Conveyor Belt Cold Bonding Adhesive Problems: Causes and Fixes

Cold bonding is the quickest way to put a belt back to work. There is no press, no steam chest, no 40-hour cooldown. The price of that speed is that every shortcut in preparation reappears later, almost always as a joint that starts to fail inside the first week. This page walks through the failure in the order a fitter should approach it: the symptoms first, then the causes behind each symptom, then a field checklist, and finally the point at which cold bonding should be abandoned in favour of hot vulcanization.

The joint on a fabric belt carries the full belt tension through a glued interface. That one interface is all that stands between a running line and a torn-out splice. When a splice fails early, the cause is almost never "bad glue". It is preparation, environment, ratio, pressure, thickness, contamination, or the wrong belt for the job. The sections below take each one in turn.

Whether the belt came from a conveyor belt manufacturer or was rebuilt by a local crew, the four symptoms below are the same, and each one points at a different cause. An industrial conveyor belt is only ever as strong as its weakest splice, so it pays to learn to read the early warnings.

Cold bonding earns its keep on light- and medium-duty fabric belts that carry aggregate, sand, bagged product and similar loads. It is the default repair on a standard EP rubber conveyor belt, and on profiled covers such as a chevron conveyor belt or a sidewall conveyor belt, where a press is awkward to fit. On an abrasion resistant conveyor belt carrying sharp, heavy rock, the cover is tougher, but the splice still holds as long as the preparation is right. The failure modes that follow apply across all of them.

Send us your splice photos and belt specs and we will help you pinpoint the cause before you re-do the joint

01The Four Early-Failure Symptoms You Should Recognize First

A cold-bonded splice that is going to fail almost always announces itself within the first 50 operating hours, and often within the first 20. There are four ways it shows up. Learn to name them, because the name points straight at the cause.

Edge lifting. The splice starts to peel at one or both edges, usually at the leading edge of the joint in the running direction. The rubber cover curls up a few millimetres first, then tears back like a zipper the moment a lump of material catches it. On a 900 mm wide EP 400/3 belt in a limestone plant we measured 12 mm of edge lift inside 40 hours. The crew had ground only the middle of the step and left the outer 60 mm glossy. Once an edge lifts it does not heal; it only grows.

Center bulge. A soft raised area, or a row of small blisters, sits in the middle of the splice. It comes from trapped solvent, trapped air, or moisture that turned to vapour under load. Press the bulge and you can sometimes feel the adhesive still soft underneath, or hear a faint crackle. A bulge is a bond that never fully closed, because the two faces are separated by a gas pocket rather than by glue.

Full slippage. The splice looks intact on top but the whole joint has sheared along the bond line. The top ply has slid over the bottom ply by anything from a few millimetres to several centimetres. Sometimes you only find it because the belt's pitch marks no longer line up, or because the joint suddenly shows a step in the carrying surface. Full slippage means the adhesive never developed shear strength across the interface.

Incomplete curing. The adhesive is still tacky, soft, or rubbery days after the job. You can leave a fingerprint, or scrape a bead off with a thumbnail. This is almost always a mixing or temperature problem, and it is the easiest symptom to misread, because a tacky joint can still look finished at a glance.

02Surface Preparation and Grinding Direction

Most cold-bond failures are prepared, not glued. The adhesive does not let go; the surface it was asked to hold on to never had a proper key. A competent conveyor belt supplier will always stress preparation first, because preparation decides the joint before a drop of adhesive is mixed.

The first rule is that the adhesive bonds to a freshly buffed surface, never to the factory finish of a cover compound. A new rubber conveyor belt ships with a thin surface skin that may carry wax, bloom, or release agents from the press. Grinding removes that skin and exposes clean polymer for the adhesive to bite into.

Grind across the splice, not along it

Grinding direction matters more than most crews realise. Buff across the splice step, never along it. Grinding along the running direction leaves fine channels that behave like wicks, so solvent and adhesive flow along them and out the edges while air gets trapped inside. Grind at roughly 45 degrees to the running direction, or straight across, so the profile is random enough to hold a uniform film.

Grit, tool and the gloss test

Use a coarse cup brush or a 24–36 grit abrasive on an angle grinder, then go over it with a wire brush to lift the buffed surface. After grinding, the surface should look uniformly matt with no glossy islands. A glossy island is unbuffed skin, and it will not bond. If you can still see the original sheen anywhere, keep buffing.

Dust and solvent wipe come straight after grinding. Brush or vacuum the dust, then wipe with a clean lint-free cloth and a proper rubber solvent or the adhesive maker's own cleaner. Wipe in one direction with a fresh cloth face on every pass; wiping back and forth just smears the dust around. Let the solvent flash off completely before applying adhesive, because solvent left on the surface becomes a bond-breaker.

Step geometry matters too. A cold-bonded splice is normally a finger or bias step splice, so the step must be straight, the depth consistent, and the angle shallow enough that the joint does not create a bump in the carrying face. A ragged step of varying depth produces a varying glue line, and that is exactly where edges start to lift.

A bias step is preferred on most fabric belts because it spreads the joint across a longer line and softens the bump at the leading edge. The exact angle and finger length come from the belt maker's splice drawing, and deviating from it changes how the joint flexes over the pulleys. That flexing is a second, slower way to make the edges lift, long after the adhesive itself has cured.

On a 600 t/h lignite line we once traced an edge failure to a fine flap disc. It polished the rubber instead of keying it, and the surface looked perfect while holding nothing. A flap disc is for finishing metal, not for opening up rubber. Swap to a coarse abrasive and the same joint holds.

A cover that has been hammered in service needs a different eye than a new one. On an impact and cut resistant conveyor belt that has taken gouges from lump material, the step area may be torn too deeply to buff clean. If the rubber crumbles or the step will not cut straight, the belt needs a fresh section let in rather than a surface splice, because cold bonding cannot rebuild rubber that is no longer there.

03Temperature and Humidity Window

Cold bonding adhesives cure by solvent evaporation and chemical cross-linking, and both processes are governed by the environment. Ignore the weather and you build the failure in before the joint is even closed. The same window applies whether the belt is a fabric unit or the product line of a transmission belt manufacturer running a clean, heated shop, because chemistry does not negotiate.

Surface temperature, not air temperature

Most two-part cold bonding adhesives want a substrate between about 10°C and 40°C. Below 10°C the reaction slows sharply, and below 5°C many adhesives go dormant and stay soft. Above 40°C the solvent flashes off before the adhesive can wet the surface, leaving a dry, starved film. Measure the belt surface with an infrared thermometer rather than trusting the air, because a steel structure in a cold shed sits colder than the air suggests.

Dew point and condensation

The biggest weather killer is a film of moisture on the rubber. Condensation forms whenever the belt surface is at or below the dew point, so keep at least 3°C between the surface temperature and the dew point, and keep relative humidity below about 65–70% for most products. On a humid coastal morning the belt is often wet even when the air feels dry. Wipe it, wait, and test with a clean paper towel; if the towel comes away damp, the surface is not ready.

Humidity also enters the adhesive itself. Some adhesives are moisture-curing, which means extra humidity can skin the surface before the joint closes. If the adhesive skins over before the two faces meet, you end up gluing skin to skin instead of rubber to rubber, and the bond is weak from the start.

Open time shifts with temperature too. In heat, the window between applying adhesive and closing the joint shortens; in cold, it lengthens but the cure slows with it. Read the product data sheet for the working window and do not close the joint outside it.

We had to re-splice a tail-end joint three times in one week in a cold warehouse, where the belt surface measured 7°C and the crew kept blaming the adhesive. Warming the joint with a hot-air gun to 22°C and waiting out the dew point fixed it on the first attempt. The adhesive had never been the problem.

Site type changes the risk. A mining and quarrying face in winter runs cold and wet, while a cement plant yard can sit hot and dusty all afternoon, and the two push the adhesive outside its window in opposite directions. The remedy in both cases is the same habit: measure the surface, check the dew point, and only then open the can.

04Mixing Ratio and Maturation Time

Two-part adhesives fail the moment the mix is wrong, and the failure shows up later as incomplete curing or as a joint that is hard but brittle. Buyers who order wholesale conveyor belts and run in-house repair crews should treat mixing as the highest-risk step, because it is the one most likely to be done by eye.

Weigh, do not guess

A two-part cold bonding adhesive has a hardener ratio by weight, often around 5–6% of the base, or a stated resin-to-hardener ratio on the can. A crew that pours "a capful" per can is guessing, and a guess that lands 30% off the hardener produces a film that never cross-links. Use a small digital scale or a graduated mixing cup, and mix only the full quantity the batch can cover within its pot life.

Mix thoroughly, in one direction

Scrape the sides and bottom of the container. Unmixed hardener pools at the bottom and leaves soft spots in the joint. Stir until the colour is uniform with no streaks, which is typically two to three minutes for a hand mix. Stirring in one direction keeps excess air out of the batch.

Maturation and pot life

Many adhesives need to stand after mixing, often 5–15 minutes, before application. This maturation, or sweat-in, lets the two components reach the right viscosity and lets air bubbles escape. Applying straight after mixing is a common cause of center bulges, because those bubbles get rolled into the joint and stay there.

Once mixed, the adhesive has a finite working life, commonly 30–60 minutes at 20°C. A batch that sits too long thickens and skins, and a joint made from it will be weak. Mix only what the crew can apply and close within the window, and throw out anything that has gone past its pot life rather than thinning it or working it harder.

A splice that stayed tacky for two days turned out to be a hardener measured at 2% instead of 6%. The entire joint had to be cut out, re-ground and re-done, eight hours lost because nobody reached for the scale. That is the cheapest lesson a repair crew will ever learn, and it repeats on sites all over the world.

Keep a short log for every joint: the batch number, the ratio you weighed, the ambient temperature, and the time the joint was closed. When a splice fails three weeks later, that log turns a guess into a one-hour diagnosis. Crews that write the numbers down stop repeating the same mistake, because the log makes the mistake visible the very first time it happens.

05Pressure and Rolling Technique

Adhesive does not bond by sitting between two surfaces; it bonds by wetting both faces and then being held under pressure while the solvent leaves. If there is no pressure, or the pressure arrives at the wrong moment, the joint is weak no matter how good the glue. A local conveyor belt distributor will tell you the same thing their best fitters repeat on every call-out: the roller does half the work.

Roll from the center outward

Start at the middle of the joint and work toward the edges with a hand or pneumatic roller, overlapping each pass. Rolling from the edge inward pushes air to the center and produces exactly the center bulge described in Section 01. Work in small sections and never let an applied face skin over before the other face meets it.

Enough pressure, at the right moment

A cold-bonded splice is not held in a press, but it still needs firm, even pressure. Roll it, then go over the whole joint with a hammer or a padded block, paying special attention to the step edges. Soft hand pressure leaves a starved or gapped film. On a wide belt, have two people roll from the center line outward in opposite directions so the film closes before the open time ends.

Close the joint in one motion. Once the two prepared faces touch, the adhesive begins to grab. Peeling them apart to reposition traps air and damages the film, so dry-fit the splice before applying adhesive, mark the alignment lines, and close it once, cleanly.

Edge finishing comes last but it is not optional. Seal the exposed edges with a bead of edge sealer or a skim of the same adhesive after rolling. An unprotected edge is where moisture and dust enter, and where lift begins, so the joint is not finished until the edges are sealed.

A 1,000 mm EP 500/3 joint in a cement plant opened in the middle after three days. The fitter had rolled from the outside in "to work the air out". Rolling from the center out and re-stitching the edges solved it, and that joint is still running two seasons later. The difference was not the product, it was the direction of the roller.

Never put a fresh joint straight back under full tension. A cold bond builds strength over its cure time, and a splice that is tensioned the moment it is closed can creep before it has cured. Run the belt slow and unloaded for the first pass, keep the take-up in its service position rather than wound up, and hold off the first full load until the data-sheet cure time has passed.

06Adhesive Layer Thickness

Glue lines fail at both extremes. Too thin starves the joint, and too thick traps solvent. Getting the film right is a matter of discipline, not talent, and it is the step most often skipped when a crew is in a hurry. A conveyor belt factory will state a wet-film thickness on its data sheet precisely because both extremes are common failure modes.

Too thin

If the adhesive is spread so thin that it dries to a dry film before the joint closes, the faces never wet, and you get a weak, patchy bond that shears under load. This happens most often when a crew tries to stretch one can over too much belt, or thins the adhesive to make it go further. Thinning is almost always a mistake, because it steals the body the film needs to fill the buffed profile.

Too thick

A thick film cannot lose its solvent before the top skins over, so solvent stays trapped and later forms blisters, or the uncured interior stays soft for weeks. A thick glue line is also a soft, creep-prone layer under shear, which means the joint can slip slowly even when nothing visibly fails. Most adhesives specify a wet-film thickness around 0.8–1.0 mm, applied in two thin coats with a flash-off between them.

The standard method is a thin first coat allowed to become tacky, or to dry according to the data sheet, then a second coat, then close the joint when the second coat reaches the right tack. Skipping the flash-off between coats is a reliable way to produce bubbles, because the first coat is still giving off solvent when the second one seals it in.

Keep the film even across the full width. Use a notched spreader or a correctly sized brush, and aim for a consistent layer rather than a fast one. A thick ridge at the edges with a thin patch in the middle is a blueprint for edge lift and center bulge arriving at the same time.

Film thickness is easiest to control with the right tool. A notched spreader leaves a fixed height across the width, where a brush stroke varies with the fitter's hand. On a wide joint, two people spreading from a central line outward keep the film even and the open time short, which is why a second pair of hands is always cheaper than a second splice.

07Contamination of the Bonding Surfaces and Wrong Belt Selection

Two silent killers remain once the technique is right: what is sitting on the surface, and what the belt itself is made of. A V-belt manufacturer deals with the same two issues on a smaller scale, and the rules do not change when the belt gets wider.

Contamination you can see, and contamination you cannot

Oil, grease, coal dust, cement dust, old wax bloom, and the release agents left from a previous repair all sit on the surface and stop adhesion. A solvent wipe is not enough when the contamination is soaked in. Heavily oiled belts may need a dedicated cleaner and repeated wipes, and sometimes a light re-grind after cleaning, because grinding before cleaning just drives the oil deeper into the rubber.

The most dangerous contaminant is the one you cannot see: a thin bloom of wax or an anti-ozonant that migrates out of the cover compound. Even a new belt can carry a surface film straight from the press. Buff deep enough to remove it, then do a quick wipe test; if the cloth comes away discoloured, clean and buff again before any adhesive goes on.

When the belt body itself is the wrong choice

Cold bonding is a legitimate repair for fabric belts in the right duty, but not for every belt. A heavy steel cord conveyor belt carrying full tension is not a candidate for a cold splice, because the joint must transfer the cord tension and a glued interface cannot do it. Belts that run hot, and belts exposed to oils that attack the adhesive, will out-work any cold bond in service, so an oil resistant conveyor belt still needs a splice chemistry that can live in the same oil.

The cover compound decides whether cold bonding will key at all. Some low-energy polymers, including certain EPDM and flame-retardant covers used on a fire resistant conveyor belt, will not accept ordinary adhesives without a primer or a different chemistry. If the adhesive simply will not grab no matter how well the surface is prepared, the cover is the variable, not the crew.

Choosing the wrong belt for cold bonding is a selection error, not a gluing error. Before you blame the adhesive, confirm the belt is one that should be cold-spliced at all. The cover compound data sheet, and the original supplier's advice, will tell you what you are actually gluing to.

If the cover simply will not wet, a primer may be the missing step. Some low-energy covers need a light primer coat before the adhesive will key, and the belt maker's data sheet will say whether one is required. Skipping a required primer looks like a cleaning problem but is actually a chemistry problem, and no amount of extra buffing will fix it.

08The Field Troubleshooting Checklist and When to Switch to Hot Vulcanization

When a joint fails early, the fastest way to find the cause is to run the checklist in the same order every time, rather than arguing over which step went wrong. The table below is the order a fitter should walk the joint, because each step eliminates one family of causes before moving to the next. This is the same sequence a conveyor belt splice technician follows when they re-open a failed joint to see what it actually looks like inside.

The surrounding equipment matters as much as the glue. Misaligned conveyor rollers or a tracking fault can pull a freshly bonded splice sideways before it has finished curing, so a joint that fails on one edge is sometimes a tracking problem wearing a glue problem's clothes. A quality assurance record for the belt, together with a field service check of alignment, will often find the true cause faster than re-splicing alone.

When cold bonding must give way to hot vulcanization

Cold bonding is a field repair, and it has a ceiling. The moment the joint has to carry full cord tension, survive sustained heat, or resist oils and chemicals that attack the adhesive, cold bonding stops being a repair and becomes a gamble. The honest answer is to switch to hot vulcanization, where heat and pressure cure the splice into the belt body so the joint has the same chemistry as the rest of the belt.

Full-tension steel cord splices belong to hot vulcanization under DIN 22131, because the cords have to be laid into the splice and cured under pressure. Fabric belts running on a heat resistant conveyor belt at high cover temperature, or carrying material that attacks the glue on an chemical resistant conveyor belt, out-work a cold bond quickly. When a joint must hold under conditions like those, hot vulcanization is not the expensive option; re-doing a cold splice every few weeks is the expensive option.

The decision comes down to three questions. Is the belt under full tension or carrying hot, oily, or chemically aggressive material? Is the cover compound a polymer that rejects ordinary adhesives? Is the line critical enough that an unexpected splice failure is not acceptable? If the answer to any one of those is yes, plan for hot vulcanization, and use cold bonding only as a temporary measure to keep material moving while the press is arranged.

Cold bonding still has a place inside a hot-vulcanization plan. While the press and a certified crew are being arranged, a cold joint can keep a non-critical belt moving for a few days, provided it is watched and not loaded to full tension. The failure of that temporary joint then becomes a scheduled stop rather than an emergency, which is a far cheaper way to fail.

For the wider question of how a splice should be built, prepared and maintained across belt types, the steel cord belt maintenance guide and the conveyor belt types and grades guide cover the full picture. This article has stayed deliberately on one question: what to do when a cold-bonded joint fails, and why.

Talk to us before you re-splice — send the failure details and belt specs

09Frequently Asked Questions

Why does the edge of our splice keep peeling up a few days after we repair it?

Edge lift almost always comes back to surface preparation. If the outer edges of the step were left glossy, or the film was starved at the edge, the adhesive never bonded there in the first place. Buff the full step to a uniform matt finish, roll from the center outward, and seal the exposed edges with a bead of sealer. Once an edge lifts it will not re-stick, so cut it back and re-do the edge rather than pressing it down.

What temperature is too cold for bonding a belt on site?

Most cold bonding adhesives need a substrate above about 10°C to cure properly, and below 5°C many of them go dormant and stay soft. Measure the belt surface with an infrared thermometer, not the air, and keep the surface at least 3°C above the dew point so condensation cannot form. If the surface is cold, warm the joint with a hot-air gun before applying adhesive and hold it warm through the cure.

How do we stop air bubbles forming in the middle of a joint?

Bubbles in the middle are trapped air or solvent, and they come from three habits: rolling from the edge inward, applying adhesive the moment it is mixed, and closing the joint before the flash-off between coats. Let the mixed batch stand for its maturation time, allow each coat to flash off, and roll firmly from the center toward the edges so the air is pushed out rather than sealed in.

How long should we let the adhesive stand after mixing before we apply it?

Many two-part adhesives need a maturation, or sweat-in, time of around 5–15 minutes after mixing. This lets the components reach the right viscosity and lets air bubbles escape. Follow the data sheet for the exact figure, and respect the pot life too, which is commonly 30–60 minutes at 20°C. Applying straight after mixing or using a batch that has gone past its pot life both produce a weak film.

Can every belt be repaired with a cold joint?

No. Cold bonding works well on fabric belts in ordinary duty, but it is not for every belt. A full-tension steel cord belt, a belt running hot, or one carrying oils and chemicals that attack the adhesive will out-work a cold bond quickly. Some cover compounds, including certain low-energy polymers, will not accept ordinary adhesives without a primer. Confirm the cover compound and the duty before you decide to cold-splice at all.

When should we stop cold repairing and vulcanize instead?

Switch to hot vulcanization when the joint has to carry full cord tension, survive sustained heat, resist oils or aggressive chemicals, or when the line is too critical to risk an unexpected failure. Hot vulcanization cures the splice into the belt body under heat and pressure, so the joint has the same chemistry as the rest of the belt. Use cold bonding as a temporary hold only, while the press and the crew are arranged for a permanent splice.