# Training in operating a balancing instrument and the basics of vibration diagnostics on your own equipment

> Your mechanic measures the vibration on an induced-draft fan and gets 7 mm/s. An hour later someone else measures the same machine and gets 3. As long as the reading isn't repeatable, the instrument stays in the cupboard and balancing turns into a lottery. We come to your site and teach people to work hands-on on your own machines: fitting sensors, reading the 1x running-speed component (vibration at the rotation frequency) and the phase (the angle showing where the heavy spot sits on the rotor), calculating the correction, and bringing the machine within tolerance.

**In short:** We train your staff in the practical operation of a balancing instrument and the basics of vibration diagnostics: where to put the sensors and how to mount them, how to get a repeatable reading, how to tell unbalance apart from misalignment and resonance, and how to go from the initial run to reaching tolerance in one and two planes. Sessions run at your site, on your own machines, either on their own or alongside delivery of a Balanset-1A instrument. We agree the programme, scope and order with you around your tasks and equipment fleet: a maintenance crew with twenty fans and a reliability department with pumps and electric motors need different emphasis. Upfront and without qualification: AXILINE is not a certification body, we do not issue ISO 18436-2 certificates, and we do not award a vibration analyst category.

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

## What your people can do on their own after training

Training only pays off if it changes how your crew actually works. Not "sat through a course", but "on Monday, picked up the instrument, measured the extraction fan and fitted two weights". So we build the session around tasks your people will handle on their own, and we're just as honest about where the line sits, the point past which you still need a contractor.

- [x] Take a repeatable reading on a bearing housing: same point, same direction, same mounting, same machine condition every time.
- [x] Separate the 1x running-speed component from the overall vibration and use the ratio between them to decide whether weights are even worth trying.
- [x] Fit a reflective marker and aim the laser sensor so that speed and phase read stably from one run to the next.
- [x] Balance a rigid rotor in one plane: a belt-drive pulley, an impeller, a pump wheel.
- [x] Balance an overhung induced-draft fan or a shaft between bearings in two planes, for when one plane won't settle the second bearing.
- [x] Bring the machine within tolerance with trim balancing, a small addition to the weight already fitted, without starting the procedure over.
- [x] Save the influence coefficients (the recorded response of a specific machine to a trial weight) and service similar rotors in one run instead of three.
- [x] Fill in a report and start a log with a baseline level, so there's something to compare against six months later.

## The reading everything else depends on

Half of all failed balancing jobs go wrong not in the calculation but in the measurement chain. The instrument calculates honestly from whatever data it's given. So the first block is taught with sensors in hand at a running machine, not from slides.

We work it out right on your own bearing housing: where the load path runs and where the accelerometer (vibration sensor) goes. We show how the mounting method affects the frequency range you can trust: a stud holds a wide band, a magnet on a clean flat spot is fine for balancing work, a hand-held probe gives scatter from one reading to the next.

- Point and direction. Radial-horizontal on the bearing housing as the working direction, vertical and axial as checks. An axial component as strong as the radial one points you away from unbalance and toward misalignment.
- Mounting and surface. Clean metal, a flat spot, the sensitive axis aligned with the direction you're measuring, the cable secured and not flapping in the airflow.
- Sensor sensitivity. The value entered in the software has to match the sensor's datasheet value, or the picture will look right while the numbers are wrong.
- Marker and tachometer. Where to stick the strip on an oily or hot shaft, what distance and angle to set the laser sensor at, how to check speed stability in two minutes before the first run.
- Machine condition. The same speed, load, damper position and temperature on every run. Drifting speed kills the phase reading and makes the influence coefficient meaningless.
- Two channels on both bearings at once. That way you see both bearings in a single run and work out sooner what kind of unbalance you're dealing with: static (a heavy spot on one side of the rotor) or couple (a pair of heavy spots at opposite ends that rocks the rotor).

> Your people do all of this hands-on, on your own equipment. We stand next to them and correct as needed. The general method is covered in a separate article of ours on how to measure vibration correctly; in the session we work with your bearing housings and your access.

## The instrument's workflow: from the initial run to reaching tolerance

The second block is taught at a laptop next to a running machine. The logic is short: the instrument measures the starting condition, then learns from a trial weight, then calculates the correction. Your person goes through every step themselves.

1. **Initial run** — Bring the machine up to running speed, take the 1x amplitude and phase at both bearings, and the overall vibration at the same time. This is also where you check the spectrum, vibration broken down by frequency: does the peak dominate at running speed, or does most of the energy sit at 2x (twice running speed) and in the harmonics.
2. **Trial weight in the first plane** — We teach how to choose the mass and radius, weigh the trial weight, and enter the actual values, not estimates. The test for a usable trial run is simple: the 1x amplitude has to change by at least 20-30%, or the phase by at least 20-30°. If the change is smaller, the weight is too small and the run needs repeating.
3. **Trial weight in the second plane** — Only for two-plane jobs. One plane means two runs, two planes means three. Every extra run on site means a shutdown, cool-down time, a guard removed and half an hour lost, so from the start we teach people to think of runs in terms of money.
4. **Calculating and fitting the correction** — The software outputs a mass and an angle for each plane. We go through how the angle is measured from the trial weight's position, and the direction of measurement relative to rotor rotation. Then the fixed-position mode indexed to blades or bolt holes, splitting the mass between two neighbouring positions, and calculating a metal removal by drilling, which flips the angle by 180°.
5. **Check run** — Fit the weights, run the check, and compare against the initial reading at each bearing separately. This is also where we teach how to read the polar plot: where the vector moved and what that tells you.
6. **Trim and tolerance** — Vibration has dropped several times over, but the target hasn't been reached. The software will suggest adding a small weight to the one already fitted, without new trial runs. One trim run is normal; four in a row is itself a diagnosis, and we go through what that's telling you.
7. **Archive and coefficients** — Save the result, fill in the report, and save the influence coefficients for this machine. On the next similar rotor, your mechanic can get away with a single run, provided the bearings, speed and geometry match.

## One plane or two, and how to attach the weight on your rotor

The number of correction planes (rotor cross-sections where weight is added or removed) is set by the rotor's shape and speed, not by habit. The guide is the ratio of working length to diameter in the zone where the weight goes. A short, disc-like rotor with an L/D below roughly 0.5 is usually handled in one plane. An elongated one needs two, or a couple unbalance will remain and the second bearing will keep shaking. The higher the speed, the more often two planes are needed.

We work through mounting on your own machines separately, because this is where theory stops. On a centrifugal wheel blade, the weight is usually welded on. On a pulley or half-coupling, a bolt with washers goes into an existing hole. On a pump impeller in an aggressive environment, metal is more often removed by drilling with depth control. On a thin-sheet impeller, welding distorts the geometry, and a different solution has to be found.

| Rotor | What the design tells you | Planes | Typical weight attachment |
| --- | --- | --- | --- |
| Belt-drive pulley, half-coupling | Short disc, L/D well below 0.5 | One | Bolt into an existing hole, occasionally drilling |
| Fan impeller, pump wheel | Disc with blades, access through an inspection hatch | One, two at high speed | Welding on a blade or disc, metal removal by drilling |
| Overhung induced-draft fan, fan on an outboard bearing | Mass sits beyond the bearings, the second bearing responds strongly | Two | Welding on the wheel's front and back discs |
| Shaft between bearings, drum, electric motor rotor | Elongated rotor, L/D above 0.5 | Two | Balancing rings, welding, drilling |
| Crusher rotor, shaft with attached elements | Hammer wear and build-up change the picture during operation | Two | Check the set and the wear first, weights only after that |

> A flexible rotor and running near a critical speed aren't something one course can cover fully. We have a separate article on critical speed and flexible rotors, and in the session we show plainly where the limits of your instrument and your own level sit.

Sources: [ISO 21940-11:2016](https://www.iso.org/standard/54074.html) · [ISO 21940-12:2016](https://www.iso.org/standard/50429.html)

## Common mistakes: where runs get wasted and results get spoiled

The same mistakes repeat from site to site. We don't just list them, we reproduce them in the session: fit a weight on the mirror-image side and show how vibration rises, take a reading off the guard and compare it with a reading off the bearing housing. What you've seen with your own eyes sticks longer than notes.

### Mirrored angle

The direction of measurement is set the wrong way, the weight ends up on the opposite side of the rotor, and vibration rises instead of falling. The sign is easy to spot: after correction, the amplitude is noticeably higher than at the start. It's cured by the fixed-position mode, where there's no angle and no direction, just a position number and a mass.

### Small trial weight

The readings barely changed, and the person carries on regardless. The influence coefficient then gets calculated from noise, and the resulting correction comes out essentially random. We teach people to stop at the 20-30% or 20-30° criterion and increase the mass.

### Reading taken on the guard

A thin sheet-metal guard or casing has its own resonances and gives impressively large numbers that have nothing to do with the load on the bearing. A reading belongs on the bearing housing, as close to the bearing itself as possible.

### Unstable speed

A slipping belt, a frequency converter with a drifting setpoint, a damper position that changed between runs. The phase wanders, and three runs end up describing three different machines.

### Changing radius without recalculating

The trial weight sat at a 400 mm radius, the correction weight went on at 250 mm, and the software was never told. Unbalance is defined by mass times radius, so the mass has to be recalculated.

### Forgotten trial weight

The software says to remove it, the person leaves it on. Or the other way round. Both mistakes give a predictably wrong check run, and we show how to spot this on the polar plot in under a minute.

## When balancing won't help, and how to spot it in half an hour

The most valuable skill is knowing when not to balance. Weights only reduce the running-speed component. If overall vibration is several times higher than 1x, your person can run trials all afternoon and never move the needle. We teach six checks that are done before the first weight is picked up.

- Compare overall vibration with 1x. The cheapest test there is, done right in the initial run in two minutes.
- Resonance in the structure, frame or ductwork. On coast-down (free deceleration after the machine is switched off), the amplitude shows a narrow peak and the 1x phase swings through roughly 180°. Balancing in that zone is pointless, the result won't hold.
- Shaft misalignment. A noticeable 2x and an axial component as strong as the radial one. This calls for shaft alignment, not weights.
- Loose fastenings and soft foot. A comb of harmonics, a raised noise floor, readings that scatter when you move to a neighbouring point on the housing.
- Rolling-bearing defects. High-frequency peaks that aren't multiples of running speed, and a rise in the envelope spectrum. Balancing will only mask them.
- A rotor that changes during operation. Build-up in a crusher, erosion of a pump wheel, a piece of hardfacing that's broken off. Vibration comes back within a shift, and that's a sign, not a coincidence.

> We build this block around your own machines and the recordings taken on the first day. The theory of causes is covered in separate articles on how to identify the cause of vibration and how unbalance differs from misalignment. In the session you work through your own spectra, not textbook ones.

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

## Tolerance, report, log and baseline level

The word "tolerance" means three different things, and mixing them up gets expensive. The software reports "in tolerance" once the residual 1x has dropped below the target value that you entered yourself. That is not a condition assessment against the applicable part of ISO 20816, and it is not confirmation of a G balance quality grade under ISO 21940-11. We teach people to keep the three criteria apart and note which one they're using each time. Check the part and edition of the standard against your own machine: small fans and belt-drive units have their own separate applicability.

Then comes the real reason it's worth keeping the instrument in-house at all. A single number tells you about today. A series of numbers buys you time: order the bearing through normal channels and stop the machine in a planned window, rather than at midnight on a Friday.

- [x] Choose the G balance quality grade (the permissible-residual-unbalance standard) to suit the rotor type and speed, rather than defaulting to G6.3 for everything out of habit.
- [x] Record the applicable part and edition of ISO 20816, the frequency band and the points, so the word "limit" is backed by an actual document.
- [x] Log the measurement conditions: speed, mode, load, temperature, damper position.
- [x] Mark the points on the machines with paint or tags and never move them again.
- [x] Take a baseline reading on a healthy machine right after balancing, so any future rise has something to be compared against.
- [x] Write the report so that a different person with a different instrument would get the same numbers at the same points: vibration before and after at each bearing, weight masses and radii, angles or fixed-position numbers.

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

## Training formats and what we need from you

We work in two formats, and both run on real equipment. We don't bring training rigs: your own rotor, with its own access, temperature and grime, teaches faster than any rig.

### At your site

We come to your site and work on your own machines. We balance some rotors together with your people, they do others themselves while we watch and correct. This format suits you when you already own an instrument, or when your fleet is uniform enough to cover in-house.

### With instrument delivery

Commissioning the Balanset-1A: installing and configuring the software on your Windows laptop, checking the measurement chain and sensor sensitivity, the first balancing job on your own machine done by your own person's hands, and going through the typical scenarios in your fleet.

### Embedded OEM version

If the measuring core is going not into a carry case but inside a machine tool or a rig, we go through working in that configuration: calculating mandrel eccentricity, recalculating weights for other planes, repeatability across a batch of identical rotors.

- Who it suits: mechanics and maintenance crews, power engineers, reliability departments, service teams who maintain their own ventilation, pumps, motors and drives.
- What we need from you: access to the machines, the ability to stop and start them under your own lockout rules, mounting spots for sensors, a Windows laptop if the instrument is yours.
- What we bring: an instrument with two accelerometers and a laser phase sensor, reflective markers, trial weights and mounting material.
- What speeds things up: photos of the machines, bearing housings and access points sent in advance, plus any vibration recordings you already have. We'll use them to build a session around your own cases specifically.

> We don't publish the programme, scope or duration as a fixed figure, because they depend on the number of machines, how accessible they are, and your people's starting level. We agree it after a conversation about your fleet and your goals.

## What we don't give you, and how to order

Let's say this plainly, so there's no ambiguity. AXILINE is not a certification body and does not issue ISO 18436-2 certificates. A vibration analyst category is awarded by an independent certification body, and only after an exam. Our training does not grant that category and does not substitute for it. We have a separate article on how Categories I-IV work and what to look for when choosing a training centre: read it before choosing between the certification route and the practical one.

Who shows up: engineers who design and manufacture Balanset instruments themselves and use them for on-site balancing themselves. Based in Vila Nova de Gaia, near Porto, we work across Portugal. We don't separate training from practice, so many clients combine it with an on-site balancing visit. We do the first machine, your person does the next one, and we stand alongside.

> We price and format training around the task; we don't have ready-made packages with fixed hours. For a rough idea of a combined visit: vibration diagnostics with a report costs 300 EUR per unit, balancing adds from 250 EUR, the minimum invoice per visit is 500 EUR, and the calculator on our site gives an exact figure for your own machines.

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

## Frequently asked questions

**Do you issue an ISO 18436-2 certificate or any other qualification credential?**

No. AXILINE is not a certification body and does not issue such certificates. A vibration analyst category is awarded by an independent certification body, and only based on exam results. We teach practical work: the measurement method, operating the instrument's software, balancing in one and two planes on your own machines. How Categories I-IV work is covered in a separate article of ours.

**How long does training take and how much does it cost?**

We don't have a ready-made programme with a fixed number of hours or a price list, and we're not going to invent one. The scope depends on how many machines you have, how accessible they are, whether your people have worked with vibration measurement before, and whether you need the OEM version covered. Tell us about your equipment fleet and the task, and we'll propose a format and work out a price.

**Do you need your own instrument to learn?**

No, we bring our own kit to the session. But the real payoff comes once an instrument stays with you: the whole point is for your mechanic to handle routine machines on their own, without calling a contractor for every case. If you don't have one, training is usually combined with delivery of a Balanset-1A and commissioning it.

**Can training be combined with an on-site balancing visit?**

Yes, that's the most common arrangement. We come out for a specific problem, balance the machine ourselves, and go through every step with your people at the same time: initial reading, trial weight, calculation, fitting the correction, check run, report. The next rotor on site, your person does themselves, under supervision.

**Who's best to send on the training?**

People who'll actually have the instrument in their hands regularly: maintenance mechanics, power engineers, reliability specialists. No specialist vibration education is required, understanding your own equipment and being willing to work with a laptop is enough. Two trainees per machine work better than a group of eight crowded around one bearing housing.

**What if it turns out our machine doesn't need balancing at all?**

That's a normal and useful outcome, and we don't hide it. Working through a case where the main vibration comes from resonance, misalignment or a bearing teaches faster than a successful balancing job does. Your people leave the site with the key skill: measure first, separate the causes, then decide whether to add weights, align the shafts, or replace the component.