Feed Mill Conveyor Belts: Fat-Coated Pellets and Dust Load

01The Return Strand, Not the Top Cover, Sets the Service Life

A feed mill conveyor belt usually reaches the end of its life with most of its top cover still in place. We have walked up to belts on pellet lines where the carrying cover still measured 5.4 mm of an original 6 mm after fourteen months of running, and the belt was being booked for replacement that same week. The edge was frayed for two metres behind the tail pulley, the splice had opened at one corner, and the belt had begun to climb the side of the take-up frame on every start. Nothing had cut the top cover. The damage had arrived from underneath, on the return strand, in a place nobody photographs during an inspection.

The real killer on a pellet line is carryback, not carrying-surface wear. Pellets that have been sprayed with fat or molasses leave a thin tacky film on the cover at the discharge. That film does not dry in a feed mill, because a feed mill is a dry plant with warm air and no washdown, so the film simply stays there and collects the fines cloud that every pellet line produces. Layer by layer it builds into a mat, the mat gets pressed flat against return rolls and pulley shells, and the organic dust inside it cements into something that behaves like a filled plastic. What forms on the tail pulley is no longer a sticky deposit. It is a hard sleeve with a step at each edge, and a hard sleeve running against a rubber cover is a grinding wheel running against a rubber cover.

Send us your pellet type, coating level and belt layout for a carryback review

Once that mechanism is accepted, the order of priorities in a specification changes. A thicker cover buys time against abrasion on the top surface, which is the surface that is not failing. A cleaner that matches the texture of the cover attacks the film at the point where it is thinnest and softest, which is the only point where it can still be removed. That is why cleaner selection and cover roughness matter more than cover gauge on this duty, and why a 4 mm cover with the right surface and the right blade outlasts a 6 mm cover that is being scraped by a blade the compound was never matched to.

Most requests we receive ask for cover thickness, cover grade and ply count, in that order. The three numbers that would actually predict the outcome are rarely in the enquiry at all. How many grams of material come back on the return strand per hour, measured rather than guessed. How much sleeve has grown on the tail pulley since the last shutdown. Which cleaner is fitted, at what angle, and with what blade hardness. A buyer who supplies those three numbers gets a different belt, and usually a thinner one.

One line we were called to ran 30 t/h of 4 mm broiler pellets on a 1,000 mm belt at 2.2 m/s, with 2.5% poultry fat applied in a drum coater. Management had already approved a heavier belt, because the previous one had "worn out early". We put a tray under the head pulley for a shift and weighed what came off the return strand: 4.1 kg per hour, most of it fines bound into greasy crumbs. The tail pulley had 6 mm of sleeve at the centre and almost nothing at the edges, so the effective crown had grown to roughly 1.9% of diameter, which is more than enough to steer a belt. The top cover was fine. The belt was being consumed from the pulley side, and the heavier belt they had ordered would have been consumed in exactly the same way.

Cover gauge would not have changed a single number in that paragraph.

What follows stays on the belt and on the surfaces it touches. Pulley supply, idler selection and frame fabrication each get a section only where they interact with carryback, because a belt cannot be specified in isolation from the parts that scrape it. For the wider construction picture, the conveyor belt manufacturer overview explains how ply count, cover compound and tensile grade are put together, and the range of industrial conveyor belt constructions shows where a pellet duty sits against heavier bulk duties. What this page adds is the failure sequence specific to coated pellets, which is not visible in any catalogue table.

It is also worth saying at the start that the belt is rarely the only thing being worn. On the same line, the drive guard held two belts from the same transmission belt manufacturer, both changed twice in a year, and the grooves were packed solid with the same fat and dust mixture that was building on the pulley. A V-belt manufacturer will tell you that a packed groove slips and burns, and that is what had happened. The dust does not respect the boundary between the conveying side and the drive side of a mill.

02What a Fat or Molasses Film Does After the Cooler

A pellet leaves the die at 80 to 90°C with a glossy, semi-plastic surface, and it does not reach a belt in that state. It goes through a counterflow cooler first, where ambient air pulls the temperature down to within a few degrees of the room and the moisture down to around 12%. What arrives at the coater is a hard, dense, slightly porous pellet at 55 to 65°C. Fat is then sprayed at 40 to 55°C, and in a vacuum coater the pressure drop pulls the liquid into the surface pores before the pellet is released. Molasses is applied in a drum, at a lower solids content and a higher viscosity, and it stays on the outside almost entirely.

The result is a pellet that behaves like a dry solid in a bin and like a wet solid on a belt. Surface loading is usually between 1% and 3% by weight for fat and between 1.5% and 3% for molasses, and either way a film of 0.1 to 0.3 mm sits on the outside of every pellet. Fat crystallises as it cools, but crystallising is not drying. At a mill internal temperature of 28 to 34°C in summer, the film is greasy rather than solid, and it transfers to the cover on contact under the weight of the bed. Molasses is worse in one specific way, because it is hygroscopic and pulls moisture out of the air, so it can stay liquid for the whole run and re-wet itself overnight.

Belt temperature matters here and is almost never discussed. Directly after the cooler, a belt carrying 60°C pellets on a 30°C day will sit at 40 to 50°C on its carrying surface for the first twenty or thirty metres. A warm cover holds a fat film in a softer state for longer, so the film spreads further and thinner, and a thin film is harder to scrape than a thick one because there is nothing for the blade to catch. Where pellets leave the cooler above 70°C, a standard cover compound starts to lose hardness and the surface polish changes, which is a separate argument for a heat resistant belt on that first flight rather than a general-purpose one.

None of these films is water-soluble in the way plant operators expect. Fat needs warm water and a surfactant to lift, and molasses needs enough water to dilute the sugar before it can be rinsed. A feed mill has no washdown, by design, because water in a feed mill is a microbiological and a structural problem. So the film is never removed by cleaning in the way a food plant would remove it. It is only ever removed mechanically, at the head pulley, by a blade, at the one moment when it is thin and warm. That single fact explains most of the belt failures on this duty.

By the time the belt reaches the head pulley, the deposit is no longer a film. It is a two-layer structure with grease underneath and dust on top, and the blade meets the dust layer rather than the grease. That is the moment a cleaning problem becomes a hardening problem.

03From Tacky Film to Hard Sleeve: The Carryback Sequence

The sequence has four stages and every one of them is visible if you know where to look. Stage one is transfer, where the film leaves the pellet and settles into the microscopic valleys of the cover. Stage two is capture, where the fines cloud produced at the discharge and at every transfer point lands on that film and is held there. Stage three is compaction, where the loaded belt passes under the head pulley and over the first return idler and the mixture is pressed into a continuous mat with the air squeezed out of it. Stage four is hardening, and it is the only stage that changes the physics of the problem.

Hardening is a combination of three things working at once, which is why it is so reliable. Fat oxidises and cross-links at the surface, turning from grease into a varnish-like skin. Moisture moves in and out of the layer as the mill warms and cools each day, and every cycle draws the soluble fraction to the surface and leaves it there as a binder. And the pressure at the pulley nip does the rest, because 4 to 6 N/mm of belt tension acting through a 400 mm pulley is enough to compact a fat and dust mixture into something with the density of a hard rubber sheet. Add the mineral fraction that comes from limestone, phosphate and premix dust and the result is a composite with embedded grit. It is, functionally, a layer of abrasive bonded onto the pulley.

This is why the wear pattern is so diagnostic. When a belt is worn by its cargo, the top cover thins in the centre of the trough and the fabric impression disappears first. When a belt is worn by carryback, the top cover is intact and the underside is polished, with a glazed band running the full width where the pulley contacts it. The back cover thins, the edges feather because the sleeve step is steering the belt into the structure, and the splice wears at its leading edge from being dragged over the sleeve rather than rolling over a clean shell. Any one of those signs on its own is worth a look. All three together are a carryback problem, and no change of cover gauge will touch it.

Stages overlap and speed up together. A line running molasses in humid weather can compress the whole table into three weeks.

We measured the whole sequence once on a 20 t/h layer line that had been through two belts in a year. The pulley sleeves were 8 mm thick at the centre of the tail drum and the belt had 2 mm of back cover left against an original 2 mm, which is to say the cover had gone and the carcass was in contact with the sleeve. The maintenance team had been checking the top cover at every shutdown for a year and had recorded almost no change, which is exactly what you would expect. Their inspection was looking at the wrong surface, and the record they kept was proof of it.

Reading the wear from the pulley side changes what an inspection routine should contain. Back cover thickness at a marked station, sleeve thickness at the pulley centre and edges, and the weight of material collected under the head are three numbers that take twenty minutes to gather and that predict a failure months before the edge starts to fray. A wear resistant cover on the pulley side helps once the sleeve is under control, and does very little while the sleeve is still being allowed to form. The pulley shell and lagging sit on the receiving end of the same mechanism.

04Matching the Cleaner to the Cover Surface, Before Choosing Gauge

A cleaner removes material by shear, and shear needs an edge to work against. On a dry belt the cover provides that edge through its own texture, and almost any blade at almost any pressure will do a reasonable job. On a coated pellet belt the film fills the texture, the surface becomes effectively smooth, and the blade is trying to shear a liquid against a liquid. That is the point where the blade either lifts the film or burnishes it, which is decided by blade hardness, tip geometry and cover roughness.

Cover finish is a manufacturing choice that buyers rarely specify. A press sheet with a polished platen gives a glossy surface that looks premium and is the worst possible finish for a fat film, because there is nothing for the blade to dig into and the film spreads to a uniform thickness that the tip simply glides over. A fine fabric impression, somewhere between 0.1 and 0.3 mm deep, gives the blade hundreds of small edges along every metre and lets it start a peel. A coarse impression of 0.5 mm or more goes too far the other way, because the valleys fill with compacted material that no blade can reach, and the belt then carries a permanent grinding paste in its own texture.

Blade hardness has to be chosen against that finish rather than against the material. A soft polyurethane tip in the 75 to 85 Shore A range conforms to a fine impression and keeps contact across the full width, which is what you want on a tacky film. A harder 90 Shore A or carbide-tipped blade holds a sharp edge longer on abrasive bulk but chatters on a warm cover and leaves a pattern of polished stripes where the film has been pressed flat instead of removed. Only one of them is wrong for a given cover.

Setting matters as much as the hardware, and it is where most installations we inspect have drifted. The primary blade should contact the belt as it leaves the head pulley, before the material has separated, with an attack angle in the 15 to 30° range and just enough pressure to keep the tip closed across the width. More pressure does not remove more film on a tacky surface. It generates heat, softens the film further, and wears the cover under the blade, which is the one place you cannot afford to lose rubber. A 0.5 mm feeler gauge run along the tip tells you more about the setting than any spring scale, because it shows whether contact is even across the width or concentrated at one end.

The column that matters is the last one. Every mismatched pairing has a signature, and the signature is visible on the belt before it is visible on the blade.

A secondary cleaner is not optional on this duty, and it is the part most often left out when a mill buys a conveyor. The primary blade removes the bulk of the film at the head. What it leaves is a thin residue, maybe 20 to 40 microns, which is invisible on a walk-past and is exactly the layer that captures the first dust. A secondary blade mounted below and behind the primary, contacting the belt as it comes off the pulley, takes that residue off before the return strand starts. Two blades in series cost less than one extra millimetre of cover, and on a coated pellet line they remove several times as much material over a year.

Where the belt is supplied as part of a package, we would rather quote the cleaner with the belt than have the mill fit whatever was in the stores. The two have to be chosen together, which is why the enquiry we prefer describes the finish as well as the gauge. Any conveyor belt supplier can ship a 6 mm cover, and a supplier who also stocks wholesale conveyor belts for feed and grain accounts will usually have the cleaner sitting on the same shelf. Asking for both in one line item is how the match gets made.

05Pellets Are Not Grain: Where This Page Stops and the Grain, Fertiliser and Silage Pages Begin

This site already carries a guide to grain, fertiliser and silage duty, and a separate treatment of fertiliser handling, corrosion and caking. Both are the right place to start if your conveyor is moving wheat, maize, soya meal, urea prills or grass silage. This page deliberately does not repeat them. What follows applies to one product family only, which is compound feed that has been pressed into pellets and then coated, and the reason for keeping it separate is that the coating changes the failure mechanism rather than just the severity of it.

Clean grain is a dry solid with a dry surface. It flows, it does not stick to rubber in any meaningful way, and the dust it produces is dry enough to be blown off with an air lance or dislodged by a brush. A grain belt can run for years with a single primary cleaner and no secondary, and the return strand stays visibly clean even when the belt is near the end of its life. Silage is the opposite extreme, wet and acidic, and the belt has to survive moisture and organic acid rather than a film. Fertiliser sits in a third place, where the material itself is hygroscopic and caking is driven by moisture migration and by the chemistry of the prill, so the belt is chosen for chemical resistance and the cleaning regime is built around washdown.

A coated pellet is none of those. Its core is as dry as grain, around 12% moisture, and its surface is as sticky as silage. That combination is what makes it difficult, because every habit a mill has learned from grain tells the operator that dry product on a dry belt needs no special cleaning. The film is thin, it is the same colour as the product, and it does not look like a problem on a walk-past. It becomes a problem three weeks later on a pulley shell, by which time the material that caused it has been eaten.

The absence of washdown is the specific trap. A fertiliser plant can hose a transfer house down, and the guide linked above treats washdown as part of the routine. A feed mill cannot, because water in a feed mill creates a microbiological risk in the product and a corrosion risk in the structure, and most mills are designed so that hose points do not exist in the pellet transfer area at all. So the cleaning method that would dissolve a molasses film is unavailable by design, and the only removal mechanism left is the one at the head pulley.

Read across the middle row and the problem is obvious. The pellet line has a wet-surface failure with a dry-plant cleaning regime.

There is a practical consequence for how a mill splits its belt stock. The intake conveyors that bring raw grain and meal into the plant are genuinely grain duty, and a standard rubber conveyor belt with an abrasion-resistant cover is the correct and economical answer there. The pellet transfer conveyors after the cooler and coater are a different duty and need a different surface, a different cleaner and a different inspection interval. Mills that standardise on one belt specification for the whole plant almost always standardise on the intake specification, because that is where most of the metres are, and the pellet lines then run on a belt chosen for a job they are not doing.

Where the same conveyor handles both, which happens on lines that can bypass the coater, the belt has to satisfy the harder of the two duties. That usually means a cover compound that resists oil and fat rather than a general-purpose one, since a standard compound swells slightly and softens when it is in constant contact with fat, and a softened cover holds the film more aggressively and releases it less readily to the blade. An oil resistant conveyor belt is not a luxury on a coated pellet line. It is the difference between a surface that sheds the film and one that absorbs it.

The same logic runs in a different direction on tobacco conveyor belts, where the oil is designed into the product rather than added afterwards, and the cover is lost to staining instead of to carryback. Read the two together and the useful question stops being how thick the cover is and becomes which contaminant the compound was formulated to refuse.

06Inclines, Chevrons and the Re-Load at the Tail

Every belt that leaves our conveyor belt factory carries a declared surface finish, and on a feed mill order that declaration matters more than the tensile grade. The usual reason a mill changes from a plain belt to a profiled one is slip on an incline, and the usual angle at which 4 mm pellets begin to roll back on a smooth cover is somewhere between 10 and 13°, depending on how much fat is on the surface and how much the belt is vibrating. Above that angle the pellets do not travel, they tumble, and a tumbling pellet abrades both itself and the cover. Your conveyor belt distributor will normally see this first as a fines complaint rather than a wear complaint, because the visible symptom is a growing pile of crumbs at the tail.

The obvious fix is a chevron conveyor belt, and it works, but it creates the exact problem this page is about. A chevron profile is a set of valleys running across the belt, and a valley is a place where a fat film can sit undisturbed by any blade. On a plain belt the cleaner has one continuous surface to work on. On a chevron belt it has a surface interrupted every 200 to 400 mm by a 10 to 25 mm wall, and no scraper blade can enter the corner where the wall meets the base. The material that collects there is compacted by every pellet that passes over it, and after a few weeks the profile is effectively filled and the incline has lost the grip it was bought for.

Profile selection is therefore a compromise between grip and cleanability, and the compromise has a shape. Open V profiles with a wide base and a modest cleat height of 10 to 15 mm shed material better than tall, narrow, undercut profiles, because the corner radius is larger and the brush can reach further into it. Interrupted patterns leave a clear lane along each edge, which lets the belt track without the cleat edges running against the skirt rubber. The guide to chevron and cleated profile selection goes through the geometry in detail; the point to carry into a feed mill specification is that every millimetre of cleat height you add is a millimetre of film you will not remove.

The other incline problem is at the loading end, where most of the dust in the system is actually made. Pellets arrive at the tail with horizontal velocity, bounce, and a proportion roll back into the skirted section, where the incoming stream crushes them continuously. Fines are the raw material for the hardened sleeve, so an incline that is slightly too steep manufactures the very dust that destroys the pulley. Widening the skirt opening and lowering the feed point both reduce the recirculation without changing the belt.

Where a profile is unavoidable, pair the primary blade with a rotary brush and a return-side scraper, and keep the incline as shallow as the layout allows. A 3° reduction in angle can remove the need for a chevron entirely.

07Idlers, Pulleys and the Blinding Order

Carryback does not attack the whole conveyor at once. It works downstream in a fixed order, and the order tells you how long the problem has been running. The first components to suffer are the return idler rolls, because they are small, they run against the loaded surface, and a 3 mm build-up on a 89 mm roll changes its diameter by more than 3%. That is enough to set up a vibration at every rotation, and the vibration is what destroys the bearing seal long before the roll itself wears out. The mechanism is worth following through the bearing seal selection logic, because a seal that keeps dust out still cannot keep out a film that is pressed into it under load.

Next in line is the tail pulley, then the snub and take-up pulleys, and the sequence is determined by wrap angle rather than by position. A pulley with 180° of wrap sees twice the contact of one with 90°, so the tail drum builds a sleeve first and builds it fastest at the centre where the belt tension is highest and the film is thickest. Once the sleeve has a step at each edge, the belt is no longer running on a cylinder. It is running on a cylinder with two shoulders, and the belt is being steered by whichever shoulder it happens to touch. Tracking corrections then get applied at the training idlers, which treat the symptom and accelerate the edge wear.

The corrective column is deliberately unglamorous. Most of the repairs are cleaning and resetting, not replacing.

Two components deserve a specific warning because they are usually bought on the wrong criteria. Grooved or ceramic lagging is sold on grip, and on a dry belt that is correct. On a coated pellet line the grooves are exactly where the film collects, and once they are full the lagging behaves like a smooth drum with a smaller effective diameter, so the drive slips on start-up and the belt takes the shock at the splice. Plain rubber lagging with a machined finish packs less and cleans more easily, even though its dry grip is lower. The other warning is about self-aligning idlers, which are often fitted to cure the tracking that the sleeve caused. They will hold a belt in the frame, but they cannot remove the sleeve, and a belt held straight by a training idler while running on a stepped pulley is being worn on its edge every metre of its length.

Where the build-up is genuinely unavoidable, for instance on a line running high-fat product with no space for a secondary cleaner, the practical answer is a sealed roll that tolerates a rough surface. A dust proof idler with a labyrinth seal will survive longer than a standard roll in the same position, and a conveyor roller with a thicker shell resists the denting that comes from a lump passing over a built-up roll. Neither is a cure. They buy the weeks you need to fix the cleaning, which is the only thing that stops the sequence restarting.

Load points have their own version of the problem, because impact idlers under a chute collect material on their exposed faces and press it into the belt with every lump that lands. The troughing idler set immediately downstream is usually the cleanest part of the conveyor, because the belt is still moving fast and shedding.

08A Specification and Commissioning Routine for a Pellet Line

A specification written around carryback looks different from a catalogue order, and it is shorter. The carcass on most pellet transfers is a 315/3 or 400/3 EP conveyor belt, which is ample here, and the covers are 4 mm on the carrying side and 2 mm on the pulley side. Those numbers are ordinary. What has to be added is the finish, which should be a fine matte impression rather than a polished sheet, the compound, which should resist fat and oil rather than merely abrasion, and the cleaner package, which should be named in the same line item as the belt rather than ordered later from a different supplier.

Commissioning should then prove the specification rather than assume it. We ask for three checks. At 72 hours, weigh the material collected under the head pulley for one full shift and note the sleeve thickness at the tail drum. At four weeks, repeat both and look at the back cover at a marked station with a depth gauge. At twelve weeks, repeat again and inspect the blade tip for uneven wear across the width. If carryback is falling and the sleeve is stable, the belt and the cleaner are matched. If the weight is falling but the sleeve is still growing, the primary blade is working and the residue is not, which is a secondary cleaner problem rather than a belt problem. Records kept to this pattern are also the only useful basis for a warranty discussion, which is why our quality assurance file asks for them.

Spares policy is part of the same argument. A mill that stocks two blades and a set of return rolls for every pellet conveyor will never run a sleeve long enough to damage a belt. The full list of what to hold, and the questions to put in an enquiry, is set out in the RFQ checklist, and our service team will review a carryback record against the belt construction that produced it.

Ask for a belt and cleaner pairing review

09Frequently Asked Questions

Why is the belt worn from underneath while the top cover still looks new?

Because the wear is being done by carryback, not by the cargo. A fat or molasses film leaves the pellets at the discharge, collects fines on the return strand, and hardens into a sleeve on the tail pulley, so the belt is abraded between a hard sleeve and its own back cover. The top cover is never in that contact.

How much carryback is too much on a pellet line?

Measure it rather than judge it by eye. Put a tray under the head pulley for a full shift and weigh what comes off the return strand. On a 1,000 mm belt running coated pellets, anything above roughly 1 kg per hour is enough to build a visible sleeve within a month, and the 4.1 kg per hour figure quoted earlier is well into the range where edge damage follows.

Will a thicker cover solve the problem?

No. Extra cover adds material to the face that is not failing and adds mass to the belt that has to be accelerated at every start. It delays top-cover wear, which is not the failure mode, and it does nothing to the sleeve. Thickness is the right answer on a dry abrasive duty and the wrong answer on a coated one.

Which cleaner suits a molasses-coated product?

A primary blade in soft polyurethane followed by a secondary blade, both on a fine matte cover, with pressure set just high enough to keep the tip closed. Molasses stays tacky and re-wets overnight, so the secondary is doing more work than the primary and should be treated as a consumable with a scheduled change.

Can the belt be washed instead of scraped?

In a feed mill, rarely. Fat needs warm water and a surfactant, molasses needs enough water to dilute the sugar, and most pellet transfer areas have no hose points because water creates product and structural risk. That is why mechanical removal at the head pulley carries the whole load on this duty.

What incline can coated pellets climb without a profile?

Usually 10 to 13° on a smooth cover, less when the surface is heavily coated or the belt is vibrating. Above that, an open V profile with a shallow cleat holds the load, but every millimetre of cleat height is a place the blade cannot reach, so keep the incline shallow before adding profile.