Bearing noise from day one: the clearance was set wrong at fitment

Most bearing complaints arrive after years of service. This one doesn't. This is the fan that was noisy the week it was commissioned, and the reason is almost always that the bearing's internal clearance was set wrong when it was mounted.

Plummer-block fans in this class typically run taper-bore spherical-roller bearings on an adapter sleeve, C3 clearance, locked with a KM nut and a lock washer. The bearing does not arrive with its running clearance fixed — the fitter sets it, by how far the sleeve is driven up the taper. Drive it too far and the inner ring expands, C3 collapses into negative clearance: preload. Leave it short and the inner ring is loose on its seat. Both are wrong from the first revolution, and neither is cured by grease.

What you're seeing

• Sound. A continuous growl or grind — and critically, from both bearings. Symmetry is the tell: a single out-of-spec bearing is rarely symmetric, but a fitment method applied identically at both ends is.
• Timing. Onset at commissioning, or within days of a bearing/impeller re-assembly. Not a slow drift over months.
• Temperature. Preloaded bearings run hot: housing temperature rise above ~40 °C over ambient, or absolute housing temperature trending past ~70–80 °C on a fan that should sit far cooler. An under-driven sleeve runs cool, but noisy with an intermittent knock.
• Vibration. Broadband rise, worse under load. Under-set clearance gives looseness harmonics (1×, 2× and higher orders); preload gives a raised noise floor and heat rather than a clean discrete peak.
• Motor current. Mildly elevated or unsteady when preload drag is significant — supporting evidence only.
• Grease didn't fix it. If someone greased it correctly and the noise is unchanged, that is a positive finding: it largely rules lubrication out.

Vibration thresholds — state the mounting. Under ISO 14694 (fan application categories, e.g. BV-3), on a rigid mount the commissioning-acceptance limit is ≤ 4.5 mm/s r.m.s., alarm 7.1, shutdown 9.0. On a flexible mount: 6.3 / 11.8 / 12.5. A number quoted without the mounting condition is meaningless. Keep this separate from ISO 10816/20816, which grades overall velocity into evaluation zones A–D for machines generally — different standard, different basis. Don't mix the two on one report line.

What it usually means

Ranked by the honest prior:

1. Adapter sleeve over-driven → clearance crushed into preload. Hot, stiff, growling at both ends.
2. Adapter sleeve under-driven, or locknut not torqued/locked → residual play, fretting, looseness harmonics.
3. Impeller central locking bolt or key not torqued at re-assembly → shaft-to-impeller play that reads at the bearings as vibration.
4. Brinelling from long storage or transport shock — a look-alike, not a fitment error.
5. Grease over-pack or wrong grade at a top-up — churning heat masking or worsening a marginal clearance.
6. A bought-out bearing genuinely out of spec. Real, but the lowest prior, and it should present asymmetrically.

A fan leaves works run-tested, so causes downstream of dispatch — transport, installation, commissioning, operation, wear, or a bought-out component — carry the higher prior. That is a statement about where the clearance is physically set, not a way of pushing the problem across the fence: specifying the setting method and handing it over is the manufacturer's job.

How to diagnose it

First: isolate. Lock out and tag out, prove dead, and confirm the impeller has coasted to a full stop — a large rotor freewheels for minutes.

1. Timeline interview. When did the noise start relative to commissioning, or to the last bearing/impeller job? Has anyone greased it already? Day-one onset, both ends, no cure from grease → you are almost certainly here.
2. Eliminate lubrication. Check grease grade and quantity against the manual. If correct greasing changes nothing, stop chasing grease.
3. Thermal check, running. IR gun or contact probe on both housings. Log the rise over ambient, not just the absolute. Symmetric high heat at both ends = preload suspicion.
4. Vibration snapshot, running. Overall velocity (mm/s r.m.s.) at each bearing, horizontal / vertical / axial; take a spectrum if your meter gives one. Compare against the ISO 14694 limit for the actual mounting.
5. Stop, isolate, feel by hand. Rotate the shaft. Gritty and stiff = preload. Radial or axial play or knock = under-set sleeve or loose locknut. Check cold axial end-float.
6. Inspect the lock stack. Is the locknut actually tight? Is the lock-washer tab bent into a slot, or sitting straight? Is the impeller central locking bolt torqued, and the key sound?
7. Check the sleeve position. Look for the fitter's reference mark. Measure residual internal clearance with a feeler where you can reach it, or confirm the drive-up distance against the bearing maker's table for that exact bearing/sleeve pair.
8. Rule out storage brinelling. If the fan sat idle more than ~2 months before start-up, inspect the raceways for evenly-pitched marks at roller spacing. That is a storage cause, not a fitment error, and it takes a different fix.

Decision point. Hot + stiff + both ends → preload; re-set the clearance looser. Play + knock + fretting → looseness; re-seat and re-lock. Evenly-pitched raceway marks with near-zero run time → brinelling; replace the bearing.

The usual root causes, by lifecycle stage

Transport & handling. Static brinelling — shipping shock on a non-rotating shaft skids the rollers and indents the raceway. Confirm: marks at roller pitch, present before any meaningful run time.

Installation & commissioning (the primary stage). - Sleeve over-driven. Taper drive-up radially expands the inner ring; C3 goes negative. Confirm: both ends hot, stiff by hand, unimproved by correct greasing. - Sleeve under-driven / locknut loose. Inner ring loose on the seat → micro-motion, fretting, walk-out. Confirm: hand-felt play, fretting corrosion on the seat, backed-off nut. - Lock stack not set. Nut not torqued, or the tab never bent into a slot, so it walks loose in service. Confirm: straight tab, slack nut. - Impeller central bolt or key not torqued. Confirm: measurable shaft-to-impeller play. - Over-tightening at fitment. Excess clamp load can crack an adjacent part such as a cooling ring or heat slinger. Confirm: visible crack at strip-down.

Operation & process. Long idle before start-up — bearings sit statically loaded, grease drains or rusts. Confirm: delay beyond ~2 months; inspect grease for rust and raceways for brinelling.

Maintenance & wear. - Grease over-pack or wrong grade. Churning heat with no mechanical play. Confirm: over-full housing, high temperature, shaft turns freely. - Clearance re-set wrong at a field bearing change. The same over/under-drive error, repeated. Confirm: the noise dates from the change, not before it.

How to fix it

• Preload (over-driven). Dismount and inspect the raceways. Clean — not blued, not spalled — then re-mount to the correct clearance: back the sleeve off, re-set the drive-up per the bearing table, verify by feeler and by hand. Discoloured or rough, the bearing is already damaged: replace it and set the clearance fresh.
• Looseness (under-driven / loose nut). Re-seat the sleeve, torque the locknut to spec and bend the lock-washer tab into a slot. Inspect the shaft seat for fretting; if it is scored, deal with the seat — do not clamp a new bearing onto a damaged one.
• Impeller locking loose / key damaged. Replace a worn key, re-torque the central locking bolt, refit the lock washer ahead of the lock nut.
• Cracked cooling ring or heat slinger. Replace the part and correct the fitment torque.
• Brinelling. Replace the affected bearing. Re-greasing a brinelled raceway buys nothing.
• Close on a running re-check. Bearing temperature rise and vibration back inside the ISO 14694 commissioning-acceptance limit for the actual mounting — 4.5 mm/s r.m.s. rigid, 6.3 mm/s flexible — before anyone signs off.

Two misdiagnoses waste the most field time. "Just re-grease it" — grease will not open a clearance that a taper closed. "Bad bearing, replace it" — fit a new bearing, drive the sleeve wrong again, and you have faithfully reproduced the failure. The setting is the fault far more often than the part.

How to stop it coming back

• Set clearance to method, not by feel. Measure residual internal clearance with a feeler, or use the specified drive-up distance for that exact bearing and sleeve. Every mounting, every field change. "It felt about right" is not a method.
• Lock the stack. Correct locknut torque; a new lock washer with the tab bent into the slot (never re-use a bent tab); impeller central bolt torqued to spec.
• Commissioning discipline. Record a cold axial end-float, a first-run bearing-temperature baseline and a vibration baseline against the ISO 14694 limit for the mounting. Do not hand over a fan that is hot or noisy at start-up — you are only deferring the callout.
• Long-storage protocol. Idle beyond ~2 months: rotate the shaft periodically, inspect and renew grease, check for rust before the first start. Put it in writing to whoever is storing the unit.
• Hand the fitment method to the fitter, not just to the purchasing file. The O&M bearing-fitment section and the "do not over-tighten" note need to reach the people who will actually mount the sleeve.
• On the design side, specify C3 deliberately so a small positive clearance still remains after mounting and thermal growth, and balance the rotor to ISO 21940 (G6.3 typical) so you are not chasing a residual-unbalance signature and calling it a bearing.

When to call a specialist

Jitamitra manufactures centrifugal fans and blowers and services fans of any make — on-site vibration diagnosis, in-situ and workshop balancing, bearing and coupling replacement, adapter-sleeve clearance setting, and re-rating where the duty has drifted away from the fan. If you have a day-one bearing noise you cannot close out, our service desk can take it.

sales@jitamitrablowers.com · Jitamitra Help Desk +91 83291 72325

Jitamitra Electro Engineering Private Limited is ISO 9001:2015 certified (third party). Fan performance is tested to IS 4894 / ISO 5801 / AMCA 210 method. CE and ATEX (Zone 2/22) conformity is self-declared.

— Jitamitra Electro Engineering · Technical Services

Engineered for Every Application.