AXILINEHow to balance an electric motor rotor: the on-site procedure
The motor comes back from rewinding, you set it on the frame, aligned it, started it up — and it's louder than before the repair. Or after a bearing replacement, vibration at the non-drive end tripled. An electric motor is one of the rare cases where the same vibration can come from two completely different causes: rotor unbalance and electromagnetic forces in the air gap between rotor and stator. Balancing only fixes the first one, so you need to start not with weights but with confirming you're actually dealing with a mechanical problem.
When a motor rotor genuinely needs balancing
A motor almost never goes out of balance on its own. The rotor is cast or welded, the cooling fan blades don't wear, no product flows through it. So the unbalance appeared at the moment someone put their hands on the rotor. It's worth keeping a short list of those moments in mind: if you've just been through one, a vibration measurement after reassembly is mandatory.
Sometimes vibration increases after work that doesn't create unbalance by itself. Bearing replacement is exactly this kind of case.
After rewinding
The new winding doesn't lay the same way the factory one did: different layout, different banding, a different amount of impregnation. Both the mass and its distribution around the angle change. On top of that, factory-fitted balance weights cast into the end rings are sometimes lost along with the old insulation during teardown.
After a bearing replacement
The replacement itself doesn't create unbalance. But the compliance and clearance at the bearing change, the rotor seats differently, and an old residual unbalance that wasn't visible before starts showing up at 1x (the running frequency: vibration that repeats once for every rotor revolution). And careless press-fitting can bend the shaft.
After restoring mounting surfaces
Weld build-up, spraying, machining journals and bearing fits — metal is removed and added unevenly around the angle. This also introduces runout at the mounting surface, which makes the half-coupling or the cooling fan seat with eccentricity — offset from the rotation axis.
After fitting a fan or half-coupling
A mounted mass at the end of the shaft acts as an overhung load, so its own unbalance shows up straight away in the bearing's vibration. Add keyway clearance and mounting runout, and a new half-coupling often produces more unbalance than the entire rotor.
After shaft repair
Straightening a bent shaft, building up a keyway by welding, a new key. Here the geometry changes, so does the mass distribution, and so do the residual stresses — the shaft can 'warp' when it heats up.
After rotor damage
A broken rotor bar, a torn-out section of an end ring, a snapped fan blade. Mass is gone from one spot — the unbalance appears instantly and is usually large right away.
Balancing only removes the 1x running-speed component. If the motor is humming because of turn-to-turn shorts, an uneven air gap, or a loose stator lamination stack, weights won't help by a single gram.
First, separate the electrical cause from the mechanical one
This is the single most useful check on an electric motor, and it takes one minute. Warm up the machine, put a sensor on a bearing housing, start recording, and cut power at operating speed.
Electromagnetic vibration disappears the instant power is cut — within a fraction of a second, right along with the current. Unbalance can't do that: centrifugal force drops in proportion to the square of the speed, so vibration from unbalance fades smoothly as the rotor coasts down. Watch specifically the first instant after disconnection, not the whole coastdown curve: further on, the rotor will pass through the supports' and frame's resonances, and the picture gets more complicated. On machines with a large moment of inertia and a rigid coupling, coastdown takes minutes — it's more reliable there to record the time waveform and look at it afterward.
The second sign: dependence on load. Unbalance knows nothing about load — its 1x barely changes from no load to full load. The electromagnetic component depends on load noticeably, because it depends on slip and currents. Take two measurements and compare them.
- A peak at twice the supply frequency (100 Hz on a 50 Hz supply) — static air-gap eccentricity, a loose lamination stack or stator mounting, winding defects. Balancing is useless.
- Sidebands around 1x spaced at the pole-passing frequency (slip multiplied by the number of poles — usually a fraction of a hertz to a few hertz) — rotor bar defects or dynamic eccentricity. To see these sidebands in the spectrum (a chart that breaks vibration down by frequency) at all, you need fine frequency resolution, not a 'quick-look' measurement with a wide Fmax.
- Peaks at the rotor bar passing frequency (the number of bars multiplied by the rotational frequency), surrounded by sidebands spaced 100 Hz apart — this is the rotor again, not unbalance.
- Axial vibration more than half the radial vibration, together with a strong 2x — shaft alignment first, not balancing.
- A dominant, clean 1x, a stable phase (the position of the vibration maximum within the revolution doesn't wander from measurement to measurement), a weak 2x, nothing at 100 Hz — that's unbalance, and it's worth balancing.
More detail on reading a spectrum and choosing Fmax and the number of lines is in separate articles on reading a spectrum and on measurement settings. Here, only one thing matters: without separating electrical from mechanical, you risk spending the whole day fitting weights on a machine that needs rewinding.
Sources: ISO 13373-3:2015
What to check before you fit a trial weight
A motor is a rigid, compact, and highly predictable machine, so balancing on it usually converges from the first trial run. But only on condition that the mechanics are in order. Every item below can make the influence coefficients — how the machine's vibration responds to a fitted weight — unstable, and the result unreliable.
- The fit of the half-coupling and the cooling fan: runout with a dial indicator, play in the keyway, set-screw tightness. A weight fitted on a loose half-coupling won't fix anything — it'll shift position from run to run.
- Bearing condition: play, noise, heat, high-frequency signatures in the spectrum. You can't balance a machine with a damaged bearing: the result won't hold, and the bearing housing 'drifts' between runs. Unbalance, incidentally, cuts the bearing's actual life below its calculated rated life under ISO 281 on its own.
- Soft foot — a motor foot that doesn't sit flush against the frame and only gets pulled down by the bolt. Check: loosen the bolts one at a time with a dial indicator on the foot. The foot's rise usually shouldn't exceed a few hundredths of a millimeter; in practice, 0.05 mm is the benchmark. Soft foot makes the support nonlinear, and the influence coefficients stop being constant.
- Tightness of the bed, frame, and anchor bolts, and the condition of the shims and grouting. Looseness produces harmonics and an unstable phase.
- Shaft alignment with the driven machine. Shaft misalignment is the cause, shaft alignment is the fix for it. The order of work is always the same: alignment first, then balancing — otherwise you'll be trying to compensate with weights for something weights can't compensate for.
- Resonance: with a variable-frequency drive, you already have the tool for it — sweep the speed across the range and check for a sharp amplitude peak with a phase swing near the operating frequency.
- Thermal regime: warm up the machine for 30–60 minutes before measuring. Thermal shaft bow, growing clearances, and winding settling all change both the amplitude and the phase of 1x. Balancing a cold motor that runs hot is a classic way to waste a shift.
- The cooling fan and its shroud: dirt, a bent or broken blade, rubbing against the shroud, clogged vents.
Sources: ISO 281:2007
Where to put the sensors, and where to look for correction planes
Mount the vibration sensors on the motor's bearing housings: the drive end and the non-drive end, as close as possible to the bearing itself, where the load actually travels through the metal. Rigid mounting: a magnet on a clean, flat pad, or a stud. Direction: horizontal-radial, perpendicular to the shaft, and the same from run to run. The fan shroud, the terminal box, a protective guard, and the nameplate are no good for the sensor: they have their own resonant life that has nothing to do with the bearing.
Aim the laser phase sensor at a reflective mark — one mark per revolution. On a motor, it's convenient to stick it on an exposed section of shaft between the housing and the half-coupling, on the end face or rim of the half-coupling, or on a pulley hub. If you're working off the outer cooling fan, the mark sometimes goes right on it, seen through a window in the shroud.
A correction plane is a spot on the rotor where a balance weight can be fitted. On an assembled motor, there aren't many such spots, and that's its main limitation. Here's what's usually available.
| Correction plane | Where on the motor | How to mount the weight |
|---|---|---|
| Outer cooling fan | Non-drive end of the shaft, under the shroud | A screw with washers, a clamp, welding onto the disc or at the blade root. It's convenient to set the number of blades as fixed positions — the instrument outputs a position number and mass instead of an angle |
| Half-coupling | Drive end of the shaft | Washers under the half-coupling's bolts — ready-made fixed positions with a known radius. Or a weight on the rim |
| Pulley | Drive end, on belt drives | The pulley's hub or rim. Remember that the pulley and belts produce their own vibration |
| Balance rings and discs on the rotor | Inside the machine | Cast-in weights, welding, drilling. Accessible only with the rotor removed |
| End rings | Inside, on a squirrel-cage motor's rotor | Many designs have built-in spots for weights. Also only with the rotor removed |
The weight has to be fitted just as securely as a permanent one, even on the trial run. And enter the actual fitting radius into the software, not the fan's diameter from memory: an error in the radius carries straight through into an error in the mass.
One plane or two
By the L/D rule — the ratio of the rotor's length to its diameter — almost any induction motor calls for two planes: the rotor is noticeably longer than half its diameter, running at 1500 or 3000 rpm. Couple unbalance — where the heavy spots at the two ends of the rotor face in opposite directions — is real on a rotor like this, and one mass can't remove it: vibration at one bearing will drop, while at the other it stays put or grows.
The problem is that, fully assembled, you often have exactly one plane available — the cooling fan. In that case, be honest about it: you're removing the static component, and the couple component stays. If, after correction, 1x at the non-drive end comes within tolerance while it stays high at the drive end, that means you need a second plane: take the shroud off, use the half-coupling, or the rotor will have to come out and go to a machine.
- A short motor, shaft height up to about 132 mm, speed up to 1500 rpm — one plane on the cooling fan is often enough.
- A two-pole motor at 3000 rpm, a long rotor, shaft height 160 mm and up — two planes, or the couple component won't go anywhere.
- Only one plane available — do a single-plane correction and record the residual 1x at the second bearing in the report. The customer needs to see what was left outside what was achievable.
- Two planes cost exactly one extra trial run: three runs instead of two before the correction gets calculated.
The L/D rule and how to choose the number of planes are covered in a separate article on one versus two correction planes.
On site, fully assembled, or the rotor separately on a machine
After rewinding, the repair shop usually balances the rotor separately, on a balancing machine or stand. That's the right call, and it delivers something you can't get on site.
And yet, almost always, some trimming on site is needed after assembly. The reason is simple: on the machine you balanced the rotor, but it's the machine as a whole that vibrates. Assembly adds its own unbalance — the key, the interference fit, mounting runout, the second half-coupling, the fan, the fasteners. The motor's own bearing housings, frame, and foundation respond to residual unbalance differently than the balancing machine's soft supports do. The thermal regime is different. As a result, a rotor certified to a G grade — the residual-unbalance tolerance — can show vibration above the norm once assembled into the machine, and that isn't the repair shop's mistake.
The good news: trimming on site is short. The same instrument works both on the stand and out in the field, the influence coefficients are saved under the machine's name, and rebalancing the same motor next time takes just one starting run plus a verification run — that's trim balancing using saved coefficients.
The rotor alone on a machine
High sensitivity, access to every plane, you can drill and weld without worrying about neighboring components, and there's certification of residual unbalance in g·mm and g·mm/kg by G grades under the applicable part of ISO 21940. The price: removing the rotor, a mandrel or its own bearings, a drive, and time.
Mandrel eccentricity
If the rotor is located on a cylindrical mandrel during balancing, the mandrel's eccentricity adds to the rotor's unbalance and carries through into the calculation as an error. It's removed with index balancing: the rotor is repositioned 180° on the mandrel, an extra run is taken, and the software calculates the rotor's actual unbalance with the mandrel's contribution removed.
Fully assembled, on site
You're measuring what's actually vibrating: the rotor-bearings-frame-foundation system at operating speed, with the half-coupling, the fan, and the machine warmed up. The limitations are real too: access to the planes, not being able to remove metal, and working between shifts.
How to choose between a site visit and a shop on cost and downtime is covered in a separate article on on-site balancing versus shop balancing.
Sources: ISO 21940-11:2016 · Balanset-1A operation manual
Three details people forget about
Higher speed means a tighter grade
Centrifugal force grows with the square of the speed: the same residual unbalance loads the bearing four times harder at 3000 rpm than at 1500. That's why the tolerance is tighter too. In the table of recommended grades in the applicable edition of ISO 21940-11, motors with a shaft height of 80 mm and up and a maximum speed above roughly 950 rpm fall around G2.5, while small motors with a shaft height under 80 mm fall around G1. Check the current edition and the exact wording for your type of machine, not a table from the internet.
The rotor can stop being rigid
On very high-speed machines — spindle motors, turbo units, high-speed drives — the operating frequency approaches the first critical speed (the speed at which the rotor enters its own resonance and starts to bend) or sits above it. In that case, the rotor deforms under its own unbalance, the influence coefficients become speed-dependent, and you have to work by the rules for rotors with flexible behavior (ISO 21940-12), usually at several speeds. A single-plane 'by eye' correction doesn't work there at all.
Variable-frequency drives and unstable speed
Balancing needs repeatable speed: 1x phase is measured relative to the mark, and if the speed jitters, the phase jitters right along with it. Lock the frequency setpoint, wait for warm-up, and take every run at the same speed. If the machine genuinely operates across a range, balance at the most problematic speed and check the result at the edges of the range. Also account for two side effects: the drive adds its own components to the spectrum that have nothing to do with unbalance, and at the bottom of the range, the rotational frequency can drop toward the lower edge of the measurement band — at a 5–200 Hz band, that's around 300 rpm.
The cooling fan as a separate source
A bent or broken blade produces two things at once: unbalance, which you'll remove, and an aerodynamic component, which you won't. Noise and vibration at the blade-pass frequency (the number of blades multiplied by the rotational frequency) aren't fixed by balancing — the fan gets straightened or replaced. Rubbing against the shroud shows up as spikes in the time waveform and a fan of harmonics. And a plastic fan unevenly picks up dirt and moisture over time — then the vibration comes back months later, and balancing isn't at fault.
Sources: ISO 21940-11:2016 · ISO 21940-12:2016
The step-by-step on-site procedure
- 01
Permit, inspection, agreeing on the runs
You'll need the authority to stop and start the machine several times, access to both bearings, and to at least one correction plane. Go through the checks from the preparation section: fasteners, feet, bearings, half-coupling, fan.
- 02
Run 0: an 'as is' measurement on a warmed-up machine
Two sensors on the bearings, the mark and the laser phase sensor, operating speed. Look at overall vibration (the total level across all frequencies at once), 1x with phase on both channels, and the spectrum. This is also where you decide: 1x dominates — carry on; it doesn't — go into diagnostics, not for the weights.
- 03
Checking for an electrical cause
A measurement under load and at no load, cutting power at operating speed, a look at 100 Hz and the sidebands around 1x. One minute of work that saves a shift.
- 04
Choosing the planes, radius, and positions
Decide whether you're working in one plane or two. Measure the actual weight-fitting radius. If the weight will go on the fan blades or under the half-coupling's bolts, set them up as fixed positions — the instrument outputs a position number and a mass, and you won't get the angle's reference direction wrong.
- 05
The trial weight
Fit a weight of known mass at a known radius, start the machine. An acceptable trial run changes the 1x amplitude by at least 20–30%, or the phase by at least 20–30°. A smaller change — stop, increase the weight, enter the new actual mass, and repeat.
- 06
A second trial run for the second plane
Move the trial weight to the second plane as directed by the software, and take a third run. After that, the instrument knows the system's response on both channels and both planes.
- 07
Fitting the correction weights
The instrument outputs a mass and angle for each plane, or fixed-position numbers. Measure the angle from the trial weight's location, always in the same direction. Mount the permanent weight securely, at the same radius.
- 08
Verification run and report
Check 1x at both bearings and overall vibration in mm/s RMS (root mean square) in the 10–1000 Hz band, and compare against the zones in the applicable part of ISO 20816. Save the influence coefficients under the machine's name and issue a report: measurement points, direction, speed, masses, radius, what it was and what it became.
One plane: three runs — starting, trial, verification. Two planes: four. With saved influence coefficients for the same machine: two.
Sources: ISO 20816-1:2016 · Balanset-1A operation manual
Mistakes that undo the work
If, after honestly working through this procedure, the vibration still doesn't come within the norm, what you've got isn't a failure but a diagnosis: the accessible plane isn't enough, the rotor needs a machine, the bearing needs replacing, or the machine needs rewinding. That's a result too, and it saves money.
AXILINE comes out to the site with the Balanset-1A: we take vibration at both bearings, separate the electrical cause from the mechanical one, balance the motor fully assembled, and leave you with a report showing the before-and-after numbers. The Balanset instruments are designed and manufactured by the same engineers who do the balancing themselves on site, so questions like 'what do we do if the phase drifts' get answered from experience, not from a manual.
If you have your own repair shop and motors come through in a steady stream, it makes more sense to get the instrument yourself: the kit is portable, works both on site and as a stand's measurement system, and the Balanset-1A OEM version without the case builds right into a machine. Consulting support while you're getting up to speed is a normal part of the service — just ask.
- Balancing instead of aligning. The motor is brought to a perfect balance, and the vibration from misaligned half-couplings stays right where it was, because it was never unbalance in the first place.
- Balancing a cold machine. An hour into operation, the 1x phase has drifted, and the vibration is back.
- Fitting a weight on a half-coupling that's about to be taken apart. Any rebuild of the joint zeroes out the result. Mark the relative position of the half-couplings, weights, and bolts with paint.
- Mounting the sensor on the fan shroud or the terminal box. The numbers look nice, but they have nothing to do with the bearing.
- Skimping on the trial weight. A 5% change in amplitude means the influence coefficient was calculated from noise, and the correction will be random.
- Changing the speed between runs. On a variable-frequency drive this happens in one motion, and it invalidates every previous measurement.
- Ignoring the peak at 100 Hz and balancing electromagnetic vibration. It will never go down.
- Drilling the fan blades or welding on a weight without thinking about strength and aerodynamics. A weight that flies off at 3000 rpm stops being a vibration problem.
Sources: Balanset-1A manufacturer specification
Frequently asked questions
Can an electric motor's rotor be balanced without removing it?
Yes, and in most cases that's exactly how it's done. The motor is balanced in its own bearing housings at operating speed: sensors on the bearings, a laser phase sensor on the shaft mark, weights in an accessible plane — on the cooling fan or under the half-coupling's bolts. There's one limitation: if only one plane is accessible but the rotor needs two, a couple unbalance component will remain. Then you'll have to take the shroud off, work through the half-coupling, or remove the rotor.
Does the rotor need balancing after rewinding?
Yes. The new winding changes both the mass and its distribution around the angle, and the factory-fitted weights are sometimes lost during teardown. The correct order is: the repair shop balances the rotor separately, on a machine, in two planes, and after the machine is assembled, a verification measurement is taken on site, with a short trim if needed. Assembly adds its own unbalance, and the bearings on the foundation respond differently than the machine's own supports do.
How do we tell vibration from unbalance apart from an electrical cause?
Three signs. First: cut power at operating speed — electromagnetic vibration disappears instantly, unbalance fades smoothly along with the speed. Second: compare measurements under load and at no load — unbalance barely depends on load, an electrical cause depends on it noticeably. Third: look at the spectrum at twice the supply frequency (100 Hz on a 50 Hz supply) and at sidebands around 1x spaced at the pole-passing frequency.
How many runs are needed, and how long does it take?
Single-plane balancing is three runs: starting, with the trial weight, and verification. Two planes: four. Rebalancing the same machine using saved influence coefficients takes just two runs. Most of the time gets eaten up not by the runs themselves but by mounting the sensors, cooling down and warming up, access to the correction plane, and coordinating stoppages with production.
What balance quality grade should we choose for an electric motor?
Use the table of recommended G grades in the applicable edition of ISO 21940-11: motors with a shaft height of 80 mm and up and a speed above roughly 950 rpm usually fall around G2.5, while small motors with a shaft height under 80 mm fall around G1. Check the exact wording for your type of machine, and don't confuse the G grade with a vibration norm in mm/s: these are two different tolerances, and meeting one doesn't confirm the other.
Should the motor be balanced with the half-coupling fitted, or without it?
With whichever half-coupling it'll actually run with, and in the condition it actually runs in: with the fan, the shroud, the standard fasteners. The half-coupling sits as an overhung load and contributes its own unbalance, so balancing without it stops making sense the moment the machine is reassembled. Mark the relative position of the half-couplings, bolts, and fitted weights with paint, so the result survives the next time the joint gets taken apart.
Related content
How to Balance a Fan: The On-Site Procedure
A fan is balanced on site, in its own bearing housings. You mount two vibration sensors on the housings, a laser phase sensor on a reflective marker on the fan shaft, run the machine at its running speed, and compare overall vibration (the total level across all frequencies) with the 1x running-speed component — vibration exactly at the rotation frequency. If 1x accounts for most of it, the unbalance is real: next comes the trial weight, the calculation, mounting the weights on the blades as fixed positions, and a verification run. Before any of that, you need a clean wheel, tight fasteners, intact blades, and a belt drive in good working order.
Electrical Causes of Motor Vibration: How to Separate Them from Mechanics
Electrical vibration gives itself away through a frequency tied to the mains supply, not to the shaft: a line at twice the supply frequency (100 Hz on a 50 Hz supply) and sidebands around 1x (running speed) spaced at the pole-pass frequency F_p = s · f_line · P, where s is slip and P is the number of poles. This is checked with a single shutdown: cut the power and keep recording vibration through the coast-down, while the shaft slows under its own inertia. The electromagnetic component cuts off within a fraction of a second; the mechanical component decays smoothly along with the speed. Balancing doesn't remove an electrical cause at all, because the source of the force isn't the rotor's mass, and a weight on the rotor has no effect on that source.
On-site balancing of electric motor and generator rotors
Yes, we balance electric motor and generator rotors on site, in their own supports, at operating speed. Three conditions apply. First: the vibration has to be dominated by the 1x running-speed component — vibration at the rotor's rotational frequency, the main sign of imbalance — rather than by line frequency and its second harmonic. Second: at least one correction plane needs to be accessible, meaning a spot where a weight can be fitted: usually the cooling fan, the coupling, a balancing ring, or the rotor face. Third: the fits, bearings, and fasteners have to be sound, because weights don't cure play. If the rotor is coming off anyway (rewinding, bearing replacement, restoring the mounting surfaces), it makes more sense to balance it on a machine in the workshop, and we'll say so plainly rather than fit weights through a hatch.
Describe your equipment and the problem
We'll answer your questions, clarify the details, and let you know what's needed for an estimate and a visit.