# Balancing pulleys, couplings and drive components at the point of operation > A drive rarely picks up unbalance gradually. There's usually an event: belts were changed, a pulley was repositioned, a brake disc was skimmed true, an assembly was put back together after a repair. We bring a vibration analyser and a dial indicator, because on drives, mass isn't the only thing that decides it. Runout, the fit on the shaft and the condition of the belts all produce vibration that looks a lot like unbalance. **In short:** Yes. We balance belt-drive pulleys, multi-groove and drive pulleys, half-couplings and shaft couplings, flywheels, drive and brake discs, rotating flanges, chain-drive sprockets right on the machine, without taking them off the shaft. There are three conditions: the fit is sound, the geometry and drive are in order, and the vibration is mainly the once-per-turn component 1x — the part of the vibration that matches the shaft's rotation speed, which is exactly the part unbalance creates. If the fit has worked loose, the grooves are worn or the belts have stretched, we'll say so plainly: the drive needs a repair, not weights. Source: https://axiline.pt/en/equipment/balancing-pulleys-couplings-drive-components-at/ Publisher: AXILINE · Vila Nova de Gaia, Portugal · +351 931 831 229 · axilinegeral@gmail.com ## Symptoms: what brings people to call us A pulley or half-coupling doesn't build up deposits the way an induced-draft fan wheel does. But any assembly operation changes how the parts sit relative to each other on the shaft, and with it the unbalance of the whole assembly. If two items on the list below match your case, there's something worth discussing. - [x] Humming appeared right after the belts, pulley or half-coupling were replaced - [x] Vibration is felt at the drive's bearing support and drops if the belts are removed - [x] Belts run hot, crumble at the edges, or throw off the grooves - [x] The pulley rim visibly wobbles when the shaft is turned slowly by hand - [x] The spectrum is dominated by a peak at the rotation frequency of the shaft the pulley sits on - [x] The keyway is battered, and reddish dust is coming out from under the hub - [x] The flywheel started humming after skimming, a disc replacement or a repair - [x] The brake disc runs out and heats unevenly - [x] The drive's bearings are living shorter lives than their rated one > How it developed matters. Showed up all at once — more likely unbalance or an assembly error. Has been building for months — more likely worn grooves, stretched belts, a worn-out fit. In the second case, balancing won't stand in for a repair. ## Which drive components we balance They all have one thing in common: a short, rigid, low-mass rotor on an interference fit, a key or a taper bush. Speeds are usually from 700 to 3000 rpm, well below the first critical speed (the speed at which the shaft goes into resonance). The influence coefficients — the assembly's response to a trial weight — are therefore stable, and the calculation converges in a single trial run per plane. ### Belt-drive pulleys Drive and driven pulleys on V-belt, poly-V and flat-belt drives, tensioner rollers. Unbalance comes from casting inhomogeneity, an off-centre fit, and uneven groove wear. ### Multi-groove pulleys A wide pulley with several grooves is no longer a thin disc. Axial length comes into it, and with it a couple component. Two planes are often needed. ### Half-couplings, shaft couplings, flanges Pin-and-bush, gear and jaw couplings, both halves of a coupling, rotating flanges. Causes: a missing pin, bolts of different mass around the circle, a re-pressed fit. ### Flywheels and drive discs Engine and press flywheels, drive and driving discs, brake discs on industrial brakes and test rigs. The mass is large, speeds are high, and even a small residual unbalance produces a large force. ### Chain-drive sprockets Balanced as a thin disc, in one plane. But the chain itself produces impacts at the tooth-mesh frequency, and no weight will cure that. ### Built-up rotor assemblies A shaft with a pulley, half-coupling and flywheel is balanced only as an assembly. Parts balanced separately produce a new unbalance once assembled. > A brake disc is a special case: the friction surfaces must not be drilled or have anything welded onto them. We fit the weight on the hub or in the service holes. And on bearing life: an ISO 281 life calculation doesn't account for the rotating force from unbalance unless it's been built into the input data. Sources: [ISO 281:2007](https://www.iso.org/standard/38102.html) ## Geometry and drive first, weights after Balancing assumes the mass is fixed in place and the rotation is geometrically true. An off-centre pulley fit breaks the second condition, a loose hub breaks the first. In both cases the instrument will show a neat peak at 1x, and the result won't hold. | What we check | With what | What we do about it | | --- | --- | --- | | Radial and face runout of the rim | Dial indicator, a full turn of the shaft | Runout above the maker's tolerance means an off-centre fit or a deformed pulley. Re-fit or replace | | The fit: key, interference, taper bush | Play, fretting marks, the bush screws' tightening torque | A battered keyway and reddish dust mean a repair. With a loose fit, the 1x phase (the angle showing where in the turn the vibration peaks) wanders, and the influence coefficients don't repeat | | Groove wear and profile | A profile gauge, checking the base and walls | A shiny base, rolled-over edges, uneven depth. A worn pulley gets replaced | | Belt set and tension | A tension gauge or the maker's method | Mismatched-length and stretched belts get replaced as a matched set, tension is set before balancing | | Pulley alignment and coupling concentricity | A straightedge or laser tool, indicators on the coupling | Misalignment is fixed by lining up the pulleys and by shaft alignment, not by mass on the rotor | | Fixings, supports, frame | Foot-bolt torque, a soft-foot check | Looseness produces a comb of harmonics and makes the influence coefficients unreliable | > The order is exactly that: geometry, fit, belts, shaft alignment, then balancing. The walk-round takes 15–30 minutes on a stopped, locked-out machine. How to tell unbalance from misalignment by phase at two supports is covered in a separate article. ## One correction plane or two We work out L/D, where L is the distance between the possible correction planes and D is the diameter in the zone where the weight is fitted. The rule is covered in the article on the number of planes; below is the summary for drives. | What we're balancing | Planes | Why | | --- | --- | --- | | Thin pulley, drive and brake disc, sprocket, flange | One | L/D is well under 0.5, the part behaves as a disc, static unbalance dominates | | Wide multi-groove pulley | One or two | We decide from the measurement. If the level stays high at the second support, we move to two | | Flywheel | Usually one | The mass is concentrated in a thin rim. Thick press flywheels sometimes need two | | Half-coupling at the shaft end | One, checking the second support | The plane sits out at the end of the span, so a weight there affects the two supports differently | | Built-up assembly: shaft, pulley, half-coupling, flywheel | Two | The planes are spread along the shaft length, and a couple component is almost always present | > A two-channel instrument settles it from the data: we see the 1x amplitude and phase at both supports at once. Close phases mean static unbalance, opposite phases mean a couple. For flexible rotors near the first critical speed the approach is different — check applicability against the current part and edition of ISO 21940-12. Sources: [ISO 21940-12:2016](https://www.iso.org/standard/50429.html) ## How the work goes on site 1. **Geometry before electronics** — The machine is stopped and locked out. We check the belts, grooves, fit and support fixings. We set a dial indicator on the rim and face, turn the shaft, and take the runout reading. At the same time we work out where a weight can physically be fitted. 2. **Accelerometers and marker** — Two sensors go on the bearing supports of the shaft the part sits on: on the metal next to the bearing, radially. We stick the reflective marker onto that same shaft, not the motor shaft — with a belt drive the two shafts turn at different speeds. 3. **Baseline measurement** — At running speed, in the working condition. We record the speed, the overall vibration in mm/s RMS (the root-mean-square value — the standard way of measuring the level), the 1x amplitude and phase at both supports, and the spectrum. This is where we decide whether there's a job for weights at all. 4. **Fixed positions and radius** — We number the fitted holes, spokes or rim sectors in the direction of rotation: Z1, Z2 and onward. We enter the number of positions and the actual radius. The instrument answers with a position number, not an angle, so getting the direction of the angle wrong is impossible. 5. **One run per plane** — A weighed weight goes into position Z1, and its actual mass and radius go into the software. A usable run: 1x changes by at least 20–30 percent, or the phase by 20–30 degrees. Pulleys are light, and the response is usually clear. 6. **Weights, check, report** — The angle reference is zeroed where the trial weight stood. The software gives the mass and position, or, for metal removal, the diameter and depth of the hole. A check run at the same speed shows 1x dropping several times over. If we don't hit the target straight away, we add small weights to the ones already fitted. > A cycle on an accessible pulley in one plane usually takes two to four hours including stops. A second plane adds one more run. Access eats up the time, not the calculation. ## We balance flywheels and built-up assemblies as a unit A part balanced on its own doesn't add up to a balanced assembly. Unbalances add up as vectors and reappear at every reassembly. Contributions come from an off-centre fit, play in the key, uneven bolt masses around the circle, and inaccuracy in the shaft's keyway. So we balance the assembly as a unit, in its own bearing housings, at running speed. The pulley, half-coupling, flywheel and drive disc stay on the shaft. ### Flywheel Balanced on its own shaft, together with the drive disc and clutch. After skimming or replacing the disc, the assembly gets rebalanced. ### Taking it apart undoes the result Reassemble the parts in a different angular position and the unbalance comes back. Marking the parts' relative position with a punch mark or paint helps. ### A repeat visit is shorter The assembly's influence coefficients stay in the instrument's archive. Next time, the correction is calculated straight away, without trial runs. ## Belt and chain drives bring frequencies of their own This is the main trap on drives. The transmission itself generates vibration, and part of it looks like unbalance. Two rotation frequencies live in the spectrum at once: the motor shaft's and the driven shaft's. On top of them comes the belt-pass frequency and its harmonics — it sits below running speed, in the subsynchronous range. - Put the marker on the shaft whose pulley you're balancing. A marker on the motor shaft will make the instrument pick out the 1x of the wrong machine - Work out the drive ratio from the pulley diameters and check it against the instrument's readings - Subsynchronous peaks and their harmonics are almost always the belt. No weight will remove them - A raised 2x together with axial vibration points to misalignment, not unbalance - Peaks at the tooth-mesh frequency (the number of sprocket teeth multiplied by the speed) mean chain and tooth wear - If it's technically possible, take the belts off and turn the shaft over without the drive > That's why we record the spectrum and the time waveform, not just a single number. It separates what a weight solves from what a belt, shaft alignment or a fit repair solves. Sources: [ISO 13373-3:2015](https://www.iso.org/standard/40840.html) ## When on-site balancing won't help Sometimes the honest answer is: you don't need weights. ### A worn-loose fit The pulley or half-coupling moves on the shaft, the taper bush turns. The 1x phase wanders from run to run, and the influence coefficients don't repeat. The fit gets repaired first. ### Worn grooves and stretched belts The transmission is producing the vibration, not the mass. We replace the pulley and the belt set, set the tension, and there's usually nothing left to balance. ### A crack A crack in the spoke or rim of a cast-iron pulley, in a flywheel, in a flange. Balancing would remove the one outward sign of it. Repair by the maker's procedure, or replacement. ### Nowhere to fit a weight A brake disc with friction surfaces, a closed pulley with no holes, a half-coupling inside a tight guard. Sometimes recalculating for a different plane or drilling the hub saves the day; sometimes a rig is the honest answer. ### Nothing to balance Overall vibration is 9 mm/s, and the once-per-turn component is 2 mm/s. A perfect balance job would take off two, leaving almost nine. Diagnostics come next. ### Frame or guard resonance A narrow peak against speed, the phase flipping by around 180 degrees on coastdown. In resonance the result is unstable. Stiffness and fixings come first. > A separate case: you need a documented residual unbalance in g·mm/kg against a G balance quality grade (from ISO 21940). Acceptance like that is done on a balancing rig. On site we can work out the tolerance for a G grade and show you the residual vibration. ## What you get, what it costs, and how to book The job finishes with numbers, not the words 'it's better now'. On the money side. vibration diagnostics with a report costs 300 EUR per unit, balancing adds from 250 EUR, and the minimum invoice per visit is 500 EUR. The calculator works out the exact amount: it depends on the number of rotors, the number of correction planes, access, and how far the site is. We're based in Vila Nova de Gaia, near Porto, and we travel all over Portugal. The engineers who design and manufacture the Balanset instruments are the ones who come out and balance with them. The Balanset-1A can also be bought outright: two accelerometers, an optical tachometer, a two-channel USB module and laptop software, balancing in one and two planes, fixed positions, a drilling calculation, trim balancing (touching up with small weights), an archive and reports. - Before-and-after readings at each support: speed, overall vibration in mm/s RMS, 1x amplitude and phase - Spectrum and time waveform on the baseline and check runs - The masses, position numbers and radius of the weights fitted - Check results: runout, groove condition, fit, belt tension - An assessment of residual vibration against the zones of the applicable part and edition of ISO 20816, and a tolerance calculation against a G grade - A conclusion: what weights fixed, what needs a repair, what to check on the next round - The assembly's influence coefficients, saved for the next balancing job > What to have ready before we arrive: the ability to stop the machine several times, a drive guard that opens, access to the fitted holes on the pulley and flywheel, permission to weld if it turns out to be needed, and someone responsible for starting the machine and lockout. A detailed checklist is in a separate article. Sources: [ISO 20816-1:2016](https://www.iso.org/standard/63180.html) · [ISO 21940-11:2016](https://www.iso.org/standard/54074.html) · [Balanset-1A manufacturer specification](https://vibromera.eu/product/balanset-1/) · [Balanset-1A operation manual](https://vibromera.eu/balanset-1a-operation-manual/) ## Frequently asked questions **Can a pulley be balanced without taking it off the shaft?** Yes, that's the main method. The pulley is balanced in its own bearing housings at running speed: sensors on the supports, the optical tachometer reading a marker on that same shaft, weight fitted into the rim's fitted holes. Taking the pulley off is only needed if the fit needs repair or you need acceptance on a balancing rig. **Do the belts need to come off before balancing?** No, we balance in the working condition, with the tension set as normal. Removing the belts is something we use as a diagnostic check: if the vibration drops several times over without them, the problem isn't pulley unbalance but the transmission. **The pulley runs out when turned by hand. Is that unbalance?** No, that's geometry: the pulley's axis doesn't line up with the shaft's rotation axis. An off-centre fit, a deformed rim, a worn-out hub. The vibration sits exactly at 1x and looks like unbalance, but weights won't fix it. **Is a flywheel balanced separately or assembled on the shaft?** As an assembly. The unbalances of the shaft, flywheel, drive disc and fixings add up as vectors, and an assembly built from separately balanced parts almost always ends up with a new unbalance. **Where do you fit the weight if the pulley has no holes?** In order of preference: the fitted balancing holes, a bolt with a locking nut through a hole in a spoke, a welded plate where the maker allows welding, or removing metal by drilling into a heavy hub. Magnets are only good as trial weights. **How many correction planes does a pulley need?** A thin pulley, a sprocket, a drive or brake disc and a rotating flange usually only need one. A wide multi-groove pulley and a built-up assembly of shaft, pulley and half-coupling often need two.