# On-site balancing of screw conveyors, conveying screws, and screw shafts, where they operate

> A screw has started shaking the trough, heating up its end bearing, or humming loudly enough to hear across the shop. Before welding a counterweight onto the tube or ordering a new screw, it's worth finding out whether this is even imbalance. A screw has its own reasons to make that hard to tell: it's long, it bends under its own weight, it hangs on intermediate bearings, and it turns slowly enough that the standard measurement band simply doesn't see its running frequency — the vibration at the shaft's own rotation frequency. We come out with a two-channel analyzer, take the numbers on site, and tell you where your money will actually get results.

**In short:** Yes, we balance screws and screw shafts on site, in their own bearing supports, without dismantling them. But the honest answer here sounds different from the one about a fan. Balancing works on a short, rigid screw, on a feed screw, and on a screw shaft with access at both ends — when the geometry is sound, the flighting is intact, the trough is clean, and the rotor reaches a stable speed. A long, multi-metre screw on hanger bearings is a different problem. It behaves as a flexible multi-bearing rotor: it bends noticeably at operating speed and rests on more than two supports. Its correction planes — the places where a correction weight can physically be mounted — are accessible only at the ends, while the vibration is more often caused by worn flighting, caked-on product, a bent shaft, or worn-out bushings. Weights don't fix that, and we'll say so before the trial runs, not after three of them.

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

## Symptoms: what brings us out to a screw

Vibration on a screw is almost always mixed. Imbalance, the flighting rubbing against the trough, a worn hanger-bearing bushing, and pulsation from the product all sit in the same signal. So our first step isn't a trial weight — it's a measurement at idle.

Call us if you recognise your machine in the list below. A measurement costs less than a shift of downtime, and the repair decision that follows is made on numbers, not on sound.

- [x] The trough hums and shakes at idle; the empty screw makes about as much noise as the loaded one.
- [x] Vibration appeared after cleaning, after replacing a flighting section, after build-up welding on the flighting, or after a shaft repair.
- [x] The end bearing support runs hot, grease is being forced out, and you're replacing the bearing noticeably more often than before.
- [x] There's a rhythmic metallic sound tied to rotation: the flighting is touching the trough.
- [x] The hanger bearings inside the trough are knocking, the bushings are worn unevenly, and the shaft visibly sags in them.
- [x] The gearmotor pulls unevenly, motor current pulsates, and the drive chain or belt whips.
- [x] The screw has gotten worse at conveying at the same speed, throughput has dropped, and product backs up at the inlet.
- [x] You need a measurement to settle an argument: balance it, straighten the shaft, or replace the whole screw.

> Balancing only reduces the part of the vibration that falls at the rotation frequency. If a support shows 9 mm/s of overall vibration (the total level across all frequencies at once) and 1.5 mm/s on the running-speed component, weights will remove only a small share — the rest has to be tracked down in the mechanics. That's the first thing we work out on site.

## How a screw is built, and where imbalance is born here

A screw is an elongated rotor with a very high length-to-diameter ratio. On conveying screws it easily runs into the tens. Everything else follows from that: bowing, sensitivity to geometry, intermediate supports, and a very limited choice of places to put a weight.

### Flighting on a pipe

The most common design. The flighting is wound and welded onto a pipe. Imbalance comes from uneven flight thickness, repair inserts, build-up welding on a worn edge, and water or product residue left inside the pipe after washdown.

### Flighting on a shaft, and shaftless screws

On a screw built on a solid shaft, the mass is distributed differently, and it's stiffer for the same length. A shaftless flighting with no central shaft behaves almost like a spring: balancing doesn't apply to it, and everything there comes down to geometry and wear.

### End bearing supports

The drive-end and tail-end units. These are the only places we can physically mount vibration sensors, and usually the only places where correction planes are accessible.

### Intermediate hanger bearings

They hang inside the trough every few metres and hold the shaft's alignment. A worn bushing lets the shaft drop, changes the clearance to the trough, and produces a once-per-revolution force that looks almost like imbalance on a measurement.

### The gearmotor and the drive

Screw speed is low, usually from a few tens up to a couple hundred rpm. There's a gearbox between the motor and the shaft, and often a chain or belt as well. Each of these units contributes its own frequency, and it can't be confused with the running-speed one.

### Trough, covers, and frame

The screw's housing takes part in the vibration too. Loose trough fasteners, a soft foot under the gearmotor (a mounting foot that doesn't sit flush against the frame), and a flexible support structure can raise the level enough to make it look like the rotor is to blame.

> The parent section on shafts, drums, and rollers covers the general logic of elongated rotors. Here, we only talk about what's specific to a screw.

## Why a screw often stops being a rigid rotor

A rigid rotor holds its shape at every operating speed, and correction in one or two planes works for it at any speed. A screw six, ten, or fifteen metres long doesn't have that property. It bows under its own weight even at rest, and its natural bending frequency drops fast as length increases — roughly in proportion to the square of the span. Intermediate supports raise that frequency. But the moment a bushing wears out, the span effectively doubles, and the natural bending frequency drops into the operating speed range.

In practice, this means the following. The deflection shape depends on speed, so weights calculated at one speed can produce more vibration than before the work at another speed. There's no calculation error involved — it's just that a rotor like this behaves differently at different speeds. If your screw has a variable-frequency drive and the regime changes with line loading, that needs to be accounted for before the trial runs, not discovered afterward as a surprise.

The mechanism of critical speed — the rotation rate at which the rotor hits resonance and bows the most — and the signs of a flexible rotor are covered in a separate article. In short: a narrow amplitude peak as speed changes, a phase reversal (the angular position of the vibration relative to the mark on the shaft) of roughly 180° around the peak, and poor repeatability from run to run. On site, we check this with a run-up or a coast-down — a measurement taken as the machine comes to a free stop — as well as at several inverter setpoints.

> If a screw's operating speed sits close to the first critical speed, the result of on-site balancing will be unstable. The honest conclusion in that case is: restore the shaft alignment and the supports, change the operating regime, or reconsider the spacing of the intermediate bearings. A weight here buys you weeks, not service life.

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

## Hanger bearings turn this into a multi-bearing problem

As long as a rotor has two bearing supports, the routine is well established: two sensors, two planes, three runs. Add one intermediate support, and the problem changes fundamentally. A weight in any plane shows up on every support at once, so the planes can't be worked out one at a time: the vibration would keep shifting from support to support, and the process wouldn't converge. Three supports produce a system of nine influence coefficients (each one showing how a weight in one plane changes the vibration at one of the supports) and at least four runs; four supports already mean sixteen coefficients. A detailed breakdown is in our article on balancing multi-bearing rotors.

On a screw, two physical restrictions get added on top of this math. The intermediate supports sit inside a closed trough, so a vibration sensor can't be fitted on them. Correction planes near them aren't accessible either, and a weight can't be placed inside the trough: it would come loose and end up in the product.

- Only the end supports can be measured, so the system is inherently solved incompletely.
- Correction planes are accessible only at the ends of the rotor, so there's nothing to correct the middle with.
- A trial weight inside the trough isn't permitted, for both safety and hygiene reasons.
- Every run requires a shutdown, opening the cover, and locking out the feed again, so the number of runs has to stay low.
- A worn bushing changes the influence coefficients between runs, and repeatability disappears.

> The conclusion is simple, and not always welcome. On a long screw with intermediate hanger bearings, on-site balancing usually isn't the right tool for the job. We come out, measure, and show you exactly what's producing the vibration, and from there you restore the bushings, the shaft alignment, and the clearance to the trough.

## Causes of screw vibration that weights won't fix

This is the most useful section on this page. Based on our measurement experience with elongated rotors, everything listed below comes up more often than genuine imbalance, and we check every item before fitting a single weight.

| Cause | How it shows up in the measurement and on the machine | What to do about it |
| --- | --- | --- |
| Worn or torn flighting | The level is unstable, throughput has dropped, product backs up at the damaged section | Build-up welding on the edge or replacing the section. Balancing after the geometry is restored, if it's even needed |
| Product build-up or freeze-on | Readings drift from run to run, the level is different after washdown, the imbalance comes back within a shift | Cleaning, thawing, eliminating the cause of the build-up. Balancing a dirty rotor is pointless |
| A bent shaft or pipe | Runout at the journals on slow rotation, rubbing at one point along the length, knocking | Straightening, machining, or replacement. A weight compensates for mass, not shape |
| Worn hanger bearings and bushings | The shaft has sagged, the clearance to the trough is uneven, the spectrum (the breakdown of vibration by frequency) shows a raised noise floor and harmonics of the running frequency | Replace the bushings and restore the shaft alignment before any correction calculations |
| Flighting rubbing against the trough | A rhythmic metallic sound, rub marks on the flighting and the housing, many harmonics in the spectrum | Restore the clearance, check the bow and the bearing fit. Balancing isn't performed on a rubbing rotor |
| Skewed end supports and drive shaft misalignment | A pronounced second harmonic (a component at twice the rotation frequency), elevated axial level, fasteners repeatedly working loose | Soft foot and shaft alignment, then a repeat measurement. Shaft misalignment is the cause, shaft alignment is the fix |
| Trough clogging and uneven loading | Jolts, spikes in motor current, vibration depends on the feed rate and disappears at idle | Loading and discharge regime. Balancing has to be done on an empty screw, or the measurement means nothing |
| Wear in the gearmotor, chain, or belt drive | Peaks at the gearbox's gear-mesh frequencies, components at multiples of the belt's rotation, pulley or sprocket runout | Repair the drive, adjust tension, fix pulley runout. These frequencies aren't removed with weights on the rotor |
| Defects in the end-support bearings | High-frequency peaks not related to rotation speed as multiples, a rising crest factor (the proportion of sharp peaks in the signal), heating | Replacement. Masking a worn-out bearing with balancing means continuing to lose service life |

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

## When balancing a screw is actually the right call

There are configurations where on-site balancing gives a solid, stable result. They all have one thing in common: the rotor is rigid at operating speed, it has two supports, and there's somewhere to put a weight.

- [x] A short, rigid screw with no intermediate supports, with two end bearing units.
- [x] A feed screw or metering screw: short, often at higher speed, with tight requirements for smooth running.
- [x] A screw shaft in a machine where both ends are accessible: an extruder, a mixer, a pelletiser, a dewatering screw.
- [x] A screw after repair or build-up welding on the flighting, once the geometry has already been restored and only the mass distribution remains in question.
- [x] A short vertical screw, where we take two mutually perpendicular radial directions and work with the worse one.
- [x] A screw that reliably reaches operating speed at idle and gives repeatable readings.

> How many correction planes are needed is determined by the rotor's geometry. Anything longer than half its diameter needs two planes, or a couple unbalance will remain: the level at the near support will drop while the far one rises. A screw's length-to-diameter ratio is always greater than that, so single-plane correction is a rare exception here, not the rule.

## How we work on site

The sequence is fixed. It exists so the measurement actually means something, and so a mechanical cause doesn't get missed.

1. **Preparing the machine and agreeing what's allowed** — The screw needs to be emptied, cleaned, and washed down if necessary. The feed is locked out for the whole time we're inspecting and running it, and we open and close covers and guards following your isolation procedure. This is also where we settle the key question: is the area classified for combustible dust, is hot work permitted, and can anything even be welded onto this rotor at all.
2. **Checking the geometry before the first weight** — We take runout readings at the journals with a dial indicator on slow rotation. Along the length, we check the clearance between the flighting and the trough and look for rub marks. We check the condition of the flighting, the weld seams, the section joints, the bushings of any accessible hanger bearings, the fasteners on the end supports, and the soft foot under the gearmotor.
3. **Fitting sensors and a mark, extending the range** — Two accelerometers on the end bearing supports, right next to the bearings, magnetically mounted on a cleaned spot or on a stud. Direction is radial, usually horizontal, and it doesn't change after that. We attach a reflective mark to an exposed section of the shaft, to the half-coupling, or to the gearbox output shaft, and aim the laser phase sensor at it. We extend the lower end of the measurement range in advance, before the first run.
4. **Measurement at idle** — A run at operating speed, with no product, without exception. We record the overall level, the amplitude and phase of the running-speed component, speed, the spectrum, and the time waveform on both channels. We calculate the running-speed component's share of the overall level. If something else dominates, we stop here and move on to diagnostics instead of weights.
5. **Trial weight and choosing the planes** — We mechanically mount a mass of known size at a known radius. A screw's correction planes are the end flanges, washers and discs at the supports, the pipe end faces, and the drive half-coupling. We consider the response valid at 20–30% or more in amplitude, or 20–30° in phase. We don't use modelling clay or magnets as a trial weight on a screw: the mass has to hold just as securely as the permanent one will.
6. **Correction and verification measurement** — We fit the calculated masses, secure them reliably, and repeat the measurement at the same points, in the same direction, at the same regime, and in the same band. We hand over the before-and-after numbers side by side, and separately list what balancing didn't resolve.

> Let's be honest about low speeds. At 40 rpm, the rotation frequency is around 0.7 Hz, and the usual 10–1000 Hz assessment band simply cuts it off. The instrument picks out the running-speed component using the tachometer mark, so the balancing itself can be done. But assessing the machine's overall condition by the zones in that band at speeds like this isn't valid, and we say so up front. On a slow-running screw, we look not just at vibration velocity but at displacement amplitude at the supports too, plus the time waveform: impacts and rubbing show up there better than in the spectrum.

Sources: [Balanset-1A operation manual](https://vibromera.eu/balanset-1a-operation-manual/) · [ISO 20816-1:2016](https://www.iso.org/standard/63180.html)

## Limits: stainless steel, food production, explosion-hazard zones

Screws are installed in places where intervention is taken seriously. Here, it's not physics that determines the correction method, but the plant's regulations, and that needs to be agreed before the visit.

- Food and pharmaceutical screws made of stainless steel. Welding is usually out of the question: it compromises the hygienic finish of the surface, leaves a zone of heat-tint discoloration, and requires requalifying the weld. We fit mass with standard fasteners, or remove metal by drilling, and choose weight material from the same steel group as the rotor.
- Explosion-hazard zones at grain elevators, flour mills, groat mills, and feed plants. Grain and flour dust form a combustible mixture, and hot work is done only under a permit, with full cleaning and purging. Our instrument and laptop aren't intrinsically safe, so we work with the machine stopped, cleaned, and ventilated, and if necessary we keep the laptop outside the zone boundary. We agree the procedure with your industrial safety department in advance.
- A closed trough. There's no access to the middle of the screw, either with a sensor or with a weight. Everything we can measure and correct is at the ends, and that's a fundamental limitation, not an organisational one.
- Sanitary requirements. If opening the trough triggers a full washdown and re-acceptance of the line, the cost of access can exceed the cost of the balancing itself. That needs to be worked out before the visit.
- A drive with no variable-frequency inverter. There's only one regime, no way to check behaviour at several speeds, and the signs of a flexible rotor have to be looked for on coast-down instead.
- Ruined geometry. A bent shaft, torn flighting, worn bushings, rubbing against the trough. In cases like these, we don't perform balancing, and we say so right away.

> Declining to balance is also a result of the visit. You get a measurement, spectra, a condition assessment, and a clear list of what to repair, instead of a shift spent fitting weights that won't hold.

## What you get, and how to book a visit

The main result is numbers you can show to the maintenance mechanic and to procurement. We hand over before-and-after measurements at the same points, at the same regime, in the same frequency band, because without that, comparison doesn't mean anything.

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 site gives an exact figure: it accounts for the number of rotors, travel, and the scope of work. Diagnostics without balancing is worthwhile too: you get an answer on exactly what to repair, before you even order parts. We're the engineers who design and manufacture Balanset instruments, and we do the on-site balancing ourselves. Base in Vila Nova de Gaia near Porto, service visits across Portugal.

- [x] A report with overall vibration and the running-speed component for each end support, before and after the work.
- [x] The correction masses fitted, radii and angles, or fixed position numbers.
- [x] Spectrum and time waveform, an assessment of the bearings, fasteners, drive, and any signs of rubbing.
- [x] A geometry assessment: shaft runout, clearance to the trough, condition of the flighting and the bushings.
- [x] An assessment against the applicable part of ISO 20816 as a reference, with a caveat on speed and the measurement band, and, if needed, a residual imbalance calculation against the balance quality grade.
- [x] Saved influence coefficients for this machine, so the next trim balance goes through without trial runs.
- [x] Recommendations: what to do before the next shutdown, and what to watch going forward.

| What to send before the visit | Why we need it |
| --- | --- |
| Length, diameter, flight pitch, number of sections, and operating speed | To assess the length-to-diameter ratio and work out in advance whether this rotor is rigid or flexible |
| Number and type of supports, whether there are intermediate hanger bearings | To determine whether this is a two-support or a multi-bearing job, and whether it's worth coming out at all |
| What's being conveyed, and whether the screw can be run empty | To find out whether the measurement will be repeatable and whether the rotor will need washing |
| Photos of the end units, flanges, the half-coupling, and possible weight locations | To choose the correction planes and the weight-mounting method in advance |
| Constraints: hygiene, a welding ban, an explosion-hazard zone, the shutdown window | To agree what operations are allowed before we arrive, not on site |
| History: clogging, shaft straightening, flighting build-up welds, bushing and support replacements | To narrow down the list of causes before the first run |

> Preparation cuts the visit time almost in half: an emptied and cleaned screw, the feed locked out, access to both end supports, cleaned mounting spots for the sensors, the ability to reach operating speed and stop as many times as needed, and someone on site who's authorised to approve runs.

Sources: [ISO 20816-1:2016](https://www.iso.org/standard/63180.html) · [Balanset-1A manufacturer specification](https://vibromera.eu/product/balanset-1/)

## Frequently asked questions

**Can a screw be balanced without taking it out of the trough?**

A short screw with two end supports — often yes. We fit sensors on the bearing units, a mark for the phase sensor on the half-coupling or an exposed section of the shaft, and mass on the end flanges, washers at the supports, or the pipe end face. A long screw with intermediate hanger bearings is a different case: its middle can't be reached with either a sensor or a weight, and it's more honest there to restore the shaft alignment and the clearance to the trough.

**The screw only turns at 40 rpm. Is it even worth balancing?**

It's worth it, but the approach is different. At speeds like this, the rotation frequency is around 0.7 Hz, and the usual 10–1000 Hz assessment band cuts it off. The instrument picks out the running-speed component using the tachometer mark, so the correction itself can be done, but assessing the machine's overall condition by the zones in that band no longer works. We extend the lower end of the range and look at displacement amplitude at the supports and the time waveform. And honestly: on a slow-running screw, vibration is more often caused not by imbalance but by rubbing, build-up, or a worn bushing.

**The screw's shaft is bent and the flighting is torn. Will balancing help?**

No, and it's better to understand that before running trials. A weight compensates for mass that's off-centre. It won't straighten the shaft, restore torn flighting, or remove rubbing against the trough. The right sequence is: straighten or replace the section, restore the clearance and the bushings, then take a repeat measurement. If a running-speed component remains after that, we'll balance against the restored geometry.

**Why can't a weight be placed in the middle of a long screw?**

For three reasons at once. The middle is closed off by the trough, and there's no access to it. A weight inside the trough isn't allowed: it would come loose and end up in the product. And even if there were access, a flexible rotor's deflection shape depends on speed, so a mass calculated at one speed could make things worse at another. So we work only at the ends, and we tell you up front what can't be achieved that way.

**It's a stainless-steel food-grade screw, and welding isn't allowed. What do you do then?**

We work without any hot work. We fit mass with standard fasteners on the flange or the half-coupling, or remove metal by drilling on a thickened section, and choose weight material from the same steel group to avoid galvanic corrosion and preserve the hygienic finish. If there are no standard weight locations on the rotor at all, on-site balancing isn't performed, and we say so before the visit, not after three runs.

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

A visit for one machine usually takes a full shift. The measurements themselves are quick; what eats the time is shutdowns, opening covers, and preparation. 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 the site gives an exact figure accounting for the number of rotors, travel, and the scope of work. If the measurement shows that repair is needed rather than balancing, you'll get diagnostics and recommendations, which is cheaper than balancing a machine that doesn't actually need a weight.