Duty, environment and failure cost — get these three right and the roller diameter, seal type, tube wall and bearing class all follow. Includes the rpm check most spec sheets skip.
A maintenance team replaces a failed roller. They measure it, find a drawing, and order "the same again, please". Six months later the same position fails again. The roller was not wrong because of its size — it was wrong because nobody asked what it was dying of.
An idler is a sealed rotating assembly running in dust, water, heat or all three, for years, unattended. Every choice inside it — wall thickness, bearing clearance, seal geometry, shaft material — is a trade against one specific failure mechanism. Change the mechanism and the correct build changes with it.
Duty is not just tonnage. It is tonnage, belt speed, loading conditions and where in the conveyor the roller sits.
| Position | What it actually sees | What the build needs |
|---|---|---|
| Carrying (troughing) | Continuous load, material impact at transfer, three-roll load sharing | Wall thickness and bearing class for the radial load; correct trough angle |
| Impact (loading point) | Drop energy from lump material, belt cover damage risk | Rubber rings or discs to absorb energy; closer spacing |
| Return | Belt weight only, but carry-back and material build-up | Self-cleaning geometry (spiral, rubber disc) where material is sticky |
| Training / self-aligning | Lateral belt wander | Friction or pivoting frame response — and correct placement, not more of them |
Two conveyors with identical tonnage can need completely different idlers if one is loaded from a chute at four metres and the other from a feeder at half a metre. Drop height is a specification input.
This is the question that decides the seal — and the seal is what actually determines service life in the field.
Running a contact seal at 4 m/s in a dry cement plant is a common and expensive mismatch: it heats, the grease degrades, and the bearing fails early in an environment where a labyrinth would have lasted years.
This is the question that turns a specification from a purchase into an engineering decision.
On a short in-plant conveyor, a failed roller is a nuisance: swap it at the next shutdown. On a 3 km overland conveyor running 20 hours a day, the same failure costs a crew mobilisation, access time along the gallery, belt cover damage from the seized shell, spillage clean-up, and — if the belt is the production bottleneck — lost throughput.
When the cost of failure is high, the specification should move up, not down: thicker wall, higher-grade bearing, better seal, closer inspection interval. The component price difference is small against the avoided outage.
Bearing life is governed by speed as well as load. Roller speed follows directly from belt speed and diameter:
rpm = (belt speed in m/min) ÷ (π × roller diameter in m)
500 rpm is the CEMA basis speed — the speed at which idler load ratings and rated life are normally published, and therefore the design target. The catalogue figure is the L10 life at 500 rpm; it is not a figure that holds at any speed. 600 rpm is the general engineering limit in both Chinese and international idler standards and design handbooks — past it, rotational resistance climbs sharply and service life falls away. 750 rpm is a ceiling only for specific standards and special designs, BOS 613 among them, and never a general target.
The matrix below turns that formula into a specification table. Read down to your belt speed, move right to the first green cell, and you have the minimum diameter for that duty:
| Belt speed | 63.5 mm | 76 mm | 89 mm | 108 mm | 133 mm | 159 mm | 194 mm | 219 mm |
|---|---|---|---|---|---|---|---|---|
| 1 m/s | 301 | 251 | 215 | 177 | 144 | 120 | 98 | 87 |
| 1.5 m/s | 451 | 377 | 322 | 265 | 215 | 180 | 148 | 131 |
| 2 m/s | 602 | 503 | 429 | 354 | 287 | 240 | 197 | 174 |
| 2.5 m/s | 752 | 628 | 536 | 442 | 359 | 300 | 246 | 218 |
| 3 m/s | 902 | 754 | 644 | 531 | 431 | 360 | 295 | 262 |
| 3.5 m/s | 1053 | 880 | 751 | 619 | 503 | 420 | 345 | 305 |
| 4 m/s | 1203 | 1005 | 858 | 707 | 574 | 480 | 394 | 349 |
| 5 m/s | 1504 | 1256 | 1073 | 884 | 718 | 601 | 492 | 436 |
| 6 m/s | 1805 | 1508 | 1288 | 1061 | 862 | 721 | 591 | 523 |
Rotational speed in rpm, from rpm = belt speed × 60 ÷ (π × diameter); rounded to the nearest rpm.
Two things follow from this. A diameter that is generous at 1.5 m/s is badly over-speed at 4 m/s — which is why copying the diameter from an older, slower line is the single most common over-speed mistake. And a 6 m/s conveyor is served by no diameter up to 219 mm within the 500 rpm basis speed: that duty needs a larger roller, a lower belt speed, or a deliberate trade-off against bearing life, and only your project can say which of those can actually change.
If you would rather not read the matrix, this is the same calculation reduced to a single recommendation per belt speed:
| Belt speed | Minimum diameter — ≤500 rpm (CEMA basis) | Minimum diameter — ≤600 rpm (general limit) |
|---|---|---|
| 1 m/s | 63.5 mm (301 rpm) | 63.5 mm (301 rpm) |
| 1.5 m/s | 63.5 mm (451 rpm) | 63.5 mm (451 rpm) |
| 2 m/s | 89 mm (429 rpm) | 76 mm (503 rpm) |
| 2.5 m/s | 108 mm (442 rpm) | 89 mm (536 rpm) |
| 3 m/s | 133 mm (431 rpm) | 108 mm (531 rpm) |
| 3.5 m/s | 159 mm (420 rpm) | 133 mm (503 rpm) |
| 4 m/s | 159 mm (480 rpm) | 133 mm (574 rpm) |
| 5 m/s | 194 mm (492 rpm) | 194 mm (492 rpm) |
| 6 m/s | none up to 219 mm | 194 mm (591 rpm) |
A ø108 mm roller under a 3.5 m/s belt turns at roughly 620 rpm — past the 600 rpm general limit. Step the roller up to ø133 mm and the same belt runs it at about 500 rpm, the CEMA basis speed, for a marginal change in roller cost. On a line that has been uprated, the diameter on the old drawing is the last number you should trust.
The question every buyer asks — how long will it last? — is answered by the CEMA class. The standard ties an idler's load capacity to a minimum L10 life, and it sets one common basis for that rating: 500 rpm. The class on your specification is therefore a statement about both the load the roller can carry and the life it is expected to give at that speed.
| CEMA class | Minimum L10 life at 500 rpm | Typical application |
|---|---|---|
| CEMA B | 30,000 hours | Light to medium load, conventional duty |
| CEMA C | 30,000 hours | Medium load, moderate belt speed |
| CEMA D | 60,000 hours | Heavy load; critical process, or hard-to-access positions |
| CEMA E / F | 60,000 hours | Very heavy load, high-impact duty |
Minimum L10 life figures at the 500 rpm basis speed. The class also sets bearing series and shaft-end requirements — it is not a statement about life alone.
First, it is a floor, not a prediction. The class states the life the roller is designed to achieve at least at 500 rpm; a well-built roller in that class will often exceed it. Two rollers in the same class can differ substantially in real service life, because class fixes the minimum and the build decides everything above it.
Second, life in hours moves with speed. Bearing life is fundamentally published as L10 revolutions, and hours are those revolutions divided by rotational speed. At constant load that makes the relationship direct: run the roller slower than the basis speed and you stretch the life in proportion; run it faster and you shorten it. A roller at 350 rpm can be expected to give roughly 500 ÷ 350 ≈ 1.4 times the published figure, and the same roller pushed past 600 rpm gives proportionally less — which is why an over-speed roller fails long before its class would suggest.
The practical consequence is that the two specification questions belong together: check the rpm first, then read the CEMA class. A CEMA D roller run past its speed limit will not deliver its 60,000 hours. The class promises life at the basis speed, not at any speed.
| Parameter | What it decides | How to specify it |
|---|---|---|
| Roller diameter | Bearing speed, belt support spacing, load capacity | From belt width, speed and trough set — check rpm, not habit |
| Tube wall thickness | Shell life under abrasion and impact | In mm, stated as a value not a range; the #1 hidden downgrade |
| Shaft | Deflection under load — deflection kills bearings | Material and diameter; cold-drawn 45# is our standard |
| Bearing | Life under load and speed | Named brand (HRB / ZWZ / SKF / FAG) and clearance class — C3 default, C4 for hot duty |
| Grease | Whether the lubrication survives the duty temperature | Continuous service range, not dropping point: −40 °C to +120 °C standard, to +160 °C for clinker duty |
| Seal | Whether the bearing reaches its design life at all | By contaminant: labyrinth for dry/fast, contact for wet/slow |
| Runout | Vibration, belt tracking, noise | Total indicator reading in mm, measured after assembly |
Most quotations state the bearing as a single line — 6205-2RS — and stop there. That one line hides the two decisions that determine whether a roller runs 8,000 hours or 40,000. Below is what we build to, and the reasoning behind each choice, so you can put the same numbers into your own specification.
| Brand | When we quote it | Why it is chosen |
|---|---|---|
| HRB (Harbin Bearing) | Standard build — general-duty conveyors, bulk and repeat orders | Broad size coverage, stable supply, and the best cost-per-hour on normal duty. The default unless something in the duty says otherwise. |
| ZWZ (Wafangdian Bearing) | Standard build — mining duty, larger roller series, higher load ratings | Stronger coverage in the heavier series that mining conveyors use, with the same supply stability. |
| SKF | When the end user or the EPC names it; critical or high-speed positions | Named-specification projects, and sites where the maintenance team already standardises on SKF and stocks it. |
| FAG (Schaeffler) | Same — named-specification projects, European-owned or European-engineered sites | Where the plant's own bearing standard is Schaeffler, or the EPC's specification schedule names it. |
We build to the named brand rather than to "a well-known brand", because a substitution at assembly time is invisible on the drawing and impossible to audit later. If your specification names a brand, say so on the PO.
Radial internal clearance is the free play inside the bearing before it is fitted. It matters more on a conveyor than almost anywhere else, because two things eat into it before the roller carries a single tonne: the interference fit on the shaft, and running warmth. Fit a standard-clearance bearing into a tight housing on a warm, fast conveyor and it starts life already partly preloaded.
| Class | What it means | When to specify it |
|---|---|---|
| C0 (normal) | Standard radial clearance as supplied | Only on short, slow, cool conveyors with a light fit. On a warm or fast flight it leaves no margin for fit and thermal growth — this is the class that produces the preload failure. |
| C3 | Greater than normal | The conveyor default. Absorbs the interference fit and normal running temperature without going into preload. If your specification does not name a class, this is the one that should have been meant. |
| C4 | Greater than C3 | High-ambient duty (cement clinker, sinter, coke), a wide seasonal temperature swing, a heavy press fit, or a long overland conveyor in a hot climate — anywhere the bearing sees more thermal growth than C3 can absorb. |
Clearance values are defined by bore diameter, not by bearing designation alone — under ISO 5753-1, and GB/T 4604 in China. If a spec sheet names a class but not a bore size, ask for the actual figure in microns.
| Grease | Continuous service range | Specified for |
|---|---|---|
| General-purpose lithium, No. 2 | −20 °C to +120 °C | Standard build on normal-duty conveyors — indoor or moderate climate, continuous running |
| Low-temperature lithium / semi-synthetic | −40 °C to +120 °C | Our standard charge. Cold-climate and winter outdoor duty — Russia, Kazakhstan, northern China — and any line that starts cold |
| Lithium complex / polyurea, high temperature | −20 °C to +160 °C | Clinker, sinter, coke and other high-ambient duty, where a standard lithium charge would degrade long before the steel does |
The number that matters is the continuous service ceiling, not the dropping point. A grease that survives 180 °C on a test rig may still break down at 120 °C in continuous running, and on a sealed-for-life roller the charge has to last the whole intended service life — which is a function of speed and temperature together.
When a bearing does fail early, the failed roller itself tells you which of these choices was wrong — we work through that in why conveyor idler bearings seize before they wear out.
Calculate the rpm from belt speed and diameter. If it sits above the practical limit for your bearing and duty, the bearing will fail long before the shell wears. One diameter step usually buys several times the service life.
35° is the general default because it balances capacity against belt stress. 20° suits older or narrow installations and fragile material; 45° suits high tonnage and long overland runs where spillage control matters more than belt fatigue.
Choose by contaminant. Dry dust at speed favours labyrinth — low friction, low heat, long life. Washdown, slurry or rain favours contact. Running a contact seal at high speed generates heat and shortens life.
Usually tube wall thickness, bearing brand and clearance class, seal construction and shaft material. Ask for all four in writing, with wall thickness in millimetres rather than a range.
The roller is the smallest part of it. A seized roller damages belt cover, sheds material, creates housekeeping and fire risk, and on a long overland conveyor the access time alone exceeds the component price many times over.
A minimum L10 life at the 500 rpm basis speed. CEMA B and C carry a 30,000 hour minimum; CEMA D and CEMA E/F a 60,000 hour minimum. It is a floor rather than a prediction — the class fixes the minimum and the build decides everything above it — and it is tied to the basis speed, so a roller run past 600 rpm will not deliver the figure its class promises.
Belt width, speed, material, tonnage, temperature and environment. We come back with the build, drawings and a quotation.
A belt does not wander; it is steered. Find the cause in the right order — build-up, loading point, structure, pulleys, idlers, camber, splice, tension — before touching an idler.
The drawing is gone and the line is down. Six measurements and one photo define a roller completely — here is how to take them and what to send with the enquiry.
Lagging sets the friction that decides how much tension the drive can transmit. Rubber, diamond groove or ceramic — and why the bonding method matters more than the material.
Carrying-side support, 20°/35°/45° trough angles.
Flat carry and return duty, standard and heavy-wall tube.
Rubber-ring cushioning at the loading point.
Drive, bend, tail & snub pulleys — with rubber, ceramic or herringbone lagging to your spec.
EP/NN, steel cord, heat, oil and abrasion resistant covers. 17 types for different duties.
Belt cleaners, impact beds, skirt seals — the parts that decide whether your idlers survive.