Conveyor Belts for Paper Mills: Heat, Moisture and Chemical Duty

A paper mill does not heat a belt the way a cement kiln does. There is no glowing clinker dropping onto the cover, no radiant wall three metres away, no smell of scorched rubber at shift change. What a mill does instead is quieter and, in our experience, harder to design against. It keeps the belt at 60–90 °C for twenty hours a day, wraps it in steam, drips white liquor on the edge, and lets chips full of sand and staple wire ride on top of it. The belt does not fail in a week. It fails in month fourteen, at the splice, on a Sunday, with the stock chest full.

That difference matters when you buy. A spec written for a hot clinker line will give you a cover that shrugs off 300 °C for ninety seconds and then delaminates in a wet chip conveyor within a year, because the failure mechanism you paid for is not the one that is killing you. Paper mill duty is a wet-heat duty with a chemical component. The carcass, the skim coat between plies and the splice matter more than the top cover, and they matter for reasons that only show up when you cut a used belt open and look at the layers.

Send us your line data and we will size the belt for wet-heat duty

Below we work through the four places a belt runs inside a pulp and paper mill, what the wet heat and the liquor actually do to rubber and fabric, how chips and tramp metal cut a belt lengthwise, and what to look at on a walk-round before the next shutdown. The scope is deliberately narrow. This is about the belts that carry chips, bark, pulp, broke, rolls and stacked sheets. It is not about the forming fabric or the press felt inside the paper machine, which are engineered textiles on a completely different principle, and treating them as one family with a conveyor belt is a common and expensive mistake in purchasing documents.

01Where Paper Mill Belts Actually Fail

The first indicator is not cover hardness. It is how much ply-to-ply adhesion survives a wet-heat cycle. Take two samples from the same roll, keep one at room conditions and soak the other in water at 80 °C for seven days, then run a peel test on both. A belt that reads 6 N/mm dry and 2.3 N/mm after the soak has lost more than half its inter-ply grip, and at that point the belt is being held together by the edge gum and habit. Mills that only check hardness on arrival never see this, because a cover can stay at 62 Shore A while the layer beneath it has already let go.

We pulled a four-ply EP belt from a chip conveyor feeding a digester in Fujian after nineteen months of service. The top cover still measured 58 Shore A. The cover-to-first-ply interface peeled at 1.9 N/mm against 5.8 N/mm on a stored sample from the same production batch. The belt had been running with a steam leak at the transfer point, so one 600 mm band of it had lived at close to 95 °C with condensate running down the inside. Nothing about that belt looked wrong from the walkway. It was separating in a strip you could only find by kneeling under the chute with a probe.

Why the profile differs from a clinker line

The table below is the one we use in the first meeting with a mill's maintenance team, because it settles the argument about whether an existing heat-resistant specification can simply be carried over from another plant. It rarely can. The two duties stress different parts of the construction, and the numbers that govern belt life diverge early.

Read that table with one thing in mind. The paper mill duty is not milder than the cement duty. It is slower, wetter and more chemical, and those three properties reach parts of the belt that heat alone never touches. A cover that survives 300 °C can still be the wrong cover here, because the enemy is at the ply interface and along the cut edge, not on the top surface.

For the clinker side of the comparison, our cement plant high temperature guide covers the burst-heat case in detail. What follows stays inside the mill fence.

02Four Duty Zones Inside One Mill

Nobody buys one belt for a paper mill. A typical integrated mill runs four distinct families of conveyor, and the duty in each is different enough that a single construction is a compromise somewhere. The wood yard is dirty and abrasive. The pulp side is hot, wet and alkaline. The finishing end is clean, flat and dimensionally fussy. The recovery island is a chemical environment with heat behind it. Buyers who send out one enquiry for "mill belts" usually end up with four mediocre answers.

Wood yard and chip handling

Logs, bark, chips and sawdust arrive wet from the log pond or dry from a truck dump, and the belt sees both within the same shift. Bark carries sand and grit bound into it, chips carry knots and the occasional broken chipper knife, and both carry staple wire from the baling. The belt here runs at 1.6–2.8 m/s on troughing idlers with a 35° or 45° trough angle, and the load is dumped from height, so impact is a real factor at the transfer. Cover thickness of 6–8 mm on the carrying side is normal, and a cut-resistant carcass earns its place because a 300 mm longitudinal slit from a knife fragment will stop the line just as effectively as a burn.

Bark is the nastier of the two materials. It is acidic when it starts to rot in a pile, it holds water against the cover for hours, and it packs into the skirt rubber until the belt is running against a wall of wet fibre. Mills that run bark on the same belt as clean chips tend to see edge damage first, because the packed material forces the belt off centre and the edge runs against the chute wall.

Pulp, rejects and wet broke

This is the hot zone, and it is where the wet-heat argument is won or lost. Screen rejects and knotter rejects come out at 60–80 °C with liquor still on them. Wet broke from the machine reel and from the broke pulper is heavy, cohesive and full of water, and it sticks to a smooth cover in a way that chips never do. Belts here usually run flat or in a shallow trough with a low speed, often under 1.2 m/s, and the whole conveyor sits in a steam-laden atmosphere that never dries out between shifts.

Sludge and primary clarifier cake belong in the same family. They are not hot, but they are wet, slightly alkaline and abrasive in a fine-grained way that polishes a cover rather than gouging it. A cover that has gone glossy and hard in a sludge line is telling you the compound is being leached, not worn.

Rolls, sheets and finished stacks

After the winder and the cutter, the duty changes completely. Paper rolls weighing 1.5–4 t are moved on flat belts, chain conveyors or roll conveyors, and the belt's job is to not mark the paper and not let the roll creep. Surface finish matters more than abrasion resistance, and a hard, loaded cover with embedded grit will print a mark into the bottom wrap of every roll it touches. In finishing halls the temperature is moderate, 25–35 °C, but the air is dry and full of paper dust, and the belt is expected to track within a few millimetres over a long, flat run.

Sheet stacks coming off the cutter present a different problem. They are light, they land flat, and they slide if the belt surface is too smooth or if the two side belts are not speed-matched. On a 1,600 mm cutter we have seen stacks rotate up to 15° between the cutter and the wrapper because one of the two delivery belts had stretched 0.3% more than its neighbour.

Recovery, lime and cooking surroundings

The recovery island is small in belt metres and large in risk. Lime mud, green liquor dregs and salt cake are all alkaline, all wet, and all at 70–95 °C at the point where they leave a filter or a kiln feed bin. Belts near the recovery boiler also live in an atmosphere with sulphur compounds in it, which is why the covers in this area discolour to a dull grey-brown long before they wear. A short belt, maybe 12 m long, in this service can cost more attention than a 400 m chip conveyor, simply because everything it touches is chemically active.

Buying for a mill, not for a catalogue

Mill procurement differs from quarry procurement in one important way. The belt is a small part of the capital budget and a large part of the outage risk, so the cheapest line item is rarely the cheapest decision. A conveyor belt manufacturer who builds for wet-heat duty will ask about liquor contact and steam exposure before quoting, because those two answers change the skim coat and the edge treatment. One who quotes straight off a width and a ply count is quoting a dry-duty belt with a heat-resistant label on it.

Ask for the build sheet, not just the datasheet. The useful document lists the fabric type and finish, the number of plies, the skim compound, the cover compound and gauge for each face, and the splice method the belt was designed for. A conveyor belt factory that cannot produce that sheet for a specific order is not building to a specification, it is shipping from stock. Our own quality assurance process exists because mills kept asking for exactly those numbers after delivery, and it is easier to record them at building than to reconstruct them later.

The sourcing side is more practical than it sounds. A mill with four duty zones should not hold four unrelated spare belts in the store, and a conveyor belt supplier that understands the plant can usually rationalise those zones into two constructions plus one specialist belt for the recovery area. That is what wholesale conveyor belts should mean in a mill context, not a discount on a container, but a smaller number of part numbers covering the same duties. Working through a conveyor belt distributor with a local warehouse shortens the wait when a bark line goes down at 2 a.m., and that availability is worth more than a small difference on the purchase order.

Two more notes before we leave the specification. A standard rubber conveyor belt in EP will cover the wood yard and the finishing hall well, and it is the right default for most of the plant. For the hot, wet, alkaline core of the mill you need the heat and chemical build, which is a different article even when the width and ply count look similar on paper. If the mill also runs its own machine shop or drives, an industrial conveyor belt catalogue that includes the drive side is convenient, and the transmission belt manufacturer range covers the drive belts, and the same V-belt manufacturer supplies the matched sets that keep fans, pumps and pulper drives running.

03Continuous Wet Heat, Not Peak Heat

Rubber ages on two curves that do not look alike. Dry heat at 90 °C for a year will harden a cover slowly and raise its modulus, and the belt usually survives it with cosmetic cracking. Wet heat at 90 °C for a year does something else. Water diffuses into the compound, plasticises it, and then attacks the bonds at the interfaces where the compound meets fabric or cord. The cover may stay soft while the structure underneath it lets go. That is why a mill that measures only cover hardness believes its belts are healthy right up to the morning the splice opens.

There is a second, purely mechanical effect. A belt running through a steam box at 80 °C expands, and the same belt on the return strand outside the hood contracts. Over a 200 m centre distance, a 0.15% reversible length change is 300 mm of belt moving back and forth through the take-up, every cycle, all day. The take-up absorbs it, but the splice and the pulley lagging feel it. Splice fatigue in a paper mill is very often thermal cycling dressed up as mechanical failure.

What the 60–90 °C band really does to each layer

Polyester is the standard carcass fibre for general duty, and it has a weakness that matters here. Above roughly 80 °C in the presence of alkali, polyester hydrolyses, and the loss is not recoverable. A fabric that tests at 200 N/mm width in the warp when new can lose a fifth of that over two years in a hot alkaline pulp line while showing no external sign at all. Nylon is more hydrolysis-resistant but stretches more, which pushes you back towards a steel cord or a hybrid construction for long centres. Neither choice is free.

The skim coat between plies is the layer most buyers never ask about. It is thin, it is applied during building, and it decides whether water that enters at a cut edge can travel along the ply interface. A mill-grade belt with a proper chemical-barrier skim will show localised swelling within 50 mm of the damage. A belt without one will show water tracking a metre down the length, and by the time you cut it out the whole section is scrap instead of a patch.

Steam leaks deserve their own line. A 3 mm leak at a flange in a hood is a knife. It delivers saturated steam at 100 °C plus condensate into one spot, 24 hours a day, and it will blister a cover in weeks. Every post-mortem we have done on a belt that failed "unexpectedly" in a pulp line has started with someone finding a leak that had been reported twice and deferred twice.

None of those four is a universal answer, and anyone who offers one compound for every conveyor in a mill is selling stock rather than a solution. The useful question is narrower. Which belts sit in steam, which sit in liquor, and which sit in both? Answer that and the compound choice mostly makes itself.

04Black Liquor, White Liquor and Splash Chemistry

Black liquor is the mill's own problem child. It runs at pH 12–13, carries 12–18% dissolved solids, sits at 70–85 °C in the weak liquor circuit, and finds its way onto belts through gland leaks, sample points, tank overflows and the simple habit of hosing a spill towards the nearest drain. It is also a surfactant. Once it wets a rubber surface it keeps spreading, and it carries fine fibre with it into every crack.

White liquor is worse in a different way. It is caustic and sulphidic, pH 13–14, and a splash that dries on a cover leaves a caustic crust that re-wets every time the humidity rises. That crust keeps working on the rubber between shifts. Mills often blame the heat for edge damage that is actually a caustic burn under a dried deposit.

Three ways the chemistry gets into the belt

The first route is the cut edge. A belt is not sealed at the edge unless someone specified it, and an open fabric edge is a wick. Liquor enters, travels along the warp, and swells the fabric, which then pushes the plies apart from the inside. A 40 mm edge split becomes a 400 mm delamination because the fabric did the transport, not the cover.

The second route is the splice. A vulcanised splice has more interface area than any other part of the belt, and if the step edges are not sealed with cover stock, water sits in the step seams. On a belt that runs wet, we have measured splice adhesion falling to 40% of the body value within 14 months, with the splice still physically intact when inspected from the walkway.

The third route is through the cover itself, slowly. No rubber is fully impermeable to hot alkaline water. The question is how fast the transport happens, and that is where compound formulation and cover thickness both count. A 4 mm cover on a pulp conveyor is not enough material to slow the process down. Six millimetres is the minimum we will quote for continuous wet duty, and 8 mm where there is splash.

There is a purchasing consequence hidden in all this. Mills tend to over-specify chemistry resistance across the whole plant after one bad experience on the recovery side, then wonder why the finishing hall belts mark the paper. Resistance to strong alkali usually comes with a harder, more highly filled compound, and a harder compound transfers grit and prints. Specifying by zone is cheaper and works better than specifying the whole mill to the worst duty.

Where the chemical load genuinely dominates, our chemical resistant conveyor belt range is built around barrier compounds and sealed edges rather than a single upgraded cover. Where the contaminant is tall oil, grease or lubricating oil rather than liquor, the oil resistant conveyor belt family is the right starting point, since a nitrile blend will handle the oil and still tolerate moderate alkali.

05Chips, Sand and Tramp Metal

Chips arriving from a chipper are not a clean bulk material. They come with sand from the log yard, grit from the truck bed, knots that never broke down, and metal. The metal is the expensive part. Staple wire from bales, broken chipper knife fragments, a piece of wear plate, a hand tool left on the belt during a repair. None of it is common. All of it is inevitable over a five-year window on a busy line.

When metal gets trapped, it does not punch a hole. It gets dragged under the skirt board or caught at the loading point, and it cuts lengthwise, following the direction of travel. A 20 mm fragment can produce a 900 mm slit in under a minute because the belt itself does the cutting, running the trapped object against a stationary edge like a blade on a slitting machine. Two of these in a week and you have a belt that needs replacing rather than repairing.

What actually resists a lengthwise cut

Cover thickness helps less than people expect. A 10 mm cover of soft rubber parts easily in front of a trapped knife fragment. What resists is the fabric. A cut-resistant carcass uses a denser, higher-tenacity warp, sometimes with a breaker layer, so the slit stops within a few hundred millimetres instead of running the full length. The difference in a field repair is the difference between a 300 mm patch and a 4 m splice, and on a chip conveyor that is the difference between a night shift and a two-day outage.

Skirt board geometry matters just as much as the belt. The gap between the skirt rubber and the belt should be set so that a chip cannot wedge, but not so tight that the rubber wears the cover. On a 1,200 mm belt we would expect a 5–8 mm gap at the loading zone with the skirt rubber just touching, and a gradual flare towards the discharge end. Too many mills run a straight, tight skirt for the full length and then blame the belt for edge wear.

Inclines, cleats and the cleaning penalty

Chip bins and bark silos are usually fed on an incline somewhere between 16° and 24°, which is beyond what a smooth belt will hold with wet, rounded chips. Chevron cleats solve the carry problem and create two new ones. The cleat roots trap material, and in a wet chip line that material turns into a growing ridge of compacted fibre that eventually rubs the cover. Cleaning a chevron belt properly needs a rotating brush plus water, not a scraper, and a scraper on cleats will simply be destroyed in a month.

Our advice on inclines is to first try a steeper trough and a slower speed before adding cleats. A 45° trough at 1.8 m/s will hold more on a 16° incline than most people assume, and a plain belt is far easier to keep clean in a wet mill.

The tracking problem in a chip line usually starts with loading, not with the belt. Off-centre feed pushes the belt sideways, the belt edge runs against the chute, and the edge is where water and liquor get in. Self-aligning idlers help, but they correct a symptom that a moved chute lip would have prevented. We would rather see a properly centred loading point than a belt line fitted with six training idlers compensating for a chute that has been wrong since commissioning.

Cut resistance is not the same property as abrasion resistance, and the two are often traded against each other in compound design. A belt specified for abrasion will resist slow wear and still split open on a wire. If your mill has a history of lengthwise damage, say so at enquiry stage. It changes the fabric, not just the cover, and it changes the answer to the question of what to keep in the store as a spare.

06Stretch, Take-Up and Long Centres

Paper mills are laid out around buildings that already exist, so belts run long and often turn corners that a greenfield design would have avoided. Two hundred metres is ordinary. Four hundred metres happens when a mill adds a chip storage hall at the far end of the site and threads a conveyor back to the digester. At that length, elongation stops being a detail and becomes the constraint that decides the construction.

Here is the arithmetic that gets missed. An EP belt running at 60% of its rated tension will show roughly 0.4–0.6% elastic stretch. On a 240 m centre distance that is 960–1,440 mm of belt that has to be absorbed by the take-up before you add installation slack, thermal movement and permanent elongation from wet service. A screw take-up with 800 mm of stroke looks generous on a drawing and is out of travel in the first year. We have walked into mills where the counterweight was sitting on its stops with the belt visibly slack on the return strand, and the maintenance team was convinced the belt had "stretched beyond spec" when the take-up had simply been sized for a shorter, drier conveyor.

Permanent elongation in wet heat is real, though, and it is larger than in dry service. A wet EP belt can accumulate 0.3–0.5% irreversible growth over two to three years, on top of the elastic component. That is the reason mills re-splice a belt two or three times in its life, cutting out a metre each time. If the take-up has no reserve for those cuts, the third re-splice becomes a new belt.

When steel cord earns its place

Above roughly 350–400 m centres, or where the tension per millimetre of width climbs past what a four-ply EP can carry, steel cord becomes the sensible choice. Its working elongation is around 0.2–0.25%, so the take-up problem shrinks dramatically, and it holds tension far better through thermal cycles. The trade-offs are real too. Steel cord belts need larger pulley diameters, they do not tolerate a badly aligned splice at all, and a wet edge that lets water reach the cords will corrode them from the inside, which is a failure you cannot see until the belt breaks.

For anything under 250 m with moderate tension, a well-built EP carcass with a properly sized take-up is simpler to maintain and easier to repair in the field. The decision belongs in the layout stage, not in the spare-parts catalogue. Our steel cord conveyor belt pages set out the tension and pulley relationships, and the maintenance article on steel cord belt inspection and failure prevention covers the edge-corrosion check that matters most in a wet mill.

Pulleys are part of the same conversation. A drive pulley with worn lagging will slip before it reaches the tension the belt was designed for, and slip in a wet environment polishes the pulley and glazes the belt underside. Ceramic lagging grips in wet conditions but is unforgiving on the splice, and a diamond-groove rubber lagging is usually the better balance on a belt with vulcanised joints. Check the lagging at the same time as the take-up, because a slipping drive and a stretched belt present with almost identical symptoms.

Mill drives are not all belt-driven conveyors, and the auxiliary side is easy to forget. Fan drives, pump drives, pulper drives and the conveyor gearbox input all run on V-belts, and a mill that standardises its V-belt sizes across the site keeps fewer spares and swaps them faster. That is a separate purchasing line from the conveyor belting and belongs with the V-belts for mill drives conversation rather than the conveyor enquiry.

07Splices in a Wet-Heat Cycle

Ask a mill maintenance team where their belts fail and nine out of ten will point at the joint. The number is not superstition. A vulcanised splice is a hand-built section of belt with a stepped interface, and every step is a place where water can sit and where the ply interface is at an angle to the load. In dry service a good splice reaches 85–90% of body strength and stays there. In wet-heat service the same splice starts losing adhesion from the first month, because the moisture finds the step seams before it finds anything else.

The single biggest cause of early splice failure that we see is moisture in the fabric at the moment of splicing. A belt taken out of a wet pulp line and spliced the same day is damp inside, and no amount of surface drying changes that. Vulcanising damp fabric traps water in the splice, and the steam generated during cure creates micro-voids along the step edges. The joint looks perfect. It fails in five to nine months, usually at the edge of the first step.

We watched this play out on a 1,400 mm screen reject conveyor. The splice was made in a four-hour window with the belt still at 68 °C and visibly wet. It reached 61% of body strength when we tested a cut-out eight months later, and the failure surface was full of small voids. The replacement splice on the same belt, made over a full shift with the belt cooled and dried under forced air for six hours before the first step was cut, was still above 80% at the two-year mark. Same belt, same crew, same materials. The difference was six hours of drying.

Getting a splice to hold in a wet mill

Cool it first. A belt at 70 °C will not take a proper splice, because the rubber is soft, the solvent in the cement flashes off too fast, and the cure cycle fights the residual heat. We want the belt below 35 °C at the splice zone, which on a long line means planning the job around a shutdown rather than a lunch break.

Dry it properly. Forced air through the opened steps for at least four hours, longer if the belt has been running wet for months. A moisture check with a cheap pin meter on the fabric is worth more than any amount of experience, and it takes two minutes.

Seal the step edges. Cover stock should be used to fill the step seams and the splice edges, not just the top surface. This is the detail that separates a splice that survives a wet mill from one that does not, and it is the first thing skipped when the crew is under time pressure.

Then leave it alone. A splice needs to cool under restraint before the belt is tensioned and run. Running a hot splice at full tension pulls the steps apart while the bond is still developing. Twelve hours of restraint is not excessive on a critical belt.

One more field note that costs mills real time. A splice that has started to take water will usually show a dark line along the step long before it fails, and that line is visible from underneath the belt on the return strand. Add it to the walk-round. It takes ten seconds per belt and it is the cheapest early warning in the plant.

08Inspection, Washdown and Shutdown Practice

Everything above describes how a belt dies. This section is about catching it early, and about the two routines that do most of the catching in a mill that runs well. The first is a fifteen-minute walk-round on the wet side of the plant. The second is a washdown procedure that does not make things worse, which is more than can be said for the way many mills hose a belt down.

Start the walk-round at the drive and work backwards against the direction of travel, because that is the order in which damage reveals itself. Look at the pulley lagging for polishing and at the take-up for stroke remaining. If the take-up is past 70% of its travel on a belt that is less than two years old, the problem is elongation, not adjustment, and tightening it further only buys a few weeks. Then walk the return strand under the hood with a torch, where the underside of the belt is visible and where water sits longest.

The measurements that matter are few. Cover thickness at four fixed points across the width, marked with paint so the same points are read every quarter. Splice step condition, recorded photographically. Edge thickness against the new-belt value, because an edge that has swollen 2 mm is holding water. And cover hardness at the discharge end rather than the drive end, since the hot end reads differently and mixing the two locations produces meaningless trends.

Washdown that helps rather than hurts

Water is the right cleaning medium in a paper mill and the wrong one if it is applied badly. Cold water on a belt at 80 °C creates a thermal shock at the cover surface, and repeated shock is how crazing starts. Use water at or near the belt temperature where the process allows, or wash down at the end of a shift when the line has cooled but before the liquor dries.

Never aim a jet at a splice. A high-pressure lance held on one spot will lift a step edge in seconds, and that damage is invisible until the splice opens. Keep the lance moving, keep it at least 300 mm from the surface, and accept that a belt does not need to be spotless, only free of packed material. Mills that chase a clean belt with a high-pressure gun at close range spend more on belting than mills that clean thoroughly but gently.

Drainage is the part that gets designed once and then forgotten. Water on the return strand runs to the lowest point, and if that lowest point is a flat return idler with a worn cover, the water sits there and wicks into the edge. A slight incline on the return run, a V-return idler set at the low point, and a plough before the tail pulley will do more for belt life than an upgrade in compound. Do not drill drain holes in a fabric belt to solve this. The holes cut the cords, and the water gets in anyway.

Storage deserves a line because mills buy ahead. Belts stored in a room next to a compressor, on a concrete floor, in a coil standing on its edge, will age before they are fitted. Keep them on a pallet, out of sunlight, away from electric motors that generate ozone, and rotate stock so the oldest belt is used first. A belt that has sat for four years can arrive with a cover that is already harder than the one it replaces.

When a belt is due for replacement, the useful conversation is about the next construction rather than the current failure. Bring the cut-out section to the review. A 300 mm piece of the failed belt tells an experienced heat resistant conveyor belt supplier more than a page of description, because the failure surface shows whether the water came from the edge, the splice or the cover. Our factory test reports are written to be compared against that kind of evidence, sample by sample, rather than as a certificate to file.

Two side notes that come up in almost every mill review. Pulley condition is inseparable from belt life, so look at conveyor pulleys and their lagging at the same time as the belt. And the idler side carries the water, so troughing geometry and the choice between a troughing idler and a return idler with a V-set changes where the water collects. Where tracking is the recurring complaint, a training idler is worth fitting, but only after the loading point has been checked.

For the temperature-designation question that always arrives with a heat enquiry, we keep the explanation separate rather than repeating grade codes here, and the cover grade and temperature class guide explains how the designations map onto real service conditions. Mills in hot climates also tend to run belts harder in summer, which is why the summer operation notes are worth reading alongside this article. Neither replaces the wet-heat discussion above, and neither is a substitute for telling us the liquor contact and the steam exposure at enquiry stage.

Ask for a wet-heat belt recommendation for your mill

09Frequently Asked Questions

Why do belts in a mill fail at the joint instead of in the middle?

Because the joint is where water gets in first. A vulcanised splice has step seams running at an angle to the load, and in a steam-laden building those seams stay wet. Adhesion falls from around 85% of body strength towards 50% over a year while the outside still looks sound. The joint also carries every thermal cycle, since a belt that expands and contracts 300 mm over a 200 m run moves most at its weakest section.

Can we use an ordinary heat-resistant belt on the pulp side?

Usually not, and the reason is not the temperature rating. A cover built for dry heat can survive 90 °C for years and still fail in a wet alkaline line within months, because the damage happens at the ply interface and along the cut edge. What you need on that duty is an alkali-resistant skim, a sealed edge and enough cover thickness to lose material without exposing fabric. That is a different build, even when the width and ply count are identical.

How do we keep water out of the plies?

Three things, in order of effect. Seal the cut edges at every splice so the fabric is not an open wick. Keep the belt tracking so the edge is not abraded against a chute. And fix leaks, because a steam or liquor drip in one spot will defeat any construction eventually. Cover thickness helps slow moisture transport, but it will not stop a wetted edge from carrying water a metre along the belt.

Is a mechanical fastener acceptable as a stop-gap repair?

For a few days on a low-tension finishing belt, yes. On a pulp or chip line it is a short-lived fix that makes the next proper splice harder. The bolt holes give liquor direct access to the fabric, the plates compress the cover, and the joint loosens as the belt flexes. We would treat it as a way to reach the next planned shutdown, not as a repair.

How much take-up travel should be left in reserve?

Enough for elastic stretch plus two future re-splices. On a 240 m run, 0.4–0.6% elastic stretch alone consumes close to a metre, and wet service adds 0.3–0.5% permanent growth over two to three years. If the counterweight is sitting on its stops, adding tension is not the answer. Either shorten the belt properly or accept a shorter service interval.

What should a weekly walk-round actually cover?

The drive and take-up, the loading point, the return strand under the hood, both edges and every splice. Look for polishing on the lagging, stroke remaining on the take-up, packed material at the skirt, dark lines at splice steps, and swelling at the edges. Fifteen minutes with a torch and a gauge beats an annual inspection that arrives after the failure.