SINOCONVE conveyor belt manufacturer & supplier makes conveyor belt more efficient.

Logistics & E-commerce Conveyors: How to Specify Sorting Duty

2026-09-29

Sorting and induction duty breaks belts that would run for years on a straight haul. A parcel sorter rated at 3,600 cartons an hour can start and stop the same short belt ninety times an hour, and every one of those starts sends a stress wave down the carcass and back again. Add a knife-edge nose bar, a sixty-degree transfer, and a scanner tunnel that wants a matte dark surface under a laser, and the belt stops being a commodity roll of fabric and rubber. It becomes a system component with a duty cycle attached to it. That is the part most specification sheets still miss.

This guide is written for the people who fill those sheets in: the integrator sizing a cross-belt loop, the maintenance engineer who replaces a finger-spliced PVC belt every fourteen months and has stopped believing the two-year claim, and the buyer comparing three quotes from a conveyor belt supplier against one from a trading house that never sees the line. Everything below is deliberately narrowed to sorting and induction work. Cycle count, pulley and nose-bar geometry, splice design, camber control, static dissipation, and the acceptance tests that decide whether a delivery goes onto the line or back on the truck.

If the belt family is still open, the pages behind our industrial conveyor belt range and the application notes under logistics warehousing are the right starting point. If the family is settled and what you need is a partner who can repeat the same camber and splice tolerance order after order, the team that runs our conveyor belt manufacturer lines is the conversation to have before the drawings are frozen.

Send us your sorter layout and cycle rate for a sorting-duty belt specification review

01Sorting and Induction Duty: What the Belt Actually Sees

A belt on an induction line is asked to do something a troughed overland belt never is. It has to accept a product at an angle, hold it against lateral drift for one or two metres, survive a photoelectric or camera read window without glare, and then hand the carton off to a sorter cell that may index in as little as 200 milliseconds. The belt surface matters as much as the carcass, because the product is light and the acceleration is not. A 12 kg parcel accelerated at 1.5 m/s² generates a modest inertia load but a very high demand on surface friction, and any slip at the transfer point shows up as a mis-placed carton downstream rather than as visible belt damage.

Throughput is the first number that should appear on the specification, and it should be expressed as cartons per hour plus the maximum single-unit mass, not as a belt speed alone. Sorting and induction duty on a wholesale conveyor belts scale usually means 1,200 to 4,500 units per hour at 0.6 to 2.2 m/s belt speed, with the induction conveyor running noticeably faster than the take-away conveyor so gaps are created rather than closed. That speed differential is where a lot of premature belt wear quietly starts.

Cycle count is the second number. A merge belt that runs continuously is a different belt from an indexing belt that performs 60 to 140 starts per hour. Motors, take-up, splice and carcass all behave differently under indexed duty, and the difference in service life between the two cases is measured in years, not months.

Reading the line as four different duty zones

Most e-commerce and parcel facilities contain at least four belt zones that look identical on a layout drawing and behave nothing alike in service. The induction and merge zone sees the most starts, the sortation feed zone sees the fastest belt and the tightest nose radius, accumulation sees zero-speed accumulation and product standing still on a moving surface, and the scan and pack zone sees cleanliness and reflectivity demands. Specifying one belt across all four is the single most common cost mistake I see on site visits.

Line zone	Throughput and cycle pattern	Dominant belt stress	Failure we actually find on site	What the specification has to pin down
Induction and singulation	2,000 to 4,000 units/hour with 60 to 140 starts per hour and frequent speed changes	Repeated acceleration transients, splice peel at the leading edge, take-up cycling	Splice opening at the lap after 8 to 14 months, belt edge fretting where the carton lands	Splice type and joint efficiency, cover friction range, take-up travel for at least 1.5% elongation reserve
Sortation feed and nose transfer	Continuous run at 1.6 to 2.4 m/s with short indexed bursts at the cell	Very high flex fatigue across a 40 to 80 mm nose bar or knife edge	Carcass cracking on the back face, cover separation starting at the nose bar wraps	Minimum pulley and nose-bar diameter per carcass type, flex-fatigue grade, ply count limit
Accumulation and buffering	Zero-speed or low-speed accumulation with product stationary on a moving surface	Abrasion on the top cover, heat build-up under stationary load, edge wear at guides	Top cover worn through to the carcass within three to four years on cardboard-heavy duty	Cover abrasion class, low-friction variant if accumulation is true zero-speed, edge reinforcement
Scan, weigh and packing	Steady speed, continuous duty, cleanliness and reflectivity requirements dominate	Static charge build-up, surface contamination, dimensional stability under the scanner	Read-rate loss from static and glare, belt tracking drift after a wet floor wash-down	Surface resistivity and colour, cleaning method compatibility, camber tolerance for straight running

One more field note before we move to geometry. On a returns line in a 1,400 parcel/hour hub we logged 38 splice inspections over two years, and every single failure traced back to the same root cause: a 2-ply belt specified onto a nose-bar radius the carcass could never tolerate. The belt was purchased on price per square metre. It was replaced twelve times. Nobody re-read the nose-bar diameter on the drawing.

02Small Pulley and Nose-Bar Geometry in Induction Lines

Sorting conveyors are built around short centres and small rotating members because the product has to change direction quickly and the frame has to stay low. A diverter nose bar of 40 mm, a 60 mm terminal pulley and a backside bend roller of 50 mm are normal on this kind of line. Those diameters are hostile to any belt with a thick, stiff carcass, and the failure is not immediate. The belt simply accumulates flex fatigue with every wrap, and the crack that finally appears on the back face is the arithmetic of wrap count times small radius.

DIN 22102 and AS 1332 both tie minimum pulley diameter to carcass construction and ply count for textile belts, and that relationship should be the first line of defence in the specification. The practical numbers on logistics duty are tighter than the standard minimums, because standard minimums assume a moderate wrap count. On a sorter nose bar running at 1.8 m/s, a 2-ply belt with a high-tenacity polyester warp and a thin 0.8 mm skim between plies will outlast a 3-ply belt of the same rated strength by a wide margin, simply because it bends without generating the same inter-ply shear.

This is where a rubber conveyor belt mindset has to give way. Heavy rubber covers of 6 mm top and 2 mm bottom, correct for a quarry, add mass and stiffness that a nose bar does not want. On induction duty, a thin PVC or polyurethane top cover of 1 to 2 mm over a 2-ply synthetic carcass, or a light EP construction where a rubber cover is unavoidable, is the geometry-friendly answer. If you need the hydrocarbon and oil behaviour of rubber but the flexing of a thin belt, the constructions on our EP rubber belt page show where that compromise sits.

Matching carcass to radius, not to a brochure headline

Ask for the carcass data sheet rather than the promotional one. What matters is fabric type and weave, ply count, total thickness under 4 mm, and the manufacturer's own stated minimum pulley diameter for a finger-spliced or vulcanised joint. A supplier who cannot state that minimum in writing for the exact construction being quoted is telling you something useful.

Belt family	Carcass and cover build	Workable nose-bar or pulley diameter	Joint normally used on sorting duty	Consequence when the radius is undersized
PVC 2-ply, monofilament warp	Two plies of high-tenacity monofilament, 0.8 to 1.2 mm skim, 1 to 2 mm PVC top cover, 2 to 3 mm total	Down to 25 to 40 mm on a nose bar, 60 mm on a terminal pulley when the joint is finger-spliced	Hot finger splice with a 100 to 150 mm finger length and a press cure at 145 to 155 degrees C	Finger tips lift within a year, then the joint peels progressively from one edge
Polyurethane 2-ply	Polyester or polyamide carcass with a cast or laminated PU cover, 0.6 to 1.5 mm cover, often with a fabric backing	40 to 50 mm minimum, better at 60 to 80 mm if the cover is above 1.2 mm	Finger splice or a stepped splice where the line cannot be shut for a press cure	Cover cracking on the wrap side, then water ingress into the carcass on wash-down lines
EP 2-ply, 125 to 160 N/mm	Polyester warp with polyamide weft, 1.0 to 1.5 mm inter-ply skim, rubber cover 2 to 4 mm top	50 to 80 mm on a pulley; a nose bar below 50 mm shortens carcass life sharply	Vulcanised stepped splice at 1.5 to 2 times the belt width in step length	Back-face cracking between plies, visible as a ladder pattern across the belt
EP 3-ply or heavier rubber	Three or more plies, covers from 3 mm, total thickness above 8 mm	100 mm and above on a pulley; not suitable for knife-edge or 40 mm nose bars	Hot vulcanised splice only, mechanical fasteners are unreliable at this cycle count	Rapid joint failure, tracking instability from the extra thickness, motor overload on indexed starts

Note the tolerance wording in that table and keep it in your own documents. A belt that runs straight enough for a truck-loading conveyor can still be a tracking nuisance on a narrow sorter with three minutes of centre distance. Short belts amplify every millimetre of camber.

Two structural items finish this section, and both sit outside the belt itself. The first is pulley crown and lagging: on short centres, a light 2 mm crown at the driven pulley does more for tracking than any guide roller, and it costs nothing at the design stage. The second is frame squareness — a sorter frame welded out of square by 4 mm across a 1,200 mm width will fight every belt you install on it. Look at our conveyor pulleys and conveyor components pages for the hardware side of that geometry, and keep the conveyor belt factory that cut the belt in the loop when the drawing is revised, because a nose-bar change usually invalidates the old construction.

Logistics & e-commerce conveyors sorting duty belt with thin 2-ply carcass and finger-spliced joint

03High Start-Stop Cycles: Acceleration, Sag and Take-Up Behaviour

Every indexed start converts motor torque into a stretch wave that travels down the belt at a few hundred metres per second and returns as a reflection. On a 12 m induction conveyor that journey takes a fraction of a second, and the belt never fully settles before the next start. The visible result is a belt that appears slack at the tail pulley during acceleration, then snaps back and rides high on the carry idlers when the carton lands. Maintenance teams usually describe this as "the belt is jumping" and try to solve it with more take-up stroke. That helps, but it treats the symptom.

The real variables are the acceleration ramp and the belt's elastic modulus. A sorter feed belt ramped from zero to 1.8 m/s in 0.35 s accelerates at roughly 5 m/s², which is aggressive for a light carcass and gentle for a stiff one. Stretch a 2-ply PVC belt of 12 m centre distance by 0.4% and you have consumed 48 mm of slack; do the same with a stiff 3-ply rubber belt and you consume a similar distance but with a much higher peak tension at the driven pulley. Take-up travel must therefore be sized from the belt's actual elastic stretch at the specified acceleration, not from a rule of thumb based on belt length alone.

On a 1,200 carton/hour cross-belt feed line we measured 0.32% elastic stretch against a manufacturer figure of 0.20%, purely because the commissioning ramp had been left at the drive default of 0.2 s. Lengthening the ramp to 0.6 s cut the measured stretch to 0.21%, removed the visible sag at the tail, and added nothing to the cycle time because the sorter cell was already gated by its own index. That single parameter change is often the cheapest reliability improvement available on an existing line.

Drive transmission hardware deserves the same attention on indexed duty. The V-belts or synchronous belts that drive the sortation gearboxes see the same start-stop count as the conveyor belt, and a lightly loaded drive belt that is fine on continuous duty can glaze and slip when it is indexed a hundred times an hour. If you are selecting drive elements for a sortation module, the range behind our transmission belt manufacturer pages and the matched sets from our V-belt manufacturer line are sized for exactly that kind of duty, and the synchronous alternatives on the timing belts page are worth considering where position accuracy at the cell matters more than cost.

Sizing take-up from the duty, not from habit

A screw take-up with 100 mm of stroke is a common default. Whether it is enough depends on throughput, ramp time, centre distance and carcass modulus, and those four values interact. The table below is the pattern we use when a customer sends us a sorter layout with cycle data.

Duty case on the line	Ramp and acceleration	Elastic stretch to plan for	Take-up travel worth building in	Symptom when this is ignored
Continuous take-away, no indexing	Single ramp at start-up over 2 to 3 s, then steady speed	0.2 to 0.3% of centre distance, plus 0.5% installation reserve	60 to 80 mm on centres under 10 m	Nothing dramatic; the belt simply needs re-tensioning every few months
Moderate indexing, 20 to 40 starts/hour	Ramp of 0.6 to 0.8 s, acceleration around 2 to 3 m/s squared	0.25 to 0.4%, rising where the carcass has a low modulus	100 to 120 mm, preferably gravity take-up with visible travel marks	Progressive slip at the drive pulley during the acceleration phase only
Heavy indexing, 60 to 140 starts/hour	Ramp compressed to 0.3 to 0.5 s unless the VFD is deliberately softened	0.35 to 0.6%; the figure must come from the carcass data sheet, not a rule of thumb	150 to 200 mm, with the mid-position marked so drift is visible at a glance	Splice peeling and edge damage that grows steadily from one side of the joint
Nose-bar sortation feed on 3 to 5 m centres	Very short centres, so stretch is small while flex fatigue is extreme	Under 0.2%, but tension uniformity across the width becomes the limiting factor	50 to 80 mm, with a pneumatic or spring take-up that reacts faster than a screw	Snatch at each start, product toppling on the nose transfer, uneven edge wear

Where the drive cannot be softened — some OEM sortation modules lock the ramp in firmware — the belt construction is the only lever left. A thinner carcass with a higher-modulus warp lowers both the stretch and the reflected tension peak, and a vulcanised splice resists the cycling far better than any mechanical fastener.

04Lateral Guiding, Tracking and Edge Damage at Merge Points

Tracking on a long overland conveyor is a slow problem measured in weeks. Tracking on a sorter merge is a fast problem measured in cartons. The belt is short, the loads arrive off-centre, and the frame members that would normally restrain lateral movement are absent by design so operators can reach the transfer. Everything therefore rests on the belt's own dimensional quality and on the small amount of guidance hardware the layout can accommodate.

Camber is the number that decides this, and it is a manufacturing property, not an installation one. A belt that curves when laid flat on a floor has camber, and camber cannot be corrected by tensioning one side. For sorting duty on widths of 400 to 1,200 mm we work to a camber of no more than 12 mm over a 10 m length, which is inside the general 0.5% of length guidance but tight enough for short centres. Splice squareness matters just as much: the cut edge of the joint should be square to the belt centreline within 1 mm per 100 mm of width, so a 1,000 mm belt tolerates 10 mm un-squareness and no more. A badly squared splice on a 4 m merge belt produces a lateral shove on every revolution of the belt, which at 1.8 m/s is once every 4.4 seconds.

I have watched a team replace three belts on the same merge before anyone measured the splice angle on the delivery. The fourth belt ran straight with no other change. That is a twenty-minute measurement with a steel square and a chalk line, and it is the single most useful incoming inspection step on a sorter belt order.

Hardware then supports the belt rather than fighting it. A light crown on the driven pulley, side guide rollers set with 3 to 5 mm running clearance, and a snub roller at the merge where the carton lands off-centre will between them handle most of the lateral load. Guides set with zero clearance do the opposite of what is intended: they hold the belt edge, wear it, and prevent the belt from finding its own centre. Look at training idlers and self-aligning idlers for the heavier end of this problem, and see the field notes on self aligning rollers and tracking downtime for how much of belt tracking actually comes from the idler set rather than the belt.

Logistics & e-commerce conveyors merge and sortation hall with induction lines and packing benches

Merge and transfer types and what each one demands

Transfer configuration drives the belt edge loading more than belt width or speed does, and the four cases below need genuinely different edge protection.

Transfer or merge type	Guiding hardware that works	Belt edge requirement	Tracking consequence if it is wrong
Side-load merge at 30 to 45 degrees	Snub roller plus a short side guide of 600 to 900 mm on the loaded side only	Edge reinforcement or a fabric-wrapped edge so side thrust does not fray the cover	Belt walks to the loaded side and climbs the guide on every heavy parcel
Ninety-degree transfer with a pop-up wheel	Nose bar of 25 to 40 mm plus a gap of 3 to 5 mm to the next belt, no side guides in the product path	Camber under 12 mm in 10 m and a splice square to within 1 mm per 100 mm of width	Product snags in the gap, then edge damage starts at the same point every shift
Tilt-tray or cross-belt sorter infeed	Crowned drive pulley of 2 to 3 mm crown, precision-mounted carrier rollers on 150 to 200 mm pitch	Dimensional stability and width tolerance within plus or minus 2 mm across the roll	Cell mis-loads, because the belt position varies by more than the tray alignment window
Wash-down or cold-store transfer	Stainless guide rollers with sealed bearings on 800 to 1,000 mm spacing, no exposed plain bearings	Edge and cover that tolerate alkaline cleaning and stay flexible below minus 20 degrees C	Belt stiffens, tracking correction disappears, and edge cracking follows within a season

Frame and bracket hardware is the last piece of this. A guide roller mounted on a bracket that deflects 3 mm under load is a guide roller that will not hold its clearance, and on a standardised sorter frame the bracket is easy to overlook because it costs a few euros. Details on this sit with the rest of our conveyor brackets range. The same principle applies in parcel and material recovery facilities where a merge serves mixed cartons and film-wrapped packs; the sorting stream behind our recycling notes shows how much harder edge protection has to work when the product mix is unfriendly.

05Static Control and Barcode Read Zones

Static is the quietest source of sorting errors on an e-commerce line. A belt running at 1.8 m/s over dry rollers generates charge continuously, and the charge has to go somewhere. In a dry hall at 25% relative humidity we have measured surface potentials high enough to lift a light polybag off the belt and to leave the next parcel trailing a visible arc to a conveyor frame. The consequences are rarely a fire. They are a missed read, a parcel routed to the exception lane, and an operator who spends the shift hand-scanning because the failure looks intermittent.

The measurable target is surface resistivity, and the accepted benchmark for antistatic conveyor belting is a figure at or below 3 x 10^8 ohms, which is the limit set out in ISO 284 for electrically conductive belting. In practice a belt in the 10^7 to 10^8 ohm range will dissipate charge fast enough for a scanning zone, provided the charge has a path to earth. That path is the part that gets forgotten. A conductive belt on an insulated frame with plastic carrier rollers is a conductive belt with nowhere to go.

Grounding therefore has to be designed with the belt, not after it. A spring-loaded carbon or brass brush riding the return side within a metre of the first scan zone, bonded to the frame with a resistance under 10 ohms, handles most of the charge. Metal carrier rollers and a bonded frame do the rest. Where a line is built with plastic components for weight or wash-down reasons, a dedicated earth strap and one brush per belt section is the minimum, and the brush has to be scheduled for replacement — a worn brush reads as a static fault that maintenance cannot find.

Read-zone requirements that end up in the belt specification

Read technology drives belt surface choice more than most buyers expect. Laser scanners dislike gloss and light colours, camera-based readers want contrast and consistent lighting, and both are affected by charge. The table below lists what to write into the belt specification for each zone rather than what to blame after commissioning.

Zone on the line	Surface resistivity target	Bonding and grounding method	Read or sort consequence
Induction scan tunnel with a laser head	10^7 to 10^8 ohms, matte black or dark green cover, no glossy finish	Brush on the return strand within 1 m of the scan window, frame bonded to earth under 10 ohms	Glare and specular reflection cause no-reads that appear random and shift with parcel orientation
Camera-based multi-side reader	10^7 to 10^8 ohms, uniform colour across the full width with no patchy cover	Conductive rollers or brushes on both strands, plus bonding of the reader frame itself	Charge-drawn dust on the cover degrades contrast, so read rate falls through the shift
Weighing and dimensioning station	Antistatic belt with stable dimension under load; no significant static requirement beyond the general target	Single brush per belt section and a bonded load cell frame, kept clear of the weighing zone	Static discharge into load cell wiring shows up as weight drift rather than a clean fault code
Wash-down and chilled packing zone	Food-grade antistatic cover, resistance verified wet as well as dry after cleaning	Stainless brushes and bonding that survive alkaline cleaning and daily hosing	Grounding paths corrode, static returns, and the failure is wrongly attributed to the scanner

A buyer checking all of this needs a supplier who can provide resistivity test values per batch and not just a claim on a datasheet. That is a reasonable thing to ask of a conveyor belt distributor with technical staff, and an unreasonable thing to expect from a reseller moving pallets. Ask which one you are talking to before the order, not after the first no-read investigation.

06Splice Design for Sorting Duty: Joint Type, Geometry and Cure

The splice is where a sorting belt fails, and it is also the only part of the belt that is made by hand rather than by a machine. That combination explains most of the variation in service life between two identical belts from the same factory. Two belts can leave the same press, run on the same line, and differ by two years in splice life, because one splice was squared within 5 mm and cured at 150 degrees C for the full cycle while the other was assembled under time pressure at 135 degrees C.

For sorting duty the practical choice is between a hot vulcanised splice and a cold-bonded one, and occasionally a mechanical fastener on a non-critical short belt. A hot vulcanised finger splice is the default on thin 2-ply PVC and PU belts below 4 mm total thickness, because the interlocking fingers distribute load across the full width and produce a joint that is barely thicker than the belt itself. Thickness at the joint matters on this duty: every extra millimetre at the splice becomes a bump that trips the nose bar, opens the gap to the next conveyor, and adds a shock load once per revolution.

Finger geometry is not arbitrary. Fingers of 8 to 12 mm width and 100 to 150 mm length are typical for a 2-ply logistics belt, cut so the tip angle is consistent and the finger ends sit square. The press cure sits between 140 and 160 degrees C depending on the material, held long enough for the bond to develop properly rather than to reach the surface temperature suggested by the platen. On a 2-ply PVC belt, vulcanising a 150 mm finger joint to full bond takes around 12 to 18 minutes of press time at temperature, and cutting that to eight minutes to reopen a line faster is how premature joint failures are manufactured.

Logistics & e-commerce conveyors belt finger splice being prepared and press cured at the belt factory

Joint efficiency against duty, in numbers

A joint that holds 65% of belt strength is fine on a reclaimer and marginal on a sorter that reverses torque twice a second in the worst case. Joint efficiency figures should be requested from the supplier in writing with the test standard named, and the joint type should be chosen from the duty rather than from what the installation crew has on the truck.

Joint type	Preparation and geometry	Assembly and cure conditions	Typical retained strength	Where it belongs on a sorting line
Hot vulcanised finger splice	Fingers 8 to 12 mm wide, 100 to 150 mm long, tip angle consistent, ends square to the centreline	Press at 140 to 160 degrees C for 12 to 18 minutes at full temperature, then controlled cool under press	85 to 90% of belt strength when geometry and cure are both within specification	Default for 2-ply PVC and PU induction and nose-bar belts under 4 mm total thickness
Hot vulcanised stepped splice	Ply-by-ply steps, each step 150 to 250 mm on a 2-ply build, stepped so no two joints align across the width	Press cure with pressure applied across the full splice length, checked with a pressure gauge not by feel	90% and above, verified to DIN 22110 or an equivalent splice test method	Rubber-covered EP belts on sorter feed conveyors where a rubber cover is required
Cold-bonded finger splice	Same finger geometry as the hot version, but the surfaces are buffed, cleaned and primed immediately before bonding	Two-part cement, clamped with even pressure for the full cure window, ambient above 15 degrees C	70 to 80% when the ambient conditions are right, and unpredictable when they are not	Maintenance repairs on lines that cannot be stopped long enough for a press cure
Mechanical fastener, plate or bolt type	Holes punched square to the centreline, fastener 10 to 20 mm narrower than the belt width on each side	Bolted or crimped to the supplier's torque figure, with the plate bedded evenly on both edges	40 to 60% of belt strength, and always below the belt cover surface unless recessed	Emergency repair only; on a nose bar it will scrape the transfer plate and snag small parcels

Two splice habits are worth building into any installation procedure. First, mark the belt centreline on both ends before cutting and check the joint with a square after assembly, before the press is closed. Second, write the press temperature, cure time and operator on the belt edge or in the maintenance log; when the joint fails three years later, that record is the only way to know whether the belt or the process was at fault.

07Surface Friction, Product Stability, Noise and Cold Modes

Cover choice on a sorting belt is a friction and cleanliness decision, not a wear decision, and it is where the widest range of opinions exists in the industry. A coefficient of friction that is too low lets a carton slide on a 10 degree inclined take-away. Too high, and a parcel tips forward when the belt stops instead of settling. For parcel and tote work we generally aim at a coefficient of friction of 0.3 to 0.4 static against a clean cardboard surface, verified with the actual pack types rather than with a laboratory sample.

Accumulation is the case that exposes the wrong choice fastest. If product is meant to stop and sit while the belt continues to move, a high-friction cover will simply scrub the underside of every carton and generate both dust and static. Low-friction or fabric-backed accumulation belts exist for that reason, and they behave very differently from a standard carton-handling surface. Asking whether the zone is true accumulation or a soft buffer is one question that prevents a great deal of cover wear.

Noise is the second practical driver, and it is measurable. On a packing line regulated to keep operator exposure below 80 dB(A) over an eight-hour shift, the combination of a hard PVC cover, a steel transfer plate and a nose bar with no radius can push a single station close to the limit. A slightly softer cover, a polymer transfer strip and a 2 mm radius blended onto the nose bar edge can take 3 to 5 dB(A) out of the same installation without any change to throughput.

Cold stores change the rules again. Standard PVC formulations begin to stiffen below about minus 10 degrees C and become genuinely brittle around minus 25 degrees C, which is why chilled and frozen packing zones need a formulation rated for the actual operating temperature and not simply a "food grade" descriptor. On a frozen goods line we saw a standard belt lose tracking correction within a month of commissioning because the belt had stiffened enough that the crowned pulley could no longer steer it. A cold-rated compound solved it. Details of the surface families we run for packing and chilled zones sit on our PVC conveyor belts pages, and the compliance side of the same question is covered under food packaging.

Surface option	Friction and product behaviour	Cleanliness and compliance position	Best fit on a sorting line	Watch-out in service
Matte PVC, antistatic, dark colour	Friction in the 0.3 to 0.4 band, stable on cardboard and film-wrapped packs, quiet at speed	Wipe-clean, antistatic to the ISO 284 benchmark, standard choice for scan zones	Induction, merge and scanner zones where read reliability matters most	Plasticiser migration can harden the cover over years and change friction without warning
Cast polyurethane cover	Higher cut and abrasion resistance, friction stable when wet, good grip on tote bases	Available in food-contact grades, tolerates alkaline cleaning better than PVC	Wash-down packing and chilled zones, and lines handling tote bottoms with sharp corners	Repair splices demand precise buffing; a poor PU joint will show water ingress quickly
Fabric-backed low-friction accumulation surface	Deliberately low friction so cartons stop and hold while the belt keeps running underneath	Easy to clean, but the backing can trap dust if the zone is not on a cleaning schedule	True zero-speed accumulation buffers and slug-forming zones before the sorter	Wrongly used on an incline it will not hold the load and product will slide back
Rubber cover, RMA grade or DIN 22102 class	High friction and high cut resistance, but heavier and stiffer than a light logistics belt	Grades available for oil contact and for cold flexibility down to minus 30 degrees C	Heavy parcel feed and mixed industrial sorting where cut damage dominates	Thickness rules out small nose bars; check the minimum diameter before ordering

If cut damage rather than friction is the dominant threat on your particular feed, the constructions on our impact and cut resistant belt page are worth reviewing alongside the plain logistics surfaces. The trade-off is always weight and flexing ability against cut resistance.

08Writing the Specification and Accepting the Delivery

A sorting-duty belt specification is short if it is written properly. It needs the duty data, the geometry limits, the surface requirements, and the joint definition, and it should state the acceptance tests and the tolerance for each one. Anything else is padding that hides the four numbers that matter. When a customer sends us a layout, cycle counter data and the nose-bar drawing, we can commit to a construction in a single technical exchange. When a customer sends a part number and a quantity, we can only quote what that part number was last made from.

Incoming inspection is where the money is saved. Camber, splice squareness and joint integrity are all measurable in under an hour with a floor, a chalk line, a steel square and a tape, and each one correlates directly with how many times the belt will be adjusted on the first shift. Surface resistivity needs a meter and a humid day to be meaningful, but the reading is worth taking before installation because it is much harder to argue afterwards.

The seven checks that decide acceptance

The checks below are the ones we recommend to buyers receiving a sorting-duty belt, and the ones our own quality assurance process applies before shipment. They are written to be quick to perform on a site floor.

Acceptance check	How to measure it	Accept criterion we work to	Consequence if it is out	Action before installation
Camber and curve	Lay the belt flat on a clean floor over 10 m and measure the maximum bow from a chalk line	No more than 12 mm deviation over 10 m for sorting widths of 400 to 1,200 mm	Persistent tracking drift that no amount of guide adjustment will remove	Reject the roll and ask for a replacement cut from a different blank, not a re-cut
Splice squareness	Square the joint line against the marked centreline on both edges and compare the two gaps	Square to within 1 mm per 100 mm of belt width, matching edges within 2 mm	A lateral shove on every revolution, expressed as edge wear on one side only	Re-square and re-cure if enough length remains, otherwise reject the joint
Splice strength record	Request the splice test data or a sample joint from the same press run and standard	85% retained strength minimum for finger joints, tested to DIN 22110 or an equivalent method	Joint opens at the lap within the first year on a high-cycle line	Hold the belt, re-press the joint under supervision, and record the deviation
Surface resistivity	Measure with a surface resistance meter on both faces at three points along the belt	At or below 3 x 10^8 ohms, referenced to the ISO 284 limit for conductive belting	No-read incidents that appear intermittent and move with humidity	Add grounding brushes and re-measure before blaming the scanner, then reject if unchanged
Thickness, width and length	Gauge the belt at six points across the width and tape the overall length under light tension	Within plus or minus 0.2 mm on thickness, plus or minus 2 mm on width, length to the drawing	The take-up runs out of travel, or the belt fouls the frame on the return strand	Re-cut from stock where possible; never force a short belt onto an under-travelled take-up
Marking and traceability	Read the belt edge marking and match it to the batch record and the purchase order reference	Batch number, construction code and cure record retrievable from the supplier in writing	Repeat orders drift in cover colour and friction, so an approved belt stops being repeatable	Freeze the construction reference in your own parts system and quote it on every PO

Roll up all of it and the working specification for a sorting and induction belt is a page and a half long: duty data, geometry limits, splice definition, surface requirements, and the acceptance table above. That document, held by the buyer rather than by the supplier, is what keeps the fifth order from the same factory matching the first. For spares planning, the rolling stock and life assumptions behind our service approach are built on exactly that idea, and for roller-side spares the item structure in the conveyor idler factory RFQ checklist maps onto the same acceptance logic. Buyers who want the wider product background can start from the conveyor belt manufacturer guide or, where the line is roller-heavy, from the conveyor roller guide.

Request a sorting-duty specification and sample acceptance sheet

09Frequently Asked Questions

How many start-stop cycles should a sorting belt be rated for?

Rather than a single rating, work from your own cycle count and then check the construction against it. A merge belt with 60 to 140 starts an hour needs a thin carcass with a high-modulus warp and a vulcanised joint, because the flex fatigue accumulates with every wrap over the nose bar. If the drive ramp can be softened to 0.5 or 0.6 seconds, the same belt will see a noticeably lower peak tension on each start, and measured stretch often falls from around 0.32% to 0.21% as a result.

What is the smallest nose bar these belts can run over?

It depends on carcass type. A 2-ply PVC belt with a monofilament warp will run over a 25 to 40 mm nose bar, a polyurethane belt generally wants 40 to 50 mm, an EP 2-ply build wants 50 to 80 mm, and anything three-ply or heavier should not see a nose bar at all. Ask the supplier for the minimum pulley and nose-bar diameter for the exact construction being quoted, in writing, alongside the carcass data sheet.

Do we need a vulcanised joint or is a mechanical fastener acceptable?

On a sorting line, hot vulcanised is the default and mechanical fasteners are emergency repairs only. A finger splice reaches 85 to 90% of belt strength and stays close to belt thickness, whereas a plate or bolt fastener typically holds 40 to 60% and sits proud of the surface, where it scrapes the transfer plate and snags small parcels. Cold-bonded joints are a reasonable middle option for maintenance work, achieving 70 to 80% when the ambient temperature and clamping are both correct.

How do we check camber and splice squareness on delivery?

Lay the belt flat on a clean floor, run a chalk line down the centre and measure the maximum bow over 10 m; for widths of 400 to 1,200 mm we work to no more than 12 mm of deviation. For the joint, square it against the marked centreline and compare the two edge gaps, working to 1 mm of un-squareness per 100 mm of width, with the two edges matching within 2 mm. Both checks take under an hour and prevent the majority of tracking problems before the belt goes onto the frame.

Why does our barcode read rate drop during the afternoon?

Humidity usually explains it. Charge builds faster on a belt in a dry hall, and once relative humidity falls, static-related no-reads rise even though nothing on the scanner has changed. Check the surface resistivity of the belt against the 3 x 10^8 ohm benchmark from ISO 284, then verify the grounding path: a brush within a metre of the scan window, bonded to the frame at under 10 ohms. A worn brush mimics a static fault and is the first thing to inspect.

Can one belt cover induction, accumulation and packing?

Physically yes, technically not advisable. Accumulation wants a low-friction or fabric-backed surface so cartons can stand still while the belt keeps moving, and a scan zone wants a matte antistatic surface at the ISO 284 conductive benchmark. A single high-friction antistatic cover used on a true zero-speed accumulation buffer will scrub the underside of every carton and generate both dust and charge. Split the duty at the zone boundaries and specify each surface separately.