# On-site balancing of cutterheads and cutter drums, where they operate > After a knife change, a thickness planer leaves a heavy wave pattern on the board, a jointer starts humming, and the bearing supports are hot by lunchtime. We come out to the plant, check the knife set and knife projection first, then balance the cutterhead in two planes right in its own bearing supports, at operating speed. Base in Vila Nova de Gaia near Porto, service visits across Portugal. **In short:** Yes, we balance cutterheads and cutter drums on site, without removing the shaft from the machine. Two accelerometers (vibration sensors) go on the shaft's bearing supports, and a laser phase sensor, which ties the measurement to the shaft's angle of rotation, is aimed at a mark on the end face or pulley of the shaft itself; the measurement runs at operating speed. But the order is strict: first the knife set, matched by mass, with equal projection and the gibs torqued to the standard value, and only after that, correction weights. Balancing a shaft with a mismatched set of knives is pointless: the very next knife change will wipe out the result. If the shaft is clean of resin, the bearings are in good shape, and the knife set is in order, we bring the shaft within the G-grade tolerance (the residual-imbalance limit) of the applicable part of ISO 21940 in a single stop, and hand over a report with before-and-after numbers. Source: https://axiline.pt/en/equipment/on-site-balancing-cutterheads-cutter-drums/ Publisher: AXILINE · Vila Nova de Gaia, Portugal · +351 931 831 229 · axilinegeral@gmail.com ## Where imbalance in a cutterhead comes from The shaft itself is balanced at the factory. Imbalance is brought in by the knives and everything that holds them. Every resharpening removes a strip of metal from a knife, and two knives from the same set end up weighing differently after a different number of sharpenings. A difference of two to three grams at a 50 mm radius produces 100–150 g·mm of imbalance. At 6000 rpm, that's already on the order of six kilograms of rotating force, sweeping around the bearing supports a hundred times a second. The second group of causes gathers around the slot. Gibs and clamping screws from different sets vary in mass, chips under a knife's seating surface change its projection at one point, and resin builds up unevenly on the shaft body and produces a wandering imbalance. A nick from meeting a nail chips out a section of the knife: it loses mass and stops cutting. - Knives from the same set with a different number of resharpenings and different remaining mass - Gibs and screws mixed up between slots or taken from a different set - Uneven knife projection along the shaft's length and between knives - Resin and caked dust on the shaft body and in the slots - Chips under a knife's seating surface - A chipped edge after a nail or a piece of embedded metal Sources: [ISO 21940-11:2016](https://www.iso.org/standard/54074.html) ## The set and the projection first, balancing second This isn't us being fussy — it's the logic of the method. The knives and gibs make up a noticeable share of the shaft's rotating mass, and how they're arranged around the circumference is exactly the mass distribution we're balancing out. Correction weights compensate for one specific configuration of the set. So we don't balance a shaft with a random assortment of knives: the result wouldn't survive the very next knife change. Projection also determines how many knives actually cut. A knife sticking out past its neighbours by three to five hundredths of a millimetre takes over all the cutting by itself, and no amount of balancing will fix that. Checking the set takes less than an hour and often resolves half the problem before the first run. - [x] Weigh the knives: keep the mass spread within the set inside the manufacturer's tolerance, usually a few tenths of a gram - [x] Don't mix gibs and screws between slots — they press against the knives and factor into the mass balance - [x] Clean the slots of chips and resin before fitting the knives, not after - [x] Set the projection with a dial indicator or a setting gauge: even along the length of each knife and consistent between knives - [x] Torque the gib screws to the standard value, working from the middle of the knife out to the edges - [x] When you send knives out for sharpening, ask the shop to grind the whole set in one setup, not knife by knife ## Waviness on the board: how projection and imbalance ruin planing A planed surface is a chain of arcs left behind by the knives. The mark's pitch equals the feed per revolution divided by the number of knives actually cutting. As long as all the knives are cutting, the arcs are small and the surface looks smooth. The moment one knife starts sticking out, it starts cutting alone: the pitch grows two to four times over, and the wave shows up under raking light and can be felt by hand. Imbalance works differently. It rocks the shaft on an orbit within its bearing supports, and the depth of every arc starts to vary. On the board, this looks like ripple with an unstable pitch, fuzzy grain on the planed face, tear-out around knots. At the same time, the rotating force wears out the bearings and the pulley fit, so an imbalance left unchecked adds humming and hot bearings to the ripple over time. | What's visible on the board | Likely cause | What to do | | --- | --- | --- | | An even, heavy wave, pitch equal to the feed per revolution | One knife is sticking out and doing all the cutting | Reset the projection, match the set by mass | | Ripple with a wandering pitch, fuzzy grain, humming machine | Shaft imbalance, orbiting in the bearing supports | Balancing in two planes | | A wave whose pitch doesn't change when the feed rate changes | Vibration not from the shaft: the drive, a neighbouring machine, table resonance | Vibration diagnostics, tracking down the source by frequency | | A periodic deep mark once per revolution | A nicked edge or a chip trapped under the knife | Inspect the set, replace the knife | ## Thickness planers, jointers, four-sided moulders: what we account for on each type ### Thickness planers The cutterhead sits above the table, belt-driven, typical speed 4000–6000 rpm. We aim the phase sensor at a mark on the pulley or end face of the shaft itself, not the motor: the belt slips, and phase taken from the motor lies. Access to the shaft is from above, after removing the cover and rerouting the dust extraction. ### Jointers The cutterhead sits between the tables, and we reach the bearing supports from below or from the ends of the bed. Correction planes are only at the ends of the shaft, but a short, rigid rotor like this balances readily. We lock the tables before measuring: a movable table on loose ways is a source of vibration in its own right. ### Four-sided moulders Four or more spindles carrying cutterheads, each at its own speed. We balance them one at a time, with the other spindles stopped. Vibration from one unit shows up on neighbouring housings too, so we separate the sources by rotation frequency and phase, not by which one sounds louder. ### Long planing shafts On a long shaft, the distance between correction planes is large compared to its diameter, so we work strictly in two planes: single-plane correction here would remove vibration at one support and raise it at the other. We have a separate article on choosing the number of planes. ### Insert-knife cutter drums Drums on calibrating and milling units fitted with insert knives. The mounting pockets wear unevenly, and identical knives end up sitting at different radii. Before balancing, we measure the projection at every pocket: worn-out pockets are a reason to skip the weights and send the drum for repair instead. ### Spiral cutterheads Shafts with spiral rows of indexable carbide inserts. Balancing here almost always comes down to getting the set right, more on that in the next section. ## Spiral cutterheads: balancing starts with getting the set right A spiral cutterhead carries dozens, and on wide machines hundreds, of indexable inserts, each weighing, screw included, just a few grams. The manufacturer balances a shaft like this as a complete assembly with the full set fitted. After that, everything depends on maintenance. A lost insert or screw creates a localised imbalance of a genuinely noticeable size: five grams at a 60 mm radius is 300 g·mm — more than the tolerance for the whole shaft. Typical findings on inspection: inserts from different batches with different masses, an insert indexed onto a worn edge with a chipped corner, chips trapped under the seat, a screw tightened by hand. So the order is this: we check the set is complete row by row, index or replace inserts per the manufacturer's procedure, torque the screws to the standard value, and clean resin off the seats. Then we take a verification measurement. Most of the time, the running-speed component (vibration at the shaft's rotation frequency, denoted 1x) is already within tolerance and no weights are needed: you're paying for diagnostics, not balancing. If 1x remains, we balance it like an ordinary shaft. ## How on-site cutterhead balancing goes 1. **Inspection, cleaning, play** — We remove the cover, inspect the knife set and the slots, and clean resin off the shaft body. You can't balance out a coating — it will flake off, taking the result with it. We check for play in the bearing supports and belt tension, with the machine stopped and locked out. 2. **Sensors and phase mark** — We fit two accelerometers on the shaft's front and rear bearing supports, horizontally, in the direction of least stiffness. We attach a reflective mark to the shaft's end face or pulley, and mount the laser phase sensor on a magnetic stand. 3. **Baseline measurement** — A run at operating speed after a short warm-up. We look at overall vibration (the total level across all frequencies), the running-speed component 1x, phase, and the spectrum — the breakdown of vibration by frequency. This is how we separate imbalance from bearings, the belt drive, and table resonance. If 1x doesn't dominate, we say honestly that weights won't help, and go looking for the real cause. 4. **Trial runs in two planes** — A trial weight, one plane at a time, in each correction plane at the ends of the shaft, with a run and a measurement after each. From the shaft's response to the trial weights (the influence coefficient method), the instrument calculates the correction masses for both planes at once. 5. **Fitting the correction masses** — There's no welding on a cutterhead: the steel is hardened, and the shop is full of wood dust. We work by selecting gib and screw mass, using standard threaded holes, and, if needed, removing metal by drilling on non-working lands. The instrument converts the weight into a drilling depth and diameter, and ties the positions to the knife slots as a fixed grid. 6. **Verification, trim, report** — A verification run, with a trim balance if needed: a precise final adjustment with a small add-on, no new trial runs required. We save your shaft's influence coefficients: at the next knife change, balancing will take one or two runs. We hand over a report with before-and-after vibration, residual imbalance, and the G grade. Sources: [Balanset-1A operation manual](https://vibromera.eu/balanset-1a-operation-manual/) · [Balanset-1A manufacturer specification](https://vibromera.eu/product/balanset-1/) ## Speed, tolerance, and the woodworking shop environment Woodworking speeds are high, and that tightens the tolerance. The allowable specific residual imbalance falls in proportion to speed: for grade G2.5 at 6000 rpm, that leaves about 4 g·mm per kilogram of rotor mass. For a 30 kg shaft, that's 120 g·mm, or roughly 2.4 g at a 50 mm radius split across both planes. We take the grade and tolerance from the applicable part and edition of ISO 21940, and assess bearing-support vibration against the applicable part of ISO 20816. How the G grade is chosen is covered in a separate article. The environment sets its own rules too. Wood dust is explosive, so there's no open flame, welding, or spark-producing tools: only mechanical mounting of weights, and drilling with local dust extraction, arranged with you in advance. We take measurements with the dust extraction running, if it's normally on during cutting: the ductwork and the extraction fan change the machine's vibration background. Sources: [ISO 21940-11:2016](https://www.iso.org/standard/54074.html) · [ISO 20816-1:2016](https://www.iso.org/standard/63180.html) ## When on-site balancing won't work - The shaft's bearings are worn: the spectrum shows bearing components, and play can be felt with a pry bar. Bearings get replaced first, balancing after — otherwise we'd be measuring play, not imbalance - The shaft is bent after hitting metal: a dial indicator shows runout along the shaft body. It needs straightening or replacement in a workshop; weights can't hide a bend - The slots or mounting pockets are worn out: the knife doesn't seat repeatably, and every clamping gives a new projection and a new imbalance - Cracks and major chipping on the knives: that's a safety issue and calls for replacing the set, not balancing - Speed won't hold stable: a worn belt or a drifting frequency inverter won't give a repeatable phase measurement - Table or bed resonance: vibration is huge for a small imbalance, and the fix is stiffness and mounting — we have a separate article on resonance > None of these cases means the visit was wasted: you get measurements, spectra, and a clear diagnosis that lets the repair be done right the first time, not by guesswork. Sources: [ISO 281:2007](https://www.iso.org/standard/38102.html) · [ISO 13373-3:2015](https://www.iso.org/standard/40840.html) ## Price, timing, and how to prepare the machine We are the engineers who design and manufacture Balanset instruments, and we do the on-site balancing ourselves. Packages start from 550 EUR per unit (diagnostics 300 EUR + balancing from 250 EUR), minimum invoice per visit 500 EUR. Several spindles on a four-sided moulder mean several rotors in a single visit, which works out cheaper per shaft. The calculator on the site gives an exact figure for your machine. On timing: the first shaft, including checking the set, usually takes two to four hours, trial runs included. A repeat balancing after a knife change, using the saved influence coefficients, fits into an hour. - [x] A freshly sharpened knife set, matched by mass, already fitted and torqued to the standard value - [x] Wrenches for the gibs and the torque figures on hand - [x] The shaft cover opens, and the shaft ends are accessible - [x] The machine can do five to eight short runs at operating speed - [x] There's a way to lock out the start switch while weights are being fitted - [x] Send photos of the shaft, the nameplate, and the speed rating plate in advance: we use them to confirm suitability and work out the price ## Frequently asked questions **Should the shaft be balanced with knives fitted or without them?** Only with the working set fitted and torqued to the standard value. A shaft without knives is a different rotor with a different mass distribution; its balance says nothing about how the unit will behave while cutting. **Why did vibration increase after the knives were sharpened?** Sharpening removed a different amount of mass from each knife, or the gibs and screws got mixed up between slots during assembly, or the projection was set unevenly. Weigh the set and check the projection with a dial indicator. If 1x is still high after that, balancing is needed. **Will balancing remove the wave marks on the board?** Depends which kind. A heavy, even wave with a pitch equal to the feed per revolution is caused by a knife sticking out, and it's fixed with projection, not weights. Ripple with a wandering pitch, fuzzy grain, and humming caused by the shaft orbiting — that's what balancing removes. **Does a spiral cutterhead with inserts need balancing?** Usually not. First, check the inserts and screws are all present, index worn edges per the procedure, torque to the standard value, and clean the seats. After that, we take a verification measurement, and in most cases the running-speed component is already within tolerance. We balance only if 1x remains. **Where do you put the weights if welding isn't allowed?** We select the mass of the gibs and clamping screws, use the standard threaded holes, and, if needed, remove metal by drilling on the shaft's non-working lands. The instrument itself converts the required mass into a drilling depth and ties the positions to the knife slots. **How often should a cutterhead be rebalanced?** A verification measurement is worth doing at every knife-set change. If the sets are matched by mass, vibration usually stays within normal limits. We save your shaft's influence coefficients, so a repeat balancing, if needed, takes one or two runs with no new trial weights.