Chapter XI: The Mule (2)

(307) It is hardly practicable to fit a motion of absolute accuracy to effect this purpose, but an approximation to it can be obtained. It is, therefore, arranged that at the beginning of a set the backing-off chain shall be slack, and during building shall be gradually tightened until at the end of it is nearly in a state of tension. The snail is proportioned so as to give a quick downward movement to the faller, and in combination with the arrangement about to be described gives very good results. Referring again to Fig. 160, attached to the snail =G= is a second chain, the other end of which is fastened to a lever =H=, hinged on the bracket shown. The other end of =H= rests on an inclined plate =N=, which slides on a bedplate fastened to the floor. The plate =N= is fastened to the copping plate connecting rod—afterwards referred to—which passes through a horn fastened on =N=. As the copping plate is moved in, =H= is also caused to assume the position indicated by the dotted lines, =N= having also moved in. The effect is that a pull is put upon the snail which gradually rotates it, and causes it to wind on the slack of the chain =E=, so that, when backing-off occurs, the faller is drawn downwards at an earlier moment. The restoration of =N= to its original position accompanies that of the copping plate, and is made at the beginning of a new set of cops.

(308) The various movements in connection with backing-off having thus been described, it is necessary to show how the traverse of the faller is obtained during the inward traverse of the carriage. This is shown in Fig. 161, page 205, which is a separate view of the copping or building mechanism. The faller “locking” lever =A= is, as has been described, raised until the shoulder =R= slips on to the slide =L=, in which position it remains until it is released at the termination of the inward run. On the underside of =L= a small bowl or runner =L^{1}= is carried, which rests upon the upper surface of a longitudinal, or “copping” rail =P=, made of a strong section. If the latter was placed in such a position that its upper surface was horizontal, it is plain that the slide =L= would receive no vertical motion during the period that the runner =L^{1}= was traversing it. In consequence the sickle =U= would remain in one position during the same time. But if the rail =P= is raised at one end so that its upper edge is inclined, the slide =L= will, during the run in of the carriage, receive a vertical traverse corresponding to the difference in the altitude of the two ends of the copping rail. That is to say, if one end of the rail was six inches from the floor line, while the other end was seven, =L= would ascend or descend to the extent of one inch while it was travelling from one end to the other of the rail =P=. The question as to whether it would ascend or descend depends entirely upon which end of the rail was highest. From this it may be inferred that by varying the angularity or profile of the copping rail any desired traverse, either regular or intermittent, could be given to the slide =L=. Now it was shown that the winding faller sickles are keyed on the shaft =B=, which is oscillated by the backing-off finger =D= fastened upon it. The latter being jointed to the “locking” lever =A=, it follows, that, as the latter is raised, the winding faller moves in an arc, which corresponds in length and direction to the length and inclination of the copping rail.

(309) It is necessary when the carriage arrives at the end of its stretch to lock it in that position during the time that backing-off is taking place, and the motions of releasing the counter faller and locking the winding faller are in operation. A reference to Fig. 162 is necessary to understand this part of the mechanism. That illustration is a diagrammatic representation of the mechanism relating to locking the carriage, and the engagement and disengagement of the taking-in gear. The parts are not in their working position, but are projected so that their operation may be better understood. The actual relative position of the various motions is shown by the diagrammatic sketch in the right hand top corner of Fig. 162. Upon the carriage =O= a bracket =O^{1}= is fixed, which carries at its outer end a pin or catch, with which the hook at the end of the horizontal arm of the =L= lever =S= can engage. The hook readily falls over the pin in =O^{1}=, as the carriage is pushed up to it near the end of its traverse. The lever =S= is coupled in the manner shown to the horizontal rod =R=, which, at its other end, is jointed to a bell crank lever =U^{1}=. The rod =R=, on account of its function, is termed the “holding-out catch rod.” The lever =U^{1}= is in turn connected with the rod =U=, jointed at its upper end to the lever =W=, which is coupled to the horizontal arm of the lever =Z^{1}= by the connecting rod =M=. A connection is thus established between the cam =Z= on the cam shaft and the “holding-out” catch lever =S=. During the run out of the carriage the friction clutch =I^{1} K= is disengaged by means of the lever =W=. The rod =R= is also locked by the small vertical slide =S^{1}=, which engages with the catch notch formed in it. The movement of the backing-off rod =X=, which is hinged to the lever =L=, causes the projecting arm in the lever =Y= to be pushed under the end of the lever =W=, thus sustaining the latter and preventing the engagement of the upper half =I^{1}= of the taking-in friction with the lower half =K=. This action occurs just before the termination of the outward run, being a little in advance of backing-off, but simultaneous with the compression of the backing-off spring on =X=. Whatever movement of =W= may take place after the arm on =Y= is thus projected into the path of the end of the lever =W=, the friction cannot fall into gear until the support of the arm is withdrawn. The whole of these parts are thus locked together, and fall into gear simultaneously. It will be noticed that the connection between the lever =S= and the rod =R= is such that the latter can make a certain movement forward before the lever falls. Further, the carriage can be arrested during its outward run by the pedal lever fixed to the floor.

(310) The action of the mechanism is as follows: When the carriage arrives at its outermost point the connecting rod =R= is unlocked, and is free to move. In this way the catch lever =S= can be easily raised by the bracket on the carriage =O=, over which it falls, and securely holds it, the slot in the rod =R= permitting this movement. In this position it remains during the whole period of backing-off, when in a way which is afterwards described, it is released simultaneously with the taking-in friction with which, as shown, it is connected. The locking of the carriage is the last operation requiring explanation before proceeding to deal with the movements, which, together, make up the fourth stage or period. This is the one in which the nicest problems require solution, and in which the mechanism used is the most ingenious.

(311) The first step in commencing to wind is, of course, to release the carriage and draw it in. Before proceeding to show how this is effected, it will be as well to recapitulate and describe the position of the various parts. The strap is entirely upon the loose pulley; the backing-off friction clutch is in gear; the spindles are revolving in the opposite direction to that normal to them; the winding faller is drawn down and locked in a position a little below the nose of the cop; the counter faller is held just out of contact with the threads, but free to rise as soon as an inward movement of the carriage occurs; the roller and back shaft clutches are disengaged; and the upper half of the taking-in friction is out of gear with the lower, but revolving with the vertical shaft on which it slides.

(312) When the chain =E= (Fig. 160) has sufficiently raised the faller locking lever =A= to permit it to lock, the swinging lever =V= is suddenly drawn back. An examination of the drawings, either Fig. 160 or Fig. 161, will show that so long as the face of the locking lever presses against the face of the slide no lateral movement of the former is possible. Further, the connection established between the locking lever =A= and the lever =V=, by means of the lever =F^{1}=, ensures that as soon as the inward movement of the lever takes place when locking occurs, the lever =V= must necessarily oscillate on its pivot. This movement of the lever =V= causes its lower jaw to exercise a pressure upon the lever =L= in the contrary direction to that previously noted, and so draws the stop =X^{1}= in contact with the bottom of the backing-off lever =D=. This action is aided by the spring on the backing-off rod, which is free to extend, and its whole force can be exerted on the lever =D=. In this way =D= is drawn back, and the backing-off clutch is disengaged.

(313) The same movement draws away the supporting piece on the vertical lever =Y=, and allows the upper half of the taking-in friction to fall into gear with the lower half, this action being aided by the spring =Q=. The slot in the end of the connecting rod =M= permits the upward movement of the left hand end of the lever =W= to be made rapidly and freely. In this way the engagement of the friction clutch is a very quick one. This upward movement of the lever =W= is communicated, in a manner described, to the holding out catch, which is also raised nearly simultaneously, and the carriage released.

(314) It is, of course, highly essential that all the three releasing motions shall be accurately “timed,” so as not to take place either before or after the proper moment. Accordingly, ample means of adjustment are provided, both on the rod =X= by the regulation of the stops =X^{1}= and =X^{2}=; on the connecting rod =M= coupling the levers =Z= and =T=; and also on the holding-out rod. In this way it is possible to secure that simultaneous movement of the three parts, which is so essential for effective working. It is obvious that the backing-off friction and holding-out catch must be released before the taking-in friction gears, but the interval between these is so slight that they occur practically simultaneously.

(315) The taking-in friction being in gear, the rotation of the loose pulley is, by the train of wheels shown, communicated to the “scroll” shaft, on which the taking-in scrolls are fixed. These have bands attached to and wrapped round them when the carriage is at the roller beam. As the carriage runs out, the bands, which are fastened to it, are drawn off the scrolls, the scroll shaft being then free to revolve. The engagement of the taking-in friction reverses this process and winds on the bands, thus drawing up the carriage. It will be observed that the scrolls vary in diameter, being about 9 inches in the largest part, and about 3 inches in the smallest. The reason of this construction is to give a varying traverse to the carriage, so as to start it easily, and bring it up to the back stops gently. The scrolls are designed so that, so long as they are revolving, they exercise a pull upon the carriage which is steady and constant. In this way, over-running is avoided, but to prevent any possibility of it a scroll =L^{1}=, shown in a detached position in Fig. 153, is fixed on =L= at an angle of 180 degrees to the others, the point of attachment of its rope being diametrically opposite that on the other scrolls. Thus when the bands on the drawing out scrolls are unwound, that on the “check scroll” is wound and vice versa. The purpose of this scroll is, as its name indicates, to check any tendency to over-running, which it effectually does. In all mules above a certain length, it is desirable to provide some means whereby the carriage shall be drawn in evenly throughout its length, and shall not be in danger of twisting or warping. The scroll shaft, it will be noticed, only extends across the headstock, so that the bands can only exercise any pull on the square, and if no other points of attachment were made, the carriage would at its extremities lag behind the centre. A considerable amount of friction would be thus caused, and the spindles at the end of the carriage would not take up the full length of yarn. It was shown that the back shaft is, during winding, disengaged, so that it is only necessary to establish a connection between it and the scroll shaft, to enable the carriage to be drawn in at several points throughout its length, instead of at one only. Accordingly, the scroll shaft is extended, and an extra scroll shown in Fig. 154 at the right hand side is fitted, from which a band is taken to a drum upon the back shaft. Thus the back shaft is converted into a taking-in shaft, and during that operation revolves of necessity at a variable speed given to it by the scrolls. In this way the carriage is kept parallel to the roller beam throughout its course, and comes up to the back stops along its entire length at one time.

(316) The arrangements for taking-in having thus been described, it now becomes necessary to describe the operation of winding. Before doing so, it will be better to deal with the problem to be solved, and it will aid in understanding it if the construction and method of building the cop be described. For this purpose a reference to the diagrams given in Figs. 163 and 164, page 207, is necessary. The cop is built, as before explained, upon the blade or taper part of the spindle, and, when finished, is of the shape shown in Fig. 163, viz., a cylinder with conical ends. The central part of the cop, =E G K F=, is cylindrical, and at the top and bottom of this part are two cones. The lower cone, =A E B C F D=, forms what is known as the “cop bottom,” and the upper one, =G H I K=, the “nose,” although the latter term is more often and strictly applied to the extreme apex at the points =H I=. As previously stated, the yarn may be wound either upon the bare spindle, upon a short paper tube, as indicated by the thick line inside the cop bottom, or upon a similar tube the whole length of the cop. The use of paper tubes of this character is preferable, especially in cases where the cop is likely to be much handled, as it prevents it from being crushed in, and enables the introduction of a skewer for subsequent winding without there being any danger of the cop being pierced or “stabbed,” this being a fruitful source of waste.

(317) In commencing to wind, the yarn is wrapped on the lower part of the spindle in close coils or spirals for a length of a little more than an inch. The whole of one stretch is wrapped upon this space, and when the next stretch requires winding, it is laid upon the previous layer, and so on until the double cone =A E B C F D= is produced. The length of the traverse of the winding faller wire, or the length of each layer vertically, is called the “chase” of the cop or faller. From this point the yarn is wound in successive layers, beginning always at a higher point, until the final traverse is obtained by which the winding is conducted upon the surface or nose represented by the letters =G H I K=. It was stated in paragraph 311 that the winding faller wire, when the winding faller is locked, is in a position a little below the point =H I=. As soon as the carriage begins to run in, the vertical movement of the winding faller locking lever begins, and is so arranged that the first movement of the wire is a rapid downward one. The effect is that the yarn is laid on the nose of the cop in coarsely-pitched descending spirals, as shown in Fig. 164, these extending downwards until the winding faller wire reaches a point opposite the base of the upper cone, in this case shown in Fig. 163 at =K=. From this point a slower ascent of the winding faller wire is made, so that the yarn is laid in the more finely pitched spirals shown, until the nose of the cop is reached. By this time the carriage has arrived at the roller beam, and the whole of the 63 inches of yarn has been wound.

(318) When the first layer of yarn is wound, and the winding faller is assuming its position to wrap on the second, the initial point of its traverse is a little raised. In this way the yarn is gradually wound in layers, which are represented by the angular lines springing from the lines =A E= and =D F= towards the spindles. During this period the enlargement of the diameter of the cop bottom is proceeding until at the points =E F= the full diameter of the cop is reached. As soon as this occurs the initial point of each layer is gradually raised, and the length of the traverse is slowly diminished as the completion of building is approached, until at the termination of a cop the angle of the layers is shown by the lines =G H= and =I K=. There are thus two adjustments shown to be necessary—first, the starting point of each traverse of the winding faller requires altering; and second, its extent also needs regulation.

(319) These two objects are attained by the regulation of the copping rail =P=, as shown in Fig. 161. The ends of this rail rest upon inclined “copping” plates =Y X=, which are fastened together by the rod =W=, and which receive, as will afterwards be described, an inward movement during the building of the cop. It was shown that the locking of the faller lever and its vertical movement leads to a corresponding movement of the faller. If, for instance, the faller locking lever fell an inch, the winding faller sector would be oscillated and the faller wire drawn upwards. The rate of the ascent of the latter is absolutely relative to the period of the descent of the locking lever. Referring now to Fig. 165, which is a small diagrammatic sketch of the copping rail and its supports, suppose the line =G H= to represent the top of former, =O P= the latter, and =L= the bowl at the foot of the locking lever, if =L=, starting from the left hand position, be supposed to travel in the direction of the arrow =V=, it will be seen that it will fall to the extent indicated by the space =Y Z=. If, on the other hand, the slides =O P= are moved into the position shown by the dotted lines, the rail =G H= will also fall into that indicated in a similar manner. The result is that if =L= now makes the same traverse as before it will rise a little as indicated by the space =W X=. The effect on the winding faller would be that in the first case it would be raised, and in the second it would be depressed to an extent corresponding to the depression of the locking lever. The extent to which this elevation or depression is made depends upon the vertical traverse of the locking lever, and the ratio of the distance of the point of junction of the sector =C= with the faller shaft and that of the faller wire from the same rod. If, for instance, this proportion was 1:2, an elevation of the locking lever half an inch would result in a depression of the faller an inch. It is therefore necessary, during the inward run of the carriage, to provide for the inclination of the carriage to such an extent as to secure the requisite traverse of the faller wire. As the amount of such traverse varies during the building of the cop, it follows that the inclination of the copping rail must be varied correspondingly.

(320) Referring again to Fig. 161, the ends of the copping rails have pins fixed in them, on which are anti-friction bowls, which run upon the edges of the copping plates. The latter are duplicated, so as to sustain the rail at each side, and thus maintain its vertical position. At one side of one of the plates =Y= is an ear =S^{1}=, which is threaded to correspond with a square threaded screw =S= passing through a fixed bracket fastened to the floor. In this way the screw =S= is free to revolve, but cannot make any longitudinal movement. On the end of the screw =S= a ratchet wheel is fixed with which a pawl =S^{2}= engages, which is oscillated so as to move the wheel one tooth at convenient times. The speed of the revolution of the screw varies according to the counts being spun, the elevation of the point of locking being more quickly effected when coarse yarns are being made than when the finer varieties are produced. Whatever may be the velocity at which this elevation is accelerated, the profile of the copping plates is such that the inner end of the copping rail =P= is lowered at a more rapid rate during the formation of the cop bottom than at a subsequent stage. The reason of this will be easily comprehended, if the description of the mode of building the latter be borne in mind. It was then shown that the traverse of the winding faller rapidly increased in extent until the full length of the cop bottom was built. It, therefore, follows that the descent of the locking lever must be largely increased at this period at a quick rate, in order to produce the result indicated. When the outer end of the copping rail begins to descend at a rate which more nearly corresponds to that of the inner end, it gradually approaches to the horizontal, and the vertical motion of the slide, locking lever, and faller is proportionately limited.

(321) The regulation of the winding faller as just described was the one which was usual until recent years. It has been found necessary, however, to obtain a more accurate regulation, so as to ensure that the faller wire shall be in its correct position when locking occurs, especially during the period between the beginning of a cop and the attainment of its full diameter. It is now customary to attach to the front end of the copping rail a loose plate =Q=, which is hinged at one end to the rail, and which carries at its outer extremity a pin and bowl resting upon a third inclined plate =Z=. By varying the profile of the plate =Z=, the regulation of the faller during the early part of its traverse can be accurately made and the proper position of the wire ensured. As a glance at the illustration will show, the upper edge of the copping rail is not straight, but is shaped so as to give a variable speed to the slide =L= in its vertical movement. The proper shaping of the copping rail gave rise to some difficulty, and it will be seen that the loose copping rail =Q= is shaped so as to produce the proper effect, while being much more easily adjusted.

(322) The actual operation of this mechanism is as follows: When the carriage is at its outermost point, and the winding faller is locked, the wire is, as previously mentioned, a little below the nose of the cop. As the inward run proceeds, the bowl first runs up the loose incline, thus raising the locking lever and depressing the winding faller wire. The distance, from the extreme outward point reached by the bowl =L^{1}= and that where the loose rail =Q= is hinged and the downward inclination of the copping rail begins, is so short that the initial depression of the winding faller is very rapid. This produces the coarsely pitched coils referred to in paragraph 317, and illustrated in Fig. 164. By the time the bowl =L^{1}= is at its highest point the winding faller wire is opposite the base of the upper cone. The subsequent downward inclination of the copping rail is much less acute, and the consequent descent of the faller locking lever less rapid. As a result the upward traverse of the winding faller wire is made more slowly, and the yarn is wound in more finely pitched spirals. It only remains to be said, in connection with this part of the subject, that owing to the shape of the copping plates their inward movement is accompanied by a gradual fall of the copping rail, and, consequently, the locking point of the faller lever is relatively elevated. In other words, the traverse of the locking lever prior to locking is gradually lessened as the trail lever slide =L= is lowered, and this is equivalent to an elevation of the winding faller lever and its locking point or shoulder =K=. This causes the depression of the winding faller wire prior to locking to be gradually diminished, so that there is an elevation of its initial point.

(323) The method of obtaining the traverse of the winding faller having been described, the equally important points relating to the mode of rotating the spindle during winding require to be dealt with. A little thought will show that so long as the surface upon which the yarn is wound remains small the spindles must revolve at a more rapid rate than when the surface is enlarged. As the extreme diameter of the cop bottom is enlarged the conditions of successful winding are continually changing. At the commencement of the cop the yarn is wound upon what is practically a parallel surface with a diameter of 5/16 inch and a circumference of ·98 inch. This implies that to wind the 63 inches of yarn 64·3 revolutions are required, these being made during the run up of the carriage. But as the diameter of the cop is enlarged the circumference of the conical surface becomes a variable one, and owing to its enlargement the number of revolutions required to wind the same length of yarn is fewer. This is quite clear and needs no demonstration. Thus when the cop bottom is formed the extreme range of variation is reached, and it follows that in the interval between the commencement of winding and the formation of the cop bottom each stretch must be accompanied by a diminution of the velocity of the spindle proportionate to the increase of diameter. In addition to this it is necessary to take into consideration the varying diameter of the conical surface on which winding takes place, which necessitates a greater terminal than initial velocity of the spindle.

(324) A further point requires elucidation. If the spindle blade were parallel, the number of revolutions necessary to wind the 63 inches of yarn properly, when the cop bottom is formed, being fixed, no further alteration would be necessary. But these conditions do not exist, and the nose of the cop is wound upon a continually diminishing diameter. It is of the utmost importance that the yarn is wound tightly at the nose during the whole of the building of the cop. The rate of the vertical traverse being practically uniform, unless an acceleration of the spindle velocity occurred, there would be slack winding during the latter part of the building of the cop. This would produce a sponginess of the nose, which, when the yarn was drawn off in the subsequent process of winding, as shown by the arrow in Fig. 164, would result in several rings or coils being pulled out in an entangled condition, thus producing waste. Technically the cop would be said to be “halched.” Illustrating this part of the subject by figures, if the diameter of the spindle at the point =B=, Fig. 163, be assumed to be 1/4 inch, its circumference would be ·7854 inch; while if the diameter at =H= be assumed to be 1/8 inch, the circumference would be only ·3927 inch. To wind, say, 10 inches of yarn in each case, would require about 12 and 25 revolutions of the spindle respectively. It is therefore clear that, if the same length is to be wound with equal tension upon the nose of the cop throughout the whole process of building, there must be a gradual acceleration of the terminal velocity of the spindle. Although this is only slight at first it is required at an earlier point as the cop is formed, and becomes of increasing importance.

(325) It will be shown, a little later, that the rotation of the spindles during winding is obtained by the pull of the carriage on a chain, which has its other end attached to an oscillating arm, being fastened to a drum on the carriage. To get a clear idea of the action of this part of the mechanism the two diagrams shown in Figs. 166 and 167 are given, a study of which will be profitable. In Fig. 166 the circles =B C D= represent three positions of the barrel or drum after it has moved in a horizontal plane in the direction of the arrow. To the drum a chain is supposed to be attached, which is held at the point =A=. It is, of course, understood that the barrel is mounted upon a shaft or axis so that it can freely revolve. If it be now assumed that the barrel is in the left hand of the three positions =B=, the chain will be wrapped completely round it. As it is moved horizontally in the direction of the arrow it is revolved, as indicated by the curved arrows, and, by the time it has reached its middle position =C=, has been rotated sufficiently to unwind about half a turn of the chain. A further horizontal motion to the right hand position =D= will complete the unwinding, and, by this time, the drum will have made one complete revolution. It will be at once seen that the rate at which the drum will be revolved will depend upon two factors—its diameter, and the speed of its horizontal traverse. If the point =A= at which the chain is held is stationary, and the horizontal movement uniform, then the rotation of the barrel will be constant. But if the barrel be traversed at a variable rate then its rotation will also be variable. In actual practice this uniformity does not exist, for, as was shown in paragraph 315, the taking-in scrolls vary considerably in diameter. Assuming this variation to be 1:3:1, it would follow that the rotation of the barrel would increase and diminish in the same ratio. In practice this is what happens, and the speed of the revolution of the barrel is quicker about the middle of the taking-in than at any other time.

(326) The assumption that the point =A= is stationary was only made to illustrate the point at issue, and is not founded upon the actual facts of the case. If now it be assumed that not only the barrel but the point at which the chain is held makes a forward movement, a new set of conditions arises. In this case the unwinding of the chain during a given time will be diminished by the amount of the advance of the point =A= in the same period. Assuming the latter to be made at a regular rate it would be easy to calculate the extent of the unwinding. If the effect of the horizontal movement of the barrel from =B= to =C= be to unwind half of one coil of chain—say a length of 7 inches—and that in the same space of time the point =A= moved 3 inches, the amount unwound would be reduced to 4 inches. But this is not the actual condition of things in practice. The point moves at a variable velocity, its forward motion gradually diminishing, so that the acceleration of the rotary velocity of the barrel is greater at the end of its horizontal traverse than at the beginning. In other words, its terminal velocity is highest.

(327) The point of the attachment of the chain at =A= is made in an oscillating arm which, during the inward run of the carriage, receives a forward movement at a speed which is controlled by the velocity of the back shaft. As the latter is, in turn, commanded by the scroll shaft during this period—see paragraph 315—it follows that the variation in the forward movement of the arm is coincident with that of the carriage. Thus the advance of the point =A= will always be in strict correspondence with the velocity of the carriage traverse.

(328) Referring now to Fig. 167, and, assuming =A B= to be the arm to which the chain is fastened, and =O J= and =H C= to represent the arcs through which the point of attachment of the chain travels at different times, it will be seen that the periods of movement are well marked. In each case the arcs are of the same number of degrees, although the chord of one is shorter than that of the other. Dealing first with the inner arc, which represents the position of the point of attachment when nearer the centre, the whole period of movement is divided into equal parts. These are represented by the letters =J K L M N O=. Now, if vertical lines are drawn from these, until they terminate in a straight line drawn parallel to a horizontal line through the point =B=, a clear idea can be formed of the effect of the oscillation of the vertical arm =A B=. The lines terminate at =J^{1} K^{1} L^{1} M^{1} N^{1} O^{1}=. It can be easily seen that the horizontal movement of the point of attachment of the chain gradually becomes less as the arm is oscillated from its most backward position =B C= to its most forward one =B H=, this diminution occurring most after the point =L= is reached. In the movement from =J= to =K= and =K= to =L= the horizontal traverse is about equal. It shows a decrease from =L= to =M=, a greater one from =M= to =N=, and a still greater one from =N= to =O=. The same thing happens if the chain be supposed to be attached at the point =D=. In this case also the decrease in the horizontal forward traverse is variable, but occurs in the same way. The periods here are marked by the letters =C= to =H=, and the extent of the forward motion by those =C^{1}= to =H^{1}=. It will be noticed that the amount of the traverse is greater than that previously noted, the total space covered being respectively =J^{1}= to =O^{1}= and =C^{1}= to =H^{1}=. That is to say, the point at which the chain is fastened moves forward in the same direction as the barrel, but at a different speed. In other words, when the chain is held at =K=, the total forward movement is comparatively small, and if it were held at a point shown by the small inner circle, it would be still less. On the other hand, its attachment at =B= implies a greater total forward movement. It therefore happens that the retardation of the chain by the arm is less in the early part of the oscillation of =A B=—or, to put it differently, the delivery of the winding chain by the arm is greater when it is fixed at =D= than when it is fixed at =K=. Therefore the barrel is more slowly rotated during the same period in the former than in the latter case, but as it completes its lateral movement it is rapidly and considerably accelerated.

(329) The application of this principle is as follows, and it can now be stated that the end of the chain is attached to a nut which slides along the arm, being actuated by the rotation of a screw upon which it fits. Remembering that an acceleration of the terminal velocity and a regulation of the revolution of the spindle is required, the demonstration just given shows that these are obtained by the removal of the nut further from the centre of oscillation. The influence of the pull of the chain upon the barrel when the nut is in the position =K= is much slighter, and shows less variation than when it is at =D=. Every inch which the nut travels outwards has an influence upon this factor, and the conditions of winding are thus accurately regulated. When the winding of the cop begins, the nut is in its lowest position, and the rotation of the barrel is then practically equal. As the nut moves away from the centre the barrel gradually rotates more slowly at the beginning of its inward movement. By the time the most outward position is reached—which, in practice, coincides with the formation of the cop bottom—the variation in the velocity has reached its greatest amount. This, it can be easily seen, is what is wanted. Referring again to Fig. 163, one revolution of the spindle when the yarn is being wound on =A D= would practically take up the same length as would be taken up at the top of the paper tube. But when the faller is guiding the yarn on the conical surface from =E= to =B=, one revolution of the spindle would wind on a greater length at =E= than it would at =B=. Therefore, the initial velocity requires to be less than the terminal. But when the point =E= has become the initial position, the conditions of winding remain thereafter constant, except in so far as is affected by the taper of the blade, and there is no further need for an outward movement of the nut.

(330) The theory underlying the method of winding having thus been dealt with, the mechanism employed can be described. This is shown in Fig. 168, which is a diagram of the whole of the apparatus, and in Fig. 169, which is an enlarged view of a portion of it. The winding arm =M= is centered at its lower end, and has formed on it a toothed quadrant =M^{1}=. The “quadrant” =M= oscillates on a short shaft, securely carried by the headstock framing, and receives its forward movement by means of a pinion =Z=, which engages with its teeth. The extent of the quadrant movement is about a quarter circle. The pinion =Z= is mounted on the same centre as a grooved pulley, over which a cord from the back shaft =H= is passed. Thus the rotation of =H= in either direction produces a similar movement in the pinion =Z=; and the effect is, that, while the back shaft is drawing the carriage out, the pinion is revolving so as to raise the arm =M= or cause it to make a backward oscillation. When the back shaft acts as a taking-in shaft, as described in paragraph 315, the pinion =Z= is revolved so as to move the arm =M= forward. The velocity at which the forward stroke is made is by this arrangement a variable one, and completely corresponds to that of the carriage traverse. Inside the winding arm a long slot is formed in which a screw =P= is placed, this being free to revolve. It may be made with a thread of equal pitch throughout, but, as shown, is provided with a thread of varying pitch, which gradually becomes finer towards the outward end of the arm. The reason of this is obvious. The effect of each layer of yarn upon the problem of winding is greater at the beginning of the formation of the cop bottom than when it is more nearly finished. That is, the enlargement of its diameter is relatively greater at the first stage than at any other. For instance, if the diameter is 3/8 inch and it be increased 1/16 inch, the ratio is 1/6th; while if the diameter is 3/4 inch, and the same increase takes place, the ratio is 1/12th only. The variation required in the speed of winding as each layer is wrapped is therefore less in the latter than in the former case. This is the purpose of the helical screw, which gives a quicker advance to the nut in the earlier stages of winding than when the cop bottom is nearly formed.

(331) The screw has fixed upon it, at its lower extremity, a small bevel pinion, gearing with a similar one placed loosely on the short shaft forming the centre for the arm. During the oscillation of the arm the pinion moves with it, and it is clear that if both remained in this position only this alternate action would occur, and no rotation of the screw would be made. If, however, the pinion on the short shaft be rotated it communicates its motion to the screw =P=, and thus traverses the nut. This is what takes place, and the precise method of effecting it will be described in detail at a little later period. The nut engages with the screw and originally had an eye or hook formed in it, to which the end of the winding chain or band =C= was fastened. The attachment is now made in a different manner, a frame =A= being fixed to the nut along with which it can slide. At the upper end of the frame a small drum is carried, round which the winding chain is wrapped, passing over a small bowl =D= at the lower end of the frame. The other end of the chain or band =C= is fastened to the drum or scroll =X^{1}= which is mounted on a shaft =X= carried in suitable bearings in the square. On the same shaft a spur wheel is geared which engages with a pinion loose upon the tin roller shaft, which it revolves by special mechanism afterwards described in detail. The use of a scroll is intended to accelerate the revolutions of the spindles during the latter part of the fallen traverse. This, like the winding arm, is a modified application of the fusee, and it will be easily understood that when the chain is being unwound from the larger diameter of the scroll, the number of revolutions given to the scroll will be less than when it is being taken off the smaller diameter.

(332) It has been previously shown that the diminishing diameter of the spindle causes it to be necessary that, as the cop is built higher upon it, a correspondingly higher rotary velocity shall be given to it, in addition to the increased terminal velocity produced in the manner described. The most usual method of doing this has been to provide at the end of the quadrant arm a bracket carrying a pin known as the “nosing peg.” The object of this device is to shorten the chain by deflecting it from a straight line about the time when the carriage nears the end of its inward run. This is equivalent to a sudden shortening of the chain, and gives a sudden acceleration to the winding drum. In some cases an automatic arrangement is fitted by which the peg is brought into contact with the chain at an earlier point every stretch, so that the acceleration of the spindle takes place sooner, as the nose of the cop is formed higher up the spindle. It is not difficult to obtain a clear notion of the action of the nose peg if a short length of string be held at one end and attached to a sliding piece at the other. If then the string be pressed down by a rod at the same point, but a little further every time, it will be seen that the sliding piece is moved to a greater extent with each depression.

(333) The arrangement used in the Platt mule is shown in detail in Fig. 169. It consists of the sliding bracket A, carrying, as described, at its upper part, a small drum on which the winding chain =C= is fastened. On the spindle of the winding drum a ratchet wheel =E= is fixed, with which the detent pawls =E^{1}= engage, thus ensuring that =E= is held in any position assumed by it. Also fastened on the spindle of the drum is the curved sector arm =F=, to which a chain =G= is secured. By means of the guide pulleys shown the chain =G= is conducted over the arm or lever =K=, and is attached to the bracket =I=. The lever =K= is hung from its upper end, and has a projecting short arm =K^{1}= attached to it, which can move upwards in the direction of the arrow. The outer end of =K^{1}= presses against a bracket =K^{2}= attached to the quadrant, so shaped that the backward movement of the quadrant pushes the lever =K= back at its lower end. In the bracket =I= a finger =I^{1}= engaging with the copping nut is fixed. The parts having been adjusted to their proper position the slide =A= is at the bottom end of the quadrant =M=, as shown, and the curved arm =F= is in such a position that it has wound upon it a certain length of the chain =G=. The latter is a little slack at first, but as the nut moves out this is rapidly taken up until the chain =G= is in tension. As soon as this happens, each of the forward oscillations of the arm =M= leads to the chain being drawn, and causes the lower end of the lever =K= to be swung forward. The return movement of the quadrant leads to the bracket =K^{2}= pressing upon the arm =K^{1}=, so as to push back the end of the arm or lever =K=. In this way the chain =G= is pulled and the curved arm =F= is drawn a little forward, thus causing the drum and ratchet wheel =E= to revolve. As the winding chain is wound on the barrel, every rotary movement of the latter in a forward direction takes up a little more chain and shortens its length. The amount of this shortening is not great up to the time of the completion of the cop bottom and the arrival of the slide =A= at the end of its traverse along the arm. The position of the parts at this period is shown in the detached view at the right hand top corner of Fig. 169. Up to this time only about the same length of chain is taken up which is needed by the increased distance of the slide =A= from the centre, and the greater forward traverse of the quadrant arm, which, in a sense, releases a certain length of winding chain. When this point is reached the finger =I^{1}= begins to be pressed against by the nut =S^{1}= of the shaper screw, and the bracket =I= commences to be drawn inward. To facilitate the correct action of this mechanism the finger =I^{1}= is adjustable, and the exact moment of its contact with the nut =S^{1}= is thus regulated. The forward movement of the shaper nut which follows gives a similar motion to the bracket =I=, and the chain =G= is thus drawn forward. In this way the drum and ratchet wheel =E= are rotated, and the winding chain gradually shortened. Thus more of it is unwound from the scrolls at each traverse of the carriage, and as it is drawn from the smaller diameter of the scroll towards the end of the run in, the velocity of the spindles is considerably accelerated. The position of the various parts when the carriage is at the back stops is shown in Figs. 170 and 171.

(334) These represent respectively the places occupied by the different portions of the mechanism immediately at the completion of the cop bottom, and at the finish of building a set of cops. The positions of the various parts connected with the slide =A= when winding is complete are shown also in Fig. 169, at the top end of the quadrant arm. Referring to Fig. 170, it will be noticed that the winding chain =C= is unwound from the large part of the scroll only, while Fig. 171 shows it almost entirely unwound from the smaller portion. As was shown, this implies a high terminal velocity of the winding scroll and spindles.

(335) It has been previously mentioned that the rotation of the quadrant screw is obtained by means of the engagement of two bevel wheels, one on the foot of the screw and the other upon the spindle, forming the centre of the quadrant. It was also stated that the last-named wheel was held so as to move round the centre with the quadrant. This is effected by means of a brake spring =P^{2}= which clips the boss of the wheel and holds it. The resistance thus created causes the bevel wheel to move with the quadrant, and prevents it from rotating on its axis. The wheel is compounded with a grooved cord pulley =P^{1}=, over which an endless band =Q= passes. The band =Q= fits the groove in the pulley, and is afterwards guided by the various carrier pulleys shown. Two of these, =S S^{1}=, are borne by brackets fixed to the carriage, and =S= is formed with teeth so as to allow of the engagement of the vertical detent catch on the lever =Y=. If the whole of the pulleys over which the band =Q= passes are free to revolve, except that on the quadrant centre, the inward run of the carriage gives no motion to the cord or band. No effect is produced beyond the rotation of the carrier pulleys, and the forward stroke of the quadrant is made without any effect being produced upon the position of the nut.

(336) It was shown that the gradual accretion of yarn by the cop results in the necessity for a graduation of the velocity of the spindle in winding. This takes place during the whole period of building, and it follows that the traverse of the nut must be governed during the whole period. After a layer of yarn has been wound the nut remains in the position occupied by it during the preceding inward run, until the carriage has made another outward run, and is again commencing to run in. At the commencement of the run in of the carriage the spindles revolve at the same speed as that at which they rotated in the preceding period of winding. If the yarn is a coarse one this is sure to be too fast, because of the increase in the diameter of the cop, owing to the yarn wound during the last inward run. The initial velocity of the spindles is, therefore, such that they take up the yarn too rapidly, and put an extra amount of tension upon it. As was shown in paragraph 303, this causes a depression of the counter faller wire. This is utilised to revolve the quadrant screw and traverse the nut and slide. In other words, the winding is said to be “governed,” and the motion is known as the “governing” or “strapping” motion.

(337) Fixed on the winding faller and counter faller shafts =B B^{1}= (Fig. 168) are two arms =U U^{1}=, to which the ends of a light chain =Y^{1}= are attached. The chain passes round a runner, or pulley, placed in the outer end of the hinged lever =Y=, which is in this way sustained. It is obvious that the vertical position of the lever will be strictly regulated by the position of the two arms =U U^{1}=. As they follow the oscillations of the winding and counter faller shafts, the elevated position of these during spinning ensures the lever =Y= being raised at its free end. This results in the tooth, or detent, being taken out of contact with the teeth on the pulley =S=. When the counter faller is depressed by reason of the tension of the yarn upon it a similar movement occurs in the lever =Y=.