# Balancing machine-tool spindles and tooling at the point of operation

> The first to report unbalance on a machine tool isn't the operator, it's the workpiece: regular waviness appears on a ground surface, size stops holding, the wheel needs truing twice as often. Nobody's going to let a machine be stripped down in the middle of an order. We come to site, measure vibration at the spindle supports, and balance whatever can genuinely be balanced on site. And we say honestly when the vibration isn't coming from mass at all, but from the bearings, the taper, or chatter.

**In short:** Yes, we balance machine-tool rotors on site, but not the spindle assembly itself. We balance whatever turns on the spindle and comes off it: a grinding wheel assembled with its flanges and arbor, a chuck, a faceplate, a tool arbor, a knife shaft, a cutter block, pulleys and the drive motor's rotor. We fit two accelerometers to the headstock casing, at the front and rear supports; the optical tachometer is aimed at a reflective marker — on the spindle, not the motor. There are three conditions: there's access to a correction plane (somewhere a mass can physically be fitted or removed), working speed is up to roughly 12,000 rpm, and the spindle bearings are in good shape. If a bearing in the assembly is worn out or the taper runs out, balancing won't give a result, and we'll say so before any work starts.

Source: https://axiline.pt/en/equipment/balancing-machine-tool-spindles-tooling-at/  
Publisher: AXILINE · Vila Nova de Gaia, Portugal · +351 931 831 229 · axilinegeral@gmail.com

## Symptoms: the workpiece suffers first, then the machine

A machine tool complains not with noise but with rejects. Waviness on a ground surface, a periodic mark after milling, faceting on a turned part, size scattering from one workpiece to the next. The operator compensates by adjusting the settings, the wheel gets trued more often, tooling gets changed early. A vibration meter is the last thing anyone reaches for.

- [x] Regular waviness on the surface, spaced at the feed per spindle revolution
- [x] Noise and vibration changed after fitting a new wheel, arbor, chuck or faceplate
- [x] The wheel needs truing more often than before, and the picture is different after every truing
- [x] The knife shaft leaves ripples, and one knife is cutting deeper than the others
- [x] Vibration rose after a repair: bearing replacement, journals reground, a new knife shaft, a rewound drive motor
- [x] Tooling, wheels and bearings are lasting noticeably less than their rated life
- [x] Vibration on the headstock casing is higher than it was six months ago, at the same settings

> A quick check that saves us half a day on site: measure the vibration on the bare spindle, then with the arbor fitted, then with the arbor and the wheel. If the level rises in steps as tooling goes on, it's a balancing job. If the vibration is already high on the bare spindle, the conversation turns to bearings and the taper.

## What machine-tool equipment we balance

Below is what people come to us about most often, and what creates unbalance in each group.

### Metal-cutting machine-tool spindles

Lathe, milling, grinding and machining-centre spindles. The shaft itself is balanced at the factory, so the source is almost always something added on: the arbor, the wheel, the chuck, the pulley, marks left by a repair. We balance the spindle assembled with the tooling it actually runs with.

### Chucks and faceplates

Three- and four-jaw chucks, faceplates, driving devices. Unbalance comes from jaws bored out of a matched set, swarf in the slots, a workpiece clamped off-centre, a lost counterweight. We balance in the same angular position the tooling is mounted on the spindle.

### Grinding wheels assembled on the spindle

The wheel, flanges, blotters and arbor are one rotor, and the whole assembly has to be balanced. The abrasive is uneven in density and soaks up coolant unevenly. We balance after fitting and again after truing.

### Machine-tool drive assemblies

Pulleys, belt drives, intermediate shafts, the drive motor's rotor. With a belt drive, two once-per-turn components live in the spectrum — the spindle's and the motor's. An off-centre pulley gives a clean 1x — vibration exactly at its own rotation frequency — and readily passes itself off as spindle unbalance.

### Woodworking-machine rotors

Knife shafts on thicknessers, jointers and four-sided planers, cutter blocks with insert knives, planer shafts, calibrating-unit rotors. Unbalance arrives with every knife change and every resinous batch of timber.

### Machine-tool rotors after a repair

After a bearing replacement, journals reground, weld build-up, a knife shaft replaced. A visit like this is short: a before reading, balancing, an after reading, and the figures go into a report for accepting the repair.

- Sharpening and polishing machines, grinding heads, wheels on arbors
- A one-off measurement, when you need to know whether it's unbalance or the bearings
- Setting up the method and marking the measurement points, so your own maintenance team can carry on unassisted

## Unbalance, runout or chatter: how we tell them apart

Balancing only reduces the once-per-turn vibration component 1x — the part that fits exactly into one revolution. Nothing else comes off with weights. A quick way to tell the causes apart by eye is to measure the pitch of the mark on the workpiece. A pitch equal to the feed per spindle revolution is left by a rotating force. A pitch unrelated to the speed is self-excited vibration at the system's natural frequency — chatter, in other words.

| What you see on the machine and the workpiece | What the instrument shows | Cause and what helps |
| --- | --- | --- |
| Waviness with a pitch matching the feed per revolution, vibration rising smoothly with speed | 1x dominates, phase repeats within a few degrees | unbalance — on-site balancing works |
| A periodic herringbone-pattern mark, squealing, the tool edge chipping | the peak isn't a whole-number multiple of speed, broadband | chatter: check cutting parameters, tool overhang, mounting stiffness |
| Faceting, size not holding, runout visible on a dial indicator | 1x plus high-frequency energy, runout with the spindle stopped | runout, or wear on the taper and the fit — cleaning and repair |
| Noise at every setting, a hot support | peaks at high frequencies, not whole-number multiples of speed | spindle bearings — the assembly needs a repair |
| Vibration rises sharply within a narrow speed range and drops above it | a narrow peak, phase flipping by around 180° | resonance of the headstock, the cover or the bed |
| Vibration on the drive, a noticeable axial component | strong 2x (twice the running-speed frequency), a phase difference of around 180° across the coupling | drive misalignment — needs shaft alignment |

> A separate trap on belt-driven machines: two running-speed frequencies sit in the spectrum. Before working out a correction, you need to know whose 1x it is — the spindle's or the motor's. The marker and the optical tachometer go on whichever shaft is being balanced. Reading a spectrum and telling unbalance apart from misalignment are covered in separate articles.

Sources: [ISO 13373-3:2015](https://www.iso.org/standard/40840.html)

## What we check before balancing on machine tools specifically

On a fan, the main enemy of a good result is dirt. On a machine tool, it's the spindle bearings and the geometry of the mounting surfaces. The walk-round takes half an hour or more and decides whether there's any point reaching for weights.

- [x] Warm-up of 15–30 minutes at running speed: a cold spindle has a different preload, and the measurement won't reproduce
- [x] Radial and face runout of the spindle, the taper and the fitted arbor — checked with a dial indicator. Runout is not unbalance, and weights don't cure it
- [x] Cleanliness of the taper, the flanges and the faces. A single chip under a wheel flange gives runout that then gets balanced against unsuccessfully
- [x] Spindle bearings: casing temperature, noise, the spectrum in acceleration. A rough guide on the outer rings is usually up to 70°C, and up to 40–45°C for precision work
- [x] Axial and radial play, signs of lost preload, marks of overheated grease
- [x] The wheel: cracks, a ring test by tapping, the condition of the flanges and blotters, tightening torque, and whether the wheel had been sitting in coolant
- [x] Belts and pulleys: pulley runout with a dial indicator, groove wear, tension to the maker's instructions, a matched set of belts
- [x] The bed: levelling, wedges and anchor bolts, no rocking. The set-up mode and door interlocks are agreed with whoever is responsible for the machine

> We don't intervene in the spindle assembly itself. Stripping it down, adjusting preload and replacing bearings is the manufacturer's service work, and it's not something to touch for the sake of balancing. If we find a worn-out bearing, we hand you the figures and a recommendation, and put balancing off until after the repair. Residual unbalance is an added rotating load on the supports; an ISO 281 life calculation won't account for it unless you build it into the input.

Sources: [ISO 281:2007](https://www.iso.org/standard/38102.html) · [ISO 13373-3:2015](https://www.iso.org/standard/40840.html)

## How the work goes on site

1. **Walking the machine and the measurement points** — We look at the operation and the rejects on the workpieces. We mark out and clean off spots for the sensors on the headstock casing, at the spindle's front and rear supports.
2. **Accelerometers and phase marker** — Two accelerometers, mounted rigidly on the metal, radially, close to the supports. Not on the cover, not on a panel. The reflective marker goes on the spindle, the wheel flange or the arbor — one only; we cross-check the speed.
3. **Baseline measurement after warm-up** — At running speed we record overall vibration, the 1x amplitude and phase at both supports, and the spectrum. Two runs back to back show whether the phase repeats. This is where we decide: balance, or look for another cause.
4. **Correction planes and radius** — The machine is stopped and locked out. We look for where a mass can physically be fitted: the spindle's balancing rings or screws, flange holes, the face of the chuck, the slots on a knife shaft, the pulley. We number the positions in the direction of rotation, and measure the actual radius.
5. **Trial weight in the first plane** — We fix the weighed trial weight just as securely as the permanent one will be fitted. A usable response is a change in 1x of at least 20–30% by amplitude, or 20–30° by phase. On a rigid spindle, a few grams is enough to give it.
6. **Trial weight in the second plane** — Only for two-plane rotors: a knife shaft, a cutter block, a wide wheel assembly, the drive rotor. One plane means two runs; two planes mean three.
7. **Calculating and fitting the masses** — The software gives the mass and position number for each plane, and can split a mass between neighbouring positions or work out a drilling depth where there's nowhere left to add weight. The angle reference is zeroed where the trial weight stood.
8. **Check run, trim and marking** — The same condition and the same speed; one or two rounds of adding small masses if needed. We also check vibration at the machine's other running speeds. We mark the tooling's angular position so the result survives being taken off and refitted.

> We work with the Balanset-1A, a portable two-channel vibration analyser and balancer: two accelerometers, an optical tachometer reading a reflective marker, a two-channel USB module and laptop software. The instrument runs one- and two-plane balancing by the influence-coefficient method (from the rotor's response to a known trial weight), measures overall vibration and 1x, phase, speed, spectrum and time waveform, works in fixed-position mode, calculates drilling and the tolerance against G grades, and keeps an archive and reports.

Sources: [Balanset-1A operation manual](https://vibromera.eu/balanset-1a-operation-manual/) · [Balanset-1A manufacturer specification](https://vibromera.eu/product/balanset-1/)

## One plane or two, and what figures we bring it down to

The rotor's shape sets the number of planes. 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 mass is fitted. A detailed look at the rule is in a separate article; below is the practical summary for machine-tool rotors.

- The target residual 1x in the instrument's software is a figure in mm/s that you set. It's an internal criterion for the procedure, and it doesn't replace the other two
- Overall vibration in mm/s RMS, for assessing the machine's condition. Fix the applicable part and edition of ISO 20816, the measurement points and the running condition separately — there are caveats there for machine tools
- Residual unbalance in g·mm/kg against a G grade from ISO 21940-11 (the smaller the grade number, the tighter the tolerance). A rough guide: G 1 for machine-tool spindles and grinding wheels, G 0.4 for precision grinding spindles, G 2.5 for less critical machine-tool rotors

| Rotor | Planes | Why |
| --- | --- | --- |
| Grinding wheel assembled with flanges and arbor | One | Behaves as a disc, L/D under 0.5. Correction through the flange's holes or slots |
| Several wheels or a wide assembly on one arbor | Two | Otherwise a couple unbalance remains — a pair of 'heavy spots' at the ends of the assembly, and the vibration at one support won't come down |
| Chuck, faceplate, driving device | Usually one | Only one face is accessible, and the mass goes into the fitted holes or slots |
| Milling-machine spindle with an arbor and tool | One, or two with a long overhang | The further the tool sits from the front support, the larger the couple component |
| Knife shaft on a thicknesser, jointer or four-sided planer | Two | A knife shaft's L/D is practically always above 0.5 |
| Cutter block with insert knives, planer shaft | Two | A long rotor, with mass spread along its whole length |
| The drive motor's rotor and its pulley | Two for the rotor, one for the pulley | The rotor is elongated, the pulley is disc-shaped |

> What these grades mean in grams. The permissible specific unbalance equals G multiplied by 9549 and divided by the speed. For G 1 at 6000 rpm that's about 1.6 g·mm/kg: on a 15 kg assembly that works out to roughly 24 g·mm, or about 0.2 g at a 120 mm radius. That's where the whole practice of machine-tool balancing comes from: masses are weighed on precision scales, a magnetic weight isn't good enough even as a trial weight, and every part of the assembly is fitted in the same angular position every time. Fix the grade and the edition of the standard in the job spec — that's a condition of acceptance, not a decision for us to make on your behalf.

Sources: [ISO 21940-11:2016](https://www.iso.org/standard/54074.html) · [ISO 20816-1:2016](https://www.iso.org/standard/63180.html)

## The wheel after truing, and a new set of knives

Two cases where unbalance comes back not from years of wear, but from a routine piece of maintenance. Balancing here is part of the routine, not a one-off event.

### The wheel is balanced as an assembly

The rotor is the wheel together with the flanges, blotters and arbor. Taking the assembly apart and putting it back together loses the result: the parts' angular position changes. So we mark the relative position of every part and of the arbor in the spindle.

### And again after every truing

Diamond truing removes a layer unevenly and exposes fresh abrasive of a different density. On precision work that shows up on the surface straight away. Using the saved influence coefficients, a repeat balance goes ahead without trial runs.

### Coolant, storage and cracks

The abrasive soaks up liquid. A wheel that's been standing with its lower part in coolant arrives with unbalance that wasn't there yesterday. A cracked wheel is never balanced, under any circumstances — it comes off.

### Knives get changed only as a set

One new knife among worn ones produces both unbalance and uneven chip thickness. Knives, wedges and screws are matched by mass as a set, otherwise balancing turns into chasing its own tail.

### A knife shaft as a ready-made grid of positions

The slots, wedge screws and fitted holes at the shaft's ends are ready-made fixed positions. The instrument gives a position number and a mass instead of an angle, so getting the direction of the angle wrong is ruled out.

- [x] Balance the wheel after fitting it to the arbor, and again after any significant truing
- [x] Don't leave the wheel sitting in coolant; turn it over during any long standstill
- [x] Mark the angular position of the flanges, arbor and chuck relative to the spindle
- [x] Save the assembly's influence coefficients: then a repeat balance takes just one run

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

It's better to know the limits before the invoice for the visit, so we state them plainly.

### Speeds above roughly 12,000 rpm

The instrument calculates the RMS vibration velocity (root-mean-square value) in the 5–200 Hz band, and 200 Hz is exactly 12,000 rpm. An electrospindle running at 18,000 or 24,000 rpm falls outside that band, and its tooling gets balanced on a specialised arbor-balancing machine.

### No access to a correction plane

A closed assembly with no balancing rings, nowhere to aim the tachometer at the shaft. Drilling or welding anything onto the spindle is out: it's a precision part under the manufacturer's warranty.

### A worn-out bearing or a taper that runs out

The vibration here isn't from mass, and weights won't remove it. Worse, balancing would mask the one outward sign of the defect. Repair comes first, with balancing as the final operation.

### Chatter and a lack of stiffness

Self-excited vibration in the machine-tool-workpiece system is fixed with cutting parameters, tool overhang, and the condition of the ways and the clamping. Masses on the rotor are useless here.

### A floating unbalance

Swarf in the faceplate's slots, resin on the knives, coolant in the wheel, play in the fit. The 1x phase wanders by tens of degrees from run to run — there's nothing to balance until the cause is removed.

- Resonance of the headstock or the cover: running speed has landed on the structure's natural frequency, and a balancing result won't hold
- Drive misalignment is fixed with shaft alignment, not weights
- Certification against a G grade, for accepting a new or repaired assembly, is done on a balancing rig with its own measurement set-up, not on site

> Seven typical cases where balancing doesn't help are covered in a separate article of ours. On a visit, we go through this list first and say so on site, not after you've paid for the weights.

## What you get, what it costs, and how to book

The result of the work isn't a single number that dropped, it's the assembly's two states, recorded under identical conditions.

On the money side, the rough guide is this: 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 website works out the exact amount for your machine: it depends on the number of rotors, correction planes, access and how far the site is. We're based in Vila Nova de Gaia, near Porto, and we travel all over the country.

- Before-and-after readings at one condition: speed, overall vibration in mm/s RMS, 1x amplitude and phase at each support and direction
- Spectrum and time waveform, showing exactly what came down and what's left
- The masses and position numbers of the weights fitted, the actual radii, and the marking scheme on the assembly itself
- An assessment against a G grade, if the rotor mass, speed and correction radius are given, plus a mechanical finding: runout, bearings, belts, fixings
- The assembly's influence coefficients, saved, so the next wheel truing or knife change fits into a single run

> The people who come out are the engineers who design and manufacture the Balanset instruments, and who balance with them themselves. The same instrument can be bought and used on your own. If you have your own maintenance team, it's usually more cost-effective to set up the method once and then balance wheels and knife shafts yourselves, calling us in only for the difficult cases. In your enquiry, tell us the type of machine and the operation, spindle speed, what's mounted on the spindle, what rejects you're seeing on the workpiece, and what access there is to the correction planes.

Sources: [Balanset-1A manufacturer specification](https://vibromera.eu/product/balanset-1/)

## Frequently asked questions

**Can a spindle be balanced without taking it off the machine?**

We don't strip down or balance the spindle itself. What gets balanced on site is the assembly turning on it: the wheel with its flanges and arbor, the chuck, the faceplate, the tool arbor, the pulley. We fit sensors to the headstock casing at the front and rear supports, the marker onto the spindle, and work at running speed. If the manufacturer has provided balancing rings or screws, we correct through those.

**Why does a grinding wheel need balancing after every truing?**

Truing removes a layer of abrasive unevenly and exposes a fresh layer of different density, so the unbalance after it is no longer what it was. On precision work that shows up as waviness on the surface and faster wheel wear. A repeat balance, using the saved influence coefficients, goes ahead without trial runs, in a single run.

**We're getting waviness on the workpiece. Is that definitely unbalance?**

Not necessarily. Measure the pitch of the mark. A pitch equal to the feed per spindle revolution points to a rotating force — unbalance or runout. A pitch unrelated to speed is chatter. A third possibility is a taper that runs out, or a worn-out bearing. We separate these causes using the once-per-turn component, phase and spectrum before we ever reach for weights.

**What balance quality grade does a machine-tool spindle need?**

A rough guide: G 1 for machine-tool spindles and grinding wheels, G 0.4 for precision grinding spindles, G 2.5 for less critical machine-tool rotors. Fix the exact grade and the applicable edition of ISO 21940-11 in the job spec — that's a condition of acceptance, not a universal constant. The instrument will work out the allowable mass once you enter the rotor mass, speed and correction radius.

**The knife shaft started humming after one knife was changed. Do you balance it?**

Restore the matched set first. One new knife among worn ones produces both unbalance and uneven chip thickness — balancing doesn't fix that. Knives, wedges and screws are matched by mass as a set. After that, the knife shaft is balanced in two planes, using the fitted slots and holes as fixed positions.

**What if our electrospindle runs at 24,000 rpm?**

We can't balance a spindle like that on site with our equipment: its once-per-turn component sits at 400 Hz — outside the band in which the instrument calculates RMS vibration velocity. The practical limit is roughly 12,000 rpm. Tooling for high-speed electrospindles is balanced on a specialised arbor-balancing machine. What we can do on site is diagnose the bearings and assess the vibration.