Pick the Location Before You Pick the Rod

A lug pulls, and that pull travels through the weld into whatever steel sits behind it. So the first real design decision is placement. Weld lugs over a stiffener, column flange, frame member, bulkhead, or other reinforced area whenever possible. Avoid thin, unsupported sheet or the middle of a large flat panel, where the base metal can dimple, oil-can, or tear locally even though the lug and its welds are perfectly sound. Lift on a lug over weak base metal and you are effectively testing the parent plate’s tearing strength with a two-ton load and a forklift.

When the base metal is thinner than the lug load demands, add reinforcement rather than hoping. Doubler plates spread the concentrated lug load over a wider area, and gussets or triangular stiffeners resist the prying action that develops when a sling pulls at an angle. A doubler should be fully welded around its perimeter, not tack-welded, and it should be sized by whoever sized the lug.

Clean the Weld Zone to Bare Metal

AWS structural welding standards require mill scale, rust, oil, paint, and coatings to be removed from the weld zone on exterior surfaces, because these contaminants cause porosity, lack of fusion, and inclusions in exactly the place a weld cannot afford them. Grinding or brushing back to bright steel an inch or so beyond the weld footprint is cheap insurance. The habit transfers across metals too: shops running aluminum work find the rule less negotiable still, as our Aluminum Cleaner for Welding guide explains.

Fit-Up, Squaring, and Tack Welds

Hold the pad eye flat against the plate with no gap. A visible gap under the lug turns the root of the fillet into an oversized, badly supported groove, and the welder ends up filling a void instead of fusing two surfaces. On assemblies with multiple lugs, use a square, template, or measured layout so the eyes sit at the same height; mismatched lug holes force the load onto one sling and defeat the point of a balanced lift.

Tack welds should never be treated as throwaway. AWS D1.1 requires tacks on structural members to be made with qualified low-hydrogen consumables or processes, and to be either incorporated into the final weld or removed by grinding before welding over them. A brittle, hydrogen-cracked tack is a crack starter, and a tack that melts into the final weld without fusion becomes a hidden defect under the toe.

Preheat: When Thickness and Temperature Demand It

Preheat requirements under AWS D1.1 scale with material thickness, steel group, and ambient temperature, and in cold-weather work thick A36 sections can still demand a minimum preheat around 0°F while thinner members may need none at all. The logic is simple: a thick or frozen plate acts as a heat sink, the weld cools fast, hydrogen stays trapped, and cold cracking forms at the toe where you cannot see it. Check the preheat table for your actual thickness and steel group instead of guessing, and hold temperature with a temp stick or infrared thermometer.

Electrode storage matters for the same reason. Low-hydrogen consumables carry hydrogen designations such as H4, H8, and H16, and an E7018 rod left out in a humid shop quietly stops being low hydrogen. For handling and drying detail, see our Alloy Steel Welding Electrodes guide.

Flat Plate Lugs Versus Chamfered or Contoured Lugs

The simplest and most common detail is a flat rectangular or teardrop-shaped lug welded flush against the plate with a fillet weld all around the contact perimeter. It is easy to lay out, weld, and inspect. A chamfered or contoured lug, with its inner edge beveled so the fillet can be welded into a corner joint, gives better throat thickness and a smoother load path, but it costs more to cut and demands tighter fit-up. Whichever you use, dress the plate edges and corners lightly: a sharp, unbroken toe on a thick lug against a thin plate is a natural stress raiser, and blending that toe slightly reduces the chance of a crack initiating there.

None of this is glamorous work. But a lug is only as good as the surface it is welded to, and five minutes of grinding, squaring, and checking a preheat table is the cheapest safety upgrade available on any lifting job.

Welding the Pad Eye: Process, Fillet Sizing, and Technique

Once the lug is fit up and tacked, the weld becomes the whole story. A pad eye can be machined perfectly and positioned over a stiffener, yet still fail because someone ran a sloppy fillet around the base. The plate almost never tears; the weld lets go. Process, weld size, and technique are where a fabricated lifting point earns or loses its rating.

Choosing the Welding Process

There is no single correct process, but each has a natural home:

• SMAW (stick) with E7018 is the classic field choice for structural fillet welds: low-hydrogen, forgiving of dirt and wind, and on every truck. If you are welding a lifting lug outdoors or in a repair situation, 7018 is usually the answer, and the advantages of stick welding explain why it still dominates structural work.
• FCAW (flux-cored) suits heavier production fabrication. It deposits fast, runs high travel speeds on thick plate, and handles out-of-position welds well when gas-shielded wire is used.
• GMAW (MIG) works for thinner fabrications and shop-controlled conditions, where clean, uncontaminated base metal and stable shielding gas are guaranteed.

Whatever process you pick, the consumable must be low-hydrogen and matched to the base metal. Hydrogen is the enemy of lug welds, and the AWS D1.1 Structural Welding Code sets the rules for filler classification, preheat, and joint details on structural steel (AWS standards). For filler classification beyond the basics, this guide to alloy steel welding electrodes covers how AWS designations translate to real performance.

Fillet Weld Sizing: Leg Size vs. Throat

Two numbers confuse people: leg size and throat. The leg is the visible face length of the fillet; the throat is the shortest distance from root to face, about 0.707 times the leg on an equal-leg fillet. Fillet weld capacity is calculated on throat area, not leg size, so a 1/4-inch leg weld has only about 0.177 inches of effective throat.

The standard practice for lifting lugs is a full fillet weld all around the contact perimeter; partial welds, tack-only corners, or a one-sided weld are not acceptable on a load-bearing lug. Leg size is commonly matched to the thickness of the thinner joined part and never falls below the AWS D1.1 fillet weld minimums, which scale required size to the thickness of the thinner material joined (AWS). A lug in 3/4-inch plate, for instance, usually gets a 1/4-inch leg fillet all around at minimum; heavier lugs step up accordingly, sized by calculation rather than habit.

Technique: Beads, Weaves, and the Undercut Problem

On thicker legs, plan for multiple passes. Use stringer beads wherever possible; if a weave is necessary, keep its width within roughly the electrode diameter plus a small margin, so the arc still cleans the root instead of washing over the surface. Each pass should tie into the previous one without cold laps.

Undercut at the weld toe is the classic crack initiator in lifting lugs. It is a stress concentrator machined by the arc itself, a notch at exactly the point where the lug meets the base plate and stress is highest. Hold a short arc, keep the electrode angle pointed into the corner, and slow slightly at the stop to fill the crater. Blend sharp toes and remove spatter so penetrant or magnetic particle inspection can read the surface honestly.

Welding sequence matters too. Balance passes around the lug, alternating sides rather than running one full side at a time, so shrinkage pulls evenly instead of tipping the lug out of alignment. Keep interpass temperature under control; stacking heat into a small lug on thick plate invites coarse grain structure and, with hydrogen present, cold cracking. Finish by checking the lug is still square and flush with the plate.

Common Defects in Lug Welds

• Lack of fusion at the root: the arc never melted into the corner, leaving a hidden void that carries no load.
• Undercut: notched toes, as above, the most common fatigue and fracture starting point.
• Porosity: usually from contaminated base metal, wet electrodes, or lost shielding gas.
• Cold cracking: hydrogen plus restraint plus a hard heat-affected zone; prevented with low-hydrogen consumables, dry electrodes, and preheat per AWS D1.1 tables keyed to thickness and temperature.

Undercut and hydrogen cracking are why pad eye welds deserve NDE even when they look fine. Visual inspection cannot see a root lack-of-fusion or a toe crack hiding under spatter.

Done right, this step yields a weld that is geometrically correct, metallurgically sound, and documented. The next section maps out which codes govern that work, and who is legally responsible when the lug goes airborne.

Codes, Standards, and Legal Responsibilities

When a welded lug fails, the investigation rarely starts with “who welded it.” It starts with “who rated it.” A pad eye sits at the intersection of four separate documents, and each one owns a different piece of the job. Confusing them is how shops end up confident about a lug nobody actually approved.

Here is the map in one view:

Standard	What it governs	What it does not cover
AWS D1.1 (Structural Welding Code—Steel)	The weld itself: procedures, preheat, consumables, inspection, qualification	How much load the lug can carry
ASME BTH-1	Design of below-the-hook lifting devices, including welded lifting lugs	Weld technique details
ASME B30.26	Rigging hardware: shackles, eyebolts, slings and their fittings	Welded lugs on your structure
OSHA 1926.251 / 1910.179	Employer duty to use rated, identified, inspected rigging	Engineering the lug for you

AWS D1.1: It Signs Off on the Weld, Not the Load

The AWS structural welding standards cover the joining side of the equation: which processes are acceptable, how consumables are classified and stored, when preheat is required, how welders and procedures are qualified, and what visual acceptance criteria apply. Run a lug weld with a low-hydrogen electrode on a qualified procedure and inspect it against D1.1 acceptance criteria, and you have proven the weld is sound. You have not proven the lug is strong enough. Those are two different certifications, and both matter. Shops in documentation-heavy sectors such as aerospace welding already understand the discipline of separating process qualification from design approval.

ASME BTH-1: The Design Standard for the Lug Itself

ASME BTH-1 is the recognized American standard for designing below-the-hook lifting devices, and welded lifting lugs fall squarely inside that scope. BTH-1 organizes design around design categories and service classes, which together set the design factor and the fatigue life you are designing toward. A lug on a one-off fabrication lift is a very different design problem from a pair of lugs on a platform picked twice a day for a decade. BTH-1 forces that distinction instead of letting it be assumed.

ASME B30.26: Everything That Connects to the Lug

ASME B30.26 covers the design, construction, testing, and use of rigging hardware: shackles, eye bolts, hoist rings, sling attachments. It does not rate your welded lug, but it governs the shackle pin that passes through it and the sling hanging off the bow. A lug rated for a 1-inch pin, paired with hardware used outside its own rating, is still an uncontrolled lift. Rated capacity is a chain, and every link is rated separately.

OSHA: The Employer’s Duty to Use Rated Gear

Two rules matter here. OSHA 1926.251 applies in construction and requires that rigging, including alloy steel chain slings and associated hardware, be rated, identified with its capacity, and periodically inspected. OSHA 1910.179 covers general industry powered hoists and overhead cranes with the same logic: lifting gear must have a published rating, be visibly identified, and be inspected on a schedule. Both require damaged or unsafe rigging to be removed from service immediately.

The practical implication is blunt. OSHA does not rate equipment for you; it requires that rated equipment be used. If your welded pad eye is a permanent part of a lifting device, someone has to own its rated capacity, and in nearly every case that someone is a professional engineer. There is no such thing as an unrated lug that is nevertheless fine to lift on.

What “Rated Capacity” Means for a Field Lug

A rated capacity is not a number you eyeball from a similar-looking catalog detail. It is a documented value backed by design calculations, and in many jurisdictions a custom below-the-hook device requires both a design calculation and a load test before service. The paper trail typically includes:

• A load rating plate or tag permanently marked on the device with rated capacity and identification.
• Design calculations or an engineering stamp showing how that capacity was derived.
• A lift plan documenting weight, center of gravity, sling angles, and hardware ratings.
• Records of any proof test and of periodic inspections.

For an improvised field lug, “rated capacity” usually means one of two things: engineer it properly, or don’t lift with it.

Think of it that way and the standards stop feeling like red tape. D1.1 tells you the weld is sound, BTH-1 tells you the lug is designed for the load, B30.26 tells you the attached hardware is fit for service, and OSHA tells you all of it must be rated, marked, and inspected before anyone hooks up. Skip one document and the lifting system has a hole in it, even if you cannot see one.

Common Pad Eye Mistakes and How They Fail

Pad eye failures rarely announce themselves in advance. The lug looks fine, the lift starts, and then a weld toe cracks, a hole elongates, or a plate rips clean out of the base metal. In most post-failure reviews, the steel eye was never the problem. The weld, the surface it was welded to, or the way it was loaded was. Nearly every failure traces back to a handful of repeatable shop-floor mistakes, each with a clear physical mechanism behind it.

Welding to Base Metal That Cannot Carry the Load

The lug may be three-quarters of an inch of A36 plate, but it is only as strong as whatever it is welded to. Welding a lifting lug onto thin sheet, an unsupported panel, or the middle of a tank wall invites local tearing: the load pulls the base metal into a cone, the plate buckles, and the lug peels away with a chunk of parent material still attached. This is the single most common reason for the shop horror story of a “lug failure” that is really a base-metal failure. Locate lugs over stiffeners, columns, or reinforced areas, and add doubler plates or gussets when the substrate is thin.

Undersized or Partial Fillet Welds

Standard practice for a lifting lug is a full fillet weld all around the contact perimeter. A weld on only the two sides parallel to the load, or a leg size smaller than the code minimum, cuts the weld group’s throat area and hands the shear load to a fraction of the joint it was designed for. Prying loads from an angled sling then concentrate on the short ends, and the weld unzips from the weakest corner.

Welding Over Paint, Mill Scale, or Galvanizing

Coatings are contaminants at the arc. Paint and oil burn into gas, producing porosity; mill scale breaks fusion at the root; zinc from galvanizing fumes off, embrittles the puddle, and leaves a weak, sooty weld. Any of these leave the lug weld with a fraction of its intended strength while looking acceptable from a distance. Grind to bare, bright metal in the weld zone before the arc is struck, and use proper aluminum cleaner for welding-style discipline on coated or non-ferrous surfaces.

Using the Wrong Electrode

E6013 and other fast-freeze, high-hydrogen-covering rods are fine for decorative brackets and sheet-metal joints. They are wrong for structural lifting lug welds. Without low-hydrogen consumables and controlled hydrogen input, the rapid cooling of a restrained lug weld can trigger hydrogen cracking, often cold-cracking hours after the weld looks finished. The classic field answer is E7018 stick, kept dry and controlled, or a qualified FCAW equivalent. Selection matters more than convenience here, as covered in our guide to alloy steel welding electrodes.

Undercut and Cold Cracking at the Weld Toe

Undercut is a notch. Under cyclic or angled lifting loads, that notch at the weld toe becomes the crack initiator, and lifting lugs see fatigue the way fixed brackets do not. Weave too wide, run too hot, or whip the rod and you cut a groove into the base plate right where the stress concentrates. Run stringer beads, keep the weave within the electrode-recommended width, and blend sharp toes after cooling.

Wrong Hole Size and Poor Edge Distance

A hole that is too tight makes the shackle pin bind and loads the eye unpredictably; one that is too loose lets the bow foul the base plate and adds bending the lug was never sized for. Worse is insufficient edge distance: with the hole center too close to the edge, the ligament of steel ahead of the pin simply tears out under load, sometimes at surprisingly modest weight. Follow the rule of thumb that the hole center sits at least 1.5 to 2 times the hole diameter from the edge, and match the hole to the shackle pin diameter, not to whatever punch is on the rack.

Lifting at Angles the Lug Was Never Designed For

A lug designed for a vertical pull sees a very different load when a sling comes in at 45 or 60 degrees. The tension in each leg rises, and the lug picks up side loading and bending that were not in the calculation. Sling-angle load multiplication is the quiet multiplier behind many “it was rated for that weight” arguments.

Reusing Damaged Lugs and Cutting Them Out Carelessly

A lug that has been bent, overheated, or visibly deformed should be cut out and replaced, not straightened and put back in service. Straightening cold-works cracked material and hides the damage; the hole may look round while microcracks run along the weld toe. Cutting a lug off with a torch leaves gouges, notches, and a fresh heat-affected zone in the parent structure. Those gouges must be ground smooth and the base metal restored to flush, then inspected, before anything else is welded or lifted there. Carbon arc gouging is a common removal method, but it is a repair operation, not a cleanup step.

A Scenario That Plays Out Every Year

A plate gets lugs welded straight onto painted steel because the job is behind schedule. The welder grabs whatever rod is in the truck, runs a quick pass in light rain, and the puddle goes milky with hydrogen and porosity. Nobody grinds, nobody inspects, and the first lift goes up in the air. The welds look dark and crusty, so nobody notices the toes. About halfway up, one lug tears free at the base and the plate swings into a rack. This sequence is not exotic. It is exactly what happens when coating prep, consumable storage, and inspection are skipped in that order. For context, OSHA 1926.251 and OSHA 1910.179 both require rated, identified, and inspected rigging; an unmarked, unengineered lug does not meet that bar.

None of these mistakes require exotic fixes. They require the same modest discipline every time: clean metal, the right low-hydrogen consumable, a full and properly sized fillet, adequate edge distance, engineered geometry, and a hard rule against reusing anything that has visibly given way.

Inspection, Testing, and Ongoing Maintenance

A pad eye that passed on Monday is not automatically trustworthy on Friday. Shock loads, sideways pulls, weather, and paint overspray accumulate between uses. Knowing whether a welded lug is still sound means looking at it, documenting what you see, and testing it when the stakes justify it. Inspection is not a one-time event at fabrication; it is a lifecycle with three distinct rhythms.

The Three Rhythms of Lug Inspection

Before every lift. This is a 30-second check done by the rigger, not the office. Look for cracks at the weld toe, where the weld meets the base plate; that is where nearly all lug cracks start. Check for a stretched or oval hole, a bent or twisted plate, gouges, corrosion pitting, illegible tags, and paint cracking in a line that might hide a fatigue crack. If the lug was painted, wire-brush the weld toe first; surface paint hides more defects than it protects.

Periodic, documented inspection. On a set schedule, someone with training records what they find: weld condition, hole geometry, plate flatness, corrosion, and tag legibility. OSHA expects employers to inspect rigging and remove damaged equipment from service, both in construction under 1926.251 and in general industry under 1910.179. A damaged, cracked, or deformed lug must be tagged and pulled from service, not discussed and reused.

Nondestructive evaluation (NDE). When a lug has seen a heavy lift, an angled lift, a side pull, or a shock load such as a dropped load, a visual check is not enough. Dye penetrant testing (PT) and magnetic particle testing (MT) are the standard methods for finding surface-breaking cracks at weld toes. PT works on any nonporous surface and is cheap and portable; MT is faster on carbon steel and can hint at near-surface indications but requires magnetizing equipment. Neither replaces an honest look at the lug first.

What Proof Testing Actually Proves

Proof testing is the practice of loading a device beyond its rated capacity in a controlled setting. ASME BTH-1 conventionally applies a proof load of 125% of rated capacity for below-the-hook lifting devices. The test confirms the device survives a known overload without visible deformation or permanent set.

What it does not prove matters just as much. A proof test says nothing about fatigue life over thousands of cycles, nothing about hidden corrosion, and nothing about a weld defect below the surface that the load happened not to open. It is a snapshot, not a warranty. That is why BTH-1 pairs the design factor with a service class and design category: the design category sets the design factor against static strength, while the service class sets the fatigue envelope for the expected number of load cycles. A lug lifted four times a year and one lifted forty times a day are not the same device.

Environment: Heat, Cold, Corrosion, and Cycles

• Corrosion. Marine and outdoor service eats edge distance and throat thickness. A lug that started at a code-minimum fillet can be below minimum after two seasons of rust. Inspect more often, coat the weld, and treat a pitted lug as a thinner lug.
• Cold. Carbon steel gets brittle. AWS D1.1 addresses toughness through Charpy V-notch considerations for weldments subject to impact loading or low temperatures, and BTH-1 service classes account for duty. If you lift in winter, toughness matters as much as strength.
• Fatigue. Repeated lifting accumulates damage that no visual inspection catches early. Track cycle counts honestly and let service class, not optimism, set the retirement point.

Retiring and Documenting a Removed Lug

Removing a lug is a repair operation, not a cleanup chore. Cut it out, grind the area smooth, inspect the base metal for embedded cracks or arc strikes, and repair gouges or undercut to the structural code before repainting. Record what was removed, why, and where the base metal stood afterward. For context on the human stakes behind these rules, NIOSH’s steelworker injury and fatality research is a sobering read, and OSHA’s rigging guidance is collected in OSHA Publication 3991.

For how cleaning and surface prep discipline translates across processes, our guide to aluminum cleaner for welding applies the same inspection mindset to a different base metal.

Alternatives to Welded Pad Eyes and When to Use Them

A welded pad eye is a permanent decision. Once it is fused to the structure, it lives there until someone cuts it out, grinds the gouges flush, and repairs the parent metal. That permanence is fine for a skid or tank that will be lifted for the next thirty years. It is a poor choice for a one-time move, a rental machine, or anything that has to pass through a paint line or a galvanizing bath afterward. Fortunately, rigging hardware offers several alternatives, each with its own strengths and its own discipline requirements.

Bolted Lifting Lugs

A bolted lug is the closest substitute for a welded pad eye. A thick plate with a hole is drilled for bolt holes, fitted against a reinforced section of the structure, and secured with high-strength bolts of a known grade, torqued to spec. The advantages are real: the lug can be removed without leaving a heat-affected zone or gouges in the parent structure, it can be inspected as a discrete piece, and it can be reused on the next job. There is no arc, no hydrogen, and no weld procedure to qualify.

The tradeoff is discipline. Bolted lugs shift the weak link from the weld to the bolt group. You must select the correct bolt grade, torque every fastener to the specified value with a calibrated wrench, and re-check torque before heavy or angled lifts. An under-torqued bolted lug carries load unevenly and can slip, and slip under load is violent. Bolted lugs also need flat, clean mating surfaces and enough structure behind them to spread the load.

Threaded Lifting Points and Swivel Hoist Rings

Threaded bushings tapped or welded into a structure, fitted with swivel hoist rings, are the standard answer when a load must be lifted from multiple angles. Hoist rings are rated for angular loading and can rotate and pivot, something a plain eyebolt cannot do. A forged eyebolt is primarily rated for straight, in-line pull; apply a side load to it and its effective capacity drops sharply, which is one of the most common rigging injuries waiting to happen. Design, testing, and use of this hardware falls under ASME B30.26.

Lifting Beams and Spreader Bars

When the geometry of a lift is the problem, a lifting beam or spreader bar solves it mechanically. By moving the sling attachment points apart, these devices reduce sling angle forces and keep side loading off the attachment points, so the load sees a more vertical pull. They also protect delicate or wide loads from crushing. The cost is a piece of engineered hardware that has to be stored, inspected, and certified in its own right. For shops that move long, wide, or fragile weldments regularly, a spreader bar earns its floor space quickly; for an occasional lift, it is usually rented.

Drilled Lifting Holes

Drilling a lifting hole directly through the structure eliminates weld risk entirely. The hole must be deburred and radiused to avoid a fatigue crack, and it must be engineered: a hole removes net section from the member, and putting it in the wrong place reduces capacity or creates a stress concentration exactly where the structure does not want one. Done correctly, a dedicated lifting hole is clean, invisible in service, and needs no hardware between the sling and the part.

Tack-Welded and Wrapped Rigging (Don’t)

Every so often a sling gets tack-welded to a plate or wrapped around a sharp corner because the job is urgent. Avoid it. A tack weld has no qualified procedure, no inspection, and no rated capacity. A wire rope wrapped around structure is damaged by the very sharp bends it is subjected to, and the load’s center of gravity is rarely where the rigger assumes. Both approaches violate the requirement that employers use rated, identified rigging under OSHA 1926.251 in construction and OSHA 1910.179 in general industry. There is no scenario where a tack weld is an acceptable lifting attachment.

Method	Permanence	Angular Loads	Removal Cost	Key Discipline
Welded pad eye	Permanent	Only if engineered for it	Cut, grind, repair HAZ	Weld procedure and NDE
Bolted lug	Removable, reusable	Yes, if designed for it	Unbolt	Bolt grade and torque
Swivel hoist ring	Removable	Yes, rated for it	Unscrew	Torque and swivel freedom
Lifting beam / spreader	Reusable hardware	Yes, by design	N/A	Storage and certification
Drilled lifting hole	Permanent	Limited	N/A	Engineered location and edge prep
Tack-welded or wrapped sling	Illegal improvisation	Unpredictable	N/A	Do not use

One more consideration swings the decision more often than cost does: removal. Cutting a welded lug off is itself a repair operation. You have to gouge or grind the weld flush, inspect the base metal for cracking and arc strikes, and rebuild any lost thickness to satisfy the structural code, which is AWS D1.1 territory for structural work. If a lift is temporary, bolting is not just simpler, it avoids creating a latent defect in a structure you have already signed off on. If you find yourself needing to compare filler metals or brush up on procedure details before committing to a welded attachment, our guides to all-position welding rods and stick welding fundamentals are good starting points.

The right choice comes from the same place every other rigging decision comes from: known weight, known center of gravity, known angles, and hardware with a rated capacity. Pick the method that fits the load and the load’s future, and the welder’s arc becomes one option instead of the default.

Practical Step-by-Step: Welding a Certified Pad Eye

Everything so far has been theory. This is the part that happens on the shop floor, in order, with a grinder in one hand and a deadline overhead. A certified pad eye is not a single clever weld; it is a twelve-step chain, and the lug is only as strong as the weakest step you skipped. Work through it in sequence and you get a lifting point you can defend in an inspection. Skip a link and you get the story from the introduction: a lug tearing out of a plate it was never really bonded to.

The Twelve-Step Workflow

1. Define the lift before you touch steel. Write down the total weight, where the center of gravity sits, the sling angles you intend to use, and how many attachment points share the load. This lift plan is the input for every decision that follows, and it is what OSHA 1926.251 expects employers to work from when selecting rated rigging.
2. Decide the geometry with an engineer or a proven standard detail. Either have a professional engineer size the lug for your load, or copy a validated detail from a qualified drawing with its rating already established. Guessing the plate size is a design decision, not a fabrication one.
3. Select the plate and make the hole properly. A36 plate covers most moderate lifts; heavier work calls for higher-strength material. Punch or drill the hole and hold edge distance at roughly 1.5 to 2 times the hole diameter so the eye does not tear out under load.
4. Locate the lug over structure that can carry it. Aim for a stiffener, column flange, web intersection, or reinforced area. Welding a lug onto thin unsupported sheet just moves the failure from the weld into the base metal.
5. Clean the weld zone to bare metal. Grind out paint, mill scale, rust, and coatings on both the lug and the parent plate. If the structure is coated or galvanized, deal with that contamination deliberately; the same discipline applies when you prep aluminum for welding, just with different chemistry.
6. Preheat when the tables say so. AWS D1.1 sets preheat and interpass requirements from material thickness and ambient temperature, and cold weather or thick sections are exactly when hydrogen cracking shows up in lug welds.
7. Fit, square, and tack. Hold the lug flush, check it with a square or template so a paired set lifts level, and make tack welds with the same procedure or a compatible low-hydrogen process as the final weld.
8. Weld all-around with a low-hydrogen consumable. E7018 stick is the classic field choice (why SMAW still wins in the field), with qualified FCAW for heavier production. Use balanced, alternating passes so heat does not pull the lug out of alignment, keep weave width in check, and never leave undercut at the weld toe.
9. Clean and inspect the weld. Chip the slag, wire-brush, and check visually against AWS D1.1 acceptance criteria. For lifting lugs, add dye penetrant or magnetic particle testing at the weld toes, which is where cracks start.
10. Tag the lug. Mark it with its rated capacity and a unique identification. An unmarked lug is, in practice, an unrated one.
11. Proof test if the design calls for it. ASME BTH-1 conventions typically apply a test load around 125% of rated capacity to below-the-hook devices, which validates the assembly but does not replace inspection.
12. Inspect before every use, and retire on damage. Look for cracks at the toe, elongated holes, and bent plates. Any one of those ends the lug’s service life immediately.

Worked Mini-Scenario: A 4,000 lb Platform on Two Lugs

Take a 4,000 lb platform carried by two lugs on a two-leg sling. With the lugs level and the load shared evenly, each lug starts with 2,000 lb. Now change the sling geometry. At 60° from horizontal, tension in each leg is (4,000 ÷ 2) ÷ sin 60°, or about 2,310 lb. Drop the legs to 30° from horizontal and each leg tension climbs to 4,000 lb. Same platform, same lugs, nearly double the force, plus a horizontal squeeze component the vertical case never saw.

That is why hook or sling loading at an angle reduces effective capacity and increases local stress in the lug. If you apply a 3:1 design factor to the 2,310 lb case, each lug should be rated at roughly 7,000 lb before you hook up the first time.

The whole sequence sounds slow until you compare it with the alternative. Under OSHA 1910.179 and related rigging rules, a welded lug with no rating, no record, and no inspection is an uncontrolled lifting device, and the person who welded it quietly became the manufacturer. Run the twelve steps, tag the lug, keep the paperwork, and you have something you can lift with and defend afterward.

Conclusion: Treat Every Pad Eye Like a Lifting Device

Go back to the fabrication shop from the start of this guide. A two-ton platform, two small lugs, a forklift, and a storage rack that got hit on the way down. Nothing about that repair was complicated, but nothing about the failure was an accident. The lug did not tear out because pad eyes are hard. It tore out because nobody treated it as what it actually was: a permanent, welded-in lifting device.

That is the whole argument of this guide in one sentence, and it carries a corollary experienced riggers will confirm from the floor: the plate almost never fails, and the eye geometry almost never fails. What fails is the weld, or the thin, unsupported, badly prepped base metal it is stuck to. Most pad eye failures trace to three causes, and all three are preventable with modest discipline:

• Weld quality porosity, undercut, cold lap, or a bead laid with whatever rod was already in the oven.
• Base metal inadequacy lug welded to sheet, to a zone with no stiffener behind it, or to a surface that was never cleaned.
• Load assumptions nobody checked a lug sized by eye, rigged at an angle, or reused after a shock load.

The Checklist That Prevents Almost Every Failure

If you remember nothing else from the preceding sections, remember this sequence:

• Engineer the geometry. Hole diameter matched to the shackle pin, edge distance of at least 1.5 to 2 times the hole diameter, plate thick enough for net-section tension and bearing, or a validated standard detail someone already stamped.
• Place and prep honestly. Weld over a stiffener, column, or reinforced area, not over thin unsupported sheet. Clean the weld zone to bare metal. Preheat thick sections or cold-weather work per the AWS D1.1 tables.
• Weld all-around with low-hydrogen consumables. E7018 stick is the classic field answer, qualified FCAW the production one, with balanced passes, controlled interpass temperature, and no undercut at the toe. If your shop buys its own rod and wire, our Airgas welding supplies buyer’s guide covers what to stock, the all-position welding rod guide explains why a rod that runs flat is not automatically right for an overhead lug weld, and our notes on alloy steel welding electrodes cover the metallurgy.
• Inspect and NDE. Visual inspection plus dye penetrant or magnetic particle testing at the weld toes, especially after heavy or angled lifts and after any shock load.
• Tag it. Rated capacity and inspection date on the lug itself, recorded in a log.
• Inspect every lift and retire on damage. A cracked toe, an elongated hole, or a bent plate ends the lug’s service life, no exceptions.

Documentation Is Compliance, Not Paperwork

There is a legal dimension here that separates a professional shop from an improvised one. OSHA’s rigging rules for construction, 1926.251, and for general industry overhead cranes, 1910.179, require employers to use rated, identified, inspected lifting equipment, and the general duty clause backs that up. An unmarked, unengineered welded lug is an uncontrolled lifting device and an open OSHA exposure. More importantly, it puts whoever stands under the load at direct risk, a hazard NIOSH tracks closely in steel erection and fabrication work.

The encouraging part is the economics. A tagged, engineered, inspected pad eye costs little more than an improvised one: a short design review by a PE, the correct low-hydrogen rod instead of whatever is in the oven, an hour of NDE, a capacity tag. That trivial increment buys the elimination of the dominant failure risks. Rated capacity tags, inspection records, and lift plans are not optional paperwork around the real work. They are the work, as much as the weld itself.

Where to Go From Here

This guide is an orientation, not a substitute for the codes. When a lift matters, go to the primary sources: ASME BTH-1 for designing the device, ASME B30.26 for the shackles and hoist rings that connect to it, and the AWS standards catalog for D1.1 and the welding procedure side. OSHA’s own rigging and signal person guidance is free reading for the whole crew.

Most pad eye failures are entirely preventable. Nobody in this trade sets out to drop a load; they accept a small shortcut, once, on a light lift, and the shortcut becomes the shop’s standard practice. Treat every pad eye like the lifting device it is, engineered, welded right, inspected, and tagged, and the load comes home every time. Pull the checklist above, put it on the wall, and check your current practice against it this week.

Frequently Asked Questions

Can I weld a lifting eye directly to a steel structure?

Yes, welded lifting lugs are standard practice, but they must be designed for the load, welded with low-hydrogen consumables to a qualified procedure, and rated. An improvised lug welded on-site without a design calculation is an uncontrolled lifting device and an OSHA exposure.

What steel and electrode should I use for a pad eye?

Commonly A36 plate (36 ksi yield, 58–80 ksi tensile) for general lugs, welded with AWS E7018 low-hydrogen stick electrode or a qualified flux-cored process. Avoid non-low-hydrogen rods like 6013 for lifting welds because of hydrogen cracking risk.

How big should the fillet weld around a pad eye be?

As a rule of thumb, weld all around the contact perimeter with a leg size matched to the thinner joined part and not less than AWS D1.1 minimum fillet sizes for that thickness. Capacity is calculated on weld throat area, and undercut at the toe must be avoided.

How far from the edge should the pad eye hole be?

Keep the hole center at least about 1.5 to 2 hole diameters from the plate edge, and check against AWS D1.1 minimum edge distance tables. Insufficient edge distance lets the eye tear out under load even when the weld is sound.

Do welded pad eyes need to be load tested?

Custom below-the-hook devices are commonly proof tested at 125% of rated capacity per ASME BTH-1 practice, and the rated capacity should be documented by a qualified designer. Testing proves strength at that load; it does not replace inspection or proper design.

What is the difference between a pad eye and an eyebolt?

A pad eye is a flat plate with a hole welded to the structure; an eyebolt is a removable threaded or forged fastener. Swivel hoist rings rated per ASME B30.26 handle angular loads better than plain eyebolts, while pad eyes are permanent and welded in place.

How do I know when to retire a pad eye?

Retire and replace any lug with cracks at the weld toe, an elongated or out-of-round hole, bending or deformation, heavy corrosion section loss, or evidence of overload. Do not straighten or re-weld a deformed lug in service.

Can I weld a pad eye to galvanized or painted steel?

No. Coatings must be removed from the weld zone; welding over paint, oil, or galvanizing causes porosity, fuming (zinc oxide fume for galvanized steel), and lack of fusion. Grind or blast the weld area to bare metal first.