METHOD OF A RING SET FOR A DRIVE BELT

The present disclosure relates to a manufacturing method for drive belt, in particular a ring circumference length calibration process therein. The drive belt is mainly used as the means for power transmission between two adjustable pulleys of the well-known continuously variable transmission that is mainly applied in motor vehicles.

The present type of drive belt is generally known and is composed of a multitude of relatively thin transverse elements that are slidably incorporated on one or more ring sets that are each composed of a number of mutually nested, flexible rings. In the present type drive belt the transverse elements are not connected to the ring set, but rather can slide along the circumference thereof, in particular during operation in the transmission. Also the individual rings of the ring set can slide relative to one another.

The rings of the drive belt, which are alternatively denoted hoops, loops or endless bands, are produced from steel, in particular a maraging steel, that combines -amongst others- the mechanical characteristics of great tensile and (bending) fatigue strength with a relatively favourable possibility to process the steel from sheet-shaped base material towards the desired shape and material properties of the end-product rings, which ideally should not vary along the circumference of the rings. These desired material properties comprise a fair hardness of the ring core material, for combining the characteristics of a great tensile strength together with a sufficient elasticity and ductility to allow longitudinal bending of the ring, and a much harder outer or surface layer, for providing wear resistance to the ring. Further, a considerable residual compressive stress is applied to the surface layer of the ring, for providing a high resistance against metal fatigue. This latter feature is of particular significance, since the ring is subjected to numerous (tension) load and bending cycles during the service life of the drive belt in the transmission.

The overall manufacturing method for such drive belts has become well known in the art. Specifically for the ring set component of the drive belt, EP1815160-A1 for example discloses a manufacturing method that departs from a plate-shaped base material that is bent and welded into a cylindrical shape, or tube. The tube is annealed to restore the original material properties thereof, i.e. to largely remove the internal/residual stress and/or inhomogeneous microstructure introduced in plate bending and tube welding.

The annealed tube is cut into a number of hoops, e.g. in a slitting or laser cutting process. The edges of the hoops (between its radially oriented main surfaces and its axially oriented side faces) are typically processed to remove burrs therefrom and/or for rounding these off, i.e. in a tumbling or laser melting process. The hoops are rolled and elongated to a desired thickness, which is typically about 0.185 mm in the end product. After rolling the hoops are flexible in their circumference direction and are referred to as rings or bands, that are annealed to restore the original material properties thereof, i.e. to largely remove the internal/residual stress and/or inhomogeneous microstructure introduced in hoop rolling.

The annealed rings are calibrated, in which process step these rings are individually wrapped around two rollers and stretched, by forcing these calibration rollers apart while being rotated. In particular, each ring is stretched to a respective predefined diameter or circumference length that corresponds to its intended radial position within the ring set. In this process step of ring calibration, also a residual stress distribution is imposed on the rings. In particular the diameter of the calibration rollers to a large extend determines such residual stress distribution after calibration. In fact, such diameter is carefully selected to realise that a maximum stress level occurs -during operation of the drive belt in the transmission- essentially equally at the radially inner surface of the rings (in a straightened section thereof), as it does at the radially outer surface of the rings (in a maximally bend section thereof), whereby such maximum stress level is minimised. These principles and their technical background are discussed in the European patent application EP1403551 -A 1 and, in more detail, in the 2002 publication article titled “Stress reduction in push belt rings using residual stresses: An approach towards increased power density for Push belt CVT's”. According to these documents, ideally the residual stress distribution defines a pre-bending or so-called curling radius of the ring, which is a radius of curvature defined by a separated circumference section of the ring, that is greater than 2 times a smallest bending radius of the ring occurring during operation of the drive belt in the transmission.

After ring calibration, the ring set is assembled by mutually nesting the rings, typically from the radial outside inwards, i.e. in order of decreasing ring circumference length. Only a small play is allowed between adjacent rings of the ring set in radial direction, i.e. between the outer radius of a respectively inner ring and the inner radius of a respectively outer ring, in order to maximise the fatigue strength of the ring set as a whole in the drive belt application thereof. Typically, such radial ring play varies between a couple of micron negative play to about 10 micron or so of positive play.

After being assembled, the ring set is heat treated by precipitation hardening to increase ring toughness, and (surface) nitriding to introduce residual compressive stress in the outer surface layer of the rings.

Precipitation hardening is also known as aging and is realised through heating the rings to a temperature exceeding 400 degrees Celsius (°C), at which temperature microscopic metallic precipitates incubate and grow at random locations throughout the ring material. As the precipitates grow, the hardness of the ring material increases until, generally speaking, a maximum hardness value is reached, after which the hardness of the ring material typically starts to decrease again (so-called over-aging). To prevent (severe) oxidation of the surface of the rings, precipitation hardening is normally performed in an inert or a reducing process atmosphere, such as nitrogen gas or nitrogen gas with some hydrogen gas mixed-in.

Nitriding provides the rings with an additionally hardened and, moreover, compressively stressed surface layer. In nitriding, at least in the typically applied gas-soft nitriding variant thereof, the rings are kept in an ammonia gas (NH3) containing process atmosphere at a temperature of more than 400°C. At such temperature, the ammonia molecules dissociate at the surface of the rings, forming hydrogen gas and nitrogen atoms, which latter nitrogen atoms enter into the crystal lattice of the ring material. As the nitriding process continues, the nitrogen atoms move away from the surface into the ring material by diffusion, thus providing the ring with a nitrided surface layer of increasing thickness. The thickness of the nitride layer, as well as closely related material properties, such as the hardness and the residual compressive stress at the ring surface, that are obtained in/by the nitriding process thus depend on the composition of the nitriding process atmosphere, in particular the ammonia concentration therein, as well as on the temperature and duration of the nitriding process.

For the practical use of the drive belt in a given transmission application thereof, the thickness of the nitride layer largely determines the mechanical performance and service life of the drive belt in the transmission. In particular, if the nitride layer is too thin, the wear and fatigue properties of the ring are suboptimal, or, if the nitride layer is too thick, the ring material will be too brittle and ring stress levels could exceed the elastic limit during operation. In either case, the ring -and hence the drive belt as a whole- will not perform to its full potential or it may even fail prematurely. Therefore, once a target value has been determined for the nitride layer thickness, it is highly desirable that such target thickness is accurately and consistently realised when mass-manufacturing the drive belt.

When the ring set is assembled after ring calibration, the radially oriented surfaces of the rings that are located inside the ring set must be in close proximity to each other with only a small gap there between in radial direction, in order to realise the desired, small radial ring play in the end product, i.e. after nitriding. Thus, considerably less ammonia gas can be supplied to and/or is available inside the ring set compared to at the outside surfaces of the ring set, which outside surfaces are constituted by the axially oriented side faces of all rings, the radially inner surface of the radially innermost ring and the radially outer surface of the radially outer most ring of the ring set. As a result, the nitride layer thickness at the inside surfaces of the ring set (i.e. at the radially oriented surfaces of the rings that face each other between the pairs of adjacent rings of the ring set) will be less than at the said outside surfaces thereof.

This latter aspect of ring set nitriding was previously considered unproblematic, since the inside surfaces of ring set do not arrive in contact with the transverse elements or the transmission pulleys, i.e. experience less contact stress than the outside surfaces of the ring set that do arrive in such contact during operation of the drive belt. However, according to a recent insight, it may still be beneficial to take this particular aspect of the preferred ring set manufacturing method into account.

In particular according to the present invention, the nitride layer thickness at the inside surfaces of the ring set must be larger than 1 .3 times a largest dimension of the non-metallic inclusions (as approximated by the diameter of a circle circumscribing a respective non-metallic inclusion in a cross-section of the ring), which non-metallic inclusions are inevitably present in the ring material. Hereby, an early fatigue fracture initiation at such inclusions near the ring surface is prevented. At the same time, the nitride layer thickness at the inside surfaces of the ring set should not exceed 2.5 times such largest inclusion dimension, in order to avoid an excessively long nitriding process (resulting in so-called over-aging) and/or an excessively thick nitride layer at the outside surfaces of the ring set. A factor of 2 appears to be optimal in this respect.

Moreover, i.e. even if its nitride layer thickness satisfies the above requirement in absolute terms, the nitride layer at the radially outer surface of the radially innermost ring of the ring set will always be thinner than at its radially inner surface by a factor of at least 1 .5 and, typically, of about 2. According to the present invention, this asymmetry in the nitride layer thickness of the radially innermost ring can be partly compensated for in the ring calibration process. In particular according to the present invention, the residual stress distribution imposed upon the radially innermost ring in ring calibration, i.e. the diameter of the calibration rollers, is selected to realise that a maximum stress level occurring at its radially inner surface exceeds a maximum stress level occurring at its radially outer surface. Effectively by this measure, this latter stress level is favourably reduced. More specifically according to the present invention, the diameter of the calibration rollers is reduced relative to the known value of more than 50 mm, in particular calibration rollers with a diameter of between 30 and 40 mm are applied. This novel setup of the calibration process results in a pre-bending radius that is equal to 2 times the said smallest bending radius of the radially innermost ring during operation of the drive belt in the transmission, or less.

As an advantageous side effect, the present invention also addresses the issue that the radially outer, i.e. outwardly facing surfaces of the rings are more easily damaged (i.e. scratched, dented, foreign particle embedded, etc.) during manufacturing than the radially inner, i.e. inwardly facing surfaces thereof. Such surface damage is known to potentially serve as an initiation location for a fatigue fracture. Therefore, the novel setup of the calibration process according to the present invention is preferably applied not only the radially innermost ring of the ring set, but also the other rings thereof.

The novel manufacturing method according to the present invention and the technical background thereof is explained hereinafter with reference to the accompanying drawing figures, whereof: figure 1 provides a schematically depicted example of the well-known continuously variable transmission provided with a drive belt, figure 2 is a section of the drive belt shown in perspective, figure 3 schematically illustrates the presently relevant part of the known manufacturing method of the ring set component of the drive belt, figure 4 provides a diagrammatic representation of the heat treatment of gas-soft nitriding in the manufacturing method according to figure 3, and figure 5 is a schematic cross-section of a radially innermost ring of a set of rings following the heat treatment of gas-soft nitriding in the manufacturing method according to figure 3.

Figure 1 shows the central parts of a known continuously variable transmission or CVT that is commonly applied in the drive line of motor vehicles between the engine and the drive wheels thereof. The transmission comprises two pulleys 1 , 2, each provided with two conical pulley discs 4, 5, where between a predominantly V-shaped groove is defined and whereof one disc 4 is axially moveable along a respective pulley shaft 6, 7 over which it is placed. A drive belt 3 is wrapped around the pulleys 1 , 2 for transmitting a rotational movement co and an accompanying torque T from the one pulley 1, 2 to the other 2, 1 . The transmission generally also comprises activation means that impose on the said at least one disc 4 an axially oriented clamping force Fax directed towards the respective other pulley disc 5 such that the belt 3 is clamped there between. Also, a (speed) ratio of the transmission between the rotational speed of the driven pulley 2 and the rotational speed of the driving pulley 1 is determined thereby.

An example of a known drive belt 3 is shown in detail in figure 2 in a section thereof, which belt 3 incorporates two ring sets 31 that are each composed of a set of -in this example- six thin and flat, i.e. band-like, flexible rings 32. The belt 3 further comprises a multitude of plate-like metal transverse elements 33 that are held together by the ring sets 31 that are each located in a respective recess of the transverse elements 33. The transverse elements 33 take-up the said clamping force Fax, such when an input torque Tin is exerted on the so-called driving pulley 1 , friction between the discs 4, 5 and the belt 3, causes a rotation of the driving pulley 1 to be transferred to the so-called driven pulley 2 via the likewise rotating drive belt 3.

During operation in the CVT the drive belt 3 and in particular the rings 32 thereof are subjected to a cyclically varying tensile and bending stresses, i.e. a fatigue load. Typically the resistance against metal fatigue, i.e. the fatigue strength of the rings 32 thus determines the service life of the drive belt 3 at a given torque T to be transmitted thereby. Therefore, it has been a long standing general aim in the development of the ring set manufacturing method to realise the required ring fatigue strength and wear resistance at a minimum combined material and processing cost.

Figure 3 illustrates the presently relevant part of the known overall drive belt 3 manufacturing method, i.e. of the manufacturing of the ring set(s) 31 thereof, wherein separate process steps are indicated by way of Roman numerals.

In a first process step I a thin sheet or plate 11 of base material that typically has a thickness in the range between 0.3 mm and 0.6 mm is bend into a cylindrical shape and the meeting plate ends 12 are welded together in a second process step II to form an open, hollow cylinder or tube 13. In a third step III of the process the tube 13 is annealed. Thereafter, in a fourth process step IV the tube 13 is cut into a number of annular hoops 14, which are subsequently -process step five V- rolled to reduce the thickness thereof to a value between 0.100 and 0.250 mm, typically to about 185 micron, while being elongated.

After rolling the hoops 14 are referred to as rings 32. The rings 32 are subjected to a further, i.e. ring annealing process step VI for removing the work hardening effect of the previous rolling process (i.e. step five V) by recovery and re-crystallisation of the ring material at a temperature considerably above 600 °C, e.g. about 800 °C.

Thereafter, in a seventh process step VII, the rings 32 are calibrated, i.e. they are each individually wrapped around two rollers and are stretched to a predefined circumference length while being rotated, by forcing the said rollers apart. In this seventh process step VII, also an internal stress distribution is imposed on the rings 32.

Thereafter, in an eighth process step VIII, ring sets 31 are assembled, each from a number of the rings 32 of suitable, mutually adapted circumference length, by stacking these rings 32 one around the other. Although in figure 3 (and in figure 4) the ring set 31 is illustrated with only two rings 32 for simplicity, the ring set 31 is typically composed of 6 to 12 rings.

Finally, the ring sets 31 are heat-treated in a ninth process step IX of precipitation hardening or aging IX-A and of gas-soft nitriding IX-N. More in particular, aging and nitriding involve heating the ring sets 31 to a temperature of between 400 and 550 °C in a furnace containing a controlled gas atmosphere that is composed of nitrogen, hydrogen and ammonia gas. The exact process settings of the heat treatment are selected in dependence on the base material of the rings 32 (i.e. the alloy composition of the maraging steel), as well as on the mechanical properties that are desired for the rings 32. In this latter respect it is remarked that, typically, it is aimed at a core hardness value of at least 500 HV1 .0, at a surface hardness value of at least 800 HV0.1 and at a thickness of the nitrided surface layer, alternatively denoted nitrogen diffusion zone, of 25 to 35 micron. The duration of the above heat treatment is then obtained as a consequence of these mechanical properties, the process temperature and the process atmosphere composition, which in practice has a value in the range from 30 to 90 minutes.

In figure 4 the nitriding part of the ninth process step IX is schematically illustrated in a cross-section B-B of the ring set 31 indicated in figure 3. In figure 4 a gap 34 between the illustrated pair of adjacent rings 32 is highly exaggerated (i.e. in reality, such gap 34 is far smaller than, i.e. considerably less than 10% the thickness of the rings 32). For providing the rings 32 with a nitrided surface layer, the ring set 31 is immersed in a process atmosphere containing gaseous ammonia molecules that are schematically represented in figure 4 by four circles each: a large circle representing a nitrogen atom and three smaller circles representing the hydrogen atoms of the ammonia molecules. At least some of the ammonia molecules will dissociate at the surfaces of the rings 32, whereby three hydrogen atoms are released to allow the one nitrogen atom to enter into the crystal lattice of the ring 32, which ammonia dissociation reaction is schematically represented in figure 4 inside the dashed ellipses. As part of the ammonia dissociation reaction, the released hydrogen atoms combine to form hydrogen gas. The ammonia dissociation reaction can thus be represented in a formula, as follows:

2NH ₃ 2[N] + 3H ₂ (1)

It is well-known that the rate at which this ammonia dissociation reaction (1 ) occurs is proportional to the process temperature and the ammonia concentration in the process atmosphere and is inversely proportional to the hydrogen concentration in the process atmosphere. However, inside the gap 34 between each pair of adjacent rings 32 in the ring set 31 , the ammonia concentration will typically be less than at the outside the ring set 31. After all, inside this gap 34, ammonia is supplied mostly by (gas) diffusion, whereas outside the gap 34 also a forced circulation of the process atmosphere is available to supply ammonia (and to remove hydrogen). Moreover, the path length of ammonia diffusion inside the gap 34 is much longer than to the outside of the ring set 31 . This means that inside the gap 34 the ammonia concentration is highly dependent on the size of the gap 34, as is the thickness of the resulting nitride layer.

Figure 5 provides is a schematic cross-section of a radially innermost ring 32i of a set of rings 31 following the nitriding heat treatment described hereinabove. As illustrated in figure 5, the thickness TNL-o of the nitride layer 35 at the radially outer surface of the ring 32i is smaller than a nitride layer thickness TNL-s at other surfaces thereof. According to the present invention, the thickness TNL-o of the nitride layer 35 at the radially outer surface of the innermost ring 32i is larger than 1 .3 times a largest dimension of the largest of the non-metallic inclusions that are inevitably present in the ring material. Hereby, an early failure of the innermost ring 32i due to fatigue fracture initiation at such inclusion near the ring surface is reliably prevented. It is noted that such largest dimension can be approximated as the diameter of a sphere encompassing the inclusion, or as the diameter of a circle encompassing a cross-section of the inclusion.

Further according to the present invention, the residual stress distribution imposed upon the innermost ring 32i in ring calibration (process step VII) is defined such that a maximum stress level occurring at its radially inner surface exceeds a maximum stress level occurring at its radially outer surface. Hereby, the said maximum stress level occurring at the radially outer surface of the innermost ring 32i is favourably reduced relative to the conventional, symmetric distribution of such maximum stress level between the radially inner surface and the radially outer surface to account for the locally small thickness TNL-o of the nitride layer thereof.

The present disclosure, in addition to the entirety of the preceding description and all details of the accompanying figures, also concerns and includes all the features of the appended set of claims. Bracketed references in the claims do not limit the scope thereof, but are merely provided as non-binding examples of the respective features. The claimed features can be applied separately in a given product or a given process, as the case may be, but it is also possible to apply any combination of two or more of such features therein.

The invention(s) represented by the present disclosure is (are) not limited to the embodiments and/or the examples that are explicitly mentioned herein, but also encompasses amendments, modifications and practical applications thereof that lie within reach of the person skilled in the relevant art.