On-Site Shredder and Granulator Balancing: Knife Rotor, Regrinding, Set Mass

You regrind the knives, reassemble the rotor, start the granulator, and the housing starts humming at running speed while the screen fasteners begin backing themselves out. That is what typical knife-rotor imbalance looks like: after sharpening, the knives lost mass unevenly, and the set stopped being symmetric. We come out with the Balanset-1A, separate imbalance from wrap-up, buildup and drive problems, and bring the rotor back into spec in its own bearing housings. Below is what is specific to shredders, plastic and waste crushers, granulators and disintegrators.

Updated 27 August 2026 · by AXILINE · Vila Nova de Gaia

Symptoms: what this looks like on a shredder or granulator

A knife machine produces imbalance in a step change tied to a specific event. Almost always that is regrinding or replacing knives. Less often the culprit is strapping tape or film wrapping around the shaft at the end seals, or a lump of melt that has welded itself to the knife holders after the chamber overheated.

A second characteristic pattern in this group of machines: the vibration level drifts from shift to shift with no work done on the rotor at all. That is the signature of buildup. Soft plastic and rubber settle unevenly on the holders, then break off in a lump, and the machine shakes differently every time. There is no point balancing that state; cleaning comes first.

• Vibration increased the same day the knives were reground, replaced, or repositioned.
• The hum is steady, once per rotor revolution, and it is the same with an empty chamber as it is under load.
• The level changes from shift to shift, drops after the rotor is cleaned, then creeps back up again.
• Film, tape or fibre can be seen wrapped around the shaft at the end seals.
• The bearing housings run hot, and the screen and bed-knife fasteners keep backing out.
• The knife-to-counter-knife gap has drifted on one side of the chamber, and the edges are wearing unevenly.
• During coast-down — free deceleration after shutdown — vibration briefly spikes in a narrow speed range: a sign of nearby resonance.

A useful figure to have before you call: overall vibration in mm/s RMS (root-mean-square value — what any vibration meter shows) at both rotor bearing housings, and the rotor's own speed, not the motor's. On a belt-driven granulator and a geared shredder these speeds differ many times over, and without them we cannot assess either the tolerance or whether the measurement is even feasible.

Knife rotor: holders, knives, shims, and where the imbalance comes from

Unlike a hammer mill, where the working elements hang loosely on pins, the knife here is bolted rigidly to the rotor. The base is a welded drum or a machined shaft with knife holders, to which the knives are bolted: solid full-length blades, segmented blades, or small V-arranged cutting bits. Shims are often fitted between the knife and the holder to set the gap to the counter-knife after regrinding. This whole assembly is the rotor, and it is only ever balanced as a complete unit — full knife set, factory bolts and shims all in place.

Rigid mounting means one simple thing: every gram of difference between knives sits permanently at its own radius and acts as a correction weight, just placed in the wrong spot. So the imbalance story on this machine is almost always a story about the mass of the set.

• Knives lost mass unevenly after regrinding: this is cause number one on granulators.
• One knife was replaced instead of a pair or a full set: the new one is heavier than its worn neighbors.
• Different fasteners and shims from position to position: bolts, washers and plates are mass at a radius too.
• Tape, film and fibre wrapping around the shaft at the rotor ends.
• Melt and rubber crumb building up on the holders and the inside of the drum.
• Worn knife seats and holders, a chipped edge, or local weld build-up from a repair on one side.

Regrinding knives: why checking the set's mass afterward is mandatory

A knife is ground on its working face, and every pass takes off tens of grams. Grinding is usually done as a set, but the metal does not come off evenly: the knives are worn to different degrees, the machine grinds down to a clean edge, and after grinding the set's mass spreads out. A spread of 50–150 grams between positions after several regrindings is normal, not a mistake by whoever did the grinding.

That gives us the rule that saves the most money on this group of machines. After every regrinding, weigh each knife to the nearest gram and true up the set: take the excess metal off the non-working end or the back face of the heavier knives. Fit knives in pairs so that diametrically opposite positions match in mass. Weigh the fasteners and shims together with the knife: after regrinding, there tend to be more shims, and their mass counts toward the balance too.

This is where set-matching and balancing part ways. If the mass spread across the set is more than a few grams per position, it is too early to fit weights: you would be using the instrument to compensate for something that scales and grinding should have removed, and it would all drift again after the next regrinding. Conversely: the set is trued up, the rotor is clean, and the running-speed component — the part of the vibration exactly at the rotor's rotational frequency, which is what imbalance produces — is still dominant. Then the imbalance sits in the rotor itself: worn holders, old weld build-up, or an asymmetric drum, and that is where our work with weights comes in.

• Grind the whole set in one machine setup, and record the amount taken off each knife.
• After grinding, weigh every knife and log the masses: a year from now that log will show you how much life is left.
• True up the masses by removing metal from the non-working side of the heavier knives.
• Arrange the knives so opposite positions match: on high-speed rotors, aim for a difference of only a few grams.
• Replace knives and mounting bolts in pairs or as a full set, with a new knife opposite a new knife.
• Fit identical shims at opposite positions.
• Do not grind a knife below the manufacturer's minimum size: a knife that has gotten too thin breaks, and a broken-off edge fragment causes both imbalance and a failure.

A benchmark figure. A granulator with a 200 kg rotor at 600 rpm, to grade G6.3, has a permissible residual imbalance of around 100 g·mm per kilogram — roughly 20,000 g·mm for the whole rotor. A single knife that is 100 g heavier than the one opposite it, at a 250 mm radius, produces 25,000 g·mm — and eats up the entire tolerance on its own. We take the grade and the tolerance calculation from the applicable part of ISO 21940, with the edition noted.

Sources: ISO 21940-11:2016

Speed and drive: what low speed and hydraulics change

The speed range in this family is wider than in any other: from 10 rpm on a twin-shaft shredder to several thousand on a disintegrator. Centrifugal force grows with the square of speed, so the same excess mass on a granulator at 600 rpm produces a hundred times more force than it would on a slow shaft at 60 rpm. The slower the rotor, the less visible the imbalance, and the less often it is even a problem.

A slow-speed shaft also runs into the physics of measurement. At 120 rpm, the running frequency is 2 Hz, already at the lower edge of the measurement band: the vibration-velocity signal is small at these frequencies, the measurement has to run longer, and the lower band limit has to be pushed down. We know how to do this, but we say honestly: below roughly 100–150 rpm, reliable on-site balancing based on the 1x running-speed component is at the edge of what is possible, and it is needed there vanishingly rarely. Choosing the band and resolution is covered in our article on measurement settings.

The hydraulic drive on large shredders is a story of its own. Hydraulic-motor speed drifts with oil temperature and pump load, and the correction calculation requires the amplitude and phase of the 1x component — the angle showing where the heavy spot sits on the rotor — to be taken at the same speed each time. So we agree in advance with the operations team on a stabilized mode for the duration of the measurements: a fixed setpoint, warmed-up oil, an empty chamber. Rotor reversal does not get in the way of the measurement; we simply work in one direction of rotation.

What we check before the first weight

The first hour of the visit goes into measurement and inspection, and some jobs get closed out without fitting a single weight. We compare overall vibration (the total level across all frequencies) with the 1x running-speed component, and we always look at the time waveform: cutting jerks, knife-to-counter-knife contact, and gearbox impacts show up in it as spikes on top of the sine wave, and are not mistaken for imbalance.

• Knife set: masses from the grinding log, edge integrity, bolt torque, and consistency of the shims.
• Wrap-up on the shaft at the end seals, and buildup on the holders: the rotor must be clean before balancing.
• Knife-to-counter-knife gap along the whole length of the chamber, and the mounting of the bed-knife beam.
• Screen and grate: cracks, loose fasteners, contact marks.
• Bearing housings: play, noise, temperature, condition of the grease and seals, and where dust and fibre get packed in.
• Drive: shaft alignment for gearbox-and-coupling drives, belt tension and pulley condition for belt drives, speed stability for hydraulic drives.
• Fastening to the frame and foundation: anchors, soft foot (a mounting foot that does not sit flush), and the condition of the vibration isolators.
• Behavior of the 1x amplitude and phase during coast-down: a check for resonance in the frame and the platform.

Misalignment after replacing a gearbox or hydraulic motor looks like imbalance in the measurement figures, but is fixed in a completely different way — shaft alignment. The order is strict: foot flatness, shaft alignment, then weights. How to tell the two apart is covered in our separate article.

Sources: ISO 13373-3:2015 · ISO 281:2007

How the work proceeds on site

A visit to one machine takes a shift. The measurements themselves are quick; the time goes into stopping the machine and gaining access: on most shredders and granulators, the rotor is reached through the feed chamber, and every iteration costs a full cycle of stopping, locking out, and restarting.

1. Step 1

   Safety and access to the feed chamber

   Before work starts, the drive is switched off and locked out under your LOTO procedure (lockout/tagout — locking with personal locks and tags), and on a hydraulic drive the accumulator pressure is bled off. This is not a formality: an unbalanced knife rotor with the lockout removed will turn on its own, heavy side down, and it does so suddenly. Before working inside the chamber we restrain the rotor against turning and keep hands clear of the knife line: freshly ground edges cut like blades.

2. Step 2

   Sensors and phase mark

   We mount two accelerometers (vibration sensors) on the rotor's bearing housings, on magnets on cleaned-down pads, as close to the bearings as possible and in one direction, usually horizontal-radial. We stick the reflective phase mark on the rotor's own shaft: on a granulator, motor and rotor speeds differ because of the belt drive, and on a shredder, because of the gearbox, so a mark on the motor would give the instrument the wrong frequency.

3. Step 3

   Baseline measurement

   Start at operating speed with an empty chamber. We record overall vibration in mm/s RMS, the 1x amplitude and phase, speed, the spectrum (vibration broken down by frequency), and the time waveform on both channels. Under load, the phase does not sit still because of the random cutting load, so all balancing runs are done at idle. This is also where we decide whether fitting weights makes sense.

4. Step 4

   Trial runs

   We bolt a temporary mass of known size into a factory hole in the end disc, at a known radius, run the machine, then move it to the second plane and repeat. The instrument calculates the influence coefficients — how much, and in which direction, a known weight changes the vibration of your particular system: rotor, bearings, frame. A valid trial run changes the 1x amplitude by 20–30 percent or the phase by 20–30 degrees.

5. Step 5

   Correction

   The software outputs a mass and an angle for each plane, or a position number if we are working from the disc's grid of bolt holes. We fit weights on thread-locked bolted fasteners, or remove metal by drilling the non-working side; we only use welding where it is acceptable given the dust hazard and the rotor's steel grade.

6. Step 6

   Check run and report

   Same speed, points and sensors. If we do not hit the tolerance on the first try, the software calculates an addition to the weights already fitted. We keep the influence coefficients on file: after the next knife regrinding, trim balancing runs off them, with no trial runs and fewer machine stops.

Sources: Balanset-1A operation manual · Balanset-1A manufacturer specification

Correction planes and weight fastening on the knife rotor

The knife drum is elongated, with a length-to-diameter ratio greater than one, so we work in two correction planes — rotor cross-sections where the weights go. A mass spread among the knives along the rotor's length produces a moment (couple) imbalance: heavy points at each end point in different directions, and one mass cannot remove it. We take the correction planes at the rotor's end discs or flanges, outside the cutting zone. Access to them is usually through the feed hopper or with the screen removed. If the drive-side end is blocked by the gearbox or hydraulic motor, we move the second plane to the nearest accessible holder disc inside the chamber: the influence-coefficient method does not require the planes to coincide with the ends. A disc-type disintegrator more often gets by with one plane, with vibration checked at both bearings.

Weight fastening on this group of machines has one hard constraint: dust. Plastic, rubber and wood dust are combustible, and hot work is most often prohibited in a processing plant. So the primary method here is bolting, not welding as it would be on mineral-material crushers.

• Bolted weights in the factory holes of the end discs: plates or washer stacks on threaded fasteners, thread-locked. The primary method.
• Weight washers under the knife-holder bolts at the end positions, with the factory fastener torque checked.
• Removing metal: drilling the non-working side of a holder or end disc at the heavy point. The software calculates the diameter and depth for the required mass. We do not drill near knife seats or on thin sections.
• Welding: only outside the shredding chamber, with a confirmed steel grade, a hot-work permit, and dust cleared away. We do not weld rotors made of quenched-and-tempered or hardened steel: the weld would create a brittle zone.
• Inside the chamber, a weight must not protrude into the material flow: a protruding plate collects wrap-up and buildup and undoes the balancing. We fit it flush, or on the outer ends.
• The zero-cost correction option is repositioning knives by mass. We check this first — it costs nothing and often closes the case.

We record the mass, radius, position and fastening method of every weight in the report. An error in the radius is a direct error in the mass. The number of planes and the L/D rule are covered in our article on choosing one plane or two.

Sources: ISO 21940-11:2016 · ISO 21940-12:2016

When on-site work will not work

On knife machines, we decline to fit weights more often than on fans, and we say so at the first measurement. The reasons are typical, and almost all of them can be fixed on your side before our visit, or instead of it.

• The knife set's mass has spread apart after regrinding. Scales and truing up the set come first, the instrument after. Weights will not fix this — it all comes back after the next sharpening.
• Wrap-up or buildup on the rotor. Balancing would just lock in a random surface condition that will change by the next shift. We clean, then measure.
• A chipped or broken knife, a torn-out holder. This is a repair job, not balancing: replacement and a chamber inspection come first, measurement after.
• Rotor speed below roughly 100–150 rpm. The running frequency is at the lower edge of the measurement band, and imbalance contributes little to the vibration: we look for the real cause in the gearbox, gear mesh, and fasteners.
• The hydraulic drive will not hold speed. Without a stable rotational frequency, the 1x phase drifts and the correction calculation is unreliable. We discuss a stabilized mode in advance, otherwise the visit turns into diagnostics only.
• Cracks in the drum, a worn rotor-to-shaft fit, holders that shift: the phase does not repeat from one run to the next.
• Worn bearings or gearbox: overall vibration is many times the 1x level, with spectrum components that are not multiples of rotor speed.
• No access to the rotor ends when stopped, or shop acceptance to a G grade is required: in that case, it is more honest to take the rotor to a shop.

If the vibration is coming from the overall condition of the machine, you get a measurement, spectra, time waveforms, and a prioritized diagnosis instead of weights. We have a separate article on cases where balancing does not help.

Sources: ISO 20816-1:2016 · ISO 13373-3:2015

What you get, and how to book a visit

We are the engineers who design and manufacture the Balanset instruments, and we do the on-site balancing ourselves. We are based in Vila Nova de Gaia, near Porto, and we travel throughout Portugal. Vibration diagnostics with a report cost EUR 300 per unit, balancing adds from EUR 250, the minimum invoice for a visit is EUR 500, and the calculator on the website gives you an exact figure for your machine. It pays to treat several machines at one site as a single visit.

After the work, you are left with a report: overall vibration and 1x at each bearing housing before and after, at the same points and speed, spectra and time waveforms, the mass, radius and position of every weight along with its fastening method, a residual-imbalance calculation to a G grade where needed, a vibration-zone assessment noting the applicable part and edition of the standard, and a list of mechanical findings, including recommendations for truing up the knife set after the next regrinding.

• Machine type: granulator, plastic crusher, single- or twin-shaft shredder, disintegrator, and what you shred.
• Rotor speed and drive type: belt, gearbox, hydraulic motor. For a hydraulic drive: whether it can hold a fixed speed.
• Approximate rotor mass, number of knives, mass of one knife, and the radius of the cutting edges.
• When the knives were last ground or replaced, and whether the set was weighed.
• Measurement figures in mm/s and the points, if you have already measured.
• Access to the rotor ends: hopper, removable screen, hinged section of the housing.
• Whether hot work is permitted on site, or whether the plant has combustible dust.
• Photos of the rotor, the bearings, and the machine's nameplate: from these we can work out the correction planes and weight-fastening method in advance.

Want to do this yourselves after every regrinding? We sell the same instrument we use ourselves. Balanset-1A: two accelerometers, a laser phase sensor working off a reflective mark, a two-channel USB module, and software on a laptop. One- and two-plane balancing by the influence coefficient method, overall vibration and 1x, phase, speed, FFT spectrum and time waveform, fixed positions and drilling calculations, saved coefficients, trim balancing, G-grade tolerances, an archive and reports. We will help you set up the sensors on your machine.

Sources: Balanset-1A manufacturer specification · Balanset-1A operation manual

Frequently asked questions

Related content