On-site balancing of a hammer crusher: the hammer set, rotor discs, welded weights

You fitted a new hammer set, started up the crusher, and the bed rolled with a low wave while the handrail on the service platform started shaking. On a hammer machine, this is almost always a story about mass. The hammers hang freely on their pins, centrifugal force swings them all out to the same radial position, and any mass imbalance between opposite pins turns into a force that both bearings feel. We bring a rotor like this back into spec right in its own housings, without removing it — but we start not with weights, but with a scale.

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

Symptoms: what this looks like specifically on a hammer machine

A hammer crusher rarely drifts out of balance gradually. The level jumps in a step, and there's almost always an event that step is tied to: a hammer-set change or flip, tramp metal in the feed, a broken hammer, a bearing repair. Think back to the last thing done to the machine, and half the diagnosis is already there.

The second sign of this design is that the heavy rotor rocks everything around it. What you hear isn't a ring, but a low hum once per revolution; you feel it through the floor grating, and you can see the frame and the housing walls move. A hammer, with its fastener, sits at a radius of 300–500 mm, so even a couple of hundred grams of extra mass produces a noticeable force at operating speed.

• Vibration increased on the same day the hammers were changed or flipped.
• The hum is low and steady, once per revolution, and it's the same at idle as it is under load.
• During coast-down, which is long for a rotor like this, the hum fades along with the speed, but spikes sharply in one narrow speed band.
• The level crept up after running wet or dirty feed material, and came partly back down after the rotor was cleaned.
• The bearing housings run hotter than usual, and grease is being forced out through the seals.
• A new impact-type sound has appeared: sounds like a hammer catching the armor plate or the grate.
• The platform handrail and the dust-extraction ducting are shaking, and the housing's anchor bolts have worked loose.

A useful figure to have before you call: overall vibration in mm/s RMS (root mean square — the standard mode on every vibration meter), at each bearing housing, in the horizontal-radial direction, taken at idle. If you kept a reading from before the set was changed, send that too. The difference between "before" and "after" is worth more than any description in words.

The hammer crusher rotor: discs, pins, hammers, and where the unbalance comes from

The rotor on this machine isn't a casting, it's an assembly. A set of discs sits on the shaft, spacer washers between them, and two, four, or six pins run through all the discs, with the hammers hanging freely on the pins, spaced along their length by washers. The stack is clamped at the ends, and the outer discs are usually the thickest. Those are exactly what become the correction planes — the rotor sections where correction weights go.

The hammers' free suspension gives a feature that matters for balancing. At operating speed, centrifugal force swings every hammer out to the same radial position, and each one acts as a mass at a known, identical radius. That means you can work out the unbalance from a mass difference almost by simple arithmetic. And that same feature makes the design merciless toward any inconsistency in the set.

One more source of vibration doesn't sit on the rotor at all. A hammer machine is protected on the inside by armor plates and lining. The armor doesn't rotate and plays no part in unbalance, but as it wears, the clearance to the hammers grows, and a plate that tears loose changes the housing's stiffness and adds an impact component that's easy to mistake for unbalance.

• Mass scatter in the hammer set. Casting or forging tolerance is still a fact of life, and a difference always exists even within a new set.
• Uneven wear. Material inside the chamber isn't distributed evenly, hammers on different pins wear at different rates, and over the set's service life the discrepancy grows to a noticeable size.
• Losing or breaking a hammer. That's no longer scatter — it's a whole part's worth of mass gone from one side of the rotor.
• Material buildup. Wet, clay-like, resinous, or greasy product settles into the pockets between the discs and onto the pins, then breaks off in a chunk at some random moment.
• One-sided wear on the discs and spacer washers themselves, especially with abrasive feed material.
• Sloppy reassembly after a repair: mixed-up positions, mismatched washers, replacing one hammer instead of a matched pair, a bent pin.
• Hardfacing on the working edges. Build up more weld metal on one side than the other, and you've created unbalance out of nowhere.

Unbalance on a hammer machine is a consequence of operation, not a manufacturing defect. It will come back with every set change, so we keep your rotor's influence coefficients on file — the machine's measured response to a trial weight — and the next calculation runs from them, with no trial runs needed. What these coefficients actually are is covered in our article on the influence coefficient method.

Scale first, weights after: arranging the hammer set by mass

This is the main thing that sets a hammer machine apart from a fan. On a fan there's nothing to arrange — there are only weights. Here it's the reverse: you remove most of the unbalance by matching, before any instrument comes into it, and for free.

The rule is simple. Weigh every hammer to the nearest gram, record the masses, and sort the hammers into groups. Then arrange the groups so the mass sums match up in three directions at once: across diametrically opposite pins, across each pin relative to its own midpoint, and between the left and right halves of the stack along the shaft. The first removes static unbalance (the rotor stops pulling to one side), and the second and third remove couple unbalance (a skew along the length that rocks the rotor from both ends).

Hammers have to be replaced in pairs and in full groups. One new hammer in a worn set guarantees a step increase in vibration: a fresh part is heavier than a worn one by tens of grams, sometimes hundreds.

• Weigh every hammer in the set, including the ones staying in service after a flip.
• Sort by mass and build groups with an equal total mass on each pin.
• Arrange the groups so diametrically opposite positions balance each other out.
• Check symmetry along the shaft's length: the right and left halves of the stack should match in mass.
• Fit identical spacer washers and identical pin hardware, and account for their mass too.
• Flip hammers to their fresh edge as a whole set at once, not selectively.
• Write down the layout scheme. It will save you half a day at the next service.

And now the honest part. If the set is carefully arranged by mass and the rotor is clean and intact, balancing is often not needed at all: the machine settles into an acceptable long-term running zone on its own. We don't hide that, and we say so plainly at the first measurement. Weights are needed where, after correct arrangement, the 1x component (vibration at rotor speed) still dominates: residual mass scatter, disc wear, the stack's fit on the shaft, or an unbalance in the shaft itself. We take the accuracy class and zone boundaries from the applicable parts of ISO 21940-11 and ISO 20816, noting the edition, the measurement points, and the regime. G class is the tolerance on residual unbalance: the lower the number after the G, the stricter the requirement.

Sources: ISO 21940-11:2016 · ISO 20816-1:2016

What we check before the weights, and why we look at the time-domain signal

A hammer crusher is an impact machine, and its vibration looks messy in the spectrum (vibration broken down by frequency). Impacts produce broadband spikes that smear across frequencies and raise the whole noise floor. Looking at the spectrum alone, you could easily mistake that mess for a bearing defect — or, just as easily, miss a real one.

That's why we always look at the time-domain signal alongside the spectrum here. Unbalance in it looks like a clean sine wave, one cycle per revolution, with a steady amplitude. Looseness, a hammer catching the armor plate, a crack, or the stack shifting all produce periodic spikes on top of that sine wave, and you can spot it by eye right away. On top of that, we compare overall vibration with the 1x component: balancing only reduces 1x, everything else is fixed a different way.

We take the measurement at idle. Under load, a random component from the feed shows up in the signal, the 1x phase (vibration's tie to the shaft's angle of rotation) stops holding still, and the correction calculation drifts.

• The hammer set: weighed, arranged by mass, all elements present and intact.
• Buildup in the pockets between the discs, on the pins, and on the spacer washers.
• The outer discs and the stack's weld seams: cracks, wear, signs of shifting.
• The stack's fit on the shaft, keys, tightness of the clamping elements, and the condition of the pins.
• The clearance from the hammers to the armor plate and to the grate, and the condition of the lining and its fasteners.
• The bearing housings: play, noise, temperature, and the condition of the grease and the housing.
• The foundation fixings: anchor bolts, how well the feet sit flush, cracks in the concrete, vibration isolators.
• The drive: belt tension and condition, any unbalance in the pulleys themselves, and, on a coupled drive, shaft alignment.
• Resonance (speed coinciding with the natural frequency of the frame or foundation): we watch the 1x amplitude and phase during coast-down. A heavy rotor's long coast-down is ideal for this.

Shaft misalignment and unbalance give similar readings at a single measurement point, but the fixes are opposite. Balance a machine with shaft misalignment, and the weight ends up compensating for a completely different force — after the shaft alignment is corrected, the vibration will come out higher than it started. The order is: foot contact, shaft alignment or belt tension, then weights. More on this in our article on telling unbalance apart from shaft misalignment.

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

How the work goes on site

A visit for one machine takes a full shift. The measurements themselves are quick — what eats up the time is the stops: coast-down of a heavy rotor, opening up the armored housing, locking out the drive, and letting a weld cool.

1. Step 1

   Agreeing safety, regime, and hot work

   We agree in advance who stops and locks out the drive, who opens the hatches and the housing's hinged section, whether a hot-work permit is needed, and whether welding is allowed given your dust. On a hammer machine, the correction planes are inside the housing, so every iteration costs a full cycle: stop, coast-down, lockout, work, close-up, restart.

2. Step 2

   Fitting sensors and the mark

   Two accelerometers on the rotor's bearing housings, right up against the bearings themselves, on a magnet on a cleaned spot or on a stud. Direction is horizontal-radial, and we keep it the same from run to run. We stick the reflective mark on the rotor shaft, not the motor: hammer crushers usually run on a belt drive, the speeds differ, and a mark on the motor would make the instrument pick out the wrong frequency.

3. Step 3

   Measurement before work

   A run at operating speed, at idle. We record overall vibration in mm/s RMS, the 1x amplitude and phase, speed, and the spectrum and time-domain signal on both channels. This is also where we decide whether fitting weights is even worthwhile.

4. Step 4

   Trial weight in the first and second plane

   We fit a temporary mass of known size at a known radius on the outer disc, run the machine, move it to the second plane, and run again. The instrument calculates the influence coefficients of your specific system: rotor, supports, frame, foundation. A valid trial run changes the 1x amplitude by at least 20–30% or the phase by 20–30°. On a rotor weighing close to a tonne, the trial weight comes out substantial, and we fit it just as securely as a permanent one.

5. Step 5

   Correction

   The software gives a mass and angle for each plane, or a position number if we're working with fixed positions. On a hammer rotor, the position grid is usually already there: the number of pins, holes in the outer discs, welded-on bosses. We fit mass by welding, with a bolt, or remove it by drilling at the heavy point.

6. Step 6

   Verification run, measurement under load, report

   Same speed, same points, same sensors. If we don't hit tolerance right away, the software calculates an addition to the weights already fitted. After that comes a measurement under load, then a report with before-and-after figures and a list of mechanical notes.

Two-plane balancing adds one full stop-start cycle compared with one plane, and we flag that in advance. At the next hammer-set change, we do trim balancing — a quick fine-tune run from the saved influence coefficients, with no trial runs and fewer stops.

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

One plane or two, and where the correction planes sit on this rotor

Geometry sets the number of planes: the ratio of the working length, L, to the diameter, D, in the zone where the weight goes. A hammer rotor is almost always elongated, with L/D at 1 or above, so we work in two planes. A single mass won't do the job here: scatter in the hammers along the stack's length produces couple unbalance, and you'll get the familiar picture where one bearing goes quiet and nothing changes on the other.

We take the planes as far apart along the length as possible — that means the outer discs of the stack. They're thicker than the rest, easier to reach through the hatch, and they give the longest lever arm, meaning the smallest mass needed. Sometimes it's more convenient to work at the pin ends protruding past the outer discs, where washer-weights go under the nuts.

The second plane costs one extra trial run — one extra stop with a full coast-down. We have a separate article on the L/D rule and choosing the number of planes. On fine grinding and at higher speeds, we also check how close the operating frequency is to the shaft's critical speed — the speed at which the shaft hits its own resonance.

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

Fitting correction weights on a hammer rotor

On a machine that works by impact, a weight stays put only because of correct fitting. A mass of a few hundred grams tearing off at operating speed punches through the armor plate and ends up in the grate or the dust-extraction system.

• Welding onto the outer disc. The main method. We take a plate in a steel weldable to the disc material, clean the spot down to bare metal, and weld it all the way around its perimeter. Tack welds alone tear off on an impact machine, so we never stop at just tacking it.
• Weld requirements. We preheat a thick disc before welding, otherwise the weld gets a hardened zone and cracks. We run the weld clear of the disc edge and the pin holes, by at least one hole diameter. After it cools, we inspect the weld: no porosity, undercuts, or unwelded sections are left.
• Where welding is off-limits. If the discs or the hardfacing are made of high-manganese steel, ordinary welding produces a brittle zone, so we fit the weight mechanically instead. Same rule in shops with combustible dust: on wood, plastic, feed, and pigment grinding lines, hot work is often simply prohibited. We check the steel grade and whether welding is permitted against the manufacturer's documentation before the visit.
• Bolted fitting. Dedicated holes in the outer discs, washer-weights under the hammer-pin nuts, factory-installed welded bosses. We always lock the thread: impacts will undo anything that isn't locked.
• Removing metal. On a thick outer disc, it's simpler to drill out metal at the heavy point — the software calculates the diameter and depth for the required mass. We don't do this on a thin disc or near the pin holes, since that's a straightforward loss of strength.
• Reshuffling the hammers. The cheapest correction option, and the first one we check. Sometimes swapping two groups is enough, and no weights are needed at all.

We record the weight's mounting radius in the report along with its mass, position, and fitting method. An error in radius is a direct error in mass. The mounting methods are covered in more detail in our article on fitting correction weights.

When on-site balancing won't give a result

Let's be direct: on hammer machines, the share of cases where weights don't solve the problem is higher than on fans and pumps. Impact damages not just mass symmetry, but fits, welds, bearings, and the frame as well.

• The set is put together with no regard for mass matching. Until that's fixed, you're using a weight to compensate for what a scale and a proper layout would remove.
• A lost or broken hammer, a bent pin, chipped-out hardfacing. That's a failure. Repair first, inspect the housing and the grate, then measure.
• The rotor is caked in material. Balance it as it stands and, within a shift, a chunk of buildup will break off and everything comes back. Rotors like this get cleaned and measured again.
• Cracks in the discs, the stack shifting on the shaft, a sheared key. The 1x phase stops repeating from run to run, and the correction calculation becomes unreliable.
• Worn-out bearings and battered housing bores. Overall vibration is several times higher than 1x, and the time-domain signal and spectrum show components that aren't multiples of rotating speed.
• Hammers catching the armor plate or the grate. In the time-domain signal, that's regular spikes, not a sine wave. What's needed is clearance, not a weight.
• Resonance of the frame, the service platform, or the housing. A heavy rotor on a steel structure falls into it easily, especially after a VFD setpoint change.
• The machine won't hold a stable speed, or there's no access to the outer discs when stopped. In the first case, we discuss the regime in advance; in the second, it's more honest to pull the rotor out to a shop.

The worst-case scenario for you sounds like this: the vibration comes from the machine's overall condition, not from rotor unbalance. In that case, you get a measurement, spectra, time-domain signals, and a prioritised diagnosis instead of weights. We have a separate article on cases where balancing doesn't help.

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

What you get, and how to book a visit

We're the engineers who design and manufacture the Balanset instruments and use them ourselves out in the field. We're based in Vila Nova de Gaia, near Porto, and travel across all of Portugal.

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 on the site gives you the exact figure for your machine. If there are several crushers at the site, count them all in: it's still one visit.

• The crusher or mill's type and model, and what you're grinding.
• Rotor operating speed, drive power, and whether the drive is by belt or coupling.
• The rotor's approximate mass, the number of discs and pins, the number of hammers, and the mass of one.
• The radius at which the hammers sit, and the thickness of the outer discs.
• What was done to the machine before the vibration appeared: a set change or flip, hardfacing, a bearing repair, a speed setpoint change.
• Any measurement figures in mm/s and the points, if you've measured yourself.
• Access to the outer discs when stopped: hatches, the housing's hinged section, the service platform.
• Whether hot work and welding weights on are permitted at your site.
• Photos of the rotor, the bearing supports, and the machine's nameplate.

• A measurement before the work: overall vibration in mm/s RMS and the 1x component at each bearing housing, speed, points, and directions.
• A measurement after, at the same regime and the same points, so the drop in 1x is visible directly.
• Before-and-after spectra and time-domain signals: they show what's left in the vibration besides unbalance.
• The mass, radius, position, and fitting method of every weight installed.
• The hammer mass-layout scheme, if the set was arranged while we were on site.
• An assessment of the machine's condition by overall-vibration zone, citing the applicable part and edition of the standard.
• A residual-unbalance and G-class tolerance calculation, if needed for acceptance.
• A list of mechanical notes, and what to check at the next set change.

"In tolerance" in the software means one specific thing: residual 1x below the set target value. It is not an assessment of the machine's overall condition based on overall vibration, and it is not confirmation of a G accuracy class. In the report, we keep these three kinds of tolerance separate. If you'd rather work with your own crew, we sell the same instrument we use ourselves. The Balanset-1A: two accelerometers, a laser phase sensor off a reflective mark, a two-channel USB module, software on a laptop, single- and two-plane balancing using the influence coefficient method, overall vibration and 1x, phase, speed, FFT spectrum and time-domain signal, fixed positions and drilling calculation, saved influence coefficients, trim balancing, a G-class tolerance calculation, and an archive with reports.

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

Frequently asked questions

Related content