Steel Cord Belts Maintenance: Inspection, Lifespan and Failure Prevention

Steel cord belts almost never fail because the wire broke first. The wire breaks because something upstream was tolerated for months — a worn skirt rubber, a misaligned chute, a seized idler, a take-up nobody re-set after the winter shutdown. Of the three components that decide belt life, the steel cord is the most forgiving. The cover, the splice and the loading point are where the money goes.

What follows is a maintenance document, not a catalogue page. It hands you a time-based inspection schedule you can paste straight into an SOP, a comparison of the four realistic ways to find broken cords, grading rules for cover wear and aging, an attribution order for splice failures, and a cost frame for choosing between repair and replacement. If you need tensile classes or carry capacity tables first, begin with our guide to conveyor belt steel cord specifications — the numbers there are the same ones our thresholds are built on. Everything below assumes the belt is already installed, tensioned and running.

Send us last quarter's inspection sheets and we will tell you what they are already saying

01Why Steel Cord Belts Fail, and What a Day of Downtime Costs

Start with the arithmetic nobody wants to do on the morning of a stoppage. A 1,200 t/h line running two ten-hour shifts moves roughly 24,000 t per day. If the material carries a contribution margin of a few dollars a tonne after processing, a single lost shift is a four- or five-figure number before anyone calls the belt supplier. Add the second shift if the repair is a splice, because vulcanizing a 1,200 mm wide ST1600 joint means press time, cooling time, and a crew that has to be found at 02:00. None of that shows up in the belt price tag, and all of it shows up in the plant P&L.

That is the practical reason maintenance matters more on steel cord belts than on fabric belts of the same tonnage. The replacement cost of the belt is a one-off. The cost of unplanned downtime recurs every time a preventable failure repeats.

Three failure families, three completely different responses

Gradual failures — cover abrasion, ozone and heat aging, slow cord fatigue — are predictable. They give you months of warning if you measure on a schedule, and they should never cause a surprise stoppage on a well-run line. Event-driven failures — a longitudinal tear from a trapped steel plate, a puncture at a transfer point, a splice pull-out — happen in seconds and are only limited by how fast your detection and shutdown systems are. Operational failures sit in between, and they come from misalignment, overloaded starts, or a chute that has quietly stopped delivering material in the centre of the belt.

The three need different countermeasures. Measurement and planning handle the first. Chute design, trap iron removal and tear detection handle the second. Discipline on start-up and loading handles the third. A maintenance program that treats all three the same way usually does none of them well.

Where failures actually cluster

Across the mining, port and cement installations we service, emergency stops are dominated by splices and by the loading zone. Cover wear is the most common finding, but it is rarely the reason a line stops without warning — worn cover usually announces itself as a planned change-out with a date attached. Cord damage in the belt body, away from a splice, is the rarest and the most expensive, because it very often ends with the belt being shortened or replaced.

One consequence of that distribution is worth stating plainly. If your maintenance budget and your inspection hours are spread evenly along the whole conveyor, they are in the wrong place. The first 20 metres after the loading point and every splice should absorb most of your attention. That is where a conveyor belt manufacturer gets most of its warranty questions from, and it is also where the answers are cheapest to implement.

The rest of this article is that program in order.

02A Time-Based Inspection Schedule for Shift, Daily, Weekly, Monthly and Quarterly Checks

A maintenance program for these belts only works when it is broken into frequencies that a normal crew can actually keep. Asking a shift fitter to measure cover thickness along 2 km of belt is how inspection programs die in month two. The split below assigns the cheap, fast checks to the shift crew, the measurement work to the weekly and monthly route, and the instrumented work to planned shutdowns.

Two rules keep it honest. Write the numbers on the same sheet every time, so a trend exists rather than an opinion, and keep one map of the conveyor with chainage marks — "cover worn near the drive" is useless when the work order needs a specific 50-metre section.

The shift-to-quarter inspection matrix

Cover stripped back on a test length: the cords are parallel and spaced evenly in one plane, which is why cord damage shows up as a local weakness rather than a gradual loss of strength.

The thresholds above are the ones we use in commissioning advice, and they are deliberately conservative for high-tonnage lines. On a lower-duty installation you can relax the monthly cover limit to 1.5 mm and still get a year of warning. What you cannot skip is the act of measuring at the same chainage every month. A single reading tells you nothing; the slope between readings tells you when to book the shutdown.

One thing we insist on before quoting any replacement: the take-up position history. If a belt has been re-shortened twice and the counterweight is still creeping up, the belt has already lost cord strength and the next failure will not be a cover problem. On a 1,200 t/h coal line last year, the monthly take-up readings alone predicted a splice failure eleven weeks ahead of the event. We have built belting since 1988 on ten lines — eight fabric and two steel cord — and the same rule holds for every industrial conveyor belt we have shipped, whatever the application.

03Finding Broken Wires and Cord Damage: Four Methods and Their Blind Spots

This is the part of maintenance where crews most often believe they are covered when they are not. A belt can lose a handful of wire strands inside a cord and keep running for a year with nothing visible from the walkway. By the time a bulge appears on the surface, the damage has usually taken the form of internal corrosion, and the corrosion has grown into a void that pushes the cover up. The visible sign is the end of the story, not the beginning of it.

So the question is not which method is best. It is which method answers which question, and how often you can afford to ask it.

What each method actually detects

The practical answer for most sites is a two-tier program. The electromagnetic scan is your early-warning net because it is the only method that covers the whole belt at a realistic pace and can show a strand breaking before the surface changes. The tap test and the visual walk then become confirmation tools aimed at whatever the scan flagged. Visual inspection on its own will always lag the damage by months.

Reading a scan result without overreacting

Not every flagged anomaly means the belt is finished. A single broken strand in a cord of 7x7 or 7x19 construction is a local event, and if nothing is corroding around it, the belt will keep running. What changes the decision is the count and the spread. Several broken strands in the same cord, or any damage spread over more than one cord pitch, is where we move from watching to planning.

Keep the geometric picture in mind when you interpret a scan. The cords sit in a single plane at a fixed pitch, typically 10 to 20 mm apart depending on the tensile class, so a scan that reports cord spacing can usually tell you whether an anomaly is one cord or several. Cord construction matters too, and that is why we specify it in writing on every order for a rubber conveyor belt as well as for steel cord. A 7x19 cord tolerates losing an outer strand differently than a 7x7 cord does, and the reporting threshold on your scan should be set accordingly.

Two habits pay for themselves. Photograph every anomaly with a tape measure in frame and file it under chainage, not under date. And reconcile the scan report against your own notes before the next shutdown, because a discrepancy usually means an access problem. That is why our quality assurance procedure asks for the inspection route map along with the belt order.

If you drive your own conveyors with V-belts or synchronous drives, the same logic applies to the transmission side. Cracks at the base of a raw-edge V-belt usually appear long after the internal cords have started to fatigue, so a transmission belt manufacturer will normally ask for running hours and pulley diameters before advising on replacement intervals rather than judging from the outside surface.

04Cover Wear and Aging — What Is Left Above the Cords, and When to Stop Running

Cover wear is the one failure mode you can measure directly, which makes it the easiest to plan around and the easiest to mismanage. The mistake we see most often is judging wear in percentages rather than millimetres. A belt with an 8 mm cover that has lost half of it is in a very different position from a belt with a 4 mm cover that has lost the same proportion, even though both read as "50 percent worn".

What matters is the rubber left above the cords. Once that residual layer gets thin, the cords are no longer protected from the material, from water, and from the impact of anything sharp. Corrosion then does the rest, and corrosion inside a cord is what turns a wear problem into a belt replacement.

Measuring wear so the number means something

Take the reading at fixed chainage points, not wherever the belt happens to look worst that day. A practical layout uses the drive, the loading point, the first idler bay after the loading point, the mid-point, and any curve or vertical bend. On each of those, record three readings across the width — left edge, centre, right edge. The centre reading is often the worst because that is where the material rides, but on a badly tracked belt the edge can lose cover far faster, and that difference tells you something about tracking rather than about abrasion.

Turn the readings into a rate. If you measured 7.8 mm in January and 6.9 mm in July, you are losing roughly 0.15 mm per month in that section, and you can now forecast the date on which you hit your limit instead of being told by an incident. Abrasive granite or clinker can run at several times that rate; a dry, fine coal duty can be gentler than the numbers your supplier assumed at the design stage. That is why we ask for a wear-rate figure before recommending a cover grade, and why we keep both abrasion-resistant and heat-resistant constructions in the range — a abrasion resistant conveyor belt used on a hot clinker duty will age differently from the same belt on wet sand.

Tap testing on the monthly route. A clean, ringing note is good news; a dull thud over a 200 mm zone means a void has started underneath the cover.

Grading cover condition from A to D

Grade C is the one people argue about, because the belt is still carrying material and the damage looks cosmetic. It is not. Once a crack reaches the cord plane, moisture follows it, and the corrosion you cannot see decides the belt's remaining life. Grade B carries a warning of its own — a network of fine cracks on a hot-duty belt usually means the compound is aging faster than the abrasion rate suggests.

Heat deserves its own note. Above roughly 100 degrees Celsius at the belt surface, ordinary compounds harden, crack and lose elongation, and the cracking starts on the side facing the hot material. If your discharge temperature readings sit in that band, the aging grade you need is not the belt you would specify from abrasion data alone. We have seen a clinker line run through three belts in four years before the compound rating was matched to the measured surface temperature instead of to the material data sheet.

Grades A and B let you choose a shutdown date. Grade C lets you choose between two. Grade D takes the choice away, because a cord-exposed belt can fail at any hour and will not respect your maintenance calendar.

05Splices Are the Weakest Link: Modes, Attribution Order and Re-Check Points

Ask any belt engineer where a high-tension line will fail and the answer comes back before you finish the question. The splice. A vulcanized joint gives up somewhere between 60 and 80 percent of the parent belt's fatigue life in most real installations, and it is the only part of the belt that is made by hand. Everything else left a factory under a controlled press.

That single fact shapes how you should investigate a splice failure. Before you accept anyone's explanation — material, workmanship or operating conditions — decide which of those three a given piece of evidence can actually support. It also explains why sites that order wholesale conveyor belts in full roll lengths, and that hold their conveyor belt supplier and their splicing crew to one drawing, end up with fewer splice failures. Every joint you never had to make is a failure site you never have to manage.

Vulcanized joints and mechanical joints fail differently

A vulcanized steel cord splice fails in one of four recognizable ways. Edge lifting starts at the outer step and peels back along the belt edge, usually after the splice has been running long enough for a small void to grow. Cord pull-out happens when the interlocking pattern is wrong or the overlap is short, and it looks sudden because it is. Internal voids and blisters show up as bulges or as a dull tap note, and they trace back to moisture, dust or oil trapped during assembly, or to a press that never reached pressure across the full plate. Under-cured splices are the quiet ones: the joint looks fine for months, then loses adhesion slowly because the temperature at the cord plane did not hold long enough. Whether the joint is made on site or in a conveyor belt factory under a controlled press, these four modes are what you will find. The same arithmetic governs the transmission drives around your plant, which is why a V-belt manufacturer answers questions about cord construction and compound before dimensions.

Mechanical splices — bolt and clamp types — are a different animal and should be treated as a temporary measure on any main haulage belt. They concentrate stress at the bolt holes, they slip and re-seat under starting torque, and the bolt line damages the cords it grips. On a long, high-tension overland conveyor they buy you the hours needed to organise a proper repair, and nothing more. Running one as a permanent joint on a 2 km line is a decision that eventually reports itself. Keeping a usable stock of clamps, filler strips and the matching cord material — the sort of local support a conveyor belt distributor can hold for you — is what makes that interim fix a plan rather than a gamble.

Attribution order when a splice fails

Working in that order keeps a warranty conversation sane and stops the industry's favourite shortcut, blaming the belt before anyone has opened the press records. In our experience most splice failures end at step 3, and most of the rest at step 2.

Re-check points after any splice is made are not optional. Verify the cured splice with a tap test along its full length, measure the thickness across the splice at three points to confirm there is no step, and record the take-up position so the next month's reading has a baseline. Then watch the splice for the first 40 running hours; that window catches almost every assembly fault that matters. The first two shifts of service tell you more about the joint than the next six months do.

Keep the paperwork honest while you are there. Record the cure chart reference, the crew, and the belt surface temperature at the moment of the splice, then file the sheet with the belt history rather than in a folder on the workshop wall. When a joint fails two years later, that sheet is the difference between a five-minute answer and a three-week argument.

06Tracking, Gouging and Longitudinal Tears: Cause Chains and Evidence Order

Sudden damage is the failure class that hurts most, because it arrives without warning and the repair is measured in shifts. Three patterns cover the majority of cases we are called out to — persistent mistracking, surface gouging under the skirt and around the loading zone, and the longitudinal tear that runs for tens of metres before anyone stops the belt.

All three have short cause chains, and the chain matters more than the symptom. Mistracking usually starts at an off-centre loading point, a seized idler, or unequal take-up tension across the width, and any of those will produce edge wear that looks like a cover problem. Gouging almost always traces to a skirt rubber gap that has opened beyond its tolerance or to a chute liner lip standing proud of the belt surface. A longitudinal tear nearly always has a physical cause you can pick up off the floor, and that object is the first thing you should collect, not the last.

Evidence order at the scene

When a tear or a serious gouge stops the line, resist the instinct to start pulling material off the belt. Photograph the damage in place with a tape and a chainage marker in frame, then mark both ends of the damage on the structure before anything is moved. Find and bag the foreign object, note where it came from and how far it travelled, and check whether the belt was stopped by a detector or by an operator. Only then start the repair.

The reason for that discipline is simple. A tear that began at the loading point because a chute grizzly bar was missing its last bolt is a structural problem, and re-splicing the belt without fixing the chute guarantees a repeat. We have walked onto sites where the same belt had been repaired three times in seven months, each time at a different chainage but always within 15 m of the same transfer. The belt was never the issue; the steel cord conveyor belt had simply been reporting, in the most expensive way available, that a piece of the loading station had fallen off.

Detection speed belongs in the maintenance budget, not on a wish list. On a belt running at 4 m/s, a tear propagates about 240 m per minute, so a loop or ultrasonic tear detector that stops the drive within seconds limits the damage to a repairable length. Without one, the same event can run the full length of the conveyor and take a splice with it. The cost of the detector is a fraction of one belt replacement, and it is the single highest-return item most sites are still missing.

07Repair or Replace: A Cost Framework You Can Actually Compute

This decision is usually made on feeling, and it is one of the few decisions in belt maintenance that can be made with arithmetic. You need four numbers — the remaining life you can defend, the downtime a repair costs, the downtime a replacement costs, and the failure probability in between.

Remaining life comes from the measurement you already have. Cover wear rate from monthly readings, plus the cord scan result, plus the splice history. If cover loss runs at 0.15 mm per month and you have 3 mm above the cords with a 2 mm action limit, you have roughly six to seven months of cover left, and the question becomes whether the cords will last longer than that. When the scan shows broken strands in one section only, a section repair buys real time. When the scan shows scattered strand damage across several hundred metres, no repair changes the belt's overall condition and you are buying a postponement, not a solution.

Shutdown inspection. Same chainage points, same three positions across the width, every month — that repetition is what turns a reading into a trend.

Scenarios and the usual right answer

Two rules turn that table into a decision. If the repair cost including the lost production window exceeds roughly a third of the delivered cost of a new belt, and the belt is past mid-life, replace rather than patch. If the expected remaining life of a repair is shorter than the lead time for a new belt, order the belt now and keep patching in the meantime — the classic mistake is to spend six months of repairs before placing an order that takes ten weeks to arrive.

When the answer is replacement, get the specification right rather than copying the old drawing. Duty, tonnage, slope and drive layout may all have changed since the belt was first fitted, and our how to size and select a steel cord conveyor belt walkthrough covers the checks that stop a replacement from carrying forward an old error at new prices.

08Field Measures That Extend Service Life

Keep the idlers turning. A seized idler stops rotating and turns into a fixed abrasive point, and it also drags the belt sideways. Replacing idlers as they seize, and using the right roll diameter and spacing in the loading zone, is the cheapest belt-life measure available. Our own conveyor rollers guidance puts the wear effect of a frozen roll in the loading bay in plain terms: the belt loses cover from the bottom side as well as the top, and the bottom cover is the one nobody measures.

Set the cleaners properly. A primary scraper doing its job removes the carryback that otherwise builds up on the return idlers and feeds mistracking. A scraper pressed too hard, or a blade that has worn to the point of a sharp edge, does more damage than no scraper at all. Check blade contact daily and tension weekly.

Fix the loading point once. This is where the largest single gain sits. Centred loading, intact skirt rubber within a sensible gap, a chute that does not pour material onto one edge, impact idlers or a cushioned bed under the drop, and a reduced drop height all work together. A station in good condition can add years to a belt; a station with a 40 mm skirt gap and an off-centre stream can halve its life. The parts that make the difference are ordinary conveyor components, and they are far cheaper than the belt they protect.

Respect the tension window. Too little tension gives slip, heat and splice pull-out. Too much shortens splice fatigue life and can lift the belt off the idlers. Keep sag at a sensible 1 to 2 percent of the idler spacing, and re-set the take-up after every belt shortening rather than leaving it where it was.

Control starts and material. Loaded starts and repeated choke-ups are the most common operational cause of splice damage we see. Soft-start drives, a start sequence that clears the belt first, and a magnet or metal detector ahead of the loading point cut the load on the joint and the belt body together. On quarry and mine duties this one discipline often explains the difference between four years and seven years on the same specification.

Ask for an inspection template built for your conveyor

09Frequently Asked Questions

How often should a steel-cord belting installation be inspected?

Split it by frequency rather than doing everything rarely. Shift crews do a visual walk, weekly crews check idlers and edges, monthly work measures cover thickness at fixed chainage points, and a quarterly scan covers cord condition across the whole belt. That mix gives you a trend instead of an opinion, and it keeps the measurement work inside a normal crew's workload.

Can a line keep running with a broken cord?

Often yes, for a while, and the decision is about count and spread rather than one strand. An isolated break in one cord with no corrosion around it can be watched, but several breaks in the same cord or damage across more than one cord pitch should trigger a repair plan. Only a scan and a cover inspection together can tell you which situation you are in.

How thin is too thin on the cover?

Measure what is left above the cords, not the percentage lost. At 2 mm remaining we plan work, and at 1.5 mm or with bare cords we replace at the next window. A belt with a thicker original cover can run longer in absolute terms, which is one reason nominal cover thickness matters as much as abrasion grade.

Why do joints fail before the belt body does?

A vulcanized joint is made by hand and typically reaches somewhere between 60 and 80 percent of the parent belt's fatigue performance. Two joints also mean two places where the cords are interrupted. That is why the inspection schedule puts splice zones first, and why cure records and splice geometry are worth more than any opinion formed after a failure.

Is a mechanical joint acceptable as a permanent repair?

On a low-tension, short conveyor it can hold up for a long time. On a long overland or high-tonnage line it is a bridge to the next planned stop, because the bolt line damages the cords it grips and starts to slip under starting torque. Treat it as a scheduled temporary fix with a date attached.

What is the fastest way to decide between repairing and replacing?

Compare the repair cost plus its lost production window against the price of a new belt plus the window to fit it, and bring the remaining life estimate into the same conversation. If a repair costs more than a third of a new belt on a belt past mid-life, or if the repair outlives the delivery time of a replacement, order the replacement.