# Overall vibration, the 1x running-speed component, and phase: what balancing actually fixes

> A fan starts humming after the impeller is cleaned. You put a sensor on the bearing housing, and the instrument shows 9.8 mm/s. Balance it, or look for a different cause? The answer is hiding in three numbers the instrument gives you at once.

**In short:** Overall vibration in mm/s RMS adds up every source at once. The 1x running-speed component is the part strictly at the rotation frequency, and that's exactly where imbalance shows up. Balancing only reduces 1x, so look at the ratio: if 1x accounts for more than 70-80% of the overall level, weights will work; if it's under half, something else is creating most of the vibration. The phase of 1x isn't there to locate the heavy spot — it's there to track how the machine responds to a trial weight.

Source: https://axiline.pt/en/articles/overall-vibration-1x-running-speed-component/  
Publisher: AXILINE · Vila Nova de Gaia, Portugal · +351 931 831 229 · axilinegeral@gmail.com

## Three numbers on the screen you shouldn't mix up

The instrument measures vibration continuously, but presents it to you in different forms. On the Balanset-1A you see four fields per channel: overall vibration, the running-speed component, its phase, and the rotation speed. Each field answers a different question.

### Overall vibration, mm/s RMS

Everything the machine is shaking with in the measured band, as one number (RMS — root mean square). Imbalance, shaft misalignment, bearings, looseness, resonance, flow effects, and vibration from a neighboring unit through the foundation are all mixed together here. This is the number you use to judge the machine's condition and compare against zones A/B/C/D. Keep the band the same every time, usually 10-1000 Hz.

### 1x running-speed component, mm/s RMS

The part of the vibration exactly at the rotation frequency, one cycle per revolution. The instrument extracts it from the laser tachometer pulse off the reflective mark. Imbalance sits here, and only here. In the interface these are the V1o and V2o fields, against the overall V1s and V2s.

### 1x phase, degrees

How far the 1x sine wave is shifted relative to the tachometer pulse. Fields F1 and F2. The value on its own says nothing about where the extra mass is sitting. What matters is the change in phase and the difference in phase between points.

### Rotation speed, rpm

Check this one first. If it jumps between 1450 and 1490, the mark is dirty or the tachometer is seeing two flashes per revolution. As long as the speed is unstable, the 1x amplitude and phase aren't trustworthy, and balancing is pointless.

> Without a phase sensor, the instrument only shows overall vibration. Neither 1x nor phase can be extracted — there's nothing to reference them against. So a clean reflective mark on the shaft isn't a formality, it's a precondition for the work.

## Why balancing only reduces 1x

Imbalance is a mass offset from the rotation axis. It creates a centrifugal force that rotates together with the rotor and therefore pushes on the bearing once per revolution. That's where the 1x peak comes from. A correction weight creates a second, matching force in antiphase, and the two cancel each other out.

That gives a simple limitation. A weight affects a force that repeats once per revolution. It has no effect at all on vibration with a different origin and a different frequency. Shaft misalignment loads the bearings twice per revolution, a bearing defect knocks at its own frequency, a loose bolt produces impacts and a whole comb of harmonics. You can balance a rotor perfectly and see no change in the overall vibration at all.

That's exactly why the instrument's manual puts a vibration-meter reading before balancing. That reading has one job: confirm that the vibration is made up mainly of the 1x component.

| Vibration source | What it looks like | Does balancing remove it |
| --- | --- | --- |
| Rotor imbalance | dominant 1x peak, stable phase | Yes, that's its job |
| Shaft misalignment | 1x plus a noticeable 2x, elevated axial vibration, phase shift around 180° across the coupling | No. Needs shaft alignment |
| Loose fasteners, soft foot | comb of 1x, 2x, 3x and beyond, raised noise floor, phase jumping by tens of degrees | No. Torque the fasteners, fix the support and shims |
| Bearing defects | high-frequency peaks not related to running speed, noise around the bearing's resonance | No. Plan a replacement |
| Support or frame resonance | sharp amplitude peak in a narrow speed range, 1x phase swings by about 180° during coast-down | No. Change stiffness, mass, or operating speed |
| Flow, rubbing, aerodynamics | blade-pass frequency (number of blades × rotation speed), broadband noise | No |

## A practical criterion: work out the 1x share

Take two numbers from the same point and work out the ratio. First example: overall 12.0 mm/s, running-speed 11.2 mm/s. The 1x share is about 93%, imbalance is creating almost all the vibration, balancing will get results. Second example: the same overall 12.0 mm/s, but running-speed 2.8 mm/s. The share is about 23%. Here you'll fit weights, bring 1x down to 0.5 mm/s, and end up with an overall of 11.7 mm/s. The customer will say nothing changed, and they'll be right.

Use the table below as a guide. It's a working rule, not a standard, but it saves you runs and reputation.

| 1x share of overall | What it means | What to do |
| --- | --- | --- |
| over 80% | imbalance is creating almost all of the vibration | Balance it. You'll see the effect on the verification run |
| 50-80% | imbalance is present, a second source is also at work | Balance it, but warn in advance that some vibration will remain |
| 30-50% | imbalance is secondary | Check the spectrum and mechanics first. Balancing comes as a second stage |
| under 30% | imbalance is almost beside the point | Don't waste runs. Look for the cause in the spectrum and the fasteners |

> Compare overall and 1x strictly at the same point, in the same direction, and at the same speed. An overall reading taken horizontally and a 1x reading taken vertically give a meaningless ratio. Keep the sensor horizontal-radial on the bearing housing and don't move it between readings.

## Phase: what it shows, and what it doesn't

The most common misconception goes like this: the instrument shows 47°, so the heavy spot is at 47° from the mark. That's wrong. The phase reading depends on where the sensor is mounted, which way it faces, which edge of the pulse the tachometer triggers on, how stiff the supports are, and how far you are from resonance. The same extra mass on different machines will give different degrees.

Phase works as an indicator of change. The instrument stores the 1x vector before the trial weight, stores it again after, and calculates the difference. That difference is the influence coefficient — the response of this specific rotor-supports-foundation system to a known mass at a known point. From there, the software solves the inverse problem and outputs the correction mass and angle. That angle is measured from the trial-weight location, not from the tachometer mark.

Phase has a second job: diagnostics. The phase difference between two points tells you more than any amplitude reading.

- [x] 1x phase is stable, spread within ±2-3° — the system can be balanced
- [x] Phase jumps by ±20-30° at constant speed — look for loose fasteners and play, not weights
- [x] Phase swings by about 180° during run-up or coast-down, amplitude gives a narrow peak — you're in resonance, balancing will give an unstable result
- [x] Horizontal-to-vertical phase difference at one bearing around 90° — the typical picture for imbalance
- [x] Phase shift of around 180° across the coupling in the axial direction — shaft misalignment, you need shaft alignment, not weights
- [x] After the trial weight, 1x amplitude changed by less than 20% and phase by less than 20° — the weight is too light, increase it and repeat the run

> Don't compare phase readings taken with different tachometer setups or after the mark has been moved. The reference point has shifted, and every reading has drifted by a constant amount. That's almost impossible to spot from the numbers alone.

## What to do when 1x is small but the machine is still shaking

This is a normal situation, and it doesn't mean the visit was wasted. It means balancing isn't the right tool. From here you work with the spectrum.

- Strong 2x and elevated axial vibration. Check the shaft alignment and the condition of the coupling. On a belt drive, check pulley parallelism and tension.
- A comb of harmonics and sub-harmonics, a raised noise floor. Check the bolt torque, soft foot, the condition of the shims, cracks in the frame and the foundation fixing. Sometimes tightening four bolts cuts vibration more than any weight ever could.
- High-frequency peaks not related to running speed, noise in the envelope (a signal-analysis mode used for bearing diagnostics). Look at the bearings: lubrication, clearance, signs of wear. Calculate life per ISO 281, and plan the replacement from the trend, not from a single reading.
- A narrow peak that shifts sharply with a small change in speed. That's resonance. Record the coast-down — the amplitude and phase while the machine freely runs down after the power is cut — find the natural frequency, and either move the operating speed away from it or change the stiffness of the supports.
- A peak at the blade-pass frequency, meaning the number of blades multiplied by the rotation speed. Check clearances, rubbing, the condition of the guard, and the flow operating point. Balancing won't help here.
- Vibration appeared after cleaning or a repair, and 1x dominates. That's exactly your case — balance it.

## The on-site procedure

1. **Reading in vibration-meter mode** — Fit the sensors to the bearing housings, aim the tachometer at the reflective mark. Start the machine at operating speed. Record the overall level, 1x, phase, and speed at both bearings. Keep the direction consistent, usually horizontal-radial.
2. **Work out the 1x share** — Divide 1x by the overall level at each point. Get above 80% at both bearings, move on. Get 20-30%, stop and switch to the spectrum.
3. **Look at the spectrum** — Look for what dominates: 1x, 2x, a comb of harmonics, or high frequencies. At the same time, check that the rotation frequency on the spectrum matches the tachometer reading.
4. **Check the mechanics** — Bolt torque, soft foot, play, the condition of the supports and foundation, belt tension. Five minutes of work that rules out half of all pointless balancing jobs.
5. **Rule out resonance** — Record the coast-down and check for a sharp amplitude peak with a phase swing near the operating speed. Don't balance in resonance — the result won't hold.
6. **Trial weight** — Fit a weighed trial mass at a known radius, enter the actual mass and radius. After the run, check that the 1x amplitude changed by 20-30% or the phase by 20-30°. Less than that, increase the weight — don't try to calculate off the noise.
7. **Correction and verification run** — Fit the calculated weights, measuring the angle from the trial-weight location. Add a trim weight if needed — a small top-up without recalibrating. Then check both values again: 1x shows the quality of the balancing, the overall level shows the condition of the machine.

> A typical result for a fan with pure imbalance: around 12 mm/s before the work and around 1.5-2 mm/s after. This is an illustrative guide for a machine where 1x dominates and the mechanics are sound, not a promise for your specific unit.

## Three different “tolerances” that are easy to mix up

When the software says “in tolerance,” it's comparing the residual 1x with the target value you entered yourself. That's not an assessment of the machine's condition, and it's not confirmation of a balance quality grade. The three criteria live separately.

| What you're assessing | By which value | Where to look |
| --- | --- | --- |
| The balancing result in the software | residual 1x and phase | comparison against the target value you entered, in mm/s |
| The overall condition of the machine | overall vibration, mm/s RMS over the 10-1000 Hz band | ISO 20816 criteria for measurements on non-rotating parts, zones A/B/C/D |
| The quality of the rotor balancing | residual imbalance, g·mm/kg | balance quality grades G under ISO 21940-11 (formerly ISO 1940-1) |

> Treat the zone and grade numbers as a working guide. Check the applicable part and edition of the standard for the specific machine: it has limits by power, speed, and support type, and exceptions by machine type. For contractual acceptance, record the part of the standard, the measurement points, the frequency band, the operating mode, and the foundation type.

Sources: [ISO 20816-1:2016](https://www.iso.org/standard/63180.html) · [ISO 21940-11:2016](https://www.iso.org/standard/54074.html) · [ISO 21940-12:2016](https://www.iso.org/standard/50429.html) · [ISO 281:2007](https://www.iso.org/standard/38102.html) · [ISO 13373-3:2015](https://www.iso.org/standard/40840.html) · [ISO 13373-5:2020](https://www.iso.org/standard/62202.html) · [Balanset-1A operation manual](https://vibromera.eu/balanset-1a-operation-manual/)

## If you need the instrument and the hands to go with it

Everything described above, you can do with one two-channel kit: two accelerometers, a laser phase sensor, a USB module, and Windows software. The Balanset-1A shows overall and running-speed vibration with phase on both channels at once, builds the spectrum, calculates influence coefficients, splits the weight across fixed positions, and stores the results in an archive for the report. Balancing is done in the machine's own bearings, with no disassembly and no taking the rotor away.

AXILINE is engineers who design and manufacture the Balanset instruments and use them ourselves on site visits. We come with the kit, take the overall and 1x readings, show you the ratio, and tell you honestly what can be solved with weights and what needs shaft alignment, retorquing, or a bearing replacement. If you already have the instrument and just need help interpreting the readings, our engineers provide consulting support.

Tell us what the machine is, what speed it runs at, and what the instrument shows. Those three numbers are the fastest way to start the conversation.

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

**How is overall vibration different from the 1x running-speed component?**

Overall is the root-mean-square value of all the vibration in the measured band, usually 10-1000 Hz. Every source is added together there. The 1x running-speed component is only the part that falls exactly at the rotation frequency. The instrument extracts it from the tachometer pulse. Imbalance shows up in 1x, so balancing changes 1x and barely touches the rest.

**Can 1x ever be higher than the overall vibration?**

Physically, no: 1x is part of the overall reading, so it's always lower than or roughly equal to it. If the instrument shows the opposite, look for a measurement error. Most often the speed is fluctuating, the tachometer is picking up an extra flash, the mark is dirty, or the signal has gone into overload. Sort this out before balancing.

**Does phase show where the rotor's heavy spot is?**

No. The phase reading depends on where and which way the sensor is mounted, on which edge of the pulse the tachometer triggers, on the stiffness of the supports, and on proximity to resonance. Phase's role is different: by comparing the 1x vector before and after the trial weight, the software works out how the system responds to a known mass, and calculates the correction mass and angle from that. The angle is measured from the trial-weight location.

**What 1x share is enough to justify balancing?**

As a working rule: above 80% of the overall level is a clear case of imbalance, 50-80% is a mixed case with a partial effect, and below 30% balancing will achieve almost nothing. Calculate the ratio at one point and in one direction, or the number is meaningless.

**Balancing reduced 1x, but the overall vibration stayed the same. What's wrong?**

Nothing went wrong — imbalance simply wasn't the main cause. Most of the vibration comes from another source. Look at the spectrum: strong 2x and axial vibration point to shaft misalignment, a comb of harmonics with a jumping phase points to loose fasteners, high-frequency peaks not related to running speed point to bearings, and a narrow peak with a phase swing points to resonance.

**Can 1x and phase be measured without a tachometer sensor?**

No. Without a reference pulse, the instrument only shows overall vibration, because there's nothing to extract the running-speed component or measure phase against. Balancing needs a laser phase sensor and a clean reflective mark on the shaft giving one stable pulse per revolution.