How to Size and Select a Steel Cord Conveyor Belt: A Practical Step-by-Step

Steel cord belts rarely fail because the cord was weak. They fail because somebody chose a grade before they knew the tension, or sized a pulley from a table written for fabric carcasses. The steel did its job. The decisions wrapped around it did not.

Most searches for conveyor belt steel cord start the same way. A buyer wants one number to drop into a requisition, and something fast to forward to the project team. We understand the instinct, which is exactly why our first reply is usually a short questionnaire instead of a price. A belt is not a part you look up in a catalogue. It is a calculation you run against your own conveyor, and it only works if the conveyor is described honestly.

Send your conveyor duty data and we will size the belt with you

We size in a fixed order, and the order is the point. Collect the duty data. Calculate running and starting tension. Turn that tension into a steel cord grade with a defensible safety factor. Fix width, cover grade and carcass. Choose the splice and the pulley diameters that the splice can live with. Check the drive and the take-up against the result. Write the specification sheet. Then verify all of it at acceptance. Each item is a link in a chain, and a chain is only as strong as the weakest link you decided to skip.

Field calls have taught us that the visible failure is almost never the real one. A splice that opens at fourteen months usually means the belt was sized on nominal tonnage with no surge factor behind it. A cord that fatigues early usually means the pulley was one diameter too small for the splice that was fitted. Torn covers usually mean the chute was loading at the wrong angle and nobody checked the impact zone before the order went out. The belt gets blamed. The sizing chain gets away with it.

This is the procedure we use in-house, written for procurement engineers and plant maintenance leads rather than for conveyor designers. It favors numbers you can obtain from a site walk and a drive nameplate over numbers that only look impressive inside a report. Seven working tables follow, and one worked example rides along the whole way, from a tonnage figure on a spreadsheet to a signed specification sheet and a belt that is actually the right one.

We have built belting since 1988 and run ten production lines today, two of them dedicated to steel cord construction. That history is useful mainly because the same mistakes keep arriving in the inbox. On a 1,800 t/h iron ore line we were asked to re-engineer a belt supplied as ST1250 on a 1,150 mm carcass. Measured peak tension at the head pulley came out 22 percent above the design figure the original supplier had used, and the splice was running at a safety factor near 3.4 instead of the 6.8 written on the sheet. Nothing about that belt was defective. It had been sized against a duty that never existed.

So the procedure below is the one we run inside the house. As a conveyor belt manufacturer we treat the sizing calculation as part of the product rather than a formality before the price. It is produced by hand at our conveyor belt factory and re-checked by a second engineer before it becomes a written commitment. If you only need to decode what a grade or a cover class means, our specification reference on steel cord belt grades carries those tables. This piece is about how those tables earn their place in a decision.

01Step 1: Collect the Duty Data Before Anyone Quotes

Nothing downstream can be recovered if this step is thin. The duty data sheet is not paperwork. It is the only version of reality the calculation will ever see, and a conveyor that runs well for twenty years is usually one that was described accurately on day one by someone who walked the structure instead of copying the tender file.

Start with the conveyor, not the belt. Belt width, trough angle and idler spacing set the capacity and the sag between idlers. Center distance and lift set the slope component of tension, which is often the largest single term on a long overland installation. Drive nameplate power and pulley wrap angle set how much of the total tension can be transferred before slip. Then add the material properties, which decide the cover more than they decide the cord.

Two numbers go missing more often than any others. The first is the surge factor, the ratio between average and peak tonnage. Chute-fed lines with a surge bin can peak 20 to 30 percent above design throughput for minutes at a time, and a belt sized on averages will be under-strength exactly when it is loaded hardest. The second is the usable take-up travel already built into the structure. Retrofitting travel is expensive. Guessing it is free until the belt runs out of stroke.

We ask for the take-up data before we quote anything. On a 900 m overland conveyor feeding a cement plant, the as-built carriage had 4.2 m of usable travel for a belt that needed close to 6 m once elastic stretch, creep and two splices were counted. The operator had already ordered a belt from another conveyor belt supplier and the belt itself was fine. The structure was not, and the rework cost more than the belt did.

Write the sheet down and get it signed by the maintenance lead and the production lead. When two departments agree on the tonnage once, you stop arguing about it in month nine.

02Step 2: Calculate the Belt Tension, Not the Belt Number

Cord pitch, cover thickness and edge condition are the three dimensions a sizing calculation actually constrains.

Tension comes from physics, not from a product table. There are only two terms that matter on a straight conveyor with a modest incline. One is friction along the whole length. The other is the weight of material and belt being lifted, and on any conveyor with real lift that second term usually dominates.

Using the DIN 22101 style average friction approach, we work with this shape of equation:

T_e = f × L × g × (2m_b + m_i + q) + (2m_b + q) × g × H

Here L is center distance in meters, H is lift, m_b is belt mass per meter, m_i is the equivalent mass of rotating idler parts per meter, q is material mass per meter, and f is the friction factor. For a well-aligned, decently maintained conveyor with proper idlers we start at f = 0.022 and adjust upward for dirty, misaligned or badly lubricated installations. Values below 0.018 belong in a salesman's spreadsheet.

The material term is easy to derive. At 3,000 t/h and 4.5 m/s, the belt carries 3,000 divided by 3,600 divided by 4.5, which is 185 kg for every meter of belt. Compare that with the belt itself at roughly 42 kg/m for a 1,200 mm ST1250 construction, and you see immediately that the load is the conveyor. A steel cord belt is chosen because it can carry that load across a long span and survive the splice, not because it is exotic.

Then split the effective tension between the tight side and the slack side. With a single drive pulley the slack side tension is T_e divided by (e^μθ − 1), and the tight side is simply T_e plus T₂. Wrap angle and lagging friction decide that ratio, and they are the reason the same belt can be comfortable on one conveyor and marginal on another with identical tonnage.

Run it on our worked example. A 1,200 mm belt on 1,200 m centers with 60 m of lift, carrying 3,000 t/h of iron ore at 4.5 m/s, gives a friction term near 84 kN and a lift term near 158 kN, so T_e lands at about 242 kN. With a 190 degree wrap and a lagged pulley at μ = 0.35, the slack side comes out near 110 kN and the tight side near 352 kN. Divide by width and you get 294 N per mm of belt.

That last number is the only figure a belt manufacturer genuinely needs. Everything before it is duty. Everything after it is compliance. If you ever receive a quotation that skips straight from tonnage to a grade without this intermediate step, you are buying the supplier's assumption instead of your own conveyor.

Keep the calculation honest on one more point. If your existing belts are a industrial conveyor belt of the fabric type, do not carry the old safety habits across to steel cord. Fabric belts tolerate a slacker take-up and a rougher approximation. A steel cord belt is a stiff, low-elongation member, and the tension at the highest point of the path is the number that will decide its life. Buyers who switch from a rubber conveyor belt to a steel cord construction are often surprised by how much of the design margin was previously hidden inside stretch.

03Step 3: Convert Tension into a Steel Cord Grade and Safety Factor

A grade is a claim about nominal cord strength in newtons per millimeter of belt width. ST2000 means the carcass is rated at 2,000 N/mm. That number is not working tension. Working tension is the grade divided by a safety factor, and the safety factor is where the practical judgement sits.

DIN 22131 and ISO 15236 form the backbone of steel cord practice in most mining and port work. In steady running we look for a factor near 6.7 between nominal strength and the tight side tension, and we accept the low end of that band only on a conveyor with good alignment, a soft start and a proven splice history. During starting we allow a factor down to about 4.5, because acceleration adds an inertial term that disappears once the belt is at speed.

Apply that to the worked example. A tight side tension of 294 N/mm needs a nominal grade of 294 multiplied by 6.7, which is 1,970 N/mm. ST1600 is short at a factor of 5.4. ST2000 gives 6.8 and that is the honest answer, not because a bigger number is better but because 6.8 sits in the band and 5.4 does not. On the same drive with a 0.15 m/s² acceleration, starting tension reaches roughly 342 N/mm and the factor at start falls to about 5.9, which is comfortable.

Read the table as a filter, not a menu. Find the first row whose safe working tension clears your calculated figure, then stop and check the splice and the pulley against that choice. The most common sizing error we see is not under-grading, it is grading up to solve a problem that grading cannot solve. If a belt is failing at the splice, a higher cord grade raises the tension the splice must carry along with the strength it adds, and the margin barely moves.

Grading up also costs you twice. A heavier carcass adds belt mass, which adds to both terms in the tension equation, so the required grade creeps upward again. Buyers comparing wholesale conveyor belts quotes across grades should expect a real price step between ST1250 and ST2000, and that step should be justified by a tension calculation rather than by a preference for a bigger number on the label.

04Step 4: Fix Width, Cover Grade and Carcass Together

Once the grade is set, four decisions remain and they constrain each other. Width, top cover, bottom cover and compound. Choosing them one at a time in the order they appear on a data sheet is how projects end up with a belt that is over-specified in one column and under-specified in the next.

A long overland flight is where the slope tension term and the take-up stroke both decide belt selection.

Width answers to capacity first. Take the design tonnage, divide by bulk density and belt speed, and you have the cross-section of material the belt must carry. Then compare that with the usable trough cross-section for your width and trough angle, and keep a sensible margin. Loading a belt past roughly 80 percent of its theoretical capacity is asking for spillage at the skirtboards and a wandering belt on the incline.

Lump size sets a hard floor under the width. A common field rule is that belt width should be at least three times the largest lump for unsized ROM material, and at least twice the largest lump for screened product. Ignore it and you get edge damage, cord damage at the load point and a cover that looks fine while the splice area quietly suffers. Troughability is the other width constraint. Very stiff steel cord constructions need enough width to form a stable trough at your idler angle, and narrow belts at 45 degrees are a known source of edge stress.

Cover thickness is a wear budget, not a style choice. Top cover takes the impact and the abrasion. Bottom cover takes the pulley contact, the idler contact and the return side spillage, so it is rarely worth going below 5 mm even when a client asks to save mass. On a 1,200 mm ST2000 belt with iron ore we would normally look at 6 mm to 8 mm on top depending on lump size and drop height, and 5 mm or 6 mm underneath.

Compound selection follows the material rather than the tonnage. Abrasive, dry, coarse ore wants a high abrasion resistance compound. Wet, sticky, fine material changes the story toward release properties and cleanliness. Hot clinker or sinter needs a heat resistant grade, and any underground or enclosed handling needs a fire resistant construction that satisfies the local code and the insurer. Our abrasion resistant conveyor belt range and the heat resistant conveyor belt range exist precisely because those two duties pull the compound in different directions.

Then check the whole set against the components that will touch it. Chute liners, skirt rubber, impact beds and idlers all interact with the belt you have just described, and a belt that is correct on paper can still be wrong against the hardware around it. The conveyor components selection guide we published covers that interface, and it is worth ten minutes before the specification is frozen.

Plants almost never buy a belt in isolation. A shutdown that replaces a steel cord belt usually replaces V-belts on crusher and fan drives at the same time, so our transmission belt manufacturer range and the V-belt manufacturer catalogue behind it are often quoted on the same enquiry. It has nothing to do with the tension calculation. It has everything to do with the fact that the maintenance crew is already on site.

05Step 5: Size the Splice, the Pulley Diameters and the Take-Up Stroke

This is the step that separates a belt that works from a belt that is merely strong enough. Three items get decided here, and all three are governed by the cord rather than by the covers.

Steel cord belts are spliced by interleaving the cords over a defined length and vulcanizing the joint. The splice is the weakest point of the belt by design, and it is also the least forgiving. Cord pitch and cord diameter set how much room the splicer has to work, and the joint length grows as the grade grows. A splice on an ST2000 construction is typically somewhere around 1.5 m to 2.2 m long, and every millimetre of that has to be cut, cleaned, laid up and cured under controlled pressure and temperature. On site we insist on a splice record rather than a verbal confirmation, because the numbers that matter are cure time, cure temperature and the pressure actually held, not the length of the finished joint.

Pulley diameter is the next constraint and the one most often missed. Steel cord carcasses are far less tolerant of tight bending than fabric belts. DIN 22131 practice ties the minimum diameter to the cord thickness, and as a working rule we look for a drive pulley diameter of at least 150 times the cord diameter, with bend and tail pulleys allowed to run somewhat tighter. A grade change quietly changes the cord diameter, so raising the grade without revisiting the pulley geometry is a real and common error.

Take-up stroke is the third item. Steel cord belts have low elongation, which is an advantage for splicing and troughing, but it does not mean the carriage never moves. Working elastic stretch of roughly 0.2 to 0.3 percent of center distance is recovered when tension is released, permanent creep adds more over the first year, and every future splice consumes a splice length of travel. Add a reserve on top and the number gets serious on long conveyors. On the worked example, 1,200 m of centers at 0.25 percent elastic, 0.1 percent creep, one 2 m splice allowance and a 40 percent reserve lands close to 8 m of usable travel.

The figures above are ranges for early layout work, not a substitute for the cord diameter quoted on the maker's drawing. Two suppliers can both offer ST2000 with different cord diameters, and the pulley requirement follows the drawing, not the grade. This is one of several reasons we prefer to see the conveyor drawing before a quotation becomes a commitment, and it is why a serious conveyor belt distributor or mill will always ask for it.

On a 1,400 m port conveyor we instrumented the take-up carriage for the first eighteen months of service. Sixty-two percent of the available travel had been used by month six, almost all of it in the first three. The stroke had been specified on elastic elongation alone, with no allowance for creep and a future splice. That belt was correctly graded and correctly spliced, and it was still going to run out of stroke inside four years for reasons that had nothing to do with the cord.

06Step 6: Check the Drive, the Starting Case and the Take-Up Behavior

A belt that is correct in steady state can still fail on the way to speed. Acceleration is where the inertial term enters, where motor torque is highest, and where a marginal drive design shows itself in slip marks on the lagging and a burnt smell near the pulley.

Start with running power. Multiply effective tension by belt speed and divide by the drive efficiency. On the worked example, 242 kN at 4.5 m/s is roughly 1,090 kW at the belt, and with a realistic efficiency chain you end up needing something near 1,250 kW installed. If the installed power is below that, the conveyor is not going to reach its design tonnage, or it will reach it with the belt running hotter and more heavily loaded than the calculation assumed.

Then test the starting case properly. Acceleration tension is the inertial mass of belt, material and rotating idler parts multiplied by acceleration. With our numbers and a 0.15 m/s² ramp, that adds about 58 kN, so the tight side climbs from 352 kN to roughly 410 kN, and the safety factor at start drops from 6.8 to about 5.9. That is still inside the band, but it is also the reason a soft start, a fluid coupling or a variable frequency drive is worth the money on any long steel cord installation. Across four or five drives the same check matters even more, because load sharing between them is rarely perfect and the worst drive sets the tension in its own section.

Do the deceleration case as well, especially on any conveyor with lift. A loaded incline that stops needs enough braking to prevent runback, and it needs the brake torque to be gentle enough not to shock the splice. We have seen inclined belts sized perfectly for running tension and destroyed by a brake that grabbed. Ask the question during sizing, not after the first incident.

Confirm the take-up arrangement while you are here. A gravity take-up holds slack side tension roughly constant, which is the friendliest behavior during starting and gives the belt a chance to absorb acceleration without losing grip. A winch or hydraulic take-up with a fixed stroke can work well, but it must be set so that the belt is not over-tensioned in the cold and under-tensioned on a hot afternoon. Stroke available and stroke usable are different numbers, and only one of them is on the drawing.

Finally, check the two geometric items that steal belt life without ever appearing on a belt data sheet. Idler spacing and sag set the minimum tension the belt must carry between supports, and the old rule of holding sag near two percent of spacing still works well for a steel cord belt with fine material. Vertical curves set a minimum radius that grows with tension, and a long overland line with a concave curve needs that radius checked against the actual tension at that point rather than an average. On tunneling and pit conveyors the same geometry check is what keeps a mining conveyor belt from wearing its edges against the structure.

We checked a 2,400 t/h phosphate conveyor where the installed drive was 1,100 kW. Running demand came out at 1,240 kW. The conveyor had been running for years, which is exactly why nobody had questioned it, but the belt was living with a permanently higher tension and a hotter pulley than the design intended. Nothing on that installation looked wrong. Everything was simply a little more loaded than anyone had calculated.

07Step 7: Write a Specification Sheet Someone Else Can Re-Derive

Every number on the sheet should trace back to a duty figure, a standard or a calculation.

A specification is finished when a second engineer who has never spoken to you can read it, redo the tension check, and arrive at the same grade and the same pulley diameters. If any number on the page cannot be traced to a duty input, a standard or a calculation, it does not belong there.

Put the duty data at the top in the same units you collected it. Design tonnage, surging figure, bulk density, lump size, belt speed, center distance, lift, incline, trough angle, material temperature and the operating standard that applies. Then give the calculated effective tension, the tight side tension, the tension per millimeter and the safety factor at both running and starting. That short block is what makes the rest auditable, and it takes ten lines.

Then the belt itself, described fully enough that a substitution cannot slip through unnoticed. Belt width and grade alone are not a specification. You need cord diameter and cord pitch, the number of cord layers or a note that it is a single plane construction, top and bottom cover thickness, the compound type and the standard it is tested against, edge construction, and the finished belt mass per meter. For steel cord work the governing documents are normally DIN 22131 and ISO 15236-1, with DIN 22102 shaping cover classes on the textile side and ISO 4649 providing the abrasion loss test behind most compound claims.

Wrap the interface requirements into the same document. Minimum drive, bend and tail pulley diameters, the pulley lagging type, the splice method and nominal splice length, cure temperature and pressure range, the usable take-up stroke required, and the direction of belt run marked on the roll. Thirty extra lines here prevent a conversation that costs three weeks later.

Add the evidence you expect to receive. A dimensional report, cover thickness measurements taken on the finished belt, adhesion values between cover and cord, and a tensile test on the finished carcass. Every belt we ship carries a test record with the batch references behind it, and those batch records are kept because a question about a belt three years into service is always answered faster with paper than with memory.

One more habit is worth copying. Keep the specification numbering consistent with your plant asset tags, and store the approved sheet next to the conveyor rather than in a project folder that gets archived. When the belt is replaced in eight years, the person doing it will not have your email, but they will have the file, and they will know exactly which numbers were assumptions and which ones came from a measurement. Our own release route runs through a documented quality assurance sequence, and the inspection plan on the sheet is what makes that sequence meaningful instead of decorative.

You can also see how this closes the loop opened in step one. The duty data that started the process is the first thing on the sheet, and the acceptance tests at the end are the only way to prove the sheet was honored. A sizing chain that skips either end is not a chain at all, it is a series of purchases.

08Step 8: Verify at Acceptance and Feed the Result Back

Acceptance is not a signature on a delivery note. It is the moment the calculation meets a tape measure, and it is the last chance to catch a substitution before the belt is buried under a load of ore.

Do the documentary check first while the rolls are still on the truck. Confirm the width, the finished gauge and the cover thicknesses against the approved sheet. Look at the length figure and the roll marking, because a belt spliced the wrong way round on a directional conveyor is an expensive correction. Then take a short sample from an end, peel it apart and count the cords, confirming diameter and pitch against the approved drawing. That single destructive check on a 300 mm offcut is the cheapest insurance available on any order.

At the splice, ask for the cure log rather than a verbal statement. Temperature, pressure and time held should all sit inside the range agreed with the belt maker. A splice that was cured short will pass every visual inspection and will still open in service, and by then the cause is unprovable.

Then keep measuring. A first-week stroke reading and a ninety-day wear reading turn a specification into a feedback loop, and a feedback loop is what makes the next order better than this one. If the drive current comes in 8 percent above prediction, that tells you the friction factor in the calculation was optimistic, and you should carry that correction into the next conveyor of the same type.

If you are buying across a language boundary, or dealing with a brokerage layer between the plant and the mill, use a written verification list rather than trust. Our buyer's checklist of twelve things to verify before ordering covers the commercial side of that process, and it pairs well with the technical steps above.

One closing thought on the chain. Every failure we get called about has an origin somewhere in these eight steps, and the origin is almost never the cord. It is a missing surge factor, an assumed take-up, a pulley left at the old diameter after a grade change, or a splice cured in a hurry. Teams that write the duty data down and check the sheet at acceptance rarely call us about a failed belt. They call about the next conveyor.

Send us your duty data for a sizing check

09Frequently Asked Questions

What tension should a steel cord belt be sized against?

Against the tight side tension at the point of maximum load, not against average tension and not against motor power. On an inclined conveyor that point is usually just after the head pulley on the loaded side, and it is where the cord sees its highest stress. Convert that figure to newtons per millimeter of width, then apply your safety factor. A belt sized on average tension will pass every meeting and fail in the field, and the splice is normally the first thing to tell you.

Which ST grade fits a 1,200 m conveyor carrying 3,000 t/h of iron ore?

Run the numbers rather than the ladder. In our worked example the effective tension lands near 242 kN, the tight side near 352 kN, and the demand near 294 N/mm of width on a 1,200 mm belt. ST1600 gives a factor of about 5.4, which is too thin for steady running, and ST2000 gives about 6.8, which sits inside the band we work to. The answer is a calculation with a grade as its output, never the other way round.

How much take-up stroke does a steel-cord belt need?

More than most people expect, even though steel cord belts stretch far less than fabric belts. Add elastic elongation at working tension, permanent creep over the first year, one splice length of allowance for every future joint you expect to make, and a reserve on top. On a 1,200 m center distance that lands near 8 m of usable travel. Check the usable figure on site, because nominal carriage length and available stroke are rarely the same number.

Can I reduce the pulley diameter to save space or cost?

Only if the cord allows it, and the cord is not negotiable once the grade is fixed. Minimum pulley diameter scales with cord diameter, and cutting it produces fatigue in the outer cords that no cover thickness can compensate for. If space forces a small pulley, the honest move is to review the grade, the drive arrangement or the conveyor route. We would rather redraw a layout than supply a belt that dies at the pulley.

Why is a steel cord belt spliced on site instead of at the factory?

Because a single-flight belt on a long conveyor cannot be transported as one piece, and because the joint has to be made in the same carcass that will carry the load. That puts the splice quality in the hands of the site crew, which is why we ask for a cure log and not a photograph. Length, temperature, pressure and time all matter, and a joint cured at the wrong temperature will look perfect for a year. Portable presses and trained splicers are what turn a good belt into a good installation.

How do I compare quotes for conveyor belt steel cord supply from different mills?

Compare the specification line by line before you compare the price line. Cord diameter and pitch, cover thicknesses, compound and compound standard, edge construction, finished belt mass, and the tests that will come with the delivery. Two offers that both say ST2000 on 1,200 mm can still differ by several millimeters of cord diameter and by an entire cover compound class. Ask what happens at acceptance and what evidence is supplied, and the cheapest offer usually stops being the cheapest.