There is a particular knock that travels around a sorter bed. Stand near the nose pulley and it sounds like a bad bearing. Move to the tail and it seems to be coming from the splice. Meanwhile the belt edge is polishing the guide beside the third divert. A timestamp alongside each symptom tells you whether they truly travel together.
Send us your sorter speed, pulley diameters and joint type for a tracking review
A carton can leave a scuff on PVC without taking the sorter out of service. A splice opening or an edge rubbing the frame is a different matter. Before choosing a harder cover, look back through the last few stops. Did the knock keep time with a mark on the joint? Was the edge already at the guide before a parcel reached the divert? Did the problem happen on an empty run? The answers are more useful than a photo of a worn patch.
Put a paint mark beside the joint. Use the same two frame marks each time you measure the moving belt edge. Do it empty and during a normal sort, with speed unchanged. Add the knock to the log when you hear it.
| What the operator notices | A useful comparison | Evidence worth keeping | Where to look next |
|---|---|---|---|
| One knock on each lap | Marked joint versus sound | Video with speed shown | Joint and contacted pulley |
| Edge creeps toward one guide | Cold versus warmed machine | Edge distances at fixed points | Roller and take-up geometry |
| Noise only during a parcel burst | Loaded versus empty | Parcel position and divert timing | Induct and transfer |
| Scuffed face, normal running | Grip before and after cleaning | Photo and transfer test | Cover specification |
Run a finger across a stopped joint. A flush joint may still resist bending. A raised corner is easier to find: it taps the nose as it wraps. Note stiffness as well as the visible step. The timing makes the joint a suspect; it still needs a stopped inspection.
That 15 m loop at 2.5 m/s puts the splice back at the nose after six seconds, ten times in a minute. A drum revolution takes far less time; its circumference sets that clock.
With power isolated, use a straightedge and feeler gauge to map the joint at the left, middle and right. Mark the high side and the leading face. On restart, does the knock arrive with that mark? A drum witness mark in the same track strengthens the case; poor timing weakens it.
The joint name on the purchase order does not tell you its finished profile. A finger splice whose fingers lie flat has room to flex. Lift at a corner says more about the press work than the name of the splice. A butt or stepped joint may suit a particular belt and drum, yet leave an audible shoulder if the two ends were not lined up. The belt maker's splice procedure and minimum pulley diameter belong beside the inspection measurements. A field splice is judged by the same finished geometry and trial as a workshop splice.
| Joint | Detail that matters here | Symptom to watch | Check against |
|---|---|---|---|
| Finger | Even fingers and flat press face | Local lift or click | Maker's pattern and press settings |
| Butt | Bond and level across the width | Shoulder at nose pulley | Approved construction and strength |
| Stepped | Feathered transition | Knock stronger on one side | Profile and pulley bend |
| Field repair | Cure and straightness record | Fault returns after restart | Finished joint criteria |
Speed does not create a new tracking rule. It leaves less time between the first touch at a guide and the first damaged edge. A loaded sorter also receives side forces that an empty trial never shows. For that reason, empty and loaded tracking should be compared at the same belt speed.
Choose a reliable frame datum and check the drive, tail and return rollers from it. A roller only slightly skewed can steer the belt lap after lap. Pushing the edge back with a guide masks that steering and creates a polished strip. The bracket and its shims are usually quicker to correct than a damaged belt is to replace.
Suppose the surveyed bed is square but the edge still snakes once a lap. Put a transverse witness line near the splice. During a guarded slow trial, see whether its ends meet the pulley together. An early arrival at one edge puts joint squareness and unequal belt-edge length on the list before any more roller adjustment.
The two take-up screws should be checked separately. A bracket bent below either screw can make a centerline reading look normal. If tightening quiets a slap but the edge still heads for the guide, the underlying steering force has not gone away.
Watch a normal burst, including the largest cartons. An angled entry can nudge the belt or load one edge of the bed. The empty run may look perfect. If the edge jumps only as a divert fires, capture its timing and actual running gap. If it starts wandering earlier, go back toward the induct. Locate that first movement before ordering belting.
| Source | Empty run | With parcels | Measurement | Corrective trial |
|---|---|---|---|---|
| Skewed roller | Walks steadily | Often still walks | Square to frame datum | Realign bracket |
| Belt or joint geometry | Change follows a lap | Same timing | Edge length and joint square | Rework joint if required |
| Uneven take-up | Position varies with tension | May worsen | Both adjusters and frame | Restore take-up geometry |
| Biased feed or divert | May run centered | Walks during load | Parcel position on video | Correct entry or timing |
“Louder than last week” is a good reason to investigate, but a poor fault location. Record a short clip at a fixed spot. Include the belt mark in view and write down speed. A joint knock follows the belt lap; a roller defect follows its own rotation; a loose guard may rattle whenever either excites the frame.
Divide loop length by belt speed. That gives seconds between joint passes. In that 15 m example, a six-second knock deserves a joint inspection. More frequent ticks within the same lap call for a different clock: roller or drum rotation.
A rough bearing need not make a neat click. It may rumble under cartons and go quiet on an empty bed. During an isolated maintenance check, feel for rough free rotation and check the temperature history. A seized return roller often leaves its own narrow polished track.
A thin guard can make a modest impact seem severe. Hold off on padding or fastening the panel until the timed event has been identified. Otherwise the sound may disappear while the joint keeps striking the nose.
| Sound heard | Clock to compare | First place to inspect | Other possibility |
|---|---|---|---|
| Single lap knock | Belt mark | Joint and nose drum | Object fixed to belt |
| Regular rumble | Roller cycle | Bearings and mounts | Lagging contact |
| Click with divert action | Parcel and actuator | Clearance and timing | Carton hitting guide |
| Guard rattle | Event that starts it | Loose panel | Upstream impact |
The divert is where tracking faults become expensive. At the divert, side load arrives in pulses. Give an already wandering edge a close guide and the damage can advance from polish to a split within a short run. Look at the parcel and the belt in the same video frame.
Metal contact concentrates work in a few millimeters of edge. Carton scuffing spreads over a much wider area. Check the running clearance during an empty run first. An edge that touches the guide with no carton present will not be saved by a harder cover.
A long polished stripe usually records steady rubbing. A tear confined to one divert calls for a frame-by-frame look at that transfer. If both edges are rubbed, measure both guide gaps and inspect the take-up. Keep the station marker in every damage photo.
Time how long the carton bears on the guide. Follow the contact from induction to exit: where does the carton weight sit when the divert begins to push? Try a heavy case, a light case and one that arrives skewed. After each single setting change, film the belt edge again.
| Edge appearance | What to capture | Question at the machine | Trial |
|---|---|---|---|
| One polished side | Empty running gap | Is contact constant? | Reset path and guide |
| Tear at one divert | Loaded video | What catches at that station? | Adjust timing or clearance |
| Both edges rubbed | Guide gaps and take-up | Is the belt being pinched? | Reset guide and frame |
The word “sorter” covers very different machines. One carton might cross a flat run, a cross-belt cell and an inclined cleated section. The splice requirements change along that route. Identify the troublesome bed by station number, and show any roller-drive or cleated construction in the replacement sketch.
The integrator knows the divert sequence; maintenance knows which brackets have moved; the belt maker knows what joint the construction permits. They should see the same marked layout and event clip. A torn-edge photo alone tends to produce three quotes and no diagnosis.
Roller alignment and load position matter on warehouse transfer beds as well. Splice clearance does not transfer unchanged to a sidewall or roller-drive layout. On those machines, check the profile and support geometry afresh.
Loop length belongs in the speed discussion. For a comparison across sorters, calculate passes per minute as 60v/L, with speed v in meters per second and loop length L in meters. A short loop can clock up plenty of joint passages at an unremarkable speed.
| Condition | What repeats | Check this component | First record | Avoid assuming |
|---|---|---|---|---|
| Slow commissioning | Step crosses pulley | Joint and drum | Stopped profile | A slow trial proves full duty |
| Loaded sorting | Drift follows cartons | Induct and divert | Empty versus loaded path | PVC grade caused it |
| Fast steady running | Knock at lap interval | Joint and nose | Sound plus belt mark | Thicker cover cures it |
| After warm-up | Edge nears guide | Frame and take-up | Cold and hot gap | Heat alone explains drift |
Start at the station where someone first noticed the change. If the fault began after a roller adjustment, that is useful history. If it began after a joint repair, inspect the joint before ordering another roll. Work back from a dated event rather than from the belt's appearance at the end of a shift.
One entry needs the operating speed, what was being sorted, when the stop occurred and where the joint and edge were. Add any recent bracket or take-up work. Even approximate entries become useful when the same event repeats.
Joint step and straightness, roller parallelism, take-up geometry, and edge position with and without parcels: these four readings separate most of the arguments. Use the machine maker's datum and its safe access procedure. A single measurement beside a parked belt cannot describe what happened at a busy divert.
A raised joint on a square bed belongs with the splice repair. A flat joint passing over a skewed drive roller points back to the machine. If drift starts only when cartons arrive, film the induct and divert. Run the loaded check again after that correction. It may uncover a second fault previously hidden by the first.
Give the supplier a marked bed sketch, belt width and loop length, construction, joint type, speed and smallest pulley. Add two joint-face photos, edge distances and a brief fault clip. That package lets them judge a belt and splice against the actual path, rather than against the cover color.
Send your line data and we will tell you which of the four is failing
Run it empty at the same speed first. If it stays centered, observe the induct and divert with the actual carton mix. Uneven take-up or frame movement under load can produce the same symptom, so record the edge path at several stations.
Mark the joint and listen at the pulley. A click synchronized with each belt lap makes the joint or its contact surface suspect. A bearing rumble tends to follow its roller and may change with load. Confirm at a stopped, isolated machine.
It cannot remove contact with a guide. Fix the running path and clearance first. Only then ask whether the belt construction needs more edge strength; extra thickness also changes bending at small pulleys.
Check after work on a roller, frame, take-up or splice, and whenever tracking changes. Set routine intervals from this line's history and record measurements against the same datum.
No. A field finger joint can suit the duty when the maker's method and tools are used. Judge its finished profile, bond and straightness, then run the loaded tracking trial.
Include width, length, construction, joint, speed, smallest pulley, carton mix and fault location. The edge readings, photos and a short sound clip are especially helpful.
Related Articles:
PRODUCTS

We are focusing on material handling, power transmission and industry application.