Method of compressing man-made vitreous fibre web

Abstract

According to the invention, there is provided a method of compressing an uncured man-made vitreous fibre web, the web having two opposed major faces. The method comprising the steps: passing the web along a path; and subjecting the web to thickness compression by applying compression to the two major opposed faces of the web. Compression of each of said major faces of the web is applied by passing the path between converging continuous or discontinuous compression surfaces. Further, the respective major face of the web that is being compressed is in contact with one of the converging compression surfaces, and said converging compression surface is inclined towards the path. Additionally, each inclined converging compression surface applies an amount of compression to the major face of the web with which the respective inclined converging surface is in contact, wherein the amount of compression applied to at least one of the two opposing major faces of the web is adjustable.

Claims

1. A method of compressing an uncured man-made vitreous fibre web conducted by a compression unit on a production line prior to entering a curing oven, the web having two opposed major faces, wherein the method comprises the steps: passing the uncured web along a path; and subjecting the uncured web to thickness compression, by the compression unit, by applying compression to the two major opposed faces of the uncured web, wherein compression of each of said major faces of the uncured web is applied by passing the path between converging continuous compression surfaces, wherein the respective major face of the uncured web that is being compressed is in contact with one of the converging compression surfaces, wherein each surface of the converging compression surfaces through which the path passes is inclined towards the path, providing an upstream end of each respective surface at a greater distance from the path than a downstream end of the respective surface, and compression is thereby applied to the two opposing major faces simultaneously, and wherein each inclined converging compression surface applies an amount of compression to the major face of the uncured web with which the respective inclined converging surface is in contact, wherein one surface is provided by a conveyor belt with at least two primary support rollers and at least one secondary support roller located between the primary support rollers, the position of the at least one secondary support roller being moveable towards and away from the path, the amount of compression applied to at least one of the two opposing major faces of the uncured web between the at least two primary support rollers is adjustable by moving the at least one secondary support roller.

2. A method according to claim 1, wherein the web is a secondary web formed from a cross-lapped primary web, which primary web is a fibre web formed by collecting mineral melt fibres on a collector.

3. A method according to claim 1, wherein the amount of compression applied by each inclined converging compression surface is proportional to an angle of incline of the respective inclined converging compression surface, and wherein the amount of compression is adjusted by altering the angle of the incline towards the path of the inclined converging compression surface that applies compression.

4. A method according to claim 1, wherein both of the converging compression surfaces are formed of at least part of a conveyor belt that transports the web.

5. A method according to claim 1, wherein each support roller that is moved has a smaller diameter than the other support rollers.

6. A method according to claim 1, wherein each support roller that is moved has a diameter of about 0.10 metres (m) to about 0.30 metres.

7. A method according to claim 1, where each support roller that is moved has a length of about 2 metres to about 4 metres.

Description

BRIEF DESCRIPTION OF FIGURES

(1) Examples of a method of compressing a man-made vitreous fibre web according to the invention are described below with reference to the accompanying drawings, in which:

(2) FIG. 1 is a side view of the prior art;

(3) FIG. 2 is a side view of an embodiment of a compression apparatus of the invention;

(4) FIG. 3 is a side view of another embodiment of a compression apparatus of the invention;

(5) FIG. 4 is a side on view of yet another embodiment of a compression apparatus of the invention;

(6) FIG. 5 is a perspective view of an embodiment similar to that shown in FIG. 4.

DETAILED DESCRIPTION

(7) The invention may be used in a process for producing a man-made vitreous fibre batt. To produce this batt, a web (a man-made vitreous fibre web) is provided in conventional manner so that it may be processed to form such a batt. At least a part of the process for providing the web is detailed above. Once the web has been provided, it is transported along a path to process the web into a batt.

(8) In more detail, the web and the subsequent batt are produced by a process in which man-made vitreous fibres are produced by fiberising a mineral melt. This is achieved by passing the mineral melt on to a centrifugal spinning apparatus, or spinner. Each spinner comprises at least one rotor attached to a motor. Each motor spins the respective rotor, which throws the mineral melt off the rotor in the form of fibres. As the fibres are thrown off the rotors, they are entrained into a cloud by an air stream produced by air blowers located in the spinner.

(9) As the fibres are entrained into a cloud, if binder or other additives are to be introduced, they are introduced by injection into the air stream. The cloud of fibres is then collected as a primary web on a collector. It is common for suction to be applied at the collector to suck the fibres from the cloud (along with any other additives in the cloud) on to the collector.

(10) Usually, the collection of the primary web is a continuous process that allows the primary web to build up over time. During the collection, the primary web is passed away from the collector at a suitable rate for maintaining the primary web at a desired thickness. Although optional, the primary web can then be cross-lapped by means of a pendulum system. The cross-lapped web forms a secondary web. The uncross-lapped primary web or the secondary web is then passed towards a curing oven. Before reaching the curing oven the web can be pressurised and compressed to shape the web for its intended use. The binder that is applied to the fibres during the fibre-formation process is usually heat-curable, and the binder is thus cured in the curing oven. This enables the binder to bond the web. Following curing, further processing can be applied to the web, and once any processing is finished, the web is cut into batts.

(11) As described above, at least a part of the processing of the uncured web (namely, a web of man-made vitreous fibres comprising uncured binder) usually comprises applying lengthwise and/or thickness compression to the web. A conventional means for applying thickness compression to a web is shown in FIG. 1. This shows a web 1, which is being transported along path A-A in the direction indicated (i.e. from left to right on the figure).

(12) To compress a web, the web is passed between a pair of conveyor belts, a lower conveyor belt 2 of which is horizontal, and an upper conveyor belt 3 that is inclined towards the path along which the web is being transported. Due to the incline of the upper conveyor belt, as the web travels along path A-A, the available space between the two conveyor belts reduces, so causing the web to be compressed. In FIG. 1, the thickness is reduced from thickness T1 to thickness T2, where T1 is greater than T2.

(13) The compression of the web shown in FIG. 1 is illustrated schematically by the fibres that are shown within the body of the web converging as the web passes through the conveyor belts and becoming more densely packed.

(14) In the arrangement shown in FIG. 1, the lower conveyor belt 2 provides a support surface onto which the web is compressed by the upper conveyor belt 3. In such an arrangement, the amount of compression applied can be easily predicted and tailored to the specific need without introducing other variables into the processing of the web.

(15) However, we have found that by replacing the support surface with a second inclined compression surface, the density variation observed through the height of a web is significantly reduced. Accordingly, in the embodiments of the invention, compression is applied to the two major faces of the web (i.e. the uppermost face and the lowermost face of the web).

(16) An example of a compression apparatus that applies compression to each of the major faces of the web, according to the invention, is shown in FIG. 2. This shows an uncured web 10 that is transported along path B-B, which passes between a lower conveyor belt 12 and an upper conveyor belt 14.

(17) Each of the upper conveyor belt 14 and the lower conveyor belt 12 are formed with a support roller 16 at each end. A single belt 18 is held taut between the two support rollers of each respective conveyor belt. One or both support rollers of each conveyor belt are driven to force the belt 18 to rotate around the path formed by the support rollers. There could of course be additional support rollers that are either driven or not driven located between the two support rollers at the ends of each conveyor belt. Any additional support rollers may be used to create bends or curves in the path or may be used only to support the belt, for example when the distance between the support rollers at the ends of the belt is too great to prevent the belt from flexing or stretching.

(18) When passing between the conveyor belts, the upper major face 20 of the web 10 is in contact with the lower surface 15 of the upper conveyor belt, and the lower major face 22 of the web is in contact with the upper surface 13 of the lower belt. This allows the web to be transported when the conveyor belts 12, 14 are driven by their respective support rollers 16.

(19) In order to transport the web 10 in the desired direction, the lower conveyor belt 12 is driven so that it rotates clockwise, and the upper conveyor belt 14 is rotated (i.e. driven by its driving support roller(s)) in an anticlockwise direction. This causes the inner surfaces of each conveyor belt to move in the same direction to transport the web.

(20) Each of the upper conveyor belt 14 and lower conveyor belt 12 is inclined towards the path B-B along which the web 10 is transported. The incline is such that the end of each conveyor belt at which the web arrives when transported along the path is at a greater distance from the path than the end from which the web departs as it is transported further along the path, potentially to undergo further processing. Thus, each conveyor belt can be described as being inclined towards the path along which the web is transported along its length in the direction of travel of the web. In this manner, the thickness of the web is compressed as the distance between the two conveyor belts is reduced, thereby applying thickness compression to the web.

(21) As well as having additional support rollers located between the two end support rollers shown in the figures, the end support rollers may just define the section of the conveyor belt that applies compression to the web. In other words, the conveyor belts may extend further along the path followed by the web in either direction to support the web and transport it along the route it is following thereby using other support rollers.

(22) Instead of using conveyor belts, it is possible to use a series of rollers to transport and compress the web. An example of this is shown in FIG. 3, which shows a similar arrangement to FIG. 2, but instead of the lower conveyor belt, there is a series of rollers 24 that are able to transport the web.

(23) The series of rollers provides a discontinuous surface (instead of the continuous surface provided by a conveyor belt), and one or more rollers, up to and including all of the rollers, may be driven. It is not necessary for all of the rollers to be driven, and it can save on the amount of energy used if only a selection of the rollers are driven.

(24) In FIG. 3, the web 10 is transported along path C-C, and the discontinuous surface of the series of rollers 24 shown in FIG. 3 is inclined towards the path along which the web is transported in the same manner as the conveyor belts shown in FIG. 2 are inclined towards the path. For this to be achieved, the distance between each successive roller 26 of the series of rollers 24 and the path reduces as the distance along the path increases. In other words, the roller that is located earliest in the path has the greatest distance from the path of all of the rollers in the series. The adjacent roller then has a lesser distance from the path, and each subsequent roller has a lesser distance than the roller before it.

(25) The difference in distance from the path of each of the rollers 26 of the series of rollers 24 is the same for each adjacent pair of rollers. This means that the discontinuous surface formed is planar. The discontinuous surface can of course be curved either to form a surface the gradient of which increases (i.e. gets steeper) as the distance along the path increases, or form a surface the gradient of which decreases (i.e. levels off) as the distance along the path increases. This is also the case for the conveyor belts 12, 14 in FIG. 2 that each form a planar surface but could have either of the two types of curved surface by locating support rollers in a suitable arrangement and applying pressure to the belts 18 (e.g. from the web) where appropriate.

(26) In FIG. 3, the web is transported between the discontinuous compression surface of the series of rollers 24 and an upper conveyor belt 14 in the same manner in which the web is transported between the conveyor belts 12, 14 shown in FIG. 2. This results in the web undergoing thickness compression.

(27) As shown in FIG. 3, this means that the lower major face 22 of the web 10 is in contact with the discontinuous surface formed by the series of rollers 24 and the upper major face 20 of the web is in contact with the lower surface 15 of the upper conveyor belt 14. It is of course possible to have a series of rollers in place of the upper conveyor belt 14, thereby forming a discontinuous surface in contact with the upper major face of the web. There can therefore be two discontinuous surfaces in contact with the web, or the series of rollers forming the discontinuous surface in contact with the lower major face of the web could be replaced with a conveyor belt. Additionally, regardless of the type of surfaces that are used to apply thickness compression to the web, at other points along the path along with the web is transported, rollers and/or conveyor belts can be used interchangeably.

(28) It is sometimes desirable to be able to modify the amount of compression applied to the web. An arrangement that allows this to be done is shown in FIG. 4.

(29) FIG. 4 shows a web 10 undergoing thickness compression by passage between two continuous compression surfaces along path D-D. The compression surfaces comprise an upper surface 27 of a lower conveyor belt 28, which upper surface is in contact with a lower major face 22 of the web, and a lower surface 15 of an upper conveyor belt 14, which lower surface is in contact with the upper major face 20 of the web.

(30) The upper conveyor belt 14 is similar to the upper conveyor belts shown in FIGS. 2 and 3 with support rollers 16 at each end. The upper conveyor belt is inclined towards the path D-D along which the web 10 is transported, in the direction in which the web is transported.

(31) The lower conveyor belt 28 has a different form from the upper conveyor belt. This is because the lower conveyor belt has at least three support rollers: at least two primary support rollers 30A, 30B, between which is located at least one secondary support roller 32.

(32) The secondary support roller 32 is located closer to the path along which the web 10 is transported than at least the primary support roller 30A located earliest along the path of the web. This location of the secondary support roller relative to the path produces an incline in the belt 34 of the lower conveyor belt towards the path between at least one of the primary rollers and the secondary roller.

(33) Due to the arrangement of the conveyor belts shown in FIG. 4, in the locations depicted in the figure, the secondary support roller 32 and the primary support roller 30B located furthest along the path of the web 10 are approximately the same distance from the path. These support rollers can of course be located so the respective distances between the support rollers and the path is different from one another.

(34) Indeed, the secondary support roller 32 is moveable towards or away from the path of the web 10. This causes the angle of incline of the belt 34 towards the path between each of the primary support rollers 30 and the secondary support roller to be modifiable.

(35) In FIG. 4, the secondary support roller 32 is moveable towards and away from the path by moving the secondary support roller vertically (as is indicated by the arrow). It would of course be possible to change the angle of incline of the belt by moving the secondary roller in other directions, such as diagonally or horizontally as long as it is moved towards or away from the path along which the web is transported. In this configuration the secondary support roller is only moveable vertically, and not in other directions.

(36) By changing the angle of incline of the belt 34 between the primary support rollers 30 and the secondary support roller 32, the amount of compression applied by the upper surface 27 of the lower conveyor belt 28 is able to be modified. This is because the amount of compression applied by said surface is proportional to the angle of incline of the belt 34 relative to the path of the web 10.

(37) Of course, due to the secondary roller 32, the upper surface 27 of the lower conveyor belt 28 is separated into two inclined surfaces 27A, 27B between the secondary support roller and each of the primary support rollers 30A, 30B. As such, each inclined surface 27A, 27B is delimited by the secondary support roller and the respective primary support roller.

(38) As an example, the secondary roller 32 may be around 3 metres in length (the distance between the axially separated ends of the roller), and may have a diameter of around 10 to 30 centimetres. The particular length used will be chosen to ensure it is suitable for supporting the width of the conveyor belt or web that is to pass over it. As such, the secondary roller may be longer or shorter than the example length. Indeed, the secondary roller may have a length of between around 2 metres and around 4 metres.

(39) The secondary roller 32 may be made of standard roller materials and can be retrofitted to a web processing line, or can be fitted when the processing line is constructed.

(40) In some embodiments, the secondary roller 32 is moveable by use of a handle, which allows it to be moved when the belt 34 is moving, and in some embodiments, the secondary roller is locked or lockable in place when the belt is moving, and is moveable, for example by unscrewing the roller from a frame, when the belt is stationary.

(41) Using an arrangement similar to the one shown in FIG. 4, the variation in the density through the thickness of the web after the web has been compressed has been shown to be reduced by between 40% and 50% over known thickness compression techniques. When the web also undergoes length compression, there is a further reduction in the density variation, which can be as much as a 70% to 80% reduction in the density variation over known thickness compression techniques.

(42) As noted above, it is possible for a conveyor belt to have other support rollers. Should there be other support rollers located between either of the primary support rollers 30 and the secondary support roller 32, each of those other support rollers will be required to be moveable in the same direction as the secondary support roller when the secondary support roller is itself moveable, with the amount of movement corresponding to an a fraction of the movement of the secondary support roller that ensures the upper surface 27 is kept planar. Of course, if the upper surface is initially curved, and the curve is to be maintained, then the other support rollers will each need to be moveable by a certain amount so that the curve may be maintained.

(43) By movement of a number of rollers, it would be possible to change the shape of at least the upper surface 27 of the lower conveyor belt 28, for example from planar to curved or vice versa. Of course, any movement of a support roller may cause the length of the route followed by the belt 34 to change, so there may be a system, such as another support roller that moves, adjusting the path length by an opposite and equivalent amount to the change caused by the movement to the secondary support roller 32, so that the belt is kept taut.

(44) Additionally, instead of the lower conveyor belt shown in FIG. 4, a series of rollers could again be used in its place and perform the same function. When a series of rollers are used, each roller of the series of rollers would be arranged to form an inclined discontinuous surface that is moveable to adjust the incline by movement of each roller of the series of rollers. Also, it would be possible to use a series of rollers in place of the upper conveyor belt.

(45) Further, in FIG. 4, there is only one conveyor belt with a portion that has an adjustable incline. It would also be possible to have both conveyor belts (or series of rollers) with at least a portion of each one having a variable incline or to have an upper conveyor belt with a portion with an adjustable incline instead of the lower conveyor belt having the portion with an adjustable incline.

(46) The process of applying compression to the web using each of the arrangements shown in FIG. 2, 3 or 4 is to pass the web along a path. The web is subjected to thickness compression by applying compression to each of the two major opposing faces of the web. This is achieved by providing converging compression surfaces and compressing each of the major faces by passage of the web along the path between the converging compression surfaces that are continuous (such as a surface of a conveyor belt) or discontinuous (such as a series of rollers). To apply the compression to a major face of the web, a respective compression surface is inclined towards the path along which the web is being transported. As such, the compression surfaces converge towards the path either simultaneously or in series.

(47) When using an arrangement where at least one of the compression surfaces has an adjustable incline, as may be the case in FIG. 2, 3 or 4, the incline can be adjusted before the web is passed through the compression surfaces or during the period in which the web is being between the compression surfaces.

(48) An alternative to using conveyor belts is to use guiding surfaces over which the web is able to slide when transported along its path. When such surfaces are used, the web may be transported by rollers or conveyor belts located at various points along the guiding surfaces, or could be pushed or pulled from an end or the sides of the web (i.e. the minor faces) could be in contact with some means of transport such as a conveyor belt or roller.

(49) Another alternative is to use a system with conveyor belts that has a lower conveyor belt and an upper conveyor belt. The lower conveyor belt is positioned substantially parallel to the direction of travel of the web, or is inclined towards the mid-point of the web. The amount of incline of the lower belt may be adjustable. The upper conveyor belt is, as a whole, rotatable around a roller that is located furthest along the path along which the web is travelling (i.e. at the downstream end of the upper conveyor belt) to adjust the amount of incline of the upper conveyor belt. Such a system is useful when multiple webs are being processed on a single processing line and there is variation in thickness between respective webs, and/or are thick, such as at least 1 metre in thickness.

(50) The upper conveyor belt has a lower surface that is in contact with the web during processing. The lower surface may be a single flat surface, or may be have at least two parts, each of which are flat and are held by one or more supporting rollers so as to produce a vertex with a reflexive angle between the two parts of the surface. This is usually provided so that compression can be applied to the web with at least one of the parts of the lower surface.

(51) When the lower surface has at least two parts, the upper conveyor belt can be positioned so that one part of the lower surface is parallel to the path along which the web travels, and another part of the surface is inclined towards the path. Usually the inclined part of the surface will be the upstream part of the lower surface (i.e. the part of lower surface that the web first comes into contact with).

(52) As the upper conveyor belt is rotatable around the most downstream roller, it is able to be raised to apply compression to a web of greater thickness, or lowered to apply compression to a web of lesser thickness depending on the initial position of the upper conveyor belt. When raised, the rotation increases the distance between the upstream end of the upper conveyor belt and the lower conveyor belt. When rotated up, both parts of the lower surface of the upper conveyor belt will be inclined towards the path along with the web travels. Hydraulics may be used to rotate the upper conveyor belt.

(53) FIG. 5 shows a similar embodiment to that shown in FIG. 4 in a setting similar to that which the compression apparatus is used. The web is not shown in FIG. 5, which allows the relative positioning of the various components to be seen.

(54) As with FIG. 4, FIG. 5 shows an upper conveyor belt 14 delimited by rollers (not shown), the underside of which provides an upper compression surface 15 in use. There is also a lower conveyor belt 28. The lower conveyor belt is delimited by rollers (not shown) and has a support roller 32 provided between the delimiting rollers that causes the upper surface 27 of the conveyor belt to be inclined relative to the lower surface (not shown) of the lower conveyor belt.

(55) The incline of the upper surface 27 of the lower conveyor belt 28 produced by the support roller 32 causes the upper surface to have a first inclined surface 27A from a upstream end of the lower conveyor belt to the support roller and a second inclined surface 27B from the support roller to a downstream end of the lower conveyor belt.

(56) When in use, the upper conveyor belt 14 and the lower conveyor belt 28 transport the web along path 40. The path is parallel to the lower surface of the lower conveyor belt from the upstream end of the compression apparatus until it reaches the support roller 32, after which it is approximately equidistant from the upper conveyor belt and the lower conveyor belt. The reason that the path is parallel to the lower surface of the lower conveyor belt at its upstream end instead of equidistant between the conveyor belts is because the path indicates the direction of travel of the web, not its mid-point. As such, since compression is applied by the compression surfaces (the lower surface 15 of the upper conveyor belt, and the first inclined surface 27A of the lower conveyor belt), although the compression surfaces are inclined, the general direction of travel is still parallel to the lower surface of the lower conveyor belt.

(57) The position of the support roller 32 is adjustable. This allows the incline of the first inclined surface 27A and the second inclined surface 27B to be adjusted. This in turn allows the amount of compression applied to the web by each of these inclined surfaces to be adjusted.

(58) The position of the support roller 32 is adjusted by an adjustment means 36. The adjustment means comprises a movement motor (not shown) and movement sensors (not shown). This enables the assembly to which the support roller is attached to be moved. In this embodiment, the position of the support roller is only moveable along a single axis (not shown). This axis is approximately perpendicular to the lower surface of the lower conveyor belt and is equivalent to the movement vertically mentioned above in relation to FIG. 4. The reason the movement is not vertical in FIG. 5 is that the upper conveyor belt 14 and the lower conveyor belt 28 are rotatable about their connections 42 to the downstream sections of the apparatus and are shown in FIG. 5 with the lower surface of the lower conveyor belt in a non-horizontal position. Otherwise, when the lower surface is horizontal, the axis along which the support roller is able to be moved is substantially vertical.

(59) The adjustment means is able to move the supporting roller 32 electrically up and down. The movement provides an adjustment in the thickness compression applied to the lower layers of the web when the web passes through the compression apparatus. The final amount of thickness compression to be applied is dictated by the product specification, and further thickness compression is able to be applied by a downstream thickness compression conveyor as well if needed. This ensures that the product has the correct final compression required for the desired product quality.

(60) The amount of compression applied by the inclined surfaces is dictated by the desired product quality relative to the deviation in density between the top and bottom layers of the web that would be obtained without the use of the invention; the greater the deviation under these circumstances, the more compression from the inclined surfaces is required to reduce the deviation. This requires the support roller to increase the incline of the upper surface 27 of the lower conveyor belt 28. However, it is important that the fibres are not damaged by the inclined surfaces so that the web material's ability to recover is not at risk.

(61) The movement sensors mentioned above are set to monitor the amount by which the support roller 32 has been moved. This is to reduce the possibility of the full stretch of the lower conveyor belt being exceeded by movement of the support roller away from the lower surface of the lower conveyor belt 28. This thereby reduces the likelihood of the lower conveyor belt being damaged by movement of the support roller. The movement sensors also monitor for movement so that the support roller being moved away from the lower surface of the lower conveyor belt too far so that it will make contact with the upper conveyor belt 14 can be restricted.

(62) As well as being able to increase the incline of the inclined surfaces 27A, 27B of the lower conveyor belt 28, the support roller 32 can reduce the incline of these surfaces to zero. This is achieved by the support roller being able to be moved in line with the rollers delimiting the lower conveyor belt. In FIG. 5, this is shown by the space between rollers downstream of the support roller and a support surface upstream of the support roller that assist in supporting the lower conveyor belt when the support roller is not raised to incline the upper surface 27 of the lower conveyor belt.

(63) As the support roller can be moved electronically, the positioning of the support roller can be precise, such as with a positioning accuracy of +/0.5 millimetres (mm) for example.

(64) Using a compression apparatus similar to that shown in FIG. 5, with a moveable support roller as shown, the data below has been produced as a comparison with a compression apparatus without a moveable support roller. This shows that for a compression apparatus as described herein with compression applied to each of the two major faces of the web there is less deviation in the density of the web between the upper major face and the lower major face than for a compression apparatus that only applies compression to one of the major faces of the web, such as an apparatus only having a flat lower conveyor belt with an inclined upper conveyor belt. The density set for the web in each set of data is shown in kilograms per cubic metre (kg/m.sup.3) and the desired web thickness for each data set is 50 mm.

(65) TABLE-US-00001 Including support roller Excluding support roller Desired web density: Desired web density: 37.0 kg/m.sup.3 35.5 kg/m.sup.3 Lower Upper Density Lower Upper Density surface surface differ- surface surface differ- density density ence density density ence 38.4 38.8 0.4 35 36.7 1.7 37.8 37.3 0.5 35.3 36 0.7 37.8 37.7 0.1 35.7 35.7 0 37.2 38 0.8 33.7 36.3 2.6 36.8 37.2 0.4 34 34.9 0.9 37.9 37.3 0.6 34.3 35.7 1.4 37.6 37.3 0.3 33.7 35.3 1.6 37.4 37.3 0.1 34.7 36.7 2 37.5 37.7 0.2 34 37 3 37.5 37.3 0.2 34.7 35.6 0.9 37.6 37.7 0.1 34.3 36 1.7 35.5 35.6 0.1 34 37.7 3.7 35.9 35.7 0.2 35.5 35.7 0.2 36.2 35.7 0.5 34 36.3 2.3 35.6 37.1 1.5 35 35.2 0.2 36.5 36 0.5 33.3 35.3 2 36.7 37.1 0.4 34.7 34 0.7 37.7 36.7 1 34.4 36.3 1.9 36.9 36.7 0.2 34.3 37.3 3 36.9 36.8 0.1 34.7 35.6 0.9 36.6 37 0.4 34.3 37.3 3 38 36.8 1.2 33 36.3 3.3 37.3 36.5 0.8 35.7 36.3 0.6 37 37.1 0.1 33.7 37 3.3 37.1 36.8 0.3 35 35.5 0.5 36.9 37.1 0.2 34.7 36.7 2 37.6 37.9 0.3 35 38 3 37.2 37 0.2 36.1 36 0.1 38.3 35.5 2.8 34.7 35.7 1 35.6 38.5 2.9 33.7 35 1.3 36.1 38.1 2 34.3 36.7 2.4 37.3 36.6 0.7 35 37.3 2.3 36.6 37.5 0.9 35.8 37.7 1.9 37.6 35.6 2 34.3 36 1.7 35.8 35.5 0.3 34 35.5 1.5 35.8 36.5 0.7 34.6 36.8 2.2 35.3 35.1 0.2 35.8 35.6 0.2 35.7 37.2 1.5 35.7 36.1 0.4 35.8 35.2 0.6 34.8 35.5 0.7 37.8 37.8 0 34.2 35.1 0.9 37.8 36.7 1.1 34.9 36.6 1.7 36 36.4 0.4 34.9 36.6 1.7 35.9 35.2 0.7 35.7 36.3 0.6 35.2 35.3 0.1 34.3 35 0.7 36.1 35.1 1 35.1 36.3 1.2 36.5 36.8 0.3 34.7 38 3.3 35.3 35.3 0 35.6 36 0.4 Average 36.8 36.7 0.05 34.66 36.17 1.52

(66) From the data it can be seen that the average difference in density between the density in the upper major face of each web and the lower major face of each web is 0.05 when a support roller is used and is 1.52 when the support roller is not used. Accordingly, the difference is almost eliminated when a support roller is used. This is surprising since it had previously been expected that even without a support roller the reaction force on the lower major face of the web from the lower conveyor belt was sufficient to provide suitable compression of the lower major face of the web.

(67) Testing has been conducted on webs with a density of about 25 kg/m.sup.3 to webs with a density of about 160 kg/m.sup.3 and has been found to have this beneficial effect. In particular, it has been found to be particularly effective on webs with a density of about 25 kg/m.sup.3 to webs with a density of about 50 kg/m.sup.3 where there have been particular problems with density variation across the thickness of the web in the past.