A belt has rubbed the left stringer again. The nearest return roller has already been skewed twice, and the edge now touches at a different station. Before turning a third bracket, mark where the belt first departs from the frame centre and compare empty and loaded running. The visible rub may be downstream of the force that started the movement.
This field guide follows an urgent tracking job from shutdown through restart.
01Define the Tracking Problem Before Touching a Roller
“Runs left” is not a location. The first station where the edge departs from its previous clearance is more useful than the guard it eventually touches.
A periodic swing needs a different notebook entry. Watch the marked splice pass the same station several times. A repeat tied to one belt section points to a crooked joint, local stiffness or repair patch. A repeat tied to one pulley revolution points toward the pulley and its buildup. The two rhythms can look alike during a hurried glance.
| Observed pattern | Fast field clue | First area to inspect | Do not assume |
|---|---|---|---|
| Local at one station | Centered entry, displaced exit | That frame, roll rotation, buildup, nearby loading | A training idler will cure contamination |
| Progressive over several bays | Edge offset grows at successive marks | String line, pulley axes, cumulative frame skew | The last roller is the cause |
| Once per belt revolution | Offset follows a marked belt section | Splice geometry, carcass damage, camber | Every idler needs steering |
| Changes with loading | Empty belt centers, loaded belt walks | Feed position, skirt drag, trough support | Unloaded acceptance is enough |
02Make the Conveyor Safe and Stable
Hands do not go near the belt to “feel” whether a roller is steering it. Isolate the drives and any other stored energy under the site procedure, then verify the conveyor cannot move. Gravity take-ups, a loaded incline and material caught at a pulley can all move after the motor has stopped. Inspection and bracket adjustment take place only in that secured condition.
There is a limit to a shift correction. An opening splice, hot bearing or damaged frame changes the work from adjustment to repair assessment.
03Choose Between Ten-Minute Triage and a Full Survey
Record the suspect station and the condition of its neighbours. A clean roll in an out-of-square frame will still be steered by that frame.

A full survey earns its time after a pulley change, splice repair, structural work or repeated tracking calls.
04Build a Baseline with Simple Tools
A bent guard and skirt rubber are poor references. Use a surveyed frame point that the next crew can find again, then compare the same edge positions after each trial.
Roll condition belongs in the baseline. A shell coated on one side has a different effective diameter from its clean neighbour; a bearing may turn by hand yet drag under load. Record obvious polish, dust wipe, noise, heat and free rotation. Those observations can justify replacing a roll without ever moving its bracket.
| Tool or check | What it establishes | Common field error |
|---|---|---|
| String or laser | A repeatable conveyor centerline and accumulated skew | Referencing a deformed stringer without checking it |
| Square and straightedge | Local frame and idler relationship | Squaring to a loose bracket |
| Tape and chalk | Edge offset at fixed stations and a belt revolution mark | Moving the station between test runs |
| Temperature screening | A comparative clue for bearing drag | Treating one infrared reading as a bearing diagnosis |
| Rotation check | Freedom of movement and roughness | Testing while stored energy is uncontrolled |
05Understand the Steering Law and Its Limits
A roller presented at an angle can steer the belt toward the end it contacts first. “Leading end” is useful shorthand only after the belt direction and actual contact have been established.
Check the shell for a narrow polish band and for unequal dust removal. A large bracket move in response can make the unloaded belt look centred and the loaded belt worse.
06Adjust One Variable in Controlled Increments
After guards are restored, run enough complete belt circuits for the changed path and marked splice to pass the observation stations. An immediate movement at the adjusted roll is only part of the result. The belt can continue to shift farther downstream. If edge clearance deteriorates, stop and return to the scribed position before trying another cause.
| Step | Controlled action | Observation gate | Stop condition |
|---|---|---|---|
| 1. Baseline | Record edge position at every fixed station | Pattern is repeatable for complete circuits | Belt approaches fixed structure |
| 2. Isolate | Select one causal frame or pulley issue | Evidence agrees with travel direction | Cause remains uncertain |
| 3. Mark | Preserve original fastener and bracket positions | A rollback reference is visible | Mount is damaged or loose |
| 4. Adjust | Make one small, measurable angular change | Fasteners are secured before testing | Slot travel or frame condition is inadequate |
| 5. Run empty | Observe several full revolutions from safe points | Response stabilizes and is logged | Rapid adverse movement or abnormal sound |
| 6. Run loaded | Increase to a documented normal condition | Clearance remains acceptable at all stations | Load changes direction or edge risk |
07Treat Carrying and Return Runs Differently
The carrying run receives the material stream. Watch where it lands, whether one skirt presses harder, and whether a damaged impact bed drops one side of the belt. A belt that stays centred empty and walks under a full chute usually needs a loading check. Steering the idlers to cancel that load makes them work continuously against the transfer.

The return path has a different nuisance: carryback. Uneven deposits on a roll change its contact surface, so clean it before deciding that its bracket is out of square.
08Diagnose Root Causes Instead of Over-Steering
Several brackets pointing in opposite directions tell a maintenance story. One correction may have hidden another bay’s error, leaving the belt central only at a particular feed rate. Return to the measured frame and pulley references. Check where the load lands, where the return first picks up carryback and whether a pulley bearing has shifted; a steering roller cannot remove any of those inputs.
| Root cause | Typical evidence | Corrective priority | Why roller steering alone fails |
|---|---|---|---|
| Off-center loading | Loaded drift; uneven burden depth | Center the stream and inspect liners | Lateral input remains at every load cycle |
| Buildup | Uneven shell diameter; repeating residue | Clean and correct cleaner performance | The contact geometry changes again |
| Seized idler | Heat, noise, flat spot, abnormal polish | Replace and verify adjacent mounts | Drag and support remain unequal |
| Skewed pulley | Broad tracking influence near pulley | Survey shaft and bearing references | Large wrap contact dominates small idler inputs |
| Crooked splice | Deviation follows splice each circuit | Measure and repair the splice | One fixed roller cannot follow a moving defect |
| Uneven tension | Take-up binding or unequal edge behavior | Restore take-up function and references | The belt enters support with unequal forces |
| Frame twist | Progressive error and inconsistent cross levels | Repair structure and foundations | Each roller inherits a faulty datum |
09Recognize Splice-Linked and Belt-Linked Drift
A periodic offset deserves a mark on the belt, not just a note that it happened “every few minutes.” Compare that mark with the edge movement at the same station on repeated circuits. If the timing follows one repaired belt section, inspect the joint geometry before spending the shutdown on roller angles.
A modest periodic swing with ample edge clearance may be monitored under a documented plan. Contact with steelwork, an opening joint, exposed plies or a growing offset requires a stop.
10Verify Reversible Conveyors in Both Directions
Reverse travel changes which end of a skewed roll meets the belt first.

The acceptance sheet needs four runs: empty and loaded in each direction. Use a comparable feed condition on the two loaded runs and record edge clearance at the same stations. If the machine feeds only at one end in reverse, note the different loading arrangement explicitly. Passing the forward empty test alone says little about reverse production.
11Know When Alignment Rollers Are the Wrong Fix
A training frame at the end of its slot or an edge rub that returns after cleaning calls for a new look at the line, loading and take-up. Selection of a different roller comes after those checks.
Another training idler may add drag and a new wear point. The same applies to a weak structure that changes shape under full load. Correcting its geometry will outlast a succession of roller settings.
12Restart, Observe, and Define Acceptance
Before restart, one person owns the stop signal and the observers stand at protected stations. Account for tools and guards. The first empty run checks gross direction, noise, immediate clearance and free rotation. Once that passes, introduce the usual material and watch the belt after the chute and along the return.
Set the acceptance limits from the actual machine: minimum clearance to fixed structure, acceptable stable offset, load condition and number of circuits. Do not substitute a single millimetre target for every conveyor. Record the measured gaps and the test condition. A statement that the belt “looks fine” cannot be compared with the next shift’s complaint.
| Acceptance item | Record | Evidence of completion |
|---|---|---|
| Edge clearance | Measured offset at every numbered station | Within the site-defined limit for repeated circuits |
| Test condition | Direction, speed, approximate feed, moisture | Representative operating state documented |
| Adjustment | Frame number, end moved, measured amount, time | Original and final marks remain traceable |
| Mechanical condition | Rotation, sound, temperature comparison, fasteners | No abnormal change after loaded test |
| Follow-up | Inspection owner and operating-hours checkpoint | Named responsibility, not an open note |
13Prepare a Useful Replacement Roller RFQ
If the suspect roll is seized or its shell is damaged, give purchasing more than a photograph. Record the carry or return position, shell diameter, face length, overall shaft length, shaft-end shape and mounting arrangement. Distinguish a dimension measured on the old roller from one read on a drawing or merely estimated. Wear may have changed the diameter.
Add belt width and speed, trough angle where relevant, material and dust or washdown exposure. A roller supplier can then check the bearing and seal arrangement against the duty. SINOCONVE supplies conveyor rollers alongside belt orders; the roller specification still needs to identify the installed position and verified geometry.
| RFQ field | Minimum useful detail | Why it matters |
|---|---|---|
| Installed position | Carry, return, impact, station number | Connects dimensions to actual loading |
| Roll geometry | Shell diameter, face, shaft, end details | Controls physical fit and contact |
| Mounting | Slot, center distance, bracket photos | Avoids shaft-end mismatch |
| Duty | Belt width and speed, load, material, hours | Supports bearing and shell review |
| Environment | Dust, water, temperature, corrosion | Guides sealing and material discussion |
| Evidence | Drawing revision and scaled photographs | Makes assumptions visible before production |
14Frequently Asked Questions
Which way should I move an idler to steer the belt?
Look at the polished contact band and confirm travel direction. Try one reversible bracket move and compare the edge gap.
Can we align rollers while the conveyor is running?
Observe from protected positions while running. Stop, isolate and verify zero energy before touching a bracket or roller.
How far should we move a roller?
The bracket, belt tension and spacing determine the response. A scribed starting position makes the trial reversible.
Why is the empty belt centred but the loaded belt drifts?
Inspect the feed position, skirt pressure, loaded sag and impact support. The material stream can introduce a side force that is absent during an empty run.
Why does the error return once per belt revolution?
Mark the splice and confirm that the swing follows the same belt section. Inspect the joint and local repairs before changing another idler.
Should several idlers be moved at once?
Keep the trial to one bracket, then compare the same stations after restart.
Will a self-aligning roller solve persistent drift?
It may help a sound conveyor with modest variation. It will not correct a seized roll, side loading, pulley skew or a damaged splice.
How many revolutions should we observe?
Enough for the marked splice to pass each critical station repeatedly and for the downstream response to settle. Record the actual count.
What changes on a reversible conveyor?
A skewed roller steers differently when travel reverses. Verify edge clearance empty and loaded in both directions.
When should the conveyor remain stopped?
An open joint, exposed fabric, hot bearing or belt-to-frame contact calls for repair assessment before restarting.










