Tobacco Conveyor Belts: Aroma Oil, Dust and Staining Duty

A tobacco primary line smells like a bakery for the first fifty metres and like a bonfire for the next fifty. Between those two smells sits the component almost nobody budgets for, the belt that carries lamina, stem, cut filler and finally the blended stream toward the maker. When that belt starts to look tired, the first instinct on the buyer's side is to ask for a thicker cover and a harder wear layer, because that is the answer that works on a quarry face. On tobacco it is mostly the wrong answer.

What ends these belts is colour. Casing and top dressing carry sugars, organic acids, essential oils and dyes that migrate into a plasticised PVC cover within weeks, and once the pigment sits inside the polymer no cleaning agent on the market will pull it back out. The belt still runs. It is still flexible and still strong. Nobody will sign it off for the next flavour change because it looks contaminated, and appearance is the only acceptance criterion the quality team can apply at the line. That is the failure this article is written around, and it changes what you should specify.

01Why These Belts Fail by Colour and Not by Wear

Start with the load. Threshed lamina has a low bulk density, soft edges and almost no free sharp mineral, so the abrasion resistance that governs a limestone or clinker belt simply is not the deciding property. Impact at the loading point is gentle because the drop height is short and the material is light. On a cut filler line we followed for three seasons, cover thickness fell by roughly 0.2 mm a year on the carrying side, which means a 3 mm cover had a wear life measured in more than a decade. The same belt was rejected by the flavour committee in its second year because the surface had gone from near-white to a dull amber that no amount of foam cleaning would shift.

That mismatch is what makes procurement arguments go in circles. Maintenance reports low wear and recommends keeping the belt. Quality reports staining and recommends replacement. Both are correct, and they are describing two different clocks. The wear clock is slow and predictable. The staining clock is fast, cumulative and irreversible, and it is the one that sets the real replacement interval.

Send us your line data and we will match a cover compound to your casing recipe

There is a second-order effect that buyers miss until it is expensive. A cover that has absorbed oil softens. The plasticiser that gave the compound its flexibility has partly left the polymer and been replaced by aromatic oil, so the surface hardness drops by several points and the surface energy changes. A softened surface is a tacky surface, and a tacky surface holds tobacco dust the way a lint roller holds lint. The dust layer then gets pressed into the cover by every idler and every pulley, and that pressed layer is what the quality team photographs. In other words the stain is not a surface deposit at all by the time anyone complains. It is filler embedded in a swollen matrix.

Once you accept that sequence, the specification priorities invert. Cover thickness stops being the headline number. What matters is how slowly the compound takes up oil, how completely the pigment can be kept at the surface where it can still be wiped away, and how few places the belt offers for dust to hide. A conveyor belt manufacturer who supplies both quarry and tobacco accounts will usually confirm the same thing privately, that the two duty classes need different compounds and not different thicknesses. We have argued this point with more than one industrial conveyor belt buyer who arrived with a specification copied from a cement plant, and the argument is always settled by a two-week immersion sample rather than by a datasheet.

So the working thesis of this page is short. On a tobacco line the belt is not worn out, it is dyed through. Oil-resistant, low-permeation cover compound plus an endless or low-step splice buys more service life than any increase in cover gauge, and it buys something thicker covers cannot buy at all, which is a belt that still passes a visual inspection after a flavour change.

02What Aroma Oil Actually Does Inside a PVC Cover

Casing is a water-based sauce built from sugars, licorice, cocoa, fruit concentrates and food acids, applied warm and absorbed into the lamina. Top dressing is the volatile part, carried in alcohol and sprayed late in the process so the aromatics survive. Both fractions reach the belt, one as a sticky film that dries, the other as a solvent-rich mist that condenses on every cool surface in the room, including the carrying cover and the return run.

The chemistry that follows is ordinary polymer physics and it is unforgiving. Aromatic oils are chemically similar to the plasticisers used in flexible PVC, so they do not sit on the surface. They dissolve into it. The cover swells, its hardness falls, and the plasticiser that moves out of the way carries dissolved dye with it deeper into the sheet. On a white or natural cover you can watch the boundary move inward across a season, from a surface tint to a band of colour that survives planing. Once that band is established, cleaning removes the top fraction only, and the belt looks dirty again within a shift.

We measure this before quoting, because a datasheet claim of oil resistance tells you very little at the compound level. The check we run is a simple immersion swell on a cured sample: fourteen days at the temperature the line actually reaches, with the casing diluted to the strength used on the leaf, then mass change, thickness change and Shore A hardness change recorded against an unexposed control. A compound that gains more than about three percent mass and loses more than five points of hardness in that window will stain early, whatever the label says. A wipe test with a white cotton cloth after the immersion tells you the rest, because the amount of colour that transfers is a direct proxy for how much pigment is still mobile near the surface.

The depth matters more than the area. A stain that reaches 0.2 mm can be dressed back with a light surface grind and returned to service on a dark-flavour line. A stain that reaches 0.6 mm cannot, because removing that much cover removes the abrasion reserve as well. This is why the same belt can look acceptable on a casing cylinder and unacceptable eight metres later, where the film is thinner and the cover is cooler, so condensation holds the oil in contact for longer.

Sample work of this kind is cheap and it settles arguments that no amount of correspondence can settle. Ask any serious conveyor belt supplier to send cured coupons rather than a finished belt, and insist the coupon comes from the same compound batch you will receive. If the supplier cannot produce a coupon, that is information too.

03Where the Colour Enters First: A Location Map

Staining is not uniform along a line, and the pattern is diagnostic. When a belt comes back to us with a complaint, the first thing we ask for is a photograph of the worst patch and a note of where it sits in the run. Nine times out of ten the darkest area is not the middle of the carrying span. It is an edge, a splice step, a fabric impression line or the point where the belt leaves a scraper.

Those locations share one property, which is that they hold a film of liquid against the cover instead of letting it drain or wipe away. An edge that has started to fray wicks casing into the carcass and the colour then appears from underneath, which is why retouching the surface never fixes an edge stain. A splice step of even a few tenths of a millimetre forms a shallow dam, and casing mist condensing on a cool belt collects behind that dam. A deep fabric impression does the same job at a smaller scale, holding oil in the weave pattern where a squeegee cannot reach.

The scraper line is the subtler case. A scraper that is set too hard polishes the cover and burnishes the embedded dust into the surface, so the belt looks darker exactly where it is being cleaned. Operators read that as the cleaner not working and increase pressure, which makes it worse. We have seen a line where three consecutive belts failed inspection in the same 300 mm band behind the head scraper, and the actual cause was a blade edge with a nick in it that had been pressing tobacco dust into the cover for months. A rubber conveyor belt with a soft, high-gloss cover is far more vulnerable to that burnishing than a matte compound, simply because there is nowhere for the dust to go.

Recording the pattern pays off because it tells you whether the fix is a compound change, a splice change or a hardware change. A belt that stains first along both edges needs different edge construction or better training, not a different polymer. A belt that stains first at every splice needs the splice rebuilt. A belt that stains in a band behind the scraper needs the blade checked before anything is ordered.

04Choosing the Cover Compound and the Carcass Behind It

Two decisions sit inside the cover specification and buyers often merge them. The first is the polymer system, which decides how much oil the sheet will absorb. The second is the surface finish, which decides how much of that oil stays where it can be wiped off. You can get the polymer right and the finish wrong, and the belt will still look bad, because a glossy surface holds a continuous film while a matte, fine-impression surface breaks the film up and lets the scraper do its job.

On the polymer side we work almost exclusively with nitrile-blended PVC for this duty, with the plasticiser loading kept low and the filler package adjusted so hardness stays in a range that still tracks well on small pulleys. A straight general purpose food compound is cheaper per metre and will do the job on a dry biscuit line. Put it on a casing applicator and it will take up aromatic oil until the surface goes tacky, and the tackiness is what starts the dust embedding described earlier. The page on PVC conveyor belts sets out the general construction family, and the food-specific page on food grade PVC belting supply covers the wider hygiene picture. Neither of them treats aroma oil as its own duty class, which is exactly the gap this page fills.

Colour of the cover is a specification item, not a cosmetic preference. A white or natural cover shows staining early, which sounds like a disadvantage and is in practice the opposite, because it lets the quality team see the transition before the belt becomes a contamination risk. A white rubber conveyor belt also makes it easy to see whether a cleaning change actually worked, since any tint that remains after drying is visible without a swab test. Lines that run dark compounds lose that early warning and usually discover the problem at an audit.

The carcass deserves equal attention and usually gets none. Tobacco belts run at modest tension and modest speed, so strength is rarely the constraint. What matters is dimensional stability, because a carcass that creeps will change the splice step height over time and reopen the dust trap. A two-ply polyester construction with low elongation is normally enough for a run under about forty metres, and a three-ply build earns its place on longer runs or where the belt has to survive repeated re-tensioning. Heavier carcass is not automatically better. Every extra ply stiffens the belt, which pushes you to larger pulley diameters and raises the load on every idler in the run.

Ordering discipline matters as much as the compound. A wholesale conveyor belts order that mixes two compounds across a line, because one section was bought on a different budget cycle, will produce a line where two belts age at different rates and the quality team cannot set a single colour reference. We would rather see one compound standardised across casing, cut filler and cooling than a mixed fleet with a better average specification. Buyers who consolidate like this usually find the conveyor belt factory willing to hold a compound code for them, which is worth more than a marginal difference in unit cost.

05Splice Geometry, Step Height and the Dust Trap

A joint is the only part of a belt that is made twice. On a quarry line the joint is judged by how long it lasts under tension. On a tobacco line it is judged by how much material it can hold, and the answer is uncomfortable, because a stepped splice is by design a set of ledges. Every ledge faces the direction of travel on one side and the scraper on the other, and casing mist plus tobacco dust will find both.

Step height is the number to control. In a well-built stepped splice the individual steps sit within a few tenths of a millimetre of the surrounding cover, and a straight edge laid across the joint will rock by no more than that. Once a step reaches half a millimetre it becomes a dam, and the deposit that forms behind it is not loose dust. It is dust bound with oil, which sets into a hard ridge that the scraper cannot lift and that grows with every pass. We have opened joints where the ridge stood a full millimetre proud of the cover on the return side, and the plies underneath were already discoloured through the weave.

An endless belt removes the problem at source, and for short primary processing runs it is almost always the right call. The belt is built as a loop with no joint, so there is no step, no ledge and no path from the surface into the carcass. The trade-off is installation. An endless belt has to go on with the frame partly opened, or with a pulley removed, and that work has to be planned into a shutdown rather than squeezed into a weekend. On lines where the frame cannot be opened, a finger splice finished flush is the next best option, and a conventional stepped splice is the last resort.

There is an additional argument for fewer joints that has nothing to do with dust. Every joint is a change in stiffness, and a stiffness change is where a belt wants to mistrack. On a run with four joints and a slight camber in the frame, tracking corrections fight the joints rather than the frame. Reducing four joints to one, or to none, often removes a tracking complaint that had been blamed on the idlers for years. That is one reason we ask for the joint count before discussing idler positions.

The joint also has to survive cleaning, which is where a lot of good splices are quietly ruined. A high-pressure lance held close to a joint will lift the edge of a step and open a gap that was tight when it was built. Once that gap exists the belt is on a countdown, because casing finds it on the same shift. The correct technique is lower pressure, wider fan and a deliberate pass along the joint rather than a blast across it. The same discipline appears in our guide to hygiene, washdown and splice selection, where the washdown pattern is treated as part of the belt specification rather than as a cleaning instruction.

Drive-side components matter here too, even though they are not the belt. A slipping drive heats the belt and softens the cover, which accelerates oil take-up exactly where the belt wraps the pulley. Matching the drive to the load, and keeping the transmission belt manufacturer recommendation for the drive ratio, avoids that particular spiral. We have also seen drives fitted with the wrong profile because a V-belt manufacturer catalogue from a different plant was used for the selection, and the resulting slip produced a polished, darkened pulley wrap that looked like an oil stain until the belt was removed.

When a line is being rebuilt, the joint count is one of the few things you can still influence after the compound has been chosen. A conveyor belt distributor who understands the duty will ask how the belt will be installed before quoting the splice, because the answer changes the scope of work and the shutdown plan. Anyone who quotes a splice without asking has not understood the job.

06Dry Dust, Static and the Ignition Question

Tobacco dust is fine, dry and combustible, and a belt is very good at generating static. Those two facts sit next to each other in every risk review, and they deserve to be treated separately rather than bundled into a single vague instruction to buy an antistatic belt.

The dust side is about accumulation. A belt that holds a layer on its return side, or in a joint, or on the frame beneath the loading point, is storing fuel. The layer that matters is not the visible pile in the corner. It is the thin, even film on horizontal surfaces, because a film is easy to disturb and easy to ignite. Belt design contributes to this in two ways. A rough or fabric-impressed surface holds more dust than a smooth matte one, and a joint with ledges holds more than an endless belt. Cleaning access is the third factor, and it is decided by the frame layout rather than the belt.

The static side is about where charge goes. A dry belt running over dry idlers in low humidity will build a surface charge, and the charge will sit there until it finds a path. If the frame is bonded and earthed, the path is to ground and the charge decays. If the frame is insulated by painted shims, rubber mounts or plastic components, the charge accumulates until it discharges across a gap, and a discharge in a dust layer is the event everyone is trying to avoid.

What we specify is a cover compound with a declared surface resistance in the dissipative band, measured on the finished belt rather than on a lab plaque, plus a carcass construction that gives a continuous path from the cover to the earthed frame. The number matters less than the method, because a supplier who cannot state how the reading was taken on a finished belt is quoting a wish. Ask for the reading, the humidity at which it was taken and the electrode arrangement used.

Earthing continuity across a joint is the detail that fails audits. A belt with a dissipative cover and a joint bonded with an insulating cement is two conductors joined by an insulator, and the downstream half floats. The fix is a conductive bridge across the splice, fitted during the build and checked with a meter after cure. It takes minutes and it is routinely forgotten.

07Cleaning, Changeover and the Flavour Sequence

Most tobacco plants do not run one product. They run a sequence, and the sequence is where a stained belt turns into a quality incident. A dark, heavily flavoured product leaves a film behind, the next product is light and delicate, and the first hundred metres of that run pick up whatever the belt was still holding. The belt is blamed, correctly, even though it was doing exactly what its compound allowed.

Cleaning has to be matched to the compound, and this is where operators and suppliers often talk past each other. A soft, oil-swollen cover responds badly to aggressive solvent, because the solvent removes plasticiser as well as the deposit and leaves the surface harder and more cracked than before. A properly specified nitrile blend tolerates a warm mild detergent, a soft brush and a rinse, and it does not need anything stronger. If a belt only comes clean with a solvent wipe, the compound was wrong from the start and no cleaning programme will recover it.

Drying matters as much as washing. Casing residue that stays wet on a cover keeps absorbing, so a belt cleaned at the end of a shift and left damp overnight is worse off than one that was merely wiped. Air movement across the return side does more good than another pass with a brush. On lines with a low ceiling and no extraction near the belt, we have recommended a simple air knife at the head end purely to shorten the wet time, and the effect on staining rate was visible within a month.

Changeover discipline is a management issue with a technical edge. A colour reference card, agreed between quality and maintenance, turns an argument into a check. Photograph the belt under fixed lighting at the same point every week, keep the images with the belt record, and the decision to replace becomes evidence-based rather than a matter of opinion. It also protects the maintenance team, who otherwise get blamed for a belt that has simply reached the end of its visual life.

Spares policy belongs in the same conversation. A line that keeps one spare belt of the correct compound can replace on a planned stop, while a line that orders on failure will accept whatever is available, which usually means a general purpose compound that stains faster than the belt it replaced. Holding a compound code with a single source, and checking the delivered cover hardness against the retained sample, is the cheapest form of quality control available to a plant. Our quality assurance process is built around retained samples for exactly this reason, and the service side can hold a build record so a repeat order arrives with the same compound rather than a near equivalent.

Small hardware details also decide how long a clean belt stays clean. Return idlers that are not free turning will polish the return cover and press dust into it, and a return idler with a worn bearing is a common cause of a dark stripe along one side. Training idlers set too aggressively will scuff an edge and open the wick described earlier. Neither is a belt fault, and both get reported as one.

08How This Duty Differs from Dairy, Beverage, Seafood and Egg Lines

Our food side already carries separate pages for the wet duties, and it is worth being explicit about why tobacco does not belong among them. Dairy, beverage, seafood and egg lines all share a cleaning assumption that tobacco cannot use, which is that water is freely available and that a wet belt is a clean belt. On a tobacco line water is a contaminant. Wetting the cover raises the tobacco moisture, changes the cut filler, and in the worst case feeds mould in a conveyor that never fully dries. So the entire cleaning strategy has to work with minimal water, and that changes what the cover must do.

Compare the soils. A dairy line deals with milk fat and protein, which are removed by warm alkaline circulation and which do not permanently dye a white cover. A beverage line deals with sugar syrup and water, both of which rinse away. A seafood line deals with salt, fish oil and ice melt, a combination that attacks hardware as much as rubber and that we treat separately in the article on salt, fish oil and ice melt duty. An egg line deals with shell dust and broken albumen, dry and wet in the same shift. Tobacco deals with an oil that is designed to penetrate, plus a dust that is designed to burn, and neither of those is on the list for the wet duties.

That is why the specification priorities differ so sharply. On a dairy line we argue about surface finish and cleanability, and cover thickness is almost irrelevant because the belt rarely wears. On a tobacco line we argue about permeation and splice geometry, and thickness is again almost irrelevant, but for a different reason, which is that the belt never gets the chance to wear out. On a seafood line the argument is about salt attacking the carcass and about cold stiffening. Three different conversations that all sound similar if you only read the cover hardness figure.

The egg and biscuit duties sit closer to tobacco, because they are dry, but they differ in one crucial respect. Their soils are not deliberately formulated to soak into a polymer. Casing and top dressing are. They contain surfactants, sugars and aromatic oils at concentrations chosen to bind to leaf, and a plasticised PVC cover is chemically similar enough to leaf cuticle that the same mechanism works on both. That is the sentence to keep in mind when a supplier offers the same compound for biscuits and for casing.

The practical consequence is that a plant running both a food line and a tobacco line should not standardise on one compound across both. It is tempting, because the belts look alike and the widths overlap, and it always ends with the tobacco line accepting a food compound that stains, or the food line paying for permeation resistance it does not need. Where the two duties meet, for example in a shared packing hall handling cartons from both processes, the food packaging duty is the governing one and the tobacco-specific requirements stop at the wall.

One more boundary is worth drawing. This page is about the belt on a tobacco line, not about the machines on it. Feeder selection, flavour drum design, cutters and makers each have their own engineering literature, and none of it changes the argument here, which is that the belt fails by absorption rather than by abrasion. The same logic drives the companion article on fat-coated pellets and dust load in feed milling, where the killer is carryback rather than staining, and the two conclusions differ because the soils differ.

Ask for a casing immersion sample before you specify the belt

09Frequently Asked Questions

Can a belt that has already gone amber be cleaned back to white?

Almost never, and it is worth understanding why before spending on cleaning chemicals. Once the pigment has moved below the surface, the deposit is no longer on the belt, it is in the belt. A detergent lifts what sits on the surface, which is why the first two or three washes appear to work and the colour returns a shift later. If a light surface grind takes the tint away and leaves an even finish, the belt can continue on a dark product. If the tint remains after a shallow grind, the pigment is deeper than the cover can afford to lose and replacement is the honest answer. Photograph the belt before and after the grind so the decision is documented rather than argued.

How thick should the cover be on a cut filler run?

Thickness is the least useful number on this duty, because the load does not abrade. A cover around two to three millimetres on the carrying side is normally more than enough for a soft, light material, and going thicker only stiffens the belt and loads the idlers. The number that decides service life is the compound's resistance to oil take-up, followed by the surface finish and the joint construction. If a supplier's proposal leads with a heavier cover and says nothing about permeation, the proposal is aimed at a different industry.

Is a stepped joint acceptable where casing is applied?

It is acceptable only as a compromise, and it should be a conscious one. A well-built stepped joint holds its steps within a few tenths of a millimetre and will run for a long time, but it gives oil and dust a ledge to sit on, and the ridge that forms behind that ledge is difficult to remove without damaging the cover. Where the frame can be opened, an endless belt removes the issue completely. Where it cannot, a flush finger splice is the better compromise. A mechanical fastener on a casing line is a temporary repair and should be recorded as one.

Do we need a dissipative cover if the room is air conditioned?

Lower humidity reduces the rate at which charge decays, so a controlled room does not remove the need, it changes the numbers. A dry belt on dry idlers will still generate charge, and an insulated frame will still hold it. What we ask for is a declared surface resistance measured on the finished belt at the humidity the room actually runs at, plus a bonded frame and a conductive bridge across every joint. That combination is what makes the reading meaningful. A datasheet figure measured at high humidity on a lab plaque tells you very little about a January night shift.

Can the same belt run casing and top dressing sections?

Often yes, provided the compound is chosen for the harsher of the two, which is normally the top dressing because of the alcohol carrier and the volatile oils. The sections that should not share a belt are the ones where the product changes from dark to light, because a belt that is acceptable on a heavily flavoured product can still bleed colour into a light one. Where a shared belt is unavoidable, sequence the runs from light to dark rather than the other way round, and clean between the two with a foam pass and a full dry.

How do we compare two suppliers without a laboratory?

Ask both for cured coupons from the compound they intend to supply and run the same simple test on each. Weigh the coupon, soak it for fourteen days in diluted casing at line temperature, then weigh it again and check the hardness with a durometer before and after. Wipe each with a white cloth and compare the transfer. Mass gain above roughly three percent, or a hardness loss beyond about five points, predicts early staining. It is a crude test and it ranks compounds correctly almost every time, which is all a plant needs before placing an order.