How to choose a balancing and vibration diagnostics contractor: questions, red flags, contract
A contractor calls and says: 'we'll come out and get rid of the vibration.' They have an instrument, the price is attractive, and a month later the machine shakes just the same. You can tell an engineer apart from someone with a vibrometer in a single ten-minute phone call, if you know what to ask. Below are nine questions, a table for decoding the answers, and a list of what you're responsible for preparing yourself.
Three things that separate an engineer from someone with a vibrometer
A vibrometer shows one number. That number lumps everything together at once: rotor unbalance, shaft misalignment, bearing wear, loose fasteners, frame resonance, a neighboring pump's vibration coming through a shared foundation. One number can't tell you what specifically to fix. It's good for a trend and for the 'investigate or don't' decision, and that's where its role ends.
Balancing only reduces one component: the running-speed one, the same 1x that fits exactly once into a shaft revolution. To isolate it and calculate a weight, the instrument needs amplitude, phase, and speed. Phase is the angle that shows at what point in the revolution the 'heavy spot' passes the sensor; it's measured by a separate sensor keyed to a reflective mark on the shaft. Balancing instruments have their phase measurement accuracy specified in degrees, and without this value, calculating a correction is impossible in principle.
That gives you three checks you can do right there on the phone.
- The instrument measures phase and speed, not just mm/s. If the shaft mark and the tachometer never come up in conversation, the contractor has a vibrometer, and they'll be fitting weights by trial and error.
- There are two channels, and they record at the same time. A correction plane is the cross-section of the rotor where a weight gets fitted. Two planes are linked by cross-effect: a weight in the first plane also changes the vibration at the second bearing. A single channel isn't disqualifying, but it sees the bearings one at a time and needs more runs.
- The contractor is willing to say 'this isn't unbalance.' Being willing to name the method's limit is worth more than a promise to fix any problem. The second kind means you'll be sold weight-fitting regardless of whether it helps.
Sources: Balanset-1A operation manual
Nine questions to ask before you order the work
Ask them in any order. What matters isn't the terminology but whether a method exists: a specialist answers specifically and quickly, because they do this on every visit.
What do you measure with, and how many channels does the instrument have?
Expect: a two-channel vibration analyzer-balancer, two accelerometers on a stud or magnet, a laser phase sensor using reflective tape. An answer of 'professional equipment' with no instrument class named doesn't cut it.
Do you measure phase and speed?
Expect: yes, the phase of the running-speed component relative to the mark on the shaft, plus the rotational speed on every run. Without phase, the instrument doesn't know what angle to fit the weight at, and balancing turns into a lottery.
How will you determine the cause is actually unbalance?
Expect: we'll compare overall vibration against 1x, look at the spectrum and the time waveform, take three directions at each bearing housing, check tightness and soft foot (a mounting foot that doesn't sit flush against the frame), check for resonance on run-up or coastdown. Vibration diagnostic methods are described in ISO 13373; use the applicable part on diagnostics.
What will you do if it isn't unbalance?
Expect: we'll stop, show you the data, name the probable cause and what to check next, and we won't fit weights. An answer like that means the contractor bills diagnostics as separate work, rather than disguising it as balancing.
What document will I get at the end?
Expect: a report with before-and-after measurements, a diagram of points and directions, speeds, overall vibration and 1x, phases, the masses and positions of the weights fitted, and the acceptance criterion. A phrase like 'we'll give you a conclusion' with no list of quantities means nothing.
What conditions do you need from us?
Expect a list: a machine in good condition, a clean impeller, access to the bearing housings and correction planes, permission for repeated stops and starts, room for the reflective tape, your own person with a wrench. A contractor with no requirements for the site isn't planning to check anything.
Who fits the weights, who welds and drills?
Expect a clear split. Either their own fitter comes out with welding gear, or your own team carries out the work to the mass and position specified. Hot work almost always needs your permit-to-work, and that gets sorted out before the visit, not standing at the machine.
How will you verify the result?
Expect: a verification run at the same regime, with the same sensors, at the same points and directions, compared against the original measurement and the acceptance criterion. Changing the point or the regime between 'before' and 'after' invalidates the comparison.
What don't you guarantee?
Expect an honest list: resonance, an eroded or cracked impeller, a bearing defect, shaft misalignment, aerodynamic pulsation. An answer of 'we guarantee everything' means the person either doesn't understand the physics or is counting on you not knowing better.
Sources: ISO 13373-3:2015
Table: what you heard, and what it means
| The contractor's answer | What it means | Your next step |
|---|---|---|
| 'No need to come look, we'll get rid of the vibration' | They're selling a result they haven't measured | Ask them to name the quantity, point, and regime the result will show up in |
| 'The instrument is single-channel, we'll go through the bearings one at a time' | This can work, but there will be more runs and stops, and the instrument won't see the cross-effect between planes at the same time | Confirm the number of stops and the window's length, and write them into the order |
| 'No need for phase, I'll fit the weight by amplitude' | That's trial and error, not a calculated correction. The timeline and the result are unpredictable | Either a different contractor, or work with no time commitment and no guarantee of meeting tolerance |
| 'We'll put the sensor on the casing, it's more convenient there' | The signal will travel through an extra joint and through the panel's own bending vibration. The measurement won't be comparable to either the norms or a previous round | Require a mounting pad on the bearing housing, as close to the bearing as possible |
| 'We guarantee G2.5 grade' | The G grade specifies the rotor's residual unbalance in g·mm/kg, not the machine's vibration | Ask where they'll get the rotor mass and correction radius from, and how they'll convert the residual into g·mm/kg |
| 'We'll get you into zone A per ISO' | Possible, but it depends on the machine group, the support type, and the condition of the components | Pin down the applicable part and edition of the standard, the points, the frequency band, and the operating regime |
| 'Here's a sample report, the before-and-after data is inside' | The contractor has a method and is used to following it | Check the sample against your own list of required fields |
| 'We'll balance it as is, no need to clean it' | Part of the buildup will come off during the runs, and the result will go with it | Clean the impeller before work starts, on your own |
| 'If it turns out to be shaft misalignment, we'll do the shaft alignment on the same visit' | The contractor has both the method and the tooling for either case | Confirm what they align with, and what goes into the alignment report |
Red flags: when it's better not to hire them
No single item here is a dealbreaker on its own. Two or more in the same conversation mean you're paying for weights to be fitted, not for a vibration reduction.
- A promise to eliminate the vibration before the first measurement. Until 1x has been compared against overall, there's nothing to promise.
- The report has no phase or speed. That means they weren't measured, and without them a correction can't be calculated.
- A refusal to show raw before-and-after data. The spectrum, the time waveform, and the run table either exist, or the work didn't happen.
- A measurement on the casing, a guard, piping, or the frame instead of the bearing housing.
- No requirements for the machine's condition at all. Balancing doesn't replace a repair and doesn't cure play in a fit.
- Willingness to balance a dirty, eroded, or cracked impeller. A rotor like that changes mass while it's running.
- A promise of a balance quality grade with no explanation of how it will be calculated. A G grade with no rotor mass and correction radius stays a word out of a catalog.
- 'We'll manage it in one run.' The bare minimum is a starting run, one or two trial runs, and a verification run.
- A price for 'balancing' with no number of machines, number of planes, or acceptance criterion stated. Afterward, there'll be nothing to argue with and nothing to argue about.
The limits of the method: what an honest contractor won't promise
A good sign of competence sounds bad for the sale and good for the job. An engineer lists, unprompted and before you sign, the cases where weights won't help.
Resonance. If the operating speed falls in the range of the rotor's, the bearings', or the frame's natural frequency, the amplitude balloons and the phase becomes unstable. Balancing under resonance gives a result that lasts until the next start-up. Detuning by stiffness, mass, or speed comes first.
Flexible rotor. The rigid-rotor method works as long as the rotor doesn't noticeably change shape at operating speed. For rotors running near or above the first critical speed, a different approach applies, along with a separate part of the standard for flexible rotors. A promise of 'we'll balance it in two planes and that'll do it' is wrong here.
Bearings. Weights don't repair a spalled raceway or a fatigue pit. Calculated bearing life is worked out under ISO 281 for given loads, and unbalance adds a dynamic component on top of them. Balancing extends the life of a sound bearing and does nothing for one that's already damaged.
Shaft misalignment, loose mounting, soft foot, aerodynamic pulsation, electromagnetic vibration. Each of these causes has its own fix, and weights only mask them. More on telling unbalance and shaft misalignment apart in a separate article on distinguishing them by symptoms and phase.
Sources: ISO 21940-12:2016 · ISO 281:2007
Qualifications: what the ISO 18436-2 categories mean
It's fair to ask about a specialist's training, but understanding the answer is more useful. ISO 18436-2 divides vibration analysts into four categories. The first collects data along a pre-set route and compares it against thresholds. The second sets up the measurement, reads a spectrum, recognizes typical defects, and balances a rigid rotor in a single plane. The third runs non-routine investigations, works with coastdowns, and balances on site in two planes. The fourth works on rotor dynamics and reliability programs.
The category is awarded by an independent certification body after an exam, not by a training center just for completing a course. A training certificate and a category certification are different documents, and they get mixed up often. We covered the scheme in more detail in a separate article on vibration analyst categories I–IV.
A certification by itself doesn't guarantee the work on your machine will be done correctly, and its absence doesn't mean you're dealing with an amateur. Look at the method instead. A specialist tells you about the 1x share, the measurement points, the criterion for an acceptable trial run, and the conditions under which they'll stop. Someone who promises everything and names not a single quantity is more dangerous than someone with no certificate.
AXILINE is not a certification body and does not issue ISO 18436-2 certifications. Our engineers design and manufacture the Balanset instruments, do the balancing themselves on site visits, and provide consulting support on measurement and calculation methodology.
What you prepare, and why it's part of the result
A significant share of site visits turn out to be failures before the crew even arrives. Balancing measures how the rotor-bearings-base system responds to a known weight. If the system changes while the work is in progress, the influence coefficients — the measured relationship between 'we fitted a weight' and 'the vibration changed' — lose their meaning, and a good instrument in good hands will produce a bad result.
Go through the checklist a day before the visit. A detailed breakdown of every item is in a separate article on preparing equipment for on-site balancing.
- The machine is in good mechanical condition: fasteners are tight, there's no play, soft foot has been checked, the bearings are quiet, and there are no cracks in the impeller or the bed.
- The impeller or fan wheel has been cleaned of built-up product, dust, and deposits.
- Both bearing housings are accessible: a clean, flat pad for the sensor near the bearing, not on the casing.
- The correction planes are accessible with the machine stopped and de-energized, and there's somewhere to fit a weight.
- There's room on the shaft or half-coupling for the reflective tape, and a spot for the magnetic stand holding the phase sensor.
- Repeated stops and starts at the same regime are permitted, and someone has been assigned to carry them out.
- Someone has been assigned to do the fitting and welding work, and a hot-work permit has been issued, if it will be needed.
- The machine's data is known: power, speed, rotor mass, support type, rotor layout, repair history, and previous measurements.
- It's been agreed who stops the work, and on what signs, if the machine turns out to be faulty.
How to document the work: scope, acceptance criterion, boundaries of responsibility
A dispute over the result almost always turns out to be a dispute over wording. A line item of 'fan balancing' on an invoice describes neither the scope nor success. Describe the work so that after the verification run, the two of you are looking at the same number and reading it the same way.
- Scope. How many machines, which ones exactly, how many correction planes on each, whether preliminary diagnostics is included, and whether a follow-up visit is included.
- Acceptance criterion. Not 'reduce the vibration,' but specifically: a value in mm/s RMS (root mean square vibration velocity), a frequency band (usually 10–1000 Hz), points and directions, operating regime, support type, and the applicable part and edition of the standard. Zones A–D under ISO 20816 only work together with these conditions.
- Residual 1x, separately. The instrument's software reports 'within tolerance' when the running-speed component is below the target value you set. That is not an assessment of overall vibration and not confirmation of a G grade. If you need a balance quality grade under ISO 21940-11, state it separately and agree how it will be calculated from the rotor mass and correction radius.
- What's included in the price. The visit, diagnostics, number of runs, consumables, fitting the weights, the report. While you're at it, decide what happens if diagnostics shows balancing would be useless: that outcome is a result too, and it's billed as diagnostics.
- Who does the fitting and welding work. The contractor's own crew, or your team following their instructions. This is also where responsibility for the quality of welding the weight on belongs.
- Who operates the machine. Starts, stops, lockout, and permits are usually carried out by your own staff under your internal rules.
- What isn't guaranteed. Resonance, bearing defects, shaft misalignment, erosion and deposits, a change in regime after handover. It's better to see this list before the work, not in correspondence afterward.
- How long the result lasts. Balancing doesn't preserve a machine in amber. If the impeller builds up deposits again or wears down, the vibration will come back, and that isn't the contractor's defect. There's a separate article on the mechanics of why it comes back.
An acceptance wording that works: 'overall vibration no higher than X mm/s RMS in the 10–1000 Hz band at bearing housings 1 and 2 in the horizontal-radial direction at an operating speed of N rpm; measured with the same instrument and at the same points as the original.'
Sources: ISO 20816-1:2016 · ISO 21940-11:2016
The report: what it absolutely has to contain
The report is the one thing you're left with after the crew leaves. It's also your evidence six months later, when the vibration climbs and you need to know what level it climbed from. Check it against the list at acceptance, not while sorting out a dispute.
- Machine, assembly, serial number, date, operating regime, and speed for every run.
- A diagram of the measurement points with directions and the sensor mounting method.
- Measurements before the work: overall vibration and 1x with phases at each point, spectrum, and time waveform.
- Trial weights: mass, radius, angular position or fixed-position number for each plane.
- The correction weights fitted: how many, where, how they're mounted, and whether the trial weight was removed or left in place.
- Measurements after: the same points, the same directions, the same regime, placed next to the originals for comparison.
- The acceptance criterion and the conclusion: whether it was met or not, and if not, what's standing in the way.
- What the contractor noticed along the way: play, bearing noise, the impeller's condition, signs of shaft misalignment, recommendations.
- The performing party's signatures and contact details, so a year from now there's someone to go back to with the archived record.
The Balanset-1A saves the results of every run, spectra, and time waveforms into an archive that the report is generated from. The saved influence coefficients later let you get by with one run instead of three on a repeat trim balancing job.
Sources: Balanset-1A operation manual
If it's simpler to call the people who make the instruments themselves
AXILINE's engineers design and manufacture the Balanset instruments and do the balancing themselves at the site where the machine operates. That's why our conversation doesn't start with a price — it starts with two questions: what share of your overall vibration is the running-speed component, and what's already been checked mechanically. If the data shows balancing won't help, we'll say so before the visit.
Send us your measurements, a spectrum, or just a description of the machine with its speed and repair history. You'll get an assessment of whether the method applies, a list of what to prepare on site, and a draft report with the acceptance criterion agreed on before the work starts. If you need the instrument more than a site visit, there's the portable Balanset-1A kit, and for building into your own machine tools and test stands, there's the OEM version without the case.
Sources: Balanset-1A manufacturer specification
Frequently asked questions
Can a machine be balanced with a vibrometer that has no phase sensor?
No, a correction can't be calculated. A vibrometer shows amplitude and has no way of knowing at what point in the revolution the heavy side passes the sensor, so the weight's fitting angle stays unknown. Working by trial and error is theoretically possible, but it means many runs and an unpredictable outcome.
Is an ISO 18436-2 certification mandatory for a contractor?
Usually not, unless your internal rules or your insurer specifically require it. The category is awarded by an independent certification body after an exam, while a course at a training center only produces a training certificate. The method matters more: ask how the contractor distinguishes unbalance from other causes, and what they'll do if the cause is something else.
The contractor is asking us to clean the impeller before they arrive. Is that normal?
It's a sign that they understand the physics of the method. The correction is tied to the rotor's current mass distribution. The buildup comes off unevenly during the runs, and the calculated weights stop matching the rotor. Cleaning before the work is part of the preparation on your side.
What do we do if diagnostics shows there's no unbalance?
That's a normal outcome of a visit, and it's worth providing for in the contract as billed diagnostics. You get a report with the measurements, a named probable cause, and a plan for checking further. A contractor who fits weights anyway in this situation also spoils the baseline measurement you'd later use to track a trend.
Why can't the contract just say 'reduce vibration to the ISO norm'?
Because there's no such thing as 'the ISO norm' without specifics. You need the applicable part and edition of the standard, the machine group, the support type (rigid or flexible), the measurement points and directions, the frequency band, and the operating regime. Without these details, acceptance turns into a dispute where each side waves its own table from the internet.
How many machine stops should we plan for a balancing job?
The minimum is a starting run, one trial run for a single plane or two for two planes, then a verification run. Plus a possible follow-up run. Add coastdown, cooling, and the time needed to gain access to the rotor. More detail in a separate article on how long on-site balancing takes.
Related content
How to Choose a Balancing Instrument: Vibrometer, Analyzer or Balancer
Name the task first, then choose the instrument class. For a route walk and trending, a single-number vibrometer is enough. To understand the cause of vibration, you need an analyzer with a spectrum and a time waveform. To balance, you need an instrument with a phase sensor and a correction calculation by influence coefficients, and for two planes, two vibration channels with simultaneous acquisition as well. You can't balance with a vibrometer: without phase, the instrument doesn't know at what angle to place the weight.
How to Choose a Balance Quality Grade: G6.3, G2.5, and Everything Else
Grade G per ISO 21940-11 (formerly ISO 1940-1) sets the allowable residual unbalance of the rotor itself, not the vibration level at the bearing housing. The number in the grade designation equals the product of the allowable specific residual unbalance and the angular velocity, expressed in mm/s: for G6.3, that product equals 6.3 mm/s. For most on-site work on fans, pumps, impellers, and general-purpose electric motors, the baseline choice is G6.3; G2.5 is used for high-speed machines, turbines, and machine-tool drives; G1.0 and G0.4 are reserved for precision spindles. At the same grade, the allowable mass falls in inverse proportion to speed: spin twice as fast, and half as many grams are allowed.
On-site balancing of a forest mulcher rotor, without removing it from the host machine
Yes, we balance forest mulcher rotors on site, in their own bearing housings, without removing the drum. The conditions: the tooth set is complete and matched, the drum has been washed clean of soil and wood pulp, the drive holds a stable speed, and the drum ends are reachable once the flap guard or access ports are opened. If the drum is bent from an impact or the teeth are inconsistent, we'll say so first, because weights won't fix that.
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.