Step 1 — Identify the Frequency: The Foundation of Gear Noise Diagnosis
Every noise source in a helical gear drive produces vibration at a characteristic frequency determined by the shaft speed and component geometry. A vibration analyser or portable spectrum analyser connected to an accelerometer on the gearbox housing produces a frequency spectrum — a plot of vibration amplitude versus frequency. The dominant peaks in this spectrum identify the source:
Mesh frequency: f_mesh = z × n / 60 [Hz, z = tooth count, n = shaft speed in RPM]
1× shaft: f_shaft = n / 60 [Hz — once per shaft revolution]
2× shaft: 2 × n / 60
Ball pass outer: BPFO = (N_balls/2) × (n/60) × (1 − d_ball cosα / d_pitch_circle)
| Dominant Frequency in Spectrum | Most Likely Source | Characteristic | Diagnostic Test |
|---|---|---|---|
| Mesh frequency (z × RPM/60) | Helical gear tooth profile or lead deviation; contamination in mesh | Tonal whine; scales proportionally with RPM; amplitude increases with load | Change RPM and confirm f shifts proportionally with z × RPM/60 |
| Mesh frequency sidebands (f_mesh ± f_shaft) | Amplitude modulation — eccentricity, bent shaft, or uneven tooth spacing | Sidebands spaced at shaft frequency around mesh peak; indicates cyclic load variation at 1× shaft | Check shaft runout and gear eccentricity; verify gear OD-to-bore concentricity |
| 1× shaft frequency | Shaft imbalance; gear eccentricity; coupling misalignment | Once-per-revolution; stronger in radial direction; independent of tooth count | Shaft balancing; gear run-out measurement; coupling alignment check |
| 2× shaft frequency | Angular misalignment; coupling misalignment | Twice-per-revolution; stronger in axial direction for misalignment | Laser alignment; coupling inspection |
| Bearing frequencies (BPFO, BPFI, BSF) | Bearing defect — pitting, spalling, cage fracture | Non-harmonic relative to tooth count; often accompanied by high-frequency broadband noise floor rise | Oil particle count; bearing temperature; replace bearing on next planned shutdown |
| Broadband noise — no dominant peak | Oil aeration (foaming); cavitation; loose housing; resonance | Broadband rumble or hiss; not proportional to RPM; often varies with oil temperature | Check oil level and viscosity; check housing fasteners; resonance test with impact hammer |
Gear Whine vs Gear Rattle — Two Different Problems
Gear whine is a tonal noise at mesh frequency — a single-pitch hum that rises with RPM. The source is transmission error at the tooth contact: the helical gear pair does not transmit perfectly constant angular velocity, and the cyclic deviation at mesh frequency excites housing resonances that radiate as audible sound. Gear whine is a precision and lubrication problem — it improves with higher DIN accuracy class, correct surface finish, and adequate EHL film.
Gear rattle is a broadband impact noise — a clatter or knocking sound that appears when the drive torque is low or reverses. The source is backlash impacts: when transmitted torque drops below a threshold, the tooth pair loses contact and the slack-side tooth flanks impact each other with a force proportional to the angular acceleration and the moment of inertia of the gear pair. Gear rattle is a backlash and dynamics problem — it is not caused by poor gear accuracy and does not improve with grinding or profile modification. It is reduced by reducing backlash (at the cost of thermal and assembly risk) or by adding torsional compliance (flexible coupling) to reduce the angular acceleration during torque reversal.
Top 10 Causes of Excessive Helical Gear Noise — With Fixes
| # | Cause | Noise Character | Fix |
|---|---|---|---|
| 1 | Profile deviation too large (DIN Class 8+ as-hobbed) | Mesh frequency whine; worse at high speed | Upgrade to ground DIN Class 5–6 helical gear |
| 2 | Lead deviation — edge loading from misalignment | Uneven mesh noise; louder under load | Laser align; specify lead crowning on replacement gear |
| 3 | Insufficient EHL film (wrong oil viscosity) | Noise increases as oil temperature rises; rough mesh sound | Upgrade ISO VG grade or switch to synthetic PAO |
| 4 | Insufficient helical gear helix angle (ε_β < 1) | Noisier than expected for the gear accuracy class | Increase β or increase face width to achieve ε_β ≥ 1 |
| 5 | Tip interference (no tip relief, torque too high for profile) | Mesh noise increases sharply at full load; smoother at part load | Apply optimised tip relief during next grinding cycle |
| 6 | Housing resonance excited at helical gear mesh frequency | Noise very sensitive to exact RPM — peak at specific speed | Add housing damping panel; change operating speed away from resonance; structural rib modification |
| 7 | Oil aeration (oil level too high; oil foam) | Broadband rushing/hissing noise; varies with oil temperature | Reduce oil fill level; use anti-foam additive; check oil change interval |
| 8 | Hard particle contamination in oil | Random impact clicks; particle count rising in oil samples | Drain and flush; replace filter; repair seal allowing external contamination |
| 9 | Gear pair eccentricity / runout (poor bore-to-OD concentricity) | 1× shaft frequency sidebands around mesh peak | Re-measure bore-to-OD runout; replace gear if above DIN class limit |
| 10 | Damaged or worn bearing (pitting, cage fracture) | High-frequency broadband rise + bearing frequency peaks | Immediate bearing replacement; oil particle count for severity assessment |
The Diagnostic Procedure — From Helical Gear Noise Complaint to Root Cause in Five Steps
Gear analyser measurement confirming DIN Class 5 profile accuracy on a replacement helical gear — eliminating transmission error as a noise source before installation, so that post-installation noise can be attributed to misalignment, lubrication, or housing resonance rather than gear quality
Record the helical gear noise complaint precisely
Tonal or broadband? At what RPM does it appear? Does it change with load? Hot or cold? Recent change (new installation, oil change, overhaul)? These answers narrow the root cause before any instrument is picked up.
Calculate all expected helical gear and bearing frequencies
Calculate f_mesh for every gear stage (z × RPM/60), f_shaft for every shaft, and the outer-race ball-pass frequency (BPFO) for every bearing. List all values before measuring — they are the lookup table for interpreting the spectrum.
Take a vibration spectrum from the helical gear drive housing
Mount accelerometer on the bearing housing (not the base or structure). Take spectra at multiple speeds and loads. The dominant peak at mesh frequency = gear whine; sidebands around mesh = eccentricity; 1× or 2× shaft = imbalance or misalignment; bearing frequencies = bearing defect; broadband = aeration or looseness.
Confirm the helical gear diagnosis with a secondary test
Mesh frequency whine → check DIN class of gear (request inspection report). Misalignment signature → laser align and re-measure. Bearing frequency → oil particle count and bearing temperature. Oil aeration → check oil level and foam; oil sample for particle count and viscosity.
Apply the fix to the helical gear drive and verify improvement
Replace the gear with a DIN Class 5–6 ground gear; realign the shafts; change the oil; replace the bearing. Re-take the spectrum after the fix at the same speed and load. Confirm the dominant peak has reduced. If the dominant peak remains, the root cause was not addressed — return to step 3.
Korea Ever-Power — Noise-Optimised Replacement Helical Gears
When the frequency diagnosis identifies the gear tooth profile or lead deviation as the noise source, the fix is a replacement helical cut gear with higher accuracy class. Korea Ever-Power’s engineering team reviews the failed gear’s inspection report (or measures the gear directly by gear analyser), confirms the noise-relevant parameters (profile deviation ff, lead deviation fHβ), and produces a replacement gear at DIN Class 5–6 with profile modification (tip relief and lead crowning) specified for the application torque and speed — eliminating the transmission error at the root cause level rather than masking the noise with sound insulation. As a direct helical gear manufacturer, Korea Ever-Power provides the complete noise engineering documentation — expected transmission error, contact ratio ε_γ, and estimated noise improvement in dB(A) — with every noise-related replacement order.
Frequently Asked Questions
How much does upgrading from DIN Class 8 to DIN Class 5 reduce helical gear helical gear noise?
The noise reduction from upgrading DIN Class 8 (hobbed) to DIN Class 5 (ground) in a helical gear drive is typically 5–10 dB(A) at the mesh frequency. The improvement comes from two mechanisms: (1) profile deviation ff is reduced from approximately 18 µm (Class 8) to 5 µm (Class 5) — a 3.6× reduction that reduces the transmission error amplitude by approximately the same factor, giving 20 × log₁₀(3.6) ≈ 11 dB. (2) Surface roughness Ra is reduced from 2.0 µm (hobbed) to 0.3 µm (ground), improving the EHL film ratio λ and reducing metal-to-metal asperity contact that contributes additional broadband noise beneath the mesh frequency peak.
Why does helical gear noise increase when the oil is cold at startup?
Cold oil is more viscous than oil at operating temperature. This might seem beneficial (thicker film = less metal contact = less noise), but in practice, cold high-viscosity oil does not flow quickly enough to flood the tooth mesh zone — causing intermittent oil starvation at startup. The rough metal contact during this starved lubrication phase produces more noise than when the correct operating temperature film is fully established. The noise typically decreases as the gearbox warms to operating temperature. If noise does not reduce after warm-up, the viscosity grade is too high for the operating temperature — likely causing excessive churning and foam rather than inadequate film.
Can housing resonance make a perfectly manufactured helical gear helical gear sound noisy?
Yes — a well-made helical gear with DIN Class 5 accuracy in a housing that has a structural resonance at the mesh frequency will radiate significantly more noise than the same gear in a well-damped housing. The gear transmits a small but finite transmission error at mesh frequency; if the housing resonance amplifies this excitation, the radiated sound is disproportionate to the actual transmission error amplitude. Diagnosis: the noise has a very strong RPM sensitivity — loud at one specific RPM, much quieter 10% higher or lower (the signature of resonance excitation). Fix: change operating RPM to avoid the resonance; add constrained-layer damping to the housing panel identified by impact hammer testing; or add structural ribs to shift the resonance frequency away from the mesh frequency at normal operating RPM.
Is there a simple field test for whether gear noise is from the helical gear itself or from a bearing?
The simplest distinguishing test is to change the shaft speed by 10–20% and observe how the noise changes. Gear mesh noise tracks exactly proportionally with speed (a 20% speed increase shifts the mesh frequency and any gear-related noise by exactly 20%). Bearing noise frequencies shift by approximately the same ratio, but the absolute frequency shifts are different numbers — and bearing noise often changes character (pitch and quality) more than gear noise across the same speed range. A portable vibration analyser with cursor measurement provides definitive identification: place the cursor on the dominant noise peak and read the frequency, then check whether it matches z × RPM/60 (gear) or the calculated bearing frequency. If neither matches, the noise is likely structural (not gear-related) (housing vibration or coupling rattle).
Noisy Helical Gear? Korea Ever-Power Can Diagnose and Fix It
Send the gear for inspection or provide the vibration spectrum. Korea Ever-Power identifies whether the noise source is gear accuracy, profile modification, lubrication, or misalignment — and produces the correctly specified replacement gear with noise improvement documentation.
DIN Class 5–6 ground · Tip relief · Lead crowning · TE calculation · dB(A) improvement estimate · MOQ 1 piece
Editor: Cxm
