Classical V-Belts: Section, Size and Interchange Guide for Buyers

A buyer who searches for classical V-belts in the United States lands on a retail page with a price tag and a shopping cart. One B-75, in stock, ships tomorrow. That page answers the question of how to replace a single belt this week, and it answers it well. It does not answer the question a maintenance planner asks when a 500 kW crusher drive eats four belts inside a month, or when a purchasing office has to release a container against a drawing that only says "B section, six grooves".

Plant buyers work from different inputs than retail shoppers do. They hold a sheave drawing, a drive horsepower, a center distance, an ambient temperature, and often an old belt with half the marking worn off. They need to know which number on that belt is the length. They need to know whether a cogged belt may take the place of the wrapped belt that failed, and whether two suppliers calling one size by different names are shipping the same object. They also want to know what paperwork should arrive with the container. None of that fits on a shopping cart page.

This article is the execution piece of our V-belt cluster, and we say so up front. It does not re-list belt types. Two overviews already cover that ground, namely the complete guide to V-belt types, sizes and applications and our own industry guide to types, grades and how to choose. For the family map, start with the resources published by this transmission belt manufacturer or the grade notes in our V-belt manufacturing notes, then come back. Neither overview decodes a part number against an ISO length table, decides whether a narrow-section belt can drop into a classical sheave, or tells you what to measure when the container is opened. Those three jobs are what the next eight sections handle, roughly in the order a buyer meets them.

Send us the drive data and get a factory answer back — pulley datum diameter, center distance, kW, ambient temperature

01Why a Retail Listing Cannot Answer a Plant Engineer's Question

Retail V-belt catalogs are built around turnover. They stock the sizes that move, they print an outside length and a price, and they leave everything else off the page because everything else does not sell a single belt. Datum length, cord material, jacket construction, antistatic rating and matched-set tolerance are all absent. Two belts with identical printing can still differ in every one of those properties, and the difference shows up eighteen months later as a second unplanned stop.

The second problem is that a belt is never a standalone part. It is one third of a drive, and the sheave pair and the tension setting carry equal weight. A belt selected from a horsepower table alone, with no datum diameter and no arc of contact, is a guess dressed up as a specification. Retail pages cannot ask you for that data. A manufacturer that also builds conveyor belting and pulleys tends to insist on it, because we see the warranty claims that come back when nobody did.

Failure patterns are usually traceable. Sidewall glazing and a polished, blistered underside point at slip from under-tensioning, not at a bad compound. Wear concentrated on one flank, with the others clean, points at sheave misalignment. Belts that have gone shiny and hard at the base while the top edge looks untouched often sat too deep in the groove, which is a section or datum-width problem rather than a rubber problem. Premature breakage in the cord, with a clean cross-section and no visible abrasion, points at a length mismatch that forced one belt in a set to carry the whole load.

We watched that last failure mode play out on a cement grinding mill in 2024. Six wrapped B-section belts on a 315 kW fan drive, all delivered in one box, all stamped B-100. Eleven weeks and they were finished. The compound was fine. What we found with a tape and a two-pulley rig was that two of the six measured 1.2 percent longer than the other four, so the tight belts carried the tension and the loose ones flapped. A matched set declared within the RMA length tolerance would not have done that. A retail order form with no matched-set field cannot prevent it.

There is a useful division of labour here. A conveyor belt manufacturer with its own mixing lines can control compound and cord, which is where the durability question lives. What the plant side must control is the drive data and the acceptance test, because those are the two things no supplier can guess on your behalf. If you want a genuine second opinion on an existing drive rather than a price, our engineering service desk will re-read the sheave drawing with you before anything is quoted.

One habit separates buyers who get 12,000 hours from buyers who get 3,000. They write the four numbers down before they order. Datum diameter of both sheaves. Center distance as built, not as drawn. Motor nameplate kW and service factor. Ambient and any radiant heat near the drive. With those four, section, length and construction follow almost mechanically. Without them, you are choosing between a $17 belt and a $34 belt on price alone, which is how a cheap order becomes an expensive year. Conveyor pulleys in a plant rarely stay exactly where the drawing puts them either, which is why we also ask for the as-built measurement of conveyor pulleys on retrofit jobs.

02What the Section Letter A, B, C, D or E Actually Controls

The letter in front of the number is a cross-section family, and it fixes four dimensions at once. Top width decides how much of the belt sits inside the groove. Pitch width sets the effective running diameter, and therefore the speed ratio. Height controls how stiff the belt is and how small a sheave it can bend around. The included angle, almost always 40 degrees for a new classical belt, is the wedge that converts tension into grip.

Buyers often treat the letter as a size label and nothing more. It is closer to a load class. Move up one letter and the usable power per belt roughly doubles at the same speed, because the cross-section area and the cord cross-section both grow. Move down and you gain flexibility and a lower minimum sheave diameter, which matters on short-center drives that were built tight.

Nominal dimensions for the ISO and DIN classical families, in millimetres, are worth keeping in the RFQ file. They are close enough across the three systems that a metric drawing and an inch catalog can be compared directly.

Height is the dimension buyers forget, and it is the one that decides whether a belt can be retrofitted. A 17 mm wide, 11 mm tall B belt and a 15.7 mm wide, 13 mm tall 5V belt look similar on a bench and behave very differently in a groove. The B belt sits on its flanks with a little room at the bottom. The 5V belt is taller, so in a groove cut for B it contacts higher up the sidewall and then bottoms out as it wears, which kills grip long before the jacket is gone. Section substitution is covered in full in section 05.

Standards differ by market and the marking must say which one applies. ISO 4184 and DIN 2215 govern the classical metric families on European drawings. The ARPM IP-20 length tables, inherited from the older RMA system, govern the inch-numbered belts sold across North America. AS 1332 covers Australian practice and GB/T 1171 covers Chinese-made classical belts, with GB/T 12730 taking care of the narrow sections. When a drawing cites one system and the warehouse stocks another, the conversion is not a straight rename, and a supplier who treats it as one should worry you.

Anyone comparing offers needs a single point of reference for what exists in the catalogue before comparing prices, and a full conveyor belt supplier range page is a reasonable place to see how the V-belt families sit alongside flat and industrial conveyor belt products. If the drive in question is actually a synchronous one and the section letter was a misreading, the right family is over in timing belts, and no amount of cross-section comparison will rescue a classical belt in that position.

03Reading the Part Number: Outside Length, Datum Length and What Ships

B-75 is not a mystery once you know which system printed it. In the inch-numbered North American system, the number after the dash is the outside length in inches, measured around the outside of the belt when it lies flat as a loop. So a B-75 measures 75 inches on the outside. Note what that does not tell you. It does not tell you the datum length, which is the length along the pitch line where the cords run, and the pitch line is what actually sets the speed ratio.

Datum length is always shorter than outside length, and how much shorter depends on the section and on whether the belt is wrapped or raw-edge. A wrapped B belt typically loses somewhere in the region of 1.5 to 2.5 inches between the outside measurement and the datum measurement. A wrapped A belt loses less. A raw-edge belt, moulded thinner at the base, loses less again. The exact offset belongs to the manufacturer's length table, not to a rule of thumb, which is why we publish ours and why we ask you to quote the table you are working against.

Metric drawings flip the convention. Under ISO 4184 and DIN 2215, the number after the section is the datum length, already, in millimetres. An SPB 2000 has a 2000 mm datum length, and its outside length is longer than 2000 mm, not equal to it. This is the single most common source of a delivered belt that cannot be fitted, and it is entirely avoidable if the RFQ says which length system the number belongs to.

Length tolerance is the next thing to pin down, because it is the difference between a set that shares load and a set that does not. The inch length tables we work to allow roughly plus or minus 0.4 inch on outside length for A and B sizes and roughly plus or minus 0.6 inch for C, D and E, measured on a two-pulley rig under a defined tension. That is a manufacturing tolerance, not a matching tolerance. For a drive carrying six belts you want the spread inside the set to sit well under that band, ideally inside 0.1 percent, and you want it stated on the packing list.

Marking on a compliant belt tells you most of what the shelf label omits. Section and length, the standard referenced, the manufacturer or the manufacturer's code, and a date or batch code. Antistatic belts carry an ISO 1813 reference or an equivalent mark. Oil and heat resistant constructions carry their own suffix in many catalogs. Photograph the old belt before it goes in the bin, because half of the disputes we settle start with a marking nobody recorded. Where the drive sits inside a larger bulk handling system, it also pays to know the belt you are replacing alongside the flat belting, and our notes on rubber conveyor belt construction explain why compound choice changes between a friction drive and a shaft-mounted one.

One more reading trap. Belt numbers are sometimes quoted without the section because the catalog assumes it. A-68 and B-68 share a number and nothing else. If a purchase order travels between sites without the letter, someone will ship the cheaper one, and the drive will let you know within a shift.

04Wrapped, Raw-Edge and Cogged: Three Ways to Build One Size

The same A-68 can be built three different ways, and the price difference between them is small compared with the difference in behaviour. Buyers who treat construction as a cosmetic detail end up paying twice, once for the wrong belt and once for the downtime while it is swapped out.

A wrapped belt has a fabric jacket folded around the whole section, top, sides and base. The jacket protects the rubber underneath from abrasion and from oil mist, and it stiffens the belt, which raises the minimum sheave diameter it will tolerate. Wrapped belts are the default in dusty, wet and moderately oily plants for the simple reason that they survive there. They also run a little hotter at the base because the jacket insulates the compression section, and on a small sheave that heat shortens cord life.

A raw-edge belt has no fabric on the flanks or the base. The compression rubber is exposed, sometimes with a moulded cog pattern, and the cord sits closer to the surface. Removing the jacket lowers bending stiffness noticeably, so a raw-edge belt of the same section will run on a smaller sheave without cooking itself. It grips harder at the start of a cold morning too, which some plants like and some over-tensioned drives punish. The trade-off is exposure. Raw edges do not shrug off sharp aggregate dust or dripping gear oil the way a jacket does.

A cogged belt is a raw-edge belt with moulded teeth along the underside. The cogs are not there to mesh with anything, unlike a synchronous belt. They exist to give the compression section somewhere to fold as the belt wraps a sheave, and that single change can allow a sheave diameter roughly a third smaller than the same section would accept in wrapped form. Airflow through the teeth also drops running temperature, which on a short-center drive is worth more than the nominal power rating suggests.

Compound choice travels with construction, and it is where a factory beats a reseller's catalog. A standard classical belt uses an SBR or CR compound with a polyester cord, good to about 80 degrees Celsius of belt temperature. Step up to EPDM and the same raw-edge geometry handles 100 to 120 degrees on a fan or a kiln-adjacent drive. Chloroprene gives the oil resistance. None of those upgrades are visible in a catalog listing that only prints a size and a price, so if the drive sits in a foundry or next to a hydraulic power unit, the compound has to be specified in writing. We keep heat resistant and oil resistant compounding notes on the flat-belt side of the catalogue, and the chemistry overlaps more than most buyers expect.

Two practical rules come out of all this. Do not fit a cogged belt to a drive that has never been aligned, because the extra grip simply finds the next weakest link. And do not fit a wrapped belt to a sheave that was sized for a cogged belt, because the stiffness will overstress the cord at the point of wrap. I have seen a single B-cogged belt outlast three wrapped belts on the same fan hub, and I have seen the reverse on a wet screen drive where the cogs packed with slurry and the belt polished itself smooth in five weeks. Construction follows the environment, not the price column.

Where a plant buys across several sites, consolidating construction and section into one schedule usually beats chasing unit price site by site, and a wholesale conveyor belts programme with fixed sections and one approved construction is easier to audit. Whoever you buy from, ask whether they extrude and cure the belt themselves. A V-belt manufacturer can tell you the cord supplier and the cord treatment, and can hold a compound specification from order to order. A trading company can only repeat what its upstream factory told it.

05Interchange Between Classical Sections and Narrow Sections

Half the interchange questions we receive are really one question wearing a costume. Can a 5V belt go into a B sheave? Usually not, and the reason is geometric rather than commercial.

Classical sections and narrow sections are different groove systems, not alternative names for one thing. A narrow belt is taller for its width, so it contacts the sheave higher on the flank and wedges at a different height. The groove angles overlap, but the datum widths and the top-width-to-height ratios do not. Dropping a 5V belt into a B groove will appear to fit, will spin the machine, and will then bottom out as the flanks wear because the extra height leaves no room to sink. Dropping a B belt into a 5V groove rides it too high, reduces contact area, and invites slip from the first hour.

The mapping between metric narrow sections and inch narrow sections is close enough for planning but not for automatic substitution. SPZ corresponds to 3V, SPB to 5V, SPC to 8V. SPA has no inch twin and lives in the middle on its own. Even for the three matched pairs the datum widths differ by fractions of a millimetre, and on a high-speed drive that fraction decides how the belt seats.

There is one interchange that does work, and it is worth knowing because it saves money. A drive originally built around several classical belts can often be re-engineered onto narrow belts on new sheaves, and the belt count drops because a narrow belt carries more power per unit of width. Fewer belts mean less re-tensioning, less matching trouble and a narrower face. That is an engineering change, not a substitution. It needs new sheaves and a fresh power calculation, and it should never be done groove by groove.

Before any of it, measure. Groove width at the datum line with a dedicated gauge, groove angle with a template, and the datum diameter from the sheave face. If the sheave is worn, a groove profile that looks correct to the eye can be two degrees out and half a millimetre wide, and no belt on the market will fix that. Where a worn pulley is the real problem, replacing it is cheaper than a year of belt failures, and the same argument applies to conveyor rollers on the handling side of the plant.

Buyers in regions without a local engineering house should build a relationship with a conveyor belt distributor who will do the measurement, and with a conveyor belt factory that will accept the resulting drawing as a specification rather than a suggestion. The two roles are different and both are useful. The distributor knows your drives. The factory knows the cord and the compound.

06Sheave Grooves, Datum Diameter and Matching a Set

The groove is where the belt either works or fails, and it gets less attention from buyers than any other part of the drive. A classical sheave groove is not a plain V. It has a datum line marked on the drawing, a specified groove angle at that line, and a tolerance on both. Wear moves the datum line outward and opens the angle, and the belt pays the bill.

Groove angle is not constant across a sheave range. Small datum diameters need a narrower angle so that the belt, which flattens as it bends, still contacts the flank properly. Larger diameters use a wider angle. That is why classical sheaves are commonly cut at 34, 36 or 38 degrees depending on datum diameter, and why a belt running in the wrong angle wears on one flank first.

Datum diameter is the number that matters in a power calculation, and it is not the outside diameter of the sheave. It is the diameter at the datum line of the groove, which sits inside the rim. Buyers who quote outside diameters into a selection sheet get a speed ratio that is wrong by a few percent, and on a conveyor drive that error propagates into belt speed and take-up travel. Measure with the groove gauge and read the datum line, or ask the sheave maker for the datum diameter directly.

Matching a set is the part of multi-belt practice that retail channels handle worst. Six belts on one sheave share load only if their lengths are close. The inch length tables allow a manufacturing band of roughly plus or minus 0.4 inch on A and B sizes and plus or minus 0.6 inch on C, D and E. A wrapped set that spans that whole band can leave one or two belts carrying most of the tension, and the tight belts fail first, usually within a few months. Ask for the length of every belt in the set, not just the nominal size, and ask for the moment of truth on the packing list.

We measured exactly that on a 200 kW quarry primary crusher in 2025. Eight C-section belts, supplied as one set from a distributor who had bought them from two production lots. The spread was 0.9 percent, well inside the tolerance band and still far too wide. The two tight belts had cord fatigue at eleven o'clock on the sheave. Replacing them as a properly matched set held 400 days without touching the drive, and the second set is still running. Nothing about the belt compound changed. Only the spread changed.

Tension matters as much as length, and the setting drifts. Classical drives lose tension in the first week as the cord beds in, and again in the first cold snap. An ultrasonic tension gauge, or the deflection method in the drive's manual, should be used at commissioning and then at a fixed interval. Over-tensioning is the more common mistake, and it shows up as sheave bearing heat, cord breakage or a bent shaft rather than as belt wear. Under-tensioning is cheaper to spot and easier to fix, but it glazes the flanks and quietly raises belt temperature.

When a drive resists every correction, the sheave itself is usually the culprit, and the fix belongs in the mechanical spare parts list rather than the belt order. Groove gauges, clearance and fit are covered in our conveyor components notes, and the same acceptance logic we apply to pulley surfaces on the handling side is documented under quality assurance.

07Tolerance, Marking and the Electrical and Oil Specifications

Three specifications quietly decide whether a belt is acceptable in an industrial plant, and none of them appears in a retail listing. Length tolerance we have covered. The other two are electrical resistance and resistance to oils and solvents, and both are testable.

An antistatic belt matters wherever a drive sits in a dusty atmosphere, and in any area classified as potentially explosive. Dust clouds and solvent vapour can be ignited by a static discharge from a belt that has built up charge on its flanks. The relevant requirement is ISO 1813, which sets a surface resistance limit for antistatic belts in the region of 300 megohms when tested to the standard, and a belt that meets it usually carries the standard number or an approved mark on the jacket. A belt without that mark is not antistatic, however similar it looks to one that is.

Oil resistance is a compound question rather than a geometric one. Standard SBR and natural rubber compounds swell in contact with mineral oil, and swelling raises internal stress in the cord, which then fails at the bond. The usual fix is a chloroprene or nitrile-based compound, sometimes with a nitrile-impregnated jacket. That belt will still be damaged by petrol, ketones and some hydraulic fluids, so a plant with a mixed chemistry should declare the fluid rather than assume oil resistance covers everything. We keep chemical resistant and fire resistant specifications on the flat-belt range, and the same test regime applies to V-belts in the same plant areas.

Marking is worth a paragraph on its own because it is the cheapest form of quality control a buyer has. A properly marked belt carries the section and length, the standard it was built to, a manufacturer or factory code, and a batch or date code. Antistatic belts carry the ISO 1813 reference or the approved equivalent. If a delivered belt has a section printed but no standard and no batch code, you have no way to trace a problem back to a production run, and no way to hold anyone accountable for the next order. Reject at goods-in rather than at failure.

Temperature limits should also be written down rather than assumed. A standard wrapped belt with an SBR compound is comfortable to about 80 degrees Celsius and unhappy above it. EPDM takes the same job to 100 or 120 degrees. Below freezing, ordinary compounds stiffen and a cold start puts a load spike through the cord that no steady-state calculation predicts. If the drive is outdoors in a cold climate, say so in the enquiry, because the right answer there is often a different compound rather than a different size.

08What to Verify Before a Batch Order Ships

A container order is where the small errors get expensive. You cannot put forty boxes on a bench and pick through them, and by the time the belts are in the warehouse the leverage is gone. So the verification work has to happen on the factory side, before the pallets are shrink-wrapped, and it has to be written into the order rather than requested by email after the fact.

Start with a specification sheet attached to the PO. Section with tolerances, length system and nominal length, construction, compound class, cord type, antistatic requirement if any, marking content, matched-set spread, packing and carton marking. One page is enough. An order without that page is an order for whatever the factory happens to be running that week, and you will find out which variant you got by measuring forty belts yourself.

Sampling is the next argument to settle. For a batch of a few hundred belts, a reasonable plan is dimensional checks on a handful from each production sub-lot, full length measurement on every belt that will be sold as a matched set, and destructive cord testing on one or two. Nobody wants to cut up belts, but one tensile sample per sub-lot is cheap insurance compared with a claim, and a factory that refuses the test is telling you something useful about its process control.

Ask for the packing to keep the set together. Belts from one matched set should travel in one carton with the individual lengths written on the label, because a set broken across two containers will be reassembled from whatever is nearest the rack when the belts are needed. Same reasoning for the batch code. If the box carries the date and the sub-lot, a failed belt can be traced and the rest of the batch quarantined while the drive is inspected. Without it, the only option left is to replace everything and hope.

Finally, plan the first fit. Even a perfectly built batch needs a commissioning routine. Re-check alignment, tension to the drive manual specification, and run the machine at no load before loading it. Take a photo of the installed set with the tension reading visible. Ninety days later, when the first belt fails, that photo is worth more than the warranty document, because it distinguishes a belt problem from a drive problem in about two minutes. On the bulk handling side of the plant we apply the same discipline to idlers and rollers, and the checklist we publish for factory RFQ preparation translates directly to V-belt orders.

Where several sites order the same sections, one approved specification and one approved supplier beats a dozen local purchases, and visiting the plant that will actually make the belts is the shortest path to confidence. Our own factory profile lists what a buyer can audit on a visit, from mixing through to the length rig. The same audit logic is standard practice in the heavy industries we serve, whether the plant runs a mining and quarrying duty schedule or a cement plant maintenance plan, and V-belt drives sit inside both.

Request a section and length table, matched-set tolerance and a sample report

09Frequently Asked Questions

Which number on the belt is the length?

In the inch-numbered North American system the number after the section letter is the outside length in inches, measured around the outside of the belt when it forms a loop. A B-75 measures 75 inches on the outside. The datum length along the cord line is shorter, typically by something between one and a half and two and a half inches on a wrapped B size. Metric DIN and ISO markings work the other way round, since the number is already the datum length in millimetres. Always state which system the number belongs to when you order.

Can a cogged belt replace a wrapped one on the same sheaves?

Often yes, and the gain is real on short-center drives, because the moulded teeth allow a sheave diameter roughly a third smaller than the same section in wrapped form. Two conditions apply. The sheaves should be in good condition and properly aligned, since a cogged belt grips harder and will find the next weakest point in the drive. And the environment should be reasonably clean, because the teeth collect fine dust and slurry where a fabric jacket would shed it.

Will a narrow-section belt work in a classical sheave?

No, not as a straight substitution. Narrow belts are taller for their width, so they seat at a different height in the groove, bottom out as the flanks wear, and lose grip long before the jacket is gone. The three matched metric and inch pairs, SPZ with 3V, SPB with 5V and SPC with 8V, are close enough for planning but their datum widths still differ slightly. A move from classical to narrow sections is an engineering change that needs new sheaves and a fresh power calculation.

How close should the lengths be inside one set?

Much closer than the manufacturing tolerance band. The tables allow roughly plus or minus 0.4 inch on A and B sizes and plus or minus 0.6 inch on C, D and E, which is a wide window on a drive carrying six or eight belts. For load sharing, aim for a spread inside 0.1 percent across the set and ask for the individual measured lengths on the packing list. A set assembled without that data is how belts start failing months early.

Do I need antistatic belts in a dusty plant?

If the area is classified as potentially explosive, yes, and the requirement should be written into the specification rather than assumed. ISO 1813 sets a surface resistance limit in the region of 300 megohms for antistatic belts, and a compliant belt normally carries the standard number or an approved mark on the jacket. Two belts with identical dimensions are not equivalent if only one of them meets that limit. Ask for the resistance test record rather than a statement of compliance.

What paperwork should come with a batch order?

At minimum, a dimensional report with actual readings, individual lengths for every belt in each matched set, cord type and breaking load for the sample tested, an oil swelling result if an oil-resistant compound was specified, the resistance value if antistatic construction was specified, and photos of the marking and carton labels. Add a production date to each belt and a shelf-life statement. A supplier who can produce all of that without being chased is a supplier worth keeping on the approved list.