Balancing production-line equipment at the point of operation

On a line, vibration shows up first in the rejects: a seam wanders on packaging, film thickness drifts, a streak appears on a web. Then noise joins in, a bearing support runs hot, belts tear more often. We come out with an instrument, separate unbalance from the other causes, and balance the rotor in its own supports — without removing it and without carting the impeller off to a workshop.

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

Symptoms: when to call, without waiting for a shutdown

Unbalance develops predictably. First the level rises at the bearing support, then product quality suffers, and only after that do bearings and seals start to fail. Weeks usually pass between the first stage and the third. Within that window, the work costs the least, because you get to choose the shutdown time.

A line has a second quirk: vibration travels along the shared frame and the floor, so it's often not the guilty machine that complains. An induced-draft fan rocks the platform, while the rejects turn up on the packaging machine in the next bay. So we take readings at several points and compare rotation frequency and phase — the angular tie between the vibration and a specific shaft's turn.

• Humming or runout appeared after washing, or after replacing an impeller, knives, hammers, or repairing blades.
• Rejects are increasing: an uneven seam, uneven film thickness, a streak on the web, paint runs, chatter marks after grinding.
• The level at the bearing support is higher than the last reading, even if it's technically still within limits.
• Belts wear out faster than usual, you're tightening the tensioner every shift, the pulley runs hot.
• One seal is leaking on one side, and the bearing in that assembly has already been changed twice this year.
• Fixings on the frame keep working loose on their own, and there's rust-coloured fretting dust visible under the washers.
• The rotor runs in a dirty or abrasive environment: product build-up, blade erosion, hardened coating material, caked clinker.
• The measurement shows that the once-per-turn component 1x accounts for most of the level, and the rest of the spectrum is clean.

If humming appeared right after assembly, don't start with weights. Fixings first, a soft-foot check (a support that doesn't sit flush on the frame and rocks the machine when tightened) and shaft alignment, then balancing. More on this in the article on telling unbalance from misalignment.

Which production-line rotors we balance

These machines all have one thing in common: the rotor turns in its own bearing supports and can be run up to working speed. That's enough to work without stripping it down. The industry and the environment change the details: where the sensors go, what we fix the correction mass with, how many correction planes are needed. Next to each subgroup is what most often creates unbalance in it.

A separate case: vibratory feeders, vibrating screens and vibrating tables. There, unbalance is designed in, and it must not be removed. On machines like these we measure vibration and assess the bearings and fixings, but we don't work out any correction masses.

Industries: typical rotors, vibration source, how the work is organised

The table is here so you can find your case faster. Read the third column most closely: it holds the reason unbalance will come back if it isn't dealt with before balancing.

Welding weights on is only mentioned for large, dusty machines. In food, pharmaceutical and chemical production we assume by default that hot work and interference in the product zone are forbidden, until you confirm otherwise.

What we check before balancing on machines like these specifically

Half the rotors that get brought in for balancing need washing and tightening up first. On line equipment this rule applies even more strictly: product builds up unevenly, knives change mass after every sharpening, and the drive is almost always a belt.

The checks take twenty minutes to an hour and decide whether there's any point going further. If there's no unbalance, we say so straight away and hand over a measurement report with a diagnosis instead of a balancing job.

• We compare overall vibration (the total level from every cause combined) with the once-per-turn component 1x: balancing only reduces 1x.
• We read the spectrum — how the vibration breaks down by frequency. A comb of harmonics (a series of peaks at whole-number multiples of speed) points to looseness, a peak at the blade-pass frequency (speed multiplied by the number of blades) points to rubbing or aerodynamics, and high-frequency peaks unrelated to speed point to the bearings.
• We check how clean the rotor is. As long as a build-up is loosely stuck on, any correction will leave with it.
• We inspect the blades, knives, hammers, shrouds: a crack, a chip or erosion isn't cured by weights.
• We check that replaceable knives and hammers are matched by mass. A mismatched set gives unbalance that will come back at the next change.
• We check fixings and soft foot, the condition of the frame and anchor bolts, and strain from pipework and ducting run in rigidly.
• We check the belt drive: belt wear, pulleys, tension. Slippage makes speed unstable, and that wrecks the phase reading.
• We check for misalignment and the coupling's condition: a noticeable 2x (vibration at twice running speed), large axial vibration, the level drifting as the machine warms up.
• We check how stable the speed is if the machine is run by a variable-frequency drive or controlled against a process parameter.
• We rule out resonance: we look at amplitude and phase as speed changes and on a coastdown (free slowdown after switch-off), and check the frame and platform separately.

One more check is specific to a line. If a crusher or compressor is running nearby, its frequency lands in your rotor's spectrum too. We note down which units were running during the measurement, and if needed we ask for anything extra to be stopped for ten minutes.

Sources: ISO 13373-3:2015 · ISO 20816-1:2016 · ISO 281:2007

How the work goes within a maintenance window

On-site balancing is made up of short runs and pauses to fit the mass. The runs are cheap; the pauses cost time. So we draw up a plan in advance, and before the visit we ask you to describe the access.

1. Step 1

   Agreeing the window and access

   Before the visit we confirm: how many rotors, the rotation speed, where the bearing supports are, whether there's a hatch or a removable cowling, how long the machine can be stopped for, and what restrictions apply to the zone and to hot work. That way we arrive with the right fixings and weights, instead of going back for them.

2. Step 2

   Inspection and preparing the rotor

   We check cleanliness, fixings, blades, belts, the condition of the supports. Washing and tightening up happen before the measurement. We stick the reflective marker onto the shaft and pick spots for the sensors: clean, flat, as close to the bearing as possible.

3. Step 3

   Baseline measurement at the working condition

   Two accelerometers on the bearing supports, the optical tachometer reading the marker, a run at working speed. We record overall vibration, the 1x amplitude and phase on both channels, speed, spectrum and time waveform. This is where it's decided whether balancing is worthwhile at all.

4. Step 4

   Trial weight and influence coefficients

   We fit a weighed trial weight in the first correction plane and run the machine. The response in amplitude and phase needs to be clear, otherwise we increase the mass and repeat. For a two-plane job we repeat the same in the second plane. That way the software gets the influence coefficients — your machine's own response to a known weight.

5. Step 5

   Fitting the correction mass

   The instrument gives a mass and angle, or a fixed-position number if we've set up the blades or holes as positions. On line equipment this is most often a bolt with washers through an existing hole, a weight on a stud, mass in a provided slot, or drilling to a calculated depth.

6. Step 6

   Check run, report, saved coefficients

   We compare the result with the baseline measurement, and do one trim pass if needed. On hot machines we take the check reading once the machine has settled at its working condition. We save the influence coefficients: after the next wash or knife change, the rotor is trimmed with a trim balance (a short touch-up using the saved data) without any trial runs.

A typical window for one rotor: two to four hours including preparation. A heavy rotor with a long coastdown, access through a hatch, and a hot machine all stretch that out. Several rotors of the same type in one visit go faster. More on this in the article on how long balancing takes.

One or two correction planes for these rotors

The rotor's geometry sets the number of planes, not a wish to save on a run. The rule of thumb is simple: the length-to-diameter ratio in the zone where the mass is fitted. A short disc behaves as a single mass. An elongated rotor almost always has a couple unbalance — the ends are heavy on opposite sides — and one mass won't remove the vibration at both supports at once.

On lines, the split comes out clearly. Disc rotors: fan and induced-draft-fan impellers, dryer spray discs, grinding wheels, faceplates. Elongated rotors: extruder screws, calender and dryer rolls, knife shafts, drums, crusher and shredder rotors, papermaking-machine rolls, mixer rotors.

• One plane suits a short disc rotor at moderate speed, especially when only one side of the impeller is accessible.
• We take two planes for any elongated rotor, and whenever the vibration at the second support stays high after a one-plane correction.
• Long shafts in paper and textile production often behave as flexible rotors: the bow shape changes with speed, and a result from one speed looks different at another. The rules for flexible rotors apply to those.
• Crusher and mill rotors formally need two planes, but the hammer set gets restored first: balancing an uneven-weight set makes no sense.

The cost of the question is measured in runs: one plane needs one trial run, two planes need two. On a continuous line, that's the main reason to settle the number of planes in advance. More on this in the article on choosing the number of correction planes.

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

What gets in the way, and when it won't work on site

Hygiene requirements can be worked around, but they need agreeing first. In food and pharmaceutical production, welding is usually out, surfaces are stainless and polished, and anything added has to withstand sanitising and not create a dead spot where product can collect. We work with bolted stainless-steel weights and washers through existing holes, or remove metal by drilling where the construction allows it and the manufacturer permits it.

Explosive and dusty zones call for a working arrangement. The measuring equipment runs off a laptop and isn't rated for explosive zones, so the work is agreed with whoever is responsible for explosion safety on your side: under a permit with the process stopped, or with the measuring equipment kept outside the zone boundary.

• No access to the correction plane. If a hatch won't open and a cowling can't come off, there's physically nowhere to fit a mass.
• The rotor can't be run up to working speed, or can't be stopped to fit a mass. Without runs and pauses, the method doesn't work.
• Speed won't hold: a slipping belt, a drifting variable-frequency drive, control against a process parameter. The phase jumps around, and the correction becomes a matter of chance.
• The rotor is worn, cracked or geometrically out of true: blade erosion, knocked-out hammers, a bent shaft. Repair first, balancing after.
• Build-up is unstable. If product falls off in chunks, a correction only lasts until the first chunk goes. Cleaning, blow-off and a coating help here, not weights.
• The machine runs in resonance. As long as the speed coincides with the natural frequency of the frame or the platform, the result is unstable.
• Contractual acceptance calls for a report from a balancing rig. Balancing in the rotor's own supports gives a different picture, and it's not worth substituting one for the other.

If the rotor does have to come out, we'll say so and explain why. The difference between working on site and in a workshop is covered in a separate article. We also do retrofits and instrumentation for balancing rigs, if you have your own rig or are planning a stand for rotors you handle repeatedly.

What you get from the work

The result has to be in numbers, otherwise there's nothing to argue from. We hand over a report showing the starting condition, the measurement settings and the result, one you can attach to the maintenance log. Measurement is done with the Balanset-1A two-channel vibration analyser and balancer: two accelerometers, an optical tachometer, a USB module and laptop software. We're the engineers who design and manufacture the Balanset instruments, and we balance with them ourselves on site visits.

• Overall vibration and the once-per-turn component 1x before and after, for each bearing support and direction.
• 1x phase, rotation frequency, spectrum and time waveform, kept in the measurement archive.
• An assessment of the machine's condition against the overall-vibration zones, stating the applicable part and edition of the standard.
• A tolerance calculation for residual unbalance against G grades (balance quality grades), if acceptance is against that criterion.
• The masses fitted: their value, radius, angle or fixed-position number, for each correction plane.
• The influence coefficients, saved for a later trim balance without trial runs.
• A finding on any defects noticed along the way: fixings, signs of misalignment, belts, bearings, blades, the coupling.

We give the figures from the standards as a working reference. For contractual acceptance, the applicable part and edition for your specific machine, the measurement points, the frequency band and the running condition need to be fixed separately.

Sources: Balanset-1A manufacturer specification · Balanset-1A operation manual · ISO 20816-1:2016 · ISO 21940-11:2016

Price and how to book

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 final figure is affected by the number of rotors, the rotor's mass and accessibility, speed, working within a shutdown window, and how far the site is. The calculator on the website gives you the exact figure for your case.

We're based in Vila Nova de Gaia, near Porto, and we travel all over the country. So we can give you a useful answer from the first email, send a short description: the type of machine, rotation speed, the rotor's approximate mass, whether the drive is direct or by belt, access to the bearing supports and the correction plane, any restrictions on the zone, and how long the shutdown window is. Two or three photos help too: an overall view, the supports, a view of the impeller or shaft through the hatch.

• One rotor with good access: usually one visit and one shutdown window.
• Several rotors of the same type on a line: more cost-effective in a single visit, once the set-up and influence coefficients are already gathered.
• Rotors that get cleaned or resharpened regularly: it's worth having your own instrument and training your own technician.
• An unclear cause of vibration: we start with on-site vibration diagnostics and decide based on the result.

If the measurement shows that balancing isn't needed or won't help, we'll say so plainly and hand over a report with a diagnosis. There's no point paying for weights that don't solve your problem.

Frequently asked questions

Related content