AXILINESavings from balancing without removal: how to calculate the full cost and avoid overpaying in downtime
A fan is shaking, and the chief mechanical engineer gets two numbers: what a site visit from an engineer with an instrument costs, and what balancing the removed wheel on a machine costs. The second number is almost always smaller, and the decision looks obvious. But the two numbers aren't comparable: one covers the entire job, the other only its middle section. Below is a calculation framework you fill in with your own figures, and a breakdown of the line items that usually decide the outcome.
Compare two routes, not two invoices
The mistake is nearly always the same one. One column gets the cost of a site visit, the other gets the cost of running the machine, and the two get compared directly. But the second column only describes three hours of an operator's time with the rotor already removed - not the machine's whole path from stopping to running normally again. Everything before and after that machine work still gets paid for by the plant, just not on that particular invoice.
So use one boundary for both options: from the moment you hit 'stop' to the moment the machine is producing again and the vibration has been confirmed by measurement. Inside that boundary, count everything: people, equipment, transport, consumables, calendar time, and risk. Only then do the numbers become comparable.
A second point: the removal option is priced in calendar days, while the on-site option is priced in shift hours. That's not the same thing, even at an identical sum of direct costs. A day of downtime is rarely compressible: a queue for the machine, a transport schedule, crane availability, waiting on consumables. An hour of work on site, you plan yourself, together with production.
A useful habit: don't ask 'how much does this cost' - ask 'how much does this cost us for every hour the machine is down.' The exact same job, on a unit with a hot standby versus on your only line, costs the plant completely different amounts of money.
The removal option: the full list of line items
Break the route down into stages and note, under each one, who pays and what it's measured in. The list below describes a typical sequence for a mid-sized impeller or rotor. Some lines will come out at zero for you, and you'll add others: what matters is going through all of them deliberately, rather than forgetting half of them.
Two line items deserve a special mention, because they're the ones most often skipped. First: shaft alignment after reassembly. A rotor certified to balance quality grade G (a residual-unbalance tolerance under ISO 21940) goes back in, the half-couplings get bolted together, and the machine goes right back to high vibration - because the shaft misalignment has come back. That's a separate job, with its own people and its own time. Second: consumables that don't survive disassembly. Seals, gaskets, fasteners, sometimes bearings you hadn't actually planned on replacing.
| Route stage | What you're paying for | What usually gets left out |
|---|---|---|
| Approval and preparation | work request, work permit, a laydown area for disassembly, lifting equipment | calendar time waiting on a crane and on people |
| Shutdown, coast-down, cooling | hours the machine is already out of production | coast-down time for a heavy wheel, and cooling time for a hot machine |
| Disassembly | fitters, hours, tools, pullers | fit surfaces and keys damaged during removal |
| Rigging and loading | a rigger, a crane or a manipulator, lifting gear | impact damage to blades and fit surfaces during slinging |
| Transport there and back | a round trip, securing the load in the vehicle | the risk of damage in transit, and of a repeat trip |
| Queue time and machine work | queueing, the operator's time, an arbor, weights | an arbor with its own runout, if the rotor has no trunnions of its own |
| Reassembly | fitters, hours, fit interference, fasteners | consumables that had to be replaced |
| Shaft alignment after reassembly | a separate job with its own time | almost always missing from the first estimate |
| Start-up and verification | a vibration measurement on the bearing housings | a repeat disassembly if the result isn't good enough |
| Lost production | the entire calendar duration, not just the hours of actual work | the single biggest line item on continuous processes |
One line deserves its own conversation: the risk of damage to the rotor in transit. It isn't hypothetical. A bent blade, a marred fit surface, a lost key, or a crushed seal turns your balancing savings into a repair job. Put your own estimate of the probability and the cost of the consequence against that line, even if you don't like the number.
The on-site option: the full list of line items
Here the route is shorter, and that's its main economic property. The machine stays assembled. The rotor never leaves its own bearing housings, the half-couplings stay bolted together, shaft alignment is undisturbed, and no crane is needed. The instrument kit arrives in a case roughly 39 × 33 × 13 cm and under 5 kg, and it needs no lifting gear of its own.
In exchange, its own line items appear, and those need to be counted honestly too. The work proceeds in runs: an initial measurement, one or two trial runs per correction plane, the correction itself, a verification run, and sometimes one or two fine-tuning runs. Every run means a shutdown, a coast-down, access to the correction plane, work with the weight, refitting the guard, authorization, run-up, stabilization, and a measurement. We break down how long the runs take in a separate article on how much time balancing takes.
- Preparation before the visit: open access to the correction plane, prepared pads for the sensors on the bearing housings, a reflective marker on the shaft, a clean impeller.
- The engineer's time with the instrument on site: usually hours within one shift, less often a full shift.
- Shutdowns and runs: 3-6 cycles, each with its own coast-down and run-up. This is your main time cost.
- Correction weights and fitting work: plates, bolts, washers, drilling, and, if needed, welding with a welder and a hot-work permit.
- The verification measurement and a report with before-and-after figures at the same points.
- Your own people on site: a fitter, an operator, the person responsible for the runs.
A zero on the 'transport,' 'rigging,' 'consumables after disassembly,' and 'shaft alignment' lines doesn't mean 'free' - it means 'structurally absent.' The savings come from exactly those missing lines, not from a lower price per labour-hour.
Sources: Balanset-1A manufacturer specification · Balanset-1A operation manual
The calculation framework: fill in your own numbers
We deliberately don't put figures into this table. Your hourly downtime rate, rigging rates, the cost of a transport run, and a fitting crew's shift rate are all specific to you, and any number we supplied would be fiction. Take the table, put two figures against every line, and add up the columns.
Work out two values per line: an optimistic one and a pessimistic one. The difference between the totals under the pessimistic scenario is usually your answer. If the optimistic scenario has both options coming out close, while the pessimistic one has them diverging several-fold, what you've actually found isn't a price - it's a risk, and the decision gets made on that basis instead.
| Line item | Unit | With removal | On-site |
|---|---|---|---|
| Hourly downtime rate for the unit or line | per hour | ___ | ___ |
| Time the machine is out of production | hours | ___ | ___ |
| Downtime losses: line 1 × line 2 | total | ___ | ___ |
| Disassembly and reassembly: people × hours × rate | total | ___ | 0 |
| Lifting equipment, rigging, gear | total | ___ | 0 |
| Transporting the rotor there and back | total | ___ | 0 |
| Machine work, or the engineer's work on site | total | ___ | ___ |
| Consumables: seals, gaskets, fasteners, bearings | total | ___ | ___ |
| Correction weights, drilling, welding | total | ___ | ___ |
| Shaft alignment after reassembly | total | ___ | 0 |
| Verification measurement and report | total | ___ | ___ |
| Risk: probability × cost of the consequence | total | ___ | ___ |
| Total: full cost of the route | total | ___ | ___ |
Fill in the risk line separately for each option. For removal, that's damage during extraction and transport, plus a repeat disassembly. For on-site work, it's the probability that the cause turns out not to be imbalance and a repair is needed instead. The difference is that with on-site work, you find that out within one shift, and before you've paid for any rigging.
How to estimate your hourly downtime cost if nobody's calculated it before
This is the one line without which the whole calculation falls apart, and it's almost always the one you don't already have on hand. It's built from simple pieces: the volume you didn't produce, multiplied by margin, not by the product's sale price. Add the people who are being paid during that time without producing anything. Add penalties and deadline obligations, if you have any. Add the cost of restarting after a shutdown, if your process needs it: warm-up, off-spec output while settling back into the process, extra energy use.
Next, work out which of three situations you're actually in. That determines whether downtime decides everything, or almost nothing.
A hot standby is available
A second pump, a second induced-draft fan, a second fan. The hourly downtime rate is close to zero, and downtime stops being the main question. In that case, the outcome comes down to rigging, transport, disassembly labour, and the calendar time you spend without a standby and without room for a second failure.
No standby, the line stops
The hourly downtime rate becomes the single biggest line in the calculation, and it usually outweighs everything else put together. Here on-site balancing wins almost every time, because its route is measured in shift hours, not calendar days.
Partial downtime
The machine isn't stopped, but it's run at reduced throughput, with a partly closed damper, with a speed limit. That's still money, just spread out. Calculate the shortfall in daily output and multiply it by however long you plan to run like that.
Don't forget to price the cost of doing nothing, too. Uncorrected imbalance is extra dynamic load on the bearings, and bearing life calculations under ISO 281 are built directly on that applied load, and are very sensitive to any increase in it. Imbalance put off 'until the next shutdown' is usually paid for in bearings, and in an unplanned stoppage at an inconvenient moment.
Sources: ISO 281:2007
The line items that usually decide the outcome
In a filled-in table, it's almost never the case that every line tips the balance a little. Two or three items decide the outcome, and the rest is just noise. Here they are.
Calendar time and whether a standby exists
Multiply your hourly rate by the duration of each route and compare it against the sum of all direct costs. If the downtime losses exceed the direct costs, you can stop calculating there: time decides the outcome, not the price of the work.
Rigging, size, and crane access
A rotor that two fitters can pull in an hour, and a rotor that needs half the housing dismantled and a crane brought in through an opening, cost wildly different amounts of money. This is the line that most often makes removal several times more expensive - not the machine work itself.
Shaft alignment and consumables after reassembly
The line item estimates lose almost every time. Reassembly requires shaft alignment, or the vibration comes back and the rotor gets blamed for it. Plus seals, gaskets, and fasteners you hadn't planned on buying.
The chance the cause isn't imbalance at all
If overall vibration is several times larger than the 1x running-speed component (vibration at the rotation frequency - the only part balancing actually removes), the problem isn't balancing at all. Being wrong about this costs differently depending on the route: on site you lose part of a shift; with removal, you've already paid for disassembly, transport, and machine time before finding out the rotor was fine all along.
The practical takeaway: the calculation almost always comes down to comparing 'downtime hours × your rate' against the sum of 'rigging + transport + disassembly and reassembly + shaft alignment.' Everything else only moves the second decimal place.
When removal is still cheaper
The economic logic shouldn't turn into a belief that a rotor should never come out. There are cases where on-site work is either impossible, or only delivers a result for one operating mode and not for long. In those cases, removal is cheaper even if the table makes it look more expensive: what you should be comparing it against is the cost of a second attempt.
- The correction plane is inaccessible. The rotor is enclosed in a housing with no removable hatches, potted, or sealed up, or the only accessible surface won't hold a weight. The correction can be calculated but not fitted.
- The rotor is flexible, or runs above its first critical speed - the RPM at which the shaft starts to bend noticeably. Calculating from influence coefficients relies on a linear rigid-rotor model. A flexible rotor needs a different methodology, and ISO 21940-12 treats rotors like this separately from rigid ones.
- Acceptance requires a report against a balance quality grade G. A requirement on the rotor in g·mm/kg is only satisfied on a certified balancing machine under ISO 21940-11. A vibration measurement on the bearings, in mm/s, doesn't substitute for that report, however good the result looks.
- The geometry is compromised: shaft runout, a crack, blade erosion, chipped-out or built-up material. Balancing adds mass - it doesn't restore shape. Inspection and repair have to come first.
- The machine is being disassembled anyway. Bearings are being replaced, or a major overhaul is under way, and the rotor is already out. In that case, machine balancing comes to you almost for free, because the expensive line items on the route are already being paid for by the repair.
- You have a fleet of identical rotors and your own tooling. Here, a balancing machine buys you repeatability, and saved influence coefficients (the machine's remembered response to a trial weight) turn every subsequent rotor into two runs instead of three.
We compared the methods on accuracy, documentation, and risk in detail in a separate article on on-site versus shop balancing. Here, only the financial conclusion matters: removal is justified not when it's cheaper on the invoice, but when there's no result to be had without it.
Sources: ISO 21940-11:2016 · ISO 21940-12:2016
Hidden benefits of on-site work that never appear in any estimate
Several advantages of working without removal never make it into the table at all, because they're hard to put a price on. They affect the result more than half the line items that do get counted.
The main one is technical. On site, the instrument measures the response of your actual assembly: the rotor in its own bearing housings, on its own foundation, with its own drive and belt, at operating speed and operating temperature. The influence coefficient comes out for this machine, not for an abstract rotor sitting in a balancing machine's supports. That's why the vibration reduction you see is exactly where it's required: on the bearing housings, in mm/s RMS (root-mean-square) over the 10-1000 Hz band.
- Fewer assembly risks. Nothing was disassembled, so nothing damaged the fit surfaces, no key got lost, shaft alignment was never disturbed, and no new problem got created in place of the old one.
- A fast check on the result. The verification run happens right after the weights go on, and within minutes you know whether you're within tolerance or not. With removal, the check happens after reassembly, by which point fixing anything is already expensive.
- Before-and-after measurements at the same points. This is your baseline for evaluating against condition zones A, B, C, D, and your reference point for trending. Without it, you can't prove an improvement, and you won't notice vibration creeping back up.
- Saved influence coefficients. On the same machine, or an identical one, the next balancing job costs two runs instead of three, provided the RPM, the support type, and the sensor mounting points stay the same.
- Diagnostics along the way. The spectrum, overall vibration, and 1x with phase at both bearings all get collected on the same visit, and often explain why the machine is noisy beyond just imbalance.
There's a flip side too, and it's more honest to say it upfront. On site, what you're fighting isn't the instrument - it's operating conditions: a wandering VFD speed, a slipping belt, changing throughput, temperature. If the 1x amplitude and phase drift by more than 10-15% between runs, the job stretches out, and that costs money too.
Sources: ISO 20816-1:2016 · Balanset-1A operation manual
Where to start: a short assessment instead of a long argument
The 'pull it or don't' argument doesn't get settled by email - it gets settled by one measurement. A single site visit with the instrument is cheaper than any mistaken disassembly, and afterward you have data in hand, not opinions: the share of 1x within overall vibration, spectra from both bearing housings, the system's response to a trial weight. If the data says the rotor needs to come out, we'll say so, instead of fitting weights onto damaged geometry.
For an initial estimate, we have a calculator: you enter your hourly downtime rate, the disassembly and transport timeline, and the cost of the work, and it shows which route is cheaper in your specific case. It's an estimate, not a quotation, but it usually settles the question right away. A site visit carries a minimum visit fee, covering travel and on-site work, so it makes sense to group smaller machines into a single visit.
AXILINE's engineers design and manufacture the Balanset instruments and use them to balance rotors on site themselves. What arrives on site is a kit in a case: two accelerometers, a laser phase sensor reading off a reflective marker, a two-channel USB module, and Windows software that calculates the correction in one and two planes, splits the mass across fixed positions, calculates drilling, and keeps an archive and grade-G tolerances. If balancing is something you need regularly, it's cheaper to keep the same instrument in-house, and we provide consulting support on using it.
- Fill in the calculation table for both routes, even roughly, and for both scenarios.
- Separately, work out your hourly downtime rate and check whether a standby unit exists.
- Don't forget the 'shaft alignment after reassembly' and 'consumables after disassembly' lines.
- Send us the machine type, power, and RPM, photos of the rotor from both sides, a description of access to the correction planes, and your own measurements, if you have any.
- Keep the 'before' measurements no matter what, even if the rotor ends up going to the shop anyway.
If you don't have a single downtime figure and there's nowhere to get one, start with the cheapest step available: a measurement at the operating condition and an estimate of the 1x share. Until proven otherwise, it's reasonable to start on site, and the proof takes just one visit to obtain.
Sources: Balanset-1A manufacturer specification
Frequently asked questions
Is balancing without removal actually cheaper, or is that just a sales pitch for a site visit?
It's cheaper not because of the price per labour-hour, but because of the lines that simply don't exist: rigging, a round trip in transport, disassembly and reassembly, shaft alignment after reassembly, consumables, and calendar days of downtime. If you have a standby unit, a crane nearby, and a rotor that comes out in an hour, the difference might turn out small. Work it out with the table from this article: the answer depends on your hourly downtime rate and on rigging, not on general arguments.
We don't have a figure for 'cost per hour of downtime.' How do we calculate it?
Build it from four pieces: the volume you didn't produce, multiplied by margin, not by the product's sale price; the wages of people not producing during that time; penalties and deadline obligations; and the cost of restarting after a shutdown, including off-spec output while settling back in. If a standby unit exists and the process doesn't stop, the rate is close to zero, and the outcome then comes down to rigging, labour, and the calendar time you spend without a backup.
Why does the removal option need to include shaft alignment?
Because reassembly changes the shafts' relative position. A rotor signed off to balance quality grade G goes back in, and the machine goes right back to high vibration - only now it's from shaft misalignment. Shaft alignment is a separate job, with its own time and its own people, and it's the line item most often missing from estimates. We cover the difference between the two defects, and the right order of operations, in a separate article on imbalance versus shaft misalignment.
On-site balancing didn't help. Is the money wasted?
Usually not. You're left with measurements taken under real operating conditions: overall vibration and 1x with phase at both bearing housings, spectra, the response to a trial weight. That data is exactly what points to the actual cause: resonance, loose fasteners, a soft foot, bearings, shaft misalignment. And it becomes the justification for disassembly, if that's genuinely what's needed. Compare that with finding out the same thing after you've already paid for rigging.
We're disassembling the machine anyway to replace the bearings. Should we balance it on a machine?
Yes, in this case the logic flips. The expensive line items on the route are already being paid for by the repair: the shutdown, disassembly, rigging, reassembly, and shaft alignment are all happening regardless of balancing. Machine balancing comes to you almost for free, and you get a residual-unbalance report as a bonus. After reassembly, be sure to take a measurement on the bearing housings: that one belongs to the machine, while the report belongs to the rotor.
How much does a site visit cost, and is there a minimum fee?
There is a minimum visit fee: it covers travel and on-site work regardless of how many machines get through in that visit. That's why it's more economical to group smaller units into a single visit, especially identical ones, where saved influence coefficients come into play. We quote the exact figure after a short questionnaire, and for a preliminary estimate there's our calculator, where you enter your own downtime and disassembly figures.
Related content
How to verify a balancing result: a verification measurement that can't be disputed
Compare four numbers at each point: overall vibration in mm/s RMS, the 1x amplitude, its phase, and the speed. An 'after' measurement is only valid if it's taken at the same points, in the same directions, with the same sensor mounting, in the same quantity and frequency band, at the same operating regime, and at the same temperature as the 'before' measurement. Signs of an honest result: it's specifically the running-speed component that dropped, the phase is stable, the amplitude repeats from run to run, and the rest of the spectrum hasn't grown. 'Within tolerance' in the instrument's software means exactly one thing: the residual 1x is below the target you entered yourself.
Balancing report: what it should contain and how to check it
A balancing report should let another person repeat your measurement and get the same numbers. The minimum content: identification of the machine and assembly, date and location, actual RPM and operating mode, support/bearing type, measurement points and directions, the measured quantity with units and frequency band, before-balancing values (overall vibration and the 1x running-speed component — vibration at the rotation frequency, which is what imbalance produces — with phase), trial-run data, the correction applied with mass, radius and angular position, after-balancing values under the same conditions, the acceptance criterion with its source, limitations, and the technician and instrument used. If the document has no RPM, no phase, no frequency band and no trial-run data, it isn't a report — it's a note confirming that someone showed up.
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.