There are five fundamental and distinct approaches to gear noise reduction available to the acoustical engineer:
1. Change the basic gear type or material.
ADVERTISEMENTS:
2. Modify the tooth profile.
3. Introduce damping or shock reducing material
4. Increase the transmission loss TL of the gear box.
5. Total enclosure.
Often a combination of selected measures from each approach is required to meet the most stringent noise criteria. With this in mind, let us consider the different measures and factors in each approach along with the corresponding range of noise reduction likely.
Basic Gear Types and Relative Noise Levels:
ADVERTISEMENTS:
Consider first the most common parallel gear sets, spur and helical. In general, the loudest of all gears is the spur gear. Here gear tooth meshing is in line or rolling contact and all mechanisms of noise production including impulse, resonance and air or oil ejection are usually present.
With respect to helical gears, it has been demonstrated that helical gears are often from 5 to 15 dB quieter in operation than spur gears for identical loading and speed. In many cases, the noise reduction is even more for the higher harmonics of gear meshing. Intuitively, the quieter operation of helical gears can be attributed to distributed gear tooth loading and reduced impulse flexure.
In nonparallel types of gear sets, the straight bevel gears are generally the loudest. Here again the straight bevel gears have line or rolling contact as opposed to the sliding contact of the spiral bevel or hypoid variety From experience, spiral bevel and hypoid gears are 5 to 15 dB quieter than straight bevel gears. The sliding contact feature is again the element responsible for the noise reduction.
By far the quietest of all the basic types of gear sets is the worm gear. Noise level reductions of 15 dB or more are often achieved with worm gears. With this configuration, the continuous sliding contact minimises impulse and corresponding tooth deflection. In addition, the nature of the continuous contact adds internal damping to the gear system. Table 21.1 summarises the typical range of noise reduction associated with basic types of gear configurations.
It is easy to see that gear design plays a dominant role. One of the most significant parameters associated with gear noise is the gear contact ratio.
One note of caution – high contact-ratio spur gears generally have more teeth and hence generate noise at higher frequencies than ordinary spur gears. Therefore, the noise reduction obtained by using a higher contact ratio may be offset by the relief associated with A-weighting.
Consider as an example a gear set with a fundamental gear meshing tone at 250 Hz. If increasing the number of teeth by a factor of 2 and thus increasing the transverse contact ratio, results in a 3 dB noise reduction, the A-weighted level would actually increase by 3 dBA. This is due to the A-scale weighted difference between 250 and 500 Hz of approximately 6 dB.
ADVERTISEMENTS:
In contrast to spur gears, there are two contact ratios associated with a meshing pair of helical gears. The transverse contact ratio Qt is the contact ratio.
The axial contact ratio Qa is associated with the number of gear teeth that are intersected by a line that is parallel to the shaft at the pith circle. Depending upon the helix angle and the thickness of the gear, this value can range from below 1 to more than 10.
By switching from a spur gear design to a helical gear design, reduction in noise level can be estimated by the following equation –
This equation helps to explain the significant noise reduction advantages of helical gears in comparison with their spur gear counterparts. For example, a helical gear set with an axial contact ratio of 2 is estimated to be 19 dB quieter than a similar spur gear set.
Noise level reductions of up to 5 dB have been reported in which the gear tooth profile was modified such that most tooth contact occurs during the recess portion of the line of contact. Despite the promising acoustical aspects of recess-action gears, little published data or design guidelines are currently available.
Within nominal values, backlash is a secondary factor in gear noise generation. It must be emphasised that too little backlash results in tight mesh and that excessive backlash can increase the noise level 10 dB or more.
With respect to lubrication, dry-running gears, can be 10 to 20 dB noisier than well lubricated gears. Studies on the effect of viscosity over a wide but realistic range of conditions indicate more viscous oil may reduce noise levels up to 2 dB. In short, viscosity and possible related oil temperature appear to be secondary factors compared to the first-order factor of deficient lubrication.
Studies on gear finish have shown that there is no significant difference between ground and milled gears. However, refinements in the teeth such as electropolishing indicate noise level reductions up to 2 dB. A gear tooth error also appears to be a secondary factor, provided a minimum quality of AGMA 12 or better is maintained.
In summary, the noise reduction achievable from minor modification of the gear teeth, gear quality or lubrication is very limited. The only exceptions are those situations where gear operation or selection includes deficient lubrication or excessive tolerances. Summarised in Table 21.2 are the gear tooth features and modifications along with the corresponding range of noise reduction available.
These measures form the basis of good design guidelines for the control of gear noise. One should not, however, expect that by including each design feature the net noise reduction is strictly cumulative.
Synthetic Gear Materials:
Gears constructed of synthetic materials such as nylon, plastic, Dacron, phenolic, etc., have been utilised extensively as a gear noise reduction measure. The obvious limitations of load, temperature and slide characteristics have not deterred engineers from ‘reaping’ an acoustical profit where noise-sensitive design goals must be met.
Where applicable, noise level reductions of 10 to 20 dB are often achieved. With these materials, tooth impulse shock loads are easily absorbed and due to the high internal damping, resultant induced vibratory response is minimal. Further, because of the high internal viscous damping, noise levels associated with harmonics of gear mesh frequency (above n = 3) are seldom significant.
Design limitations for the use of synthetic gears include the following:
1. Loads should not exceed 500 psi.
2. Temperatures should not exceed 100°C.
Application of Damping Materials:
It is generally accepted that high-strength thin-web gears are extremely susceptible to high induced vibratory response or in jargon ringing. This phenomenon is simply the basic noise mechanisms of resonance between gear meshing excitation and the natural frequencies of the gear itself. The obvious remedial approach is to stiffen the gear, which will generally increase the weight or to introduce damping or both.
With the development of highly effective light-weight viscoelastic materials and easy-to-adhere constrained layers, the damping approach has been both popular and successful. In many applications, only a thin layer of material is circumferentially applied in narrow strips in the web and rim areas. When applied in these areas to both sides of the gear, noise reduction of up to 5 dB has been reported.
Currently available damping materials can be easily applied by spraying or troweling or adhered by using precured strips backed with pressure-sensitive adhesives. In addition, these materials have a high resistance to most oils and industrial solvents.
Schlegel has also reported excellent noise reduction results from introducing a compliant material, stiff rubber or plastic, between the rim and hub. This approach essentially decouples or mechanically isolates the two units. Unfortunately, excessive dynamic and static distortion, unbalance and wear rates were also experienced.
In final summary, the basic approaches to gear noise reduction, along with the range of anticipated noise reduction, are listed in Table 21.3. Because of the discrete frequency character of gear noise, it must be emphasised again that careful measurements utilising narrow-band analysis with statistical signal conditioning are usually essential to detect differences and to evaluate applied measures.
Gearbox Enclosures:
All too often gear tooth modification or gear type changes are either too costly, involve major design changes or require long-term development programmes. Therefore, to meet immediate design goals, consideration should be given to a total gearbox enclosure.
The design of a gearbox enclosure is approached in essentially the same manner as any other enclosure and the usual problems of accessibility, visibility and maintainability must be considered along with cooling air, etc.
A marvelous feature of this approach is that, usually, there is no lower limit to the degree of noise reduction obtainable. The practical limitations are only size and weight and in most installations these are not too restrictive. Exceptions include, of course, road vehicles, boats, aircraft, etc.
Before leaving gears, one other, noise source needs mentioning. Frequently, a resonant condition associated with the gearbox occurs when the natural frequencies of the gearbox case coincide with gear tooth meshing frequencies. Under these resonance conditions, the gearbox ‘rings like a bell’.
Gearbox natural frequencies can be changed or the response reduced significantly by adding stiffening members to the case and/or by the application of damping materials. Specific design recommendations cannot be made regarding stiffening. However, for damping material application on heavy castings, a minimum layer of 0.50 inch, well distributed over 50 per cent of the case, is a good rule of thumb.