Ribbed Belt Manufacturer: How to Vet PK and PJ Multi-V Belts

01Why Multi-V Drives Punish Loose Specification

A towel-converting line we surveyed in 2024 had replaced the belts on its spindle drive three times inside eleven months, and the maintenance log blamed poor belt quality every time. The real cause sat in the deliveries, not the line. Across three shipments from the same source, the quiet drift was one rib short on two belts and about 0.4 mm low on rib height on most of the rest, which let the belt seat deeper into the pulley groove than the drive had been laid out for. Slip followed, slip became heat, and heat cooked the compound from the inside until the ribs glazed and cracked. Nobody noticed, because nothing on the incoming inspection sheet measured rib geometry at all.

That is the ordinary shape of a ribbed belt failure. It shows up in the field and the evidence is sitting in the paperwork.

PK and PJ multi-V belts carry load through a cord layer that sits above a row of wedge-shaped ribs, and the assembly is held together by a backing compound and a facing fabric. Each layer covers a job the others cannot. Get the cord wrong and the belt elongates further than the take-up slot can recover, so tension walks off the pulley within a few hundred hours. Get the backing wrong and the flat between ribs starts cracking long before the rated hour count. Get the facing wrong and a pulley that is only marginally out of alignment announces itself as a squeal at every start-stop cycle. None of these faults is visible to a tape measure. Most of them are visible on a specification sheet, if the sheet is complete enough to be read. This is also why an experienced conveyor belt manufacturer treats drive belts and conveying belts as two different engineering problems rather than two catalogues of rubber.

A reseller cannot answer the questions that matter. A maker can.

The fastest way we know to separate the two is to ask how many cords run inside the belt and what build tension holds them. A real transmission belt manufacturer replies with a number, a tolerance, and the cord supplier's grade. A trading office forwards the question, waits three days, and returns the word standard.

Send your PK or PJ drive data and get a build recommendation with rib, cord and backing detail

Standard is not a specification. It is a promise that somebody else already made the decision, usually without seeing your drive.

What the part number leaves out

A multi-V part number encodes the section, the rib count and the effective length, and that is all it encodes. Two belts with the same part number can differ in cord material and cord count, in whether the backing is plain rubber or a fabric-faced laminate, in rib flank angle while under load, and in the antistatic behaviour of the compound. On a low-power application running in a clean, temperature-controlled hall, those differences often stay invisible for years. Put the same two belts on a 45 kW drive on a quarry screen deck, where ambient swings from 5 °C at night to 48 °C under the hood by midday and grit finds every pulley groove, and the gap between them opens within a single season. We have pulled belts off aggregate screens at 900 hours that had lost 30 percent of their rib height, and pulled the equivalent section off a comparable screen at 5,200 hours with rib wear still inside tolerance, with the only structural difference being the cord construction and the compound used in the backing. Choosing a V-belt manufacturer who will disclose those two variables is most of the risk reduction, done in one conversation.

Then there is the question of who actually runs the press. Plenty of suppliers show a workshop photograph; far fewer can show the mould register that proves they made the product being quoted. Requests to visit a conveyor belt factory are not a formality in this category, because the difference between a maker and a packager is exactly the difference between the two belts described above.

Belt failure rarely arrives as a single dramatic event. It arrives as a slow argument about who was responsible. Good specification work moves that argument forward into the quoting stage, where it is cheap, and out of the plant, where it costs a shift. Buyers who also purchase timing belts alongside their ribbed belts tend to develop this discipline quickly, since toothed drives punish sloppy tension data even harder. If you are new to the wider category, the house guide to types, grades and how to choose sets out the vocabulary that the rest of this article assumes.

02Reading a PK and PJ Specification Without Guessing

The first thing to settle is the section. PK and PJ are not interchangeable, and a supplier who quotes both from the same stock sheet without asking about your pulley diameters is telling you something useful about how they work. ISO 9982 fixes the geometry of the PH, PJ, PK, PL and PM profiles, and the numbers that matter are pitch, rib height and the minimum pulley diameter the rib can wrap without buckling. When rib height shrinks but pitch stays correct, the belt still looks like a PK belt on the bench and behaves like a loose one on the machine, because the wedging action that transmits torque depends on the rib filling the groove, not merely entering it.

Pitch is measured rib to rib, and the tolerance band around it is where cheap tooling quietly fails.

The second thing to settle is effective length and its tolerance. A belt quoted as 1,400 mm with a plus-or-minus tolerance is a different purchase from a belt quoted as 1,400 mm with no band given at all. On a two-pulley drive with a fixed centre distance, a 4 mm length error across a 200 mm span is enough to move tension out of the window the bearing was chosen for, which shows up first as a whine from the motor bearing and later as a warranty claim on the driven equipment. Ask for the length tolerance in writing, then ask whether it is measured on a free belt or on a belt under the standard test tension. The two numbers can differ by more than the band itself.

Rib count is a load decision, not a size label

Rib count multiplies capacity. It does not simply match a pulley width. A three-rib and an eight-rib belt cut from the same sleeve share the same profile and the same compound, and they are nowhere near interchangeable, because the load per rib changes and so does the heat each rib has to shed. We once watched a plant step down from an eight-rib to a five-rib belt to save cost on a fan drive that had been sized for the higher count; the belts lasted 300 hours instead of the 6,000 they had been getting. The saving on the belt was real and the cost of the lost fan uptime was about forty times larger. If a conveyor belt supplier offers to trim rib count without asking for drive power, pulley diameters and duty class, treat the quote as arithmetic rather than engineering. The mechanical basis behind these trade-offs is handled in more depth in the note on roller types and selection criteria for the conveying side of the same plants.

One more measurement belongs in the specification conversation, and it is usually forgotten until a belt has already failed twice. That is pulley groove wear.

A worn groove swallows a new belt as surely as an undersized rib does, so a specification agreed without groove measurements is only half a specification. On replacements we measure groove depth and flank condition at the same time as we check centre distance, because a drive that ate one belt will eat the next one just as happily. Where the drive shares a frame with conveying equipment, it pays to keep the pulley data with the conveyor pulleys and roller records for the same line, so that a change in belt source never gets approved without a matching look at the hardware it runs on. A plant that has been ordering the same industrial conveyor belt specification for a decade often discovers during a belt review that nothing has been checked on the drive side since installation.

03Backing, Fabric and the Cord That Carries the Load

The cord is the load path and it is the least visible part of the belt. In a PK profile the cords run parallel to the belt axis under the backing, and their construction decides how the belt behaves over its whole life rather than merely at break. Polyester cord is common in general industrial PK belts because it holds length reasonably well and tolerates humid conditions; aramid appears where the drive demands high tensile capacity in a thin section, at a higher price and with sharper sensitivity to pulley diameter. What matters at the vetting stage is that the supplier can name the cord material, the cord grade and the number of cords, and can show that the build tension is controlled rather than estimated by feel. We have opened belts from two suppliers on the same day and found 18 cords in one and 14 in the other, both sold as the same size and both sold at the same price. The fourteen-cord belt had a measured permanent elongation about twice the eighteen-cord belt after 500 hours on a comparable load, and it was the one that ended up on the scrap pile with the take-up fully extended.

Backing and facing do different jobs

The backing compound protects the cord from abrasion and from the oil mist that drifts through any real plant. A chloroprene-based backing handles oily environments and moderate heat better than a plain general-purpose compound, while an EPDM-based backing suits hot, dry, ozone-rich conditions and outdoor installations, and loses ground where oil is present. This is the kind of choice that a maker will raise before you do, and that a stocksheet supplier will not raise at all. Facing fabric matters for a different reason. A belt with a low-friction facing runs quietly where a bare compound would squeal against a slightly misaligned pulley, and it holds up under the cyclic reversal of a reversing drive. On a reversing packaging drive, we have seen belt life go from about 900 hours with a plain facing to well past 4,000 hours with a purpose-chosen friction-facing fabric, on the same pulleys, at the same tension, with no other change. That single substitution is usually cheaper than one unplanned changeover.

Compound selection follows the environment, and the same logic runs through conveying products. A plant that fights oil contamination on its drives usually fights it on its belts too, which is why a belt review often turns into a review of the oil resistant conveyor belt specification for the adjacent line, and why sites near kilns and ovens end up comparing it with a heat resistant conveyor belt option at the same meeting. Abrasive loading on a dry product stream pulls the discussion toward an abrasion resistant conveyor belt grade. None of that is a detour. It is the same question about environment and duty asked twice, once of a drive belt and once of a conveying belt, and plants that answer it consistently are the ones that stop writing failure reports.

Where the specification stops being negotiable is at electrical and thermal limits. Many industrial sites require an antistatic belt, and the relevant property is defined rather than implied. ISO 1813 sets out the electrical conductivity requirement for antistatic belts and the test that demonstrates it, and a claim of antistatic with no test report is a marketing statement rather than a specification. Ask for the test method and the measured value. Then ask whether that measurement came from a belt in the batch shipped to you or from a belt made eighteen months earlier on a different compound.

04Moulds, Tooling and the Marks That Date a Lot

Ask a supplier how many moulds they hold for PK profiles and you will learn more in ten seconds than you would learn from an hour of catalogue reading. Moulds are expensive, they are specific to a rib count and a length, and they wear. A maker with a real tooling register will tell you the mould number, the press it runs on, the last time the cavity was re-cut, and how many belts have come off it since. A packager will tell you they have full capacity.

Tooling condition drives the geometry of every belt that comes out of it, which is why mould history belongs in your supplier file.

The practical reason is wear. A PK mould that has produced tens of thousands of belts has flanks that are no longer the flanks it started with, and the belts that come off it are marginally narrow at the root and marginally wide at the crown. Those belts pass a fresh-cavity template check on the first millimetre and fail a groove fit test on the line. When you audit tooling, do not ask only whether moulds are maintained. Ask what triggers a replacement, how many belts that trigger allows, and what happens to stock produced between the last inspection and the replacement. A supplier who answers that question with a number is managing tooling. A supplier who answers it with a policy statement is managing the conversation.

Markings that let you trace a belt back

A properly marked belt carries the section, the effective length, the rib count, a brand or mould identifier, and a production code that the maker can decode to a date and a press. That last item is the one buyers rarely check and the one that matters most when something goes wrong at month nine. Without a production code, a warranty discussion has nothing to stand on, because the supplier can always argue the belt is older than the claim. With one, a plant can compare the failed belt's code against the shipping documents and settle the question of whether the failure came from a stock item that had already been sitting for a year.

The same traceability discipline applies to high-load conveying products, where a lot code on a rubber conveyor belt is what allows a splice failure to be tied back to a cord shipment. Suppliers who build that system for one product line usually build it for all of them, and suppliers who do not will not build it for your drive belts either.

Buying in volume changes the arithmetic on all of this. When a plant consolidates its drive belts onto a wholesale conveyor belts and V-belt programme covering several sites, the supplier has a reason to invest in tooling for that account and to keep the moulds in condition, because the volume justifies it. The same supplier treats a one-off order of twenty belts as a sleeve cut and forgets it. If you want tooling attention, structure the purchase so that attention is worth someone's time, and put the tooling clauses in the framework agreement rather than in a single purchase order. Plants that run this properly ask for the register once a year and compare it with what they were told the year before. Everything is in the supporting quality documentation, which is why a review of quality assurance procedures should always include the mould and length-control pages, not only the certificate on the cover. It also helps to know who the plant actually is, and the background pages on the manufacturing operation are a reasonable place to start before you ask for a register you cannot interpret.

05Batch Consistency: What to Measure and How Often

Consistency is the single property that separates a supplier you keep from a supplier you tolerate, and it is measurable on the receiving dock with equipment most plants already own. The measurement plan does not need to be elaborate. It needs to be applied to every delivery with the same discipline, and it needs to have a trigger attached so that a result outside the band stops the release instead of being noted and filed.

Dimension checks come first, because they catch the most common failure early and cheaply.

Rib pitch and rib height go onto the incoming sheet with a caliper or a profile gauge, measured at three points along the belt and recorded per delivery. Effective length follows, measured under the tension stated in the specification rather than under hand tension, because a hand-tensioned length figure is not comparable between one inspector and the next. Hardness testing on the backing gives a fast read on whether the compound in this delivery matches the one that performed well last quarter, and it takes about a minute per sample. Cord tension uniformity is harder to test without destructive work, which is why it belongs in the periodic audit rather than the dock check; a simple destructive pull on one belt per quarter, with the break location and elongation recorded, will show whether the load path has stayed consistent.

The pattern that matters is direction, not any single reading. One belt sitting at the edge of tolerance is noise. Four consecutive deliveries each a little further out than the last is a tooling or compound problem in progress, and it is cheaper to raise it while the supplier still has belts on the shelf than after a shutdown. We have watched a plant accept a slow outward drift in rib height for nine months because each delivery passed on its own, then lose two shifts in one week when an entire lot finally fell outside the fit that the pulleys would accept. The data to prevent it had been collected from the beginning. Nobody had been asked to look at it as a trend.

Sampling location also matters more than most buyers expect. When a plant buys through a conveyor belt distributor rather than directly, the belts may have passed through a warehouse that has held them for a year in a rack beside a dock door, and storage conditions show up in hardness readings. Sampling the top belt in a carton tells you about the top belt in a carton. Take samples from different depths in the packaging, note the production code on each, and you will find out whether the shipment is one lot or three spliced together. Sites running mixed duty, where the same drive belts sit next to mineral handling equipment, tend to see the effect most clearly on their mining and quarrying lines, where airborne dust settles on stored stock and changes the friction behaviour of the facing fabric before the belt is ever installed. The same holds for plants with a heavy thermal load nearby, such as a cement plant, and for any site where a food packaging area shares a store with the general engineering spares. Storage is part of the specification, even when nobody writes it down.

06Twelve Checks Before You Release a PO

Everything above describes what to specify. This section describes what to verify, in the order that costs least when it goes wrong. The checks are written for a ribbed belt purchase, but they map cleanly onto any drive belt category, and most of them can be completed in two conversations and one plant visit.

Start with legal identity and scope. Confirm that the entity quoting is the entity that will appear on the invoice, the packing list and the quality certificate. Mismatched names across those three documents is the most reliable early warning of a trading layer, and it is also the document set you will need if a claim ever goes to a third party. Ask for the business licence scope in English or with a translation you can read, and check that belt manufacturing appears in it rather than general trading.

Then move to the tooling questions from the previous section, and follow them with the cord questions. Both lines of enquiry rely on the same underlying fact, which is whether the person answering has stood next to the press.

Scoring the answers instead of trusting the impression

We recommend writing the checks down and scoring them, because supplier visits produce a strong impression that fades within a week. A scored sheet survives staff changes and gives a new buyer something to work from. The weighting below reflects what has predicted service life most reliably in our own supplier reviews, with tooling and cord construction carrying the most weight because they are the hardest things to fix after the purchase.

Two more checks sit outside the factory and are frequently skipped. The first is the response to a technical question you already know the answer to. Send a drive condition that you understand well and see whether the answer engages with speed, pulley diameter and duty class or simply offers a section and a price. The second is the handling of a small, deliberate inconsistency, such as a length that sits just outside the quoted band. A maker will tell you their band and offer to remeasure; a packager will offer a discount, because a discount is the only tool they have. Buying on price alone usually produces a lower unit cost and a higher total cost, and the difference rarely appears in the same budget year. Plants that have learned this lesson often rebuild their whole approach to drive and conveying purchases at once, which is why a belt review frequently ends with a wider engineering and after-sales service agreement rather than a single line item.

If you want a single question to carry the most information, use this one. What changes in the belt if we double the drive hours per year?

A maker answers with something about heat build-up, rib selection or cord construction. A trader answers that the price will improve.

For sites that handle a variety of drives across distribution centres, the same vetting logic extends into conveying equipment selection as well, and the notes on logistics and warehousing applications are a useful companion when one supplier is being considered for both categories. Where the drive belts and the conveying belts come from the same plant, insist that the quality system covers both product families, not only the one with the larger order value, because audit fatigue is a real failure mode and it usually lands on the smaller line.

07The RFQ Pack That Exposes a Trader

The content of your request for quotation determines the quality of the answers you get, and a thin RFQ guarantees a thin reply. Buyers who send a part number and a quantity cannot reasonably complain when the response contains nothing but a price, and that is the situation most price-driven comparisons start from. Send a drive data sheet instead.

The data sheet does not have to be elaborate. It needs the driven equipment type, motor rating and speed, both pulley effective diameters, the centre distance and its adjustment range, the duty class with expected hours per year, ambient temperature range, the presence of oil mist or dust, whether the drive reverses, and whether the area is classified for electrical reasons. Add the failure history of the previous belt if there is one, including hours achieved and where it failed. That last item is the most valuable line on the page, because it turns the conversation from a transaction into a diagnosis.

A short RFQ produces a short answer, and a short answer is usually a price without a build. Send the pack and watch what comes back. A maker will return questions of their own, often about pulley groove condition and the previous failure, and those questions are worth more than the quotation itself.

The response also tells you how the supplier handles the rest of your plant. A team that engages with drive data will engage with a conveying problem in the same way, and a team that returns a bare price on a well-documented RFQ will return a bare price on a full product range enquiry too. Consistency of behaviour is the most portable piece of evidence you can gather, and it costs nothing to observe. If your plant also runs toothed drives, the comparison notes on synchronous belt drives show what a well-formed technical reply looks like in a neighbouring product family, and the section geometry and selection pages on classical and narrow V-belts give the same reference point for wrapped belts. Keeping idler and roller hardware records alongside the drive data, including the sealed idler and labyrinth design details for dusty lines, means a single review meeting can cover the belt, the pulley and the supporting hardware without anyone having to go and find a file.

08Life-Cycle Maths and the Failure Signals That Matter

Unit price is the least informative number in a belt purchase, and it stays that way until someone puts hours into the calculation. A belt that costs one figure and lasts 5,000 hours is a different product from a belt that costs considerably less and lasts 1,200 hours, and the difference is easy to express once you have a maintenance record. Work out the cost per 1,000 running hours and include the labour for the retensioning visit that the shorter-life belt requires, because the visit usually costs more than the belt.

Then add the failure mode. Early rib wear, cord elongation and cracking in the flat between ribs all end the same way, in an unplanned stop, but they have different causes and different fixes. Reading which one you are seeing is the difference between buying the same belt again and actually solving the problem.

The arithmetic that convinces finance is simple when the records exist. Take the achieved hours from the last two belt changes, divide the installed cost including labour by those hours, and compare the figure with the quoted alternative. On the failed fan drive we mentioned earlier, the five-rib substitution looked like a saving of a third on unit price and worked out at roughly twelve times the cost per thousand hours. Nobody needed a theory to accept that number.

It is worth remembering that drive belts and conveying belts fail in different ways but for the same underlying reasons, which are environment, duty and how well the specification matched both. A plant reviewing a drive belt problem will often find a parallel question on its conveying side, whether on a bulk material handling terminal, a recycling sortation line, or a line built around steel conveyor rollers and troughing idler frames. The record-keeping habits from one side carry over well to the other, and the supplier-verification work described above transfers with them. Sites that select roller hardware on the basis of documented batch testing, as set out in the notes on mining roller supplier specifications, tend to apply the same standard to the belts that drive their equipment, and the two practices reinforce each other.

One final point about replacement stock. Keep at least one failed belt from every failure, tagged with the drive, the date and the hours. Six months later, that belt is the only reliable evidence you have, and a photograph taken at the time of removal is almost as good. The cost of keeping a scrap belt is a shelf, and the cost of not keeping one is an argument you cannot win.

Ask for a PK or PJ build recommendation and the batch records behind it

09Frequently Asked Questions

How can we tell whether our supplier makes these belts or buys them in?

Ask for the mould register and the cord details in the same message, then watch which one gets answered. A manufacturer will give mould numbers, press assignments and a re-cut history, and will usually follow up with a question about your pulley diameters. A reseller tends to answer one item and deflect the other, or reply with a general statement about capacity and quality control. The identity check is the second filter. The name on the quotation, the invoice and the quality certificate should match, and the business scope should mention belt manufacturing rather than trading. If all three documents name different entities and the explanation involves group companies, treat the quote as a purchase from a trader and price the risk accordingly.

Which measurements should we check on every incoming shipment?

Rib pitch and rib height on two belts per delivery, measured at three positions along each belt and compared against the profile nominal. Effective length on one belt, measured at the stated test tension rather than by hand. Backing hardness on one belt, recorded as a running figure so that drift is visible. Sample from different depths in the packaging, because a carton can hold more than one production lot, and record the production code on every sample you keep. The whole routine takes under twenty minutes per delivery once the rig and the gauge are set up, and it is the cheapest insurance in the entire purchase.

Is a higher ribbon count always the safer choice?

No. Extra ribs add capacity and also add width, mass and cost, and a wider belt on a small pulley wraps less easily and runs hotter at the edges. Rib count should follow drive power, pulley diameter and duty class. What the count should never do is shrink to meet a price target, because the load per rib rises and the belt fails early. On the fan drive we described, cutting from eight ribs to five produced belts that lasted about 300 hours instead of roughly 6,000. If a supplier proposes a lower count without asking for the drive data, the proposal is not engineering.

What does an antistatic claim actually require?

A defined test method and a measured value, tied to the batch you are buying. The relevant industrial reference for antistatic belt conductivity sets both the requirement and the procedure, and a supplier should be able to show a report that names the standard, states the measured figure and identifies the compound family tested. The common failure here is a genuine report for a compound that is not the one quoted on your order, which is why the report should always be reviewed beside the compound declaration. Where the area is classified for electrical reasons, treat the report as mandatory rather than as supporting evidence.

Why do the same boards keep failing early on one particular drive?

Repeated early failure on one drive usually points away from the belt and toward the hardware. If the outer ribs wear far faster than the inner ribs, the cause is pulley misalignment or a groove profile that has worn out of shape, and a new belt will simply wear the same way. If the whole belt runs hot and glazes without a clear wear pattern, the tension is wrong or the rib geometry of the delivered belts sits low in the groove. Measure the grooves, check the alignment, verify the installation tension with an instrument, and then compare the failed belt against the drawing. Three of those four checks cost nothing but time.

How should we compare a cheaper quote with the incumbent supply?

Convert both to cost per thousand running hours, including the labour for retensioning visits and any downtime the drive has historically caused. Take the achieved hours from your own maintenance record rather than the hours printed on a catalogue. Add the cost of holding spare stock for the shorter-life option, because a belt that fails more often needs more stock on the shelf. Only then compare the two unit prices. In most of the cases we have reviewed, the cheaper belt was genuinely cheaper to buy and more expensive to operate, and the difference survives being questioned by anyone, including finance, once the hours are in the calculation.