Conveyor Belt Steel Cord Construction: Cord Build, Load and Splice

Two steel cord belts can leave the same factory carrying the same 1,600 N/mm stamp and then behave nothing alike. One runs four years on a 2,900 m overland curve. The other throws cords out of a splice after eleven months. The strength class is rarely what separates them. The cord build underneath is.

Steel cord is not one product. It is a family of rope structures, and every structure trades stiffness against fatigue resistance, adhesion against openness, and splice economy against belt life. A 7×7 cord and a 7×19 cord can both deliver ST1600, yet the pulley diameters they tolerate, the step length their splices need, and the take-up travel they demand from your tower are not the same numbers.

This page stays on one layer only. We read the construction code, follow the load path from a single wire up to the rated belt strength, and then look at where cord build actually shows up on site — in take-up travel, splice retention, tracking behaviour and the first signs that a cord is about to break. If you are still working out which class or cover grade you need, start with the steel cord conveyor belt range overview first. Here the class is fixed and the only question left is what sits inside the rubber.

Send us your cord drawing and duty point — we will tell you which cord build fits

01Cord Construction Codes: Reading 6×7+IW, 7×7, 7×19 and 6×19

A cord designation is a compressed description of how the wires are assembled, and it reads left to right. The first number counts the outer strands. The second counts the wires twisted into each of those strands. Anything that follows a plus sign names the core the outer strands close around.

So 7×7 means seven strands carrying seven wires each, which is forty-nine wires in one cord. A 7×19 cord means seven strands of nineteen wires and one hundred and thirty-three wires in total. Then 6×7+IW means six strands of seven wires laid around an independent wire rope core, and that core is a small rope in its own right, so it thickens and stiffens the middle of the cord rather than simply filling it.

Why anyone should care comes down to wire diameter. Push a given steel cross-section into more and smaller wires and the cord bends more easily over a drum, because every wire sees a shallower bend radius and less surface strain. Pull the same cross-section into fewer and thicker wires and the cord becomes stiffer, draws more cheaply and resists abrasion at the cord surface better, but the outer wires work harder on every wrap. It is the same tonnage of steel doing two different jobs.

On a 1,200 t/h coal line we swapped 7×7 ST1250 for 7×19 ST1250 across an eight-pulley profile with a 630 mm drive drum. Cord breakage in the drive-side splice fell from roughly one event every fourteen months to none across three years of service. The strength class never moved. The wire diameter and the wire count did.

Two further properties travel with the wire count. Openness describes how much gap exists between wires for compound to enter, and lay length describes the twist distance of the strands. Shorten the lay and the cord grips rubber better and resists bending fatigue better, but it gives up a small amount of breaking force because the wires no longer run parallel to the pull. That trade is normally worth taking when the lay sits between 6.5 and 10 times the cord diameter.

Openness is measurable rather than a matter of opinion. Cut a belt sample, polish a cross-section and count how much of each cord perimeter has compound bonded against it. Specifications written around ISO 15236 commonly ask for rubber penetration in the range of eighty to ninety percent on cords of 4 mm and above, and the test belongs in the pre-shipment file next to the cord breaking force records. We hold penetration sections and pull-out data for every batch under quality assurance, because a cord can be fully embedded in compound and still walk out of the belt if the bond line itself is weak.

Cord count per unit width is the last figure worth pulling out of the belt drawing. A 1,200 mm belt on 16 mm pitch carries 75 cords across its width. Run the same belt on 14 mm pitch and there are 85 load paths. More cords at a closer pitch gives better damage tolerance and a lower tension per cord, but it also thins the rubber bridge between neighbouring cords and raises the risk of cord contact if the belt is troughed steeply. How those numbers land in a real order is set out in the cord specification and carry capacity breakdown, and the wider commercial picture sits in the 2026 steel cord belt guide. As a conveyor belt manufacturer we see the same pattern at the drawing stage, where a customer specifies a class correctly and a cord geometry almost by accident.

02From One Cord's Breaking Force to the Belt's Rated Strength

The strength class on the belt is not a property of the rubber and it is not even a property of a single cord. It is the outcome of three numbers multiplied and divided against each other, and once you can do that arithmetic you can audit any quotation in about a minute.

Start with the cord itself. Every cord has a nominal breaking force, quoted in kilonewtons or newtons, measured on a tensile machine with the sample gripped in a way that avoids jaw failure. Then measure the spacing between neighbouring cords. That spacing is the pitch, and it behaves like the repeat distance of a fence. Divide the cord breaking force by the pitch in millimetres, and you have the belt's nominal rated strength in newtons per millimetre of width.

Take a cord with 25.6 kN breaking force running at 16 mm pitch. Divide 25,600 N by 16 mm and the answer is 1,600 N/mm, which is where the ST1600 label comes from. Change the pitch to 20 mm and the same cord only delivers 1,280 N/mm. Nothing about the cord changed. The way it was spaced in the belt changed, and the belt class changed with it.

This is why two suppliers can both quote ST2000 and ship belts with different cord diameters. One runs a heavier cord at wider pitch, the other a lighter cord at closer pitch. Both arithmetic paths land on 2,000 N/mm, and both will pass a static tensile test, but the two belts will not age the same way on a curved overland route.

DIN 22131 and ISO 15236 both frame the class in terms of the nominal belt breaking strength rather than a specific cord, which leaves the cord diameter and pitch as the manufacturer's engineering choice. That freedom is real but it is not unlimited. A shorter pitch raises the cord count and lowers the tension each cord carries, which helps fatigue life but compresses the rubber bridges between cords. A wider pitch does the opposite. Somewhere between the two sits the geometry that suits the duty.

Numbers in a table are a starting point, not a guarantee. We once tested a batch of nominally 5.5 mm cord and found individual breaking forces spread across a 1.8 kN band, with the weakest sample sitting three percent under the drawing value. Nothing failed, because the belt class had been set with headroom. Had that same spread occurred on a belt sized to the arithmetic alone, the weak cords would have been the ones to break first, and they would have broken at the splice rather than in mid-span.

Run certification at the cord supplier deserves the same attention as the finished belt. Cord breaking force records, wire tensile grade and, above all, whether the whole coil came from one heat are all worth asking for. If you are comparing suppliers, the step by step sizing walkthrough covers the demand side of the calculation, while the cord geometry is the supply side you are checking here. Buyers working through a conveyor belt supplier or a industrial conveyor belt catalogue should expect the cord diameter and pitch to be stated on the offer, not just the class.

03Adhesion: The Bond Line That Decides Whether the Cord Ever Works

A cord carries tension only if the rubber around it shares that tension over a long enough length. Copper-zinc brass plating on the wires is what makes that sharing possible, because during vulcanisation the brass reacts with the sulphur in the compound and forms a chemical bridge between steel and rubber. Get the plating wrong and you still have a belt, but you have a belt in which the cords slide inside their own tunnels.

Brass coating is specified by mass and by ratio. A typical conveyor belt cord carries a coating in the region of 3 to 5 grams of brass per kilogram of cord, with a copper to zinc ratio near 67:33 by weight. Shift that ratio and the film changes character. Too much copper and the bond becomes brittle and loses adhesion quickly under hot, humid conditions. Too much zinc and the reaction during cure is sluggish, so you get a belt with excellent initial appearance and mediocre pull-out figures.

Pull-out force is the number that gets recorded. A single cord is extracted from a cured belt sample and the force per unit length is logged. Requirements vary with cord diameter and with the standard being worked to, but a common specification band for cords between 4 mm and 8 mm sits around 45 to 60 N/mm on delivery, with aged samples held at elevated temperature for a defined period expected to retain at least seventy-five to eighty percent of that value. Cobalt salt additions in the bonding compound are one of the usual levers used to hold the aged number up.

We have measured pull-out figures on a 7×19 ST1600 belt after 168 hours at 90 °C and found retention at 82 percent, well inside the specification. The same compound on a 6×19 cord of similar diameter returned only 68 percent, because the tighter strand geometry left less compound between the wires, so the bond area per unit length was smaller. Same rubber, same cure cycle, different cord build, measurably different ageing behaviour.

Penetration and pull-out are usually tested together because they fail for opposite reasons. A closed, densely packed cord gives a high single-cord pull-out value in the laboratory because so much surface area touches compound, yet compound cannot reach the inner wires, so those wires corrode and fret. An open cord lets compound reach the core, but if the twist is loose the pull-out figure drops. The cord build is where those two effects are balanced.

Splice adhesion deserves its own line in the inspection plan. Rubber compounds formulated with rubber conveyor belt friction grades in a fabric belt are not automatically suitable for a cord splice, and the same holds for a transmission belt manufacturer grade moving into heavy cord work. Ask for the splice compound by name and check that the cure cycle matches the belt rather than the workshop's habit.

04Elongation and Take-Up Travel: What a Fraction of a Percent Costs You

Steel cord belts are chosen for long routes partly because they barely stretch. That is true in absolute terms and misleading in practice, because a very small percentage applied to three kilometres is still several metres of travel, and the take-up tower has to be built for it before the belt ever arrives.

Two separate mechanisms produce the movement. Elastic elongation comes from the steel itself stretching under load and recovers fully when tension is released, and its size is governed by the belt's elastic modulus. Structural elongation comes from the rope geometry straightening out, wire against wire, under first tension, and most of it is permanent. A belt with a coarse, few-wire cord has less structural stretch than one built from many fine wires, because there is less internal freedom to take up.

On a 2,900 m overland route we recorded 2.6 m of take-up travel with a 7×7 ST1600 belt, and the tower had been sized for 3.2 m. A later rebuild used a 7×19 cord at the same class and the same tonnage, and the stroke went to 3.9 m. Nothing about the load changed. The finer wires simply straightened more under first tension, and the take-up had to give back the difference every time the belt cycled.

Creep adds a slow, one-way component on top. Steel cord belts creep far less than fabric belts, but on a hot overland route it still accumulates, and it interacts with splice settlement. Design practice for long steel cord conveyors usually leaves a stroke reserve of twenty to thirty percent above the calculated figure for exactly that reason.

Where the take-up sits matters as much as how big it is. A gravity tower placed before the drive sees a different tension than one placed after it, and on a two-flight overland system the stroke figures must be read per flight rather than added together. The conveyor pulley diameters on the route set the other half of the picture, because a belt strained over undersized drums loses cover and cords long before the take-up ever runs out of travel. If you are pricing a replacement through an wholesale conveyor belts channel rather than direct, put the required stroke in writing, since it is the one number that turns a cheap offer into an expensive one. The cost calculation and comparison guide shows how stroke and cord geometry feed into total cost of ownership.

05Same Rated Strength, Different Cord Build: The On-Site Difference

This is the comparison most quotations never show, and it is the one that decides how a belt behaves in year three. Hold the class constant, hold the belt width and the tonnage constant, and let the cord build change. Everything that follows on site changes with it.

The first consequence is the drum diameter the belt can survive. Fine-wire cords tolerate a smaller drum because each wire bends through a gentler arc at the same drum diameter. Coarse cords need a bigger drum or they start to show wire fatigue from the inside, which is invisible until a cord finally lets go and the belt edge kicks.

The second consequence is splice length. A fine-wire cord needs a shorter step to develop its bond, because the bond is distributed over more, smaller wires. A coarse cord needs a longer step, and on wide high-class belts the difference runs to several hundred millimetres per step, multiplied across every step in the splice. That is a real cost in rubber, cure time and downtime.

The third consequence is damage tolerance. Hit a coarse-cord belt with a sharp lump and one cord may lose most of its section, taking a disproportionate share of the tension with it. Hit a fine-wire belt in the same spot and the load redistributes across a denser field of cords, so the belt keeps working while you plan the repair.

There is a fourth and less obvious one. Tracking behaviour depends on how evenly the cord field is laid across the belt width, and a coarse six-strand cord is harder to lay evenly than a fine-wire seven-strand cord. Over a curved overland route and a few hundred metres between idler sets, that evenness shows up as a slow, persistent drift that no amount of skewing fixes.

Choosing between them is not a matter of always buying the finest cord available. A short plant conveyor with a fixed drum package and low cycle count gains nothing from a 7×19 cord except a higher price and a longer commissioning stretch. The finer cord earns its money where cycles are high, drums are small, the route curves, or the load is dirty and impact-prone. That is the same logic a conveyor belt distributor is expected to be able to argue through with the numbers, and it is the sort of thing a conveyor belt factory normally knows but rarely writes down. A useful cross-check on the ageing side sits in the inspection and failure prevention notes, where field failure patterns are grouped by what actually failed first.

Duty point, not preference, should settle it. A quarry face conveyor moving 900 t/h over 380 m with a 500 mm drum is a poor candidate for fine-wire cord. A 2,800 m curved overland line on a 1,000 mm drum is a poor candidate for a coarse one. The mining and quarrying duty notes and the cement plant notes both show how much the answer swings between those two extremes. Talk it through with a conveyor belt distributor who has the profile drawings, or with the conveyor belt factory directly if the profile is unusual.

06Cord Build Decides How the Splice Is Cut and How Much Strength It Keeps

A steel cord belt has no mechanical fastener worth trusting. Every metre of it is joined to the next by hand, in the field, with heat and pressure, and the joint only develops strength through the length of cord that is embedded alongside its neighbour. Cord construction therefore sets the splice geometry directly, and the splice is where a large share of cord belt failures begin.

In a step splice the cords of one belt end are laid alongside the cords of the other within the same rubber envelope. The distance over which the two cord sets overlap is the step length, and the distance between successive cord ends across the belt width is the step pitch. Both are functions of cord diameter and cord construction. A fine-wire cord develops its working bond over a shorter distance, so the steps can be shorter and the whole splice more compact. A coarse cord needs more length for the same bond force, and it also needs wider spacing between adjacent steps to avoid cord congestion.

At a cement plant we sectioned a step splice on a 1,400 mm ST2000 belt that had run twenty-six months and failed at the drive pulley. The step length had been cut at 250 mm where the cord geometry called for 400 mm. Cord ends were clean and brass was still visible on the exposed wires, which told the whole story. The bond had never been the limiting factor. There simply was not enough embedded length to transfer the tension, and the splice had been quietly running at a fraction of its design capacity from day one.

Physical size is not the only variable. The arrangement of cord ends matters too, and reputable practice alternates the free ends in a defined pattern so that no cross-section of the splice loses too many cords at once. Where a belt must be joined to a different cord build, a transition splice with an intermediate geometry is the only sound approach. Mixing a coarse cord end against a fine cord end in one step leaves the coarse cords carrying more than their share.

Vulcanisation is where a good geometry can still be ruined. Splice compound is usually cured in the region of 145 to 150 °C, under press pressure of about 1.4 to 1.8 MPa, with the holding time set by the full thickness of the joint rather than the belt itself. A splice that is under-cured looks finished and feels solid, then loses adhesion at the first hot day. One that is over-cured goes brittle at the bond line and fails in fatigue. Neither shows up in a visual check.

Read the endurance column carefully. A static tensile test on a spliced belt sample can return ninety percent of the parent belt strength and still leave you with a joint that fatigues out at sixty percent after a few tens of thousands of drum cycles. Long, heavily loaded conveyors live in the second column, not the first, which is why the drum endurance figure belongs in the specification alongside the static number.

Two practical rules follow. First, never let a splice be cut to the workshop's habitual step length when the cord build has changed, because the old numbers are tied to the old cord. Second, photograph and record the step geometry of every splice made on your site, together with the belt batch it joined. When a splice eventually fails, that record is what tells you whether the geometry or the cure was at fault. Structural records of this kind sit alongside the belt specification work collected in the tensile strength and thickness reference, and the ordering side is covered item by item in the twelve point verification checklist.

For cord belts running in matched drive systems, the same discipline applies to the drive side. A V-belt manufacturer tunes belt pairing to keep load shared, and the principle is identical on a twin-drive conveyor. Both motors must see the same tension, or one drum does more work and one section of splice does more work with it. Belt monitoring and on-site splice support can also be arranged through our field service team, and the full cord belt range is listed on the steel cord belt specification page.

07The Cover Gauges and Compounds That Protect the Cord Line

Cord build sets what the belt can carry. Cover rubber decides how long the cord stays worth carrying it. The two are specified separately and they interact more than most order sheets admit.

Cover compound on a cord belt is selected against DIN 22102 style grades where abrasion resistance is the driver, and against RMA cover grades in markets that work to that convention. The DIN abrasion figure is quoted as a maximum volume loss in cubic millimetres from a standard test, and the RMA grades are quoted as a maximum volume loss in cubic inches, with Grade I the tighter limit and Grade II the looser one. A cord belt hauling sharp granite wants the tightest compound you can justify. A cord belt moving dry sand at moderate temperature may not need it.

Top cover gauge is the other half. Thick cover protects the cord at the loading point, where lumps arrive and where impact is concentrated. It also adds weight, adds bending stiffness and adds cost on every metre of a two-kilometre belt. We have seen a 3,000 m overland line specified with 12 mm top cover purely out of habit, when the loading point occupied forty metres of the route and a 6 mm cover with a heavier cord would have done the same job for a lower total mass and a lower drive power. The cord does not need the cover in mid-span. It needs it under the chute.

Bottom cover is usually overlooked and it should not be. On a cord belt the pulley side takes the wraps, and it is the surface that wears against the drum and against return idlers. A 4 mm bottom cover on a 2,500 m belt with a tight pulley package will wear through at the drum before the top cover wears anywhere. Under-specifying the bottom cover to save weight is one of the more expensive habits in cord belt purchasing, and it shows up as cord exposure on the pulley side with the top surface still looking healthy.

Edge construction deserves a line of its own because edges are where cord belts die. The edge rubber holds the outer cord in place and absorbs the wandering contact with the structure. When it wears, the outer cord is exposed and starts to corrode and fret, and the belt begins to drift toward that side. That step from edge wear to tracking drift to cord damage is the normal sequence, and the grades and gauges that slow it down are set out across the abrasion resistant, heat resistant and impact and cut resistant cover options. Component hardware that dresses the belt correctly on the return run is covered under conveyor components, and compound selection sits inside industrial conveyor belt cover families rather than being made up on the spot.

08Early Signals of Tracking Drift and Cord Breakage

Cord damage almost never announces itself. The wires are buried under cover rubber, and by the time a broken cord becomes visible on the surface, the section around it has usually been overloaded for weeks. What you can see earlier is the belt's behaviour changing, and those behavioural changes are worth reading systematically.

Tracking drift is the most misread of them. A belt that has always run true and then begins to pull to one side through one particular pulley, while running normally for the rest of the route, is rarely a frame alignment problem. A localised loss of cord tension on one edge, caused by a cord break or by an exposed outer cord corroding, changes the strain distribution across the width and the belt steers toward the weaker side. Adjusting the idlers will mask it for a shift and it will come back.

We tracked a 1,600 mm belt through four months of exactly that pattern. The drift appeared only through the head pulley, and only at above 800 t/h. Cord inspection with a magnetic scanner found three broken cords within a nine-metre section some forty metres back from the head. Once that section was spliced out, the drift disappeared and no idler had been touched. The tell was that the drift was load dependent and pulley specific, which no alignment fault produces.

A second class of signal comes from the surface. A slight ripple or a soft, raised line running across the belt at an angle to the travel direction usually marks where the cord field underneath has changed shape. Press the area with a hand or a probe and compare it with the neighbouring metre. A local soft spot that springs back slowly, or a bulge that appears only when the belt is under tension, is worth marking and watching on every inspection round.

Noise and vibration belong on the list too. A dull repetitive knock at one pulley, timed to belt speed rather than pulley rotation, points at a cord end that has migrated out of position inside a splice, or at a broken cord end lifting the cover as it passes the drum. Timed the other way, to pulley rotation, it is a pulley or bearing problem instead. Getting that distinction right early saves a lot of unnecessary belt cutting.

Instrumentation is worth the money on a belt that matters. Magnetic and, on some sites, X-ray scanning can map broken cords along the whole belt and hand you a list of positions rather than a suspicion. On a 2,600 m overland we ran a scan every six months and used the cord count in each anomaly to schedule repairs, which turned an unplanned failure pattern into a planned one. The inspection depends on the carrying side being clean and reasonably dry, so book it against a shutdown rather than a production day.

Everything outside the belt has to be right as well, or you will keep chasing cord damage that is really hardware damage. Return side conveyor rollers that have seized flatten a spot into the belt on every revolution. A training idler set too aggressively will wear the edge rubber and expose the outer cord within a season. Self aligning idlers should correct drift, not create it, and if the drift returns after adjustment that is your cue that the cause is inside the belt. At the loading point, impact idlers and a correctly set skirt take the strike energy before it reaches the cords, and a steel conveyor roller of the right wall thickness keeps the return strand from sagging into the structure. On a return idler spacing that is too wide, the belt sags, the edges fold and the outer cords work against each other.

Tell us your belt profile and we will recommend a cord build

09Frequently Asked Questions

How do I read a cord construction code such as 7×19 or 6×7+IW?

The first number is the count of outer strands and the second is the number of wires inside each strand, so 7×19 means seven strands of nineteen wires and one hundred and thirty-three wires in the cord. A suffix after a plus sign names the core. In 6×7+IW there are six strands of seven wires laid around an independent wire rope core, which adds stiffness and diameter in the middle. More wires of smaller diameter make the cord bend more easily over a drum. Fewer, thicker wires make it stiffer and usually cheaper.

If two belts share the same strength class, does the cord diameter have to match?

No, and this is the single most useful thing to check on a quotation. Take the cord breaking force in newtons and divide it by the cord pitch in millimetres, and the result is the rated strength per millimetre of width. A 25.6 kN cord on 16 mm pitch gives 1,600 N/mm. A lighter cord on a closer pitch can reach the same class, and a heavier cord on a wider pitch can also reach it. Both pass a static test, but they behave differently on fatigue, splice length and take-up travel.

Why does take-up travel run out earlier than the drawing predicted?

Two effects are usually behind it. Structural stretch from the rope geometry straightening under first tension is larger on fine-wire cords than on coarse ones, and it is mostly permanent. Creep adds a slow one-way component on top, and splice settlement adds a little more in the first year. On a 2,900 m route we measured 2.6 m of stroke with 7×7 cord and 3.9 m after rebuilding with 7×19 at the same class. A stroke reserve of twenty to thirty percent above the calculated figure covers that gap.

How long should the splice steps be on a fine-wire cord build?

Shorter than most workshops assume for a coarse cord and, more importantly, tied to the cord geometry rather than to habit. On a 1,200 mm belt at 1,600 N/mm, a 7×19 or 6×19 build typically needs a step pitch near 250 mm and an overlap around 1,200 to 1,250 mm, while a 6×7+IW build needs a pitch of 300 to 350 mm and roughly 1,500 mm of overlap. Static retention after cure can reach ninety percent or more, but retention after drum endurance cycling typically sits between fifty-five and seventy percent, and that is the figure long conveyors should be sized on.

Can I splice two belt ends with different cord constructions together?

Not in a standard step. If one end carries coarse cord and the other fine wire, the coarse cords take more than their share of the tension across the joint and the splice fails early. The sound approach is a transition splice with an intermediate geometry that transfers load gradually between the two cord fields. The same rule applies when a rebuild changes cord design part way through a route, which is why we ask for the existing belt drawing before quoting a replacement, not only the strength class.

What is the earliest sign that a cord has broken inside the belt?

Usually a change in behaviour rather than anything visible. A belt that starts drifting only through one pulley and only when heavily loaded is showing a local loss of cord tension on one edge. A soft spot that bulges under tension, a raised line running across the belt, or a repetitive knock timed to belt speed rather than pulley rotation all point the same way. Magnetic scanning confirms it by listing actual broken cord positions, and it is worth booking against a shutdown rather than a production day.