Perfect Layer Coil Winding Apparatus

The present invention relates to perfect layer coil winding apparatuses for making perfect layer coils in which a wire is wound round a bobbin so as to form a large number of parallel turns.

The term "perfect layer coil" used here and hereinafter is defined to mean a coil that is formed in such a manner that, placing in close contact with the turn already wound on a coil bobbin, the next turn is wound one after another successively along the axial direction of the bobbin.

As a winding apparatus for forming the perfect layer coils there has been hitherto proposed a winding machine in which a pair of rollers for holding a wire therebetween and for defining a supply position of the wire are forced to move in such a manner that the supply position of the wire follows a wire winding position on a bobbin with the delay of a predetermined amount to supply the wire to the bobbin with a predetermined angle of lag and to wind the wire along a coil formed by this time on the bobbin so that a desired coil is continuously formed.

In such a winding machine, a pair of rollers for defining the supply position of wire are moved usually by feed means (formed by a combination of screw and nut) in synchronism with the rotation of the bobbin, and are instantaneously displaced by a desired amount and reversed in their moving direction with a solenoid or the like at the end of each layer of the coil to conduct the winding operation for the next layer. The proposed winding machine has a drawback that, when the number of turns per layer is varied due to the non-uniformity in the width of bobbin or the diameter of wire, the position of wire which is wound round the bobbin falls into disorder. Further, since the wire has to be wound closely along a flange of the bobbin at the initial stage of the winding operation in order to form a perfect layer coil, it is required to manually form first one or two turns prior to an automatic winding operation, and thus the winding machine is very low in operation efficiency.

An object of the present invention is to provide a perfect layer coil winding apparatus which can automatically form perfect layer coils without being affected by the non-uniformity in the diameter of wire or the winding width of bobbin to which the wire is wound.

In order to attain the above and other objects, according to the present invention, there is provided a perfect layer coil winding apparatus in which the width between flanges provided on both ends of a bobbin, which is supported by winding means and whose winding width is adjustable, is measured together with the diameter of a wire wound on the bobbin, the number (an integer) of turns per layer is calculated from the diameter of wire and the width of bobbin which is obtained from the positions of the flanges, a winding width of bobbin required for forming a perfect layer coil having the calculated number of turns per layer is calculated, the position of at least one of the flanges is adjusted to obtain the calculated winding width, the initial portion of the wire is fixed to one of the flanges along which the winding operation starts, the supply position of wire is so set as to supply the wire to the bobbin along the above-mentioned flange and in a direction perpendicular to the axis of the bobbin, the bobbin is rotated while detecting the degree of rotation to start the winding operation simultaneously with the detection of a predetermined degree of rotation made by the bobbin, the supply position of wire is moved in the direction parallel to the axis of the bobbin in synchronism with the rotation of the bobbin in order to conduct the winding operation in such a manner that the supply position of wire is moved with the delay of a predetermined amount for the position at which the wire is wound on the bobbin, and when the above-mentioned number of turns is given to a layer of the coil and the winding operation is turned to the next layer, the supply position of wire is moved to such a position as supplying the wire in a direction perpendicular to the axis of the bobbin at a position on the bobbin at which the first turn of the next layer is placed, and then is reversed in its moving direction in order to continue the winding operation of the next layer in a manner that the supply position of wire follows the position at which the wire is wound on the bobbin, with the delay of the predetermined amount.

FIG. 1 is a perspective view showing a main part of an embodiment of a perfect layer coil winding apparatus according to the present invention.

FIG. 2 is a block diagram for showing a control circuit included in the above-mentioned embodiment.

FIG. 3 is an elevational view showing an example of a bobbin having a variable width used in the present invention.

FIG. 4 is an elevational view for showing an initial state of the winding operation.

FIG. 5 is an elevational view for showing an intermediate state of the winding operation.

FIG. 6 is an elevational view for showing an final state of the winding operation for a layer of a coil.

FIG. 7 is an enlarged view for showing a manner in which a wire is wound to form a perfect layer coil.

An embodiment of a perfect layer coil winding apparatus according to the present invention will be explained hereinafter in conjunction with the accompanying drawings.

Winding means

Referring to FIG. 1, the winding means which support and rotate a bobbin to perform the winding operation, are provided with a rotable shaft 3 supported by a bearing 2 which is fixed on a base 1. A timing pulley 4 fixed to one end of the shaft 3 is connected through a timing belt 6 to a timing pulley 7 fixed to the shaft of a motor 8 which is placed beneath the base 1, and thus the shaft 3 is rotated with the rotation of the motor 8. A table 11 is fixed on the base 1 face to face with the bearing 2, and supports a tail stock 10 in such a manner that the tail stock 10 is free to slide on the table 11. A rotatable tail stock shaft 9 is supported by the tail stock 10 with the axis thereof coincident with the axis of the shaft 3. A nut 12 fixed to one end of the tail stock 10 is threaded onto a screw 13 connected with the rotary shaft of a stepping motor 14 which is fixed to the table 11, in order to slide the tail stock 10 in the axial direction of the shaft 3 by the rotation of the stepping motor 14.

The axially end surface of the shaft 3 facing the tail stock shaft 9 is preferably formed with a diametrically extending groove (not shown) and the axially end surface of the tail stock shaft 9 facing the shaft 3 is preferably provided with a tongue-like protrusion (not shown) which may fit into the groove when the tail stock shaft 3 is moved to the shaft 3. The tongue-like protrusion is diametrically slidable in the groove. Under the thus coupled condition between these shafts 3 and 9, the rotation of the shaft 3 is transmitted to the tail stock shaft 9 through the engagement of the groove and the tongue-like protrusion.

A bobbin

A bobbin 15 is provided with flanges 15a and 15b at both ends of its drum portion on which a wire 19 is wound. A plurality of groups of approximately doglegged holes 15c are alternately formed in the drum portion of the bobbin 15, as shown in FIG. 3, and a thin frame 15d is provided between the holes 15c. Accordingly, when the flanges 15a and 15b are pushed from the outside in the axial direction of the bobbin 15, the frame 15d is deformed and the spacing between the flanges 15a and 15b can be changed.

A case that the winding operation starts from the side of the flange 15, will be described hereinafter.

Push means for the initial stage of winding operation

The means for pushing the wire 19 at the initial stage of winding operation is made up of a rotary pneumatic cylinder 17 (hereinafter simply referred to as a cylinder ) which is fixed through support member 18 to a predetermined position on the base 1, and a push plate 16 fixed to the rotary shaft of the cylinder 17. When the cylinder 17 is rotated, the push plate 16 is interposed between the flanges 15a and 15b of the bobbin 15 which is supported by the winding means, and pushes the wire 19 against the flange 15a to wind the wire 19 along the flange 15a.

Feed means

For the feed means is provided a bracket 28 which is fixed to the base 1 at a predetermined position. On the upper surface of the bracket 28 is mounted a table 24 which is interposed through steel balls 25 between a pair of parallel guides 26 fixed to the upper surface of the bracket 28 and is free to slide on the upper surface in a direction parallel to the axis of the bobbin 15 supported by the winding means. A female screw formed in a side surface of the table 24 is threaded onto a screw 27 connects with the shaft of a stepping motor 29 which is fixed to one end of the bracket 28, so that the table 24 can slide in the direction parallel to the axis of the bobbin 15 with the rotation of the stepping motor 29. A supporting plate 22 is fixed to the table 24, and a pair of rollers 20a and 20b for holding the wire 19 therebetween are supported with a predetermined spacing therebetween by pins 21a and 21b fixed to one end of the supporting plate 22. The supply position of the wire 19 is defined by the rollers 20a and 20b. Accordingly, the supply position of wire can follow the position at which the wire 19 is wound on the bobbin 15, by moving the table 24 in the direction parallel to the axis of the bobbin 15 in synchronism with the rotation of the bobbin 15 which is given by the winding means.

Diameter measuring means for wire

For the means for measuring the diameter of the wire 19 there is provided a stand 33 fixed to the base 1 and a base 32 which is fixed to the stand 33 and is placed above the bobbin 15 supported by the winding means. The diameter measuring means are arranged between a pair of rollers 30a supported in a rotatable manner by pins 31 at an upper end of the base 32 and another pair of rollers 30b supported in a rotatable manner by pins 31 at a lower end of the base 32. Two rollers included in each pair are placed opposite to each other with a predetermined spacing therebetween. The diameter measuring means include a reference roller 34, a differential transformer or potentiometer 39 and a measuring roller 36. The reference roller 34 is supported in a rotatable fashion by a pin 35 fixed to the base 32, the differential transformer or potentiometer 39 is supported by the base 32 face to face with the reference roller 34, and the measuring roller 36 is supported in a rotatable fashion by a pin 37 at one end of an anvil 38 coupled with a movable member of the differential transformer or potentiometer 39 and moreover is placed face to face with the reference roller 34. The wire 19 which is held between the rollers 30a and between the rollers 30b and moves on a predetermined path while making contact with the outer circumference of the reference roller 34, is pushed at a predetermined pressure by the measuring roller 36 and is thus held between the reference roller 34 and the measuring roller 36 to measure the diameter of the wire 19. Further, a plate 45 is fixed to the base 32, and a guiding roller 43 is supported through a pin 44 by the plate 45 in a rotatable fashion. The roller 43 leads the wire 19 which is to be supplied to the winding means, to the rollers 30a.

Means for detecting the degree of rotation of bobbin

The means for detecting the degree of rotation of the bobbin 15 are formed by a rotary encoder 5 which is supported by the base 1 face to face with one end of the shaft 3 and whose rotary shaft is coupled with the shaft 3 to detect the degree of rotation of the shaft 3 equivalent to the degree of the rotation of the bobbin 15.

Means for detecting an origin for the supply position of wire

The origin detecting means for detecting the standby position of the rollers 20a and 20b, which are included in the feed means and define the supply position of the wire 19, are given by a detector, for example, a magnetic scale, made up of a storage element 23a and a read-out head 23b. The storage element 23a is fixed to one end portion of the supporting plate 22 and stores therein predetermined magnetic signals, and the read-out head 23b is supported face to face with the storage element 23a by the bracket 28 and reads the magnetic signals out of the storage element 23a.

Position detecting means

The position detecting means for detecting the position of each of the flanges 15a and 15b on both ends of the bobbin 15 which is held by the winding means, are placed face to face with the bobbin 15. In more detail, on a table 48 fixed to the base 1 is mounted a slide base 47 which is free to slide on the table 48 in the direction perpendicular to the axis of the bobbin 15. A pair of detectors 46a and 46b are fixed to the slide base 47 with a predetermined spacing therebetween, and each of the detectors is provided with a contact which projects forward from the slide base 47. The slide base 47 is connected with a rod of a cylinder 49 fixed to the table 48, and the actuation of the cylinder 49 causes the slide base 47 to slide on the table 48. When the cylinder 49 is actuated, the slide base 47 moves forward and thus respective contacts of the detectors 46a and 46b make contact with the flanges 15a and 15b of the bobbin 15 supported by the winding means to detect the position of each of the flanges 15a and 15b. An electric micrometer or the like can be used for each of the detectors 46a and 46b.

Control means

Referring to FIG. 2, the control means 100 include, on the input side thereof, a number-of-turn setting circuit 101 for setting the total number of turns in which the wire 19 is wound round the bobbin 15, a position detecting circuit 102 connected to the detector 46a for detecting the position of the flange 15a of the bobbin 15 which is supported by the winding means, a position detecting circuit 103 connected to the detector 46b for detecting the position of the flange 15b of the bobbin 15, a diameter-of-wire measuring circuit 105 connected to the differential transformer or potentiometer 39 of the diameter measuring means, a degree-of-rotation detecting circuit 106 connected to the rotary encoder 5 for detecting the degree of rotation of the bobbin 15, an origin detecting circuit 107 connected to the read-out head 23b of the origin detecting means which detect the origin for the supply position of the wire 15, and a pulse generating circuit 122 for generating pulses having a predetermined pulse interval. An arithmetic circuit 104 is connected to the position detecting circuits 102 and 103 and calculates the width of the bobbin 15 on the basis of the results of the position detection of the flanges 15a and 15b supplied from the circuits 102 and 103. An arithmetic circuit 108 is connected to the arithmetic circuit 104 and the diameter-of-wire measuring circuit 105, and calculates the number (an integer) of turns per layer which is allowed on the bobbin 15 in the case of perfect layer coil winding, from the width of the bobbin 15 and the diameter of the wire 19 which are delivered from the circuits 104 and 105, respectively. An arithmetic circuit 109 is connected to the diameter-of-wire measuring circuit 105 and the arithmetic circuit 108, and calculates the winding width of the bobbin 15 required to conduct the perfect layer coil winding, from the diameter of the wire 19 and the number of turns per layer which are delivered from the circuits 105 and 108, respectively. A comparator circuit 115 is connected to the arithmetic circuits 104 and 109, and compares the width of the bobbin 15 delivered from the circuit 104 with the winding width of the bobbin 15 delivered from the circuit 109 to deliver an amount of adjustment by which the width of bobbin 15 has to be varied. A pulse generating circuit 116 is connected to the comparator circuit 115, and the number of pulses delivered from the circuit 116 corresponds to the amount of adjustment for the width of the bobbin 15 which is delivered from the comparator circuit 115. An amplifier circuit 117 is connected to the pulse generating circuit 116, and amplifies the amplitude of the pulses supplied from the circuit 116 to an amplitude capable of driving a stepping motor 14 in order to apply the amplified pulses to the stepping motor 14. Upon the application of the pulses, the stepping motor 14 begins to turn and rotates the screw 13 threaded onto the nut 12, to move the tail stock 10 toward the shaft 3, push the flange 15b of the bobbin 15 with the tail stock shaft 9, and adjust the winding width of the bobbin 15. At this time, the frames 15d of the bobbin 15 are deformed and the width of each hole 15c becomes narrower. An arithmetic circuit 110 is connected to the diameter-of-wire measuring circuit 105, and calculates, on the base of the diameter of the wire 19 delivered from the circuit 105, a rotational frequency dividing ratio which indicates a ratio of the number of pulses delivered from the rotary encoder 5 per one turn (or one revolution) of the bobbin 15 to the number of pulses required to move the rollers 20a and 20b of the feed means by the diameter of the wire 19. A rotational-frequency-dividing-ratio counter 111 is connected to the degree-of-rotation detecting circuit 106 and the arithmetic circuit 110, and sets therein the rotational frequency dividing ratio delivered from the circuit 110 to convert the number of pulses supplied from the circuit 106 into the number of pulses for moving the rollers 20a and 20b on the basis of the rotational frequency dividing ratio. A number-of-turn counter 124 is connected to the degree-of-rotation detecting circuit 106 and the arithmetic circuit 108, compares the number of turns supplied from the circuit 108 with the number of turns which is converted from the number of pulses delivered from the circuit 106, and delivers a signal indicating that the winding operation for one layer is completed when the above-mentioned two kinds of numbers of turns are coincident with each other. A counter 112 is connected to the degree-of-rotation detecting circuit 106 and the number-of-turn counter 124, counts up pulses delivered from the circuit 106, delivers a single pulse signal and ceases to count up the pulses when the number of pulses from the circuit 106 becomes equal to a predetermined number, and is reset by a signal delivered from the counter 124 to start again the counting operation. A number-of-turn counter 125 is connected to the number-of-turn setting circuit 101 and the degree-of-rotation detecting circuit 106, compares the total number of turns set in the circuit 101 with the number of turns which is converted from the number of pulses supplied from the circuit 106, and delivers an end signal indicating that the winding operation is completed when the total number of turns and the converted number of turns coincide with each other. A gate circuit 113 is connected to the rotational-frequency-dividing-ratio counter 111, the counter 112 and the number-of-turn counter 125, is enabled by the pulse signal delivered from the counter 112 to pass therethrough pulses supplied from the counter 111, and is disabled by the end signal supplied from the counter 125 to stop the passage of the pulses supplied from the counter 111. Incidentally, when the number-of-turn counter 125 delivers the end signal indicating that the winding operation is completed, the motor 8 and the winding machine are turned off. An inverter circuit 114 is connected to the counter 112 and the gate circuit 113, and reverses the polarity of the pulses supplied from the counter 112. An amplifier circuit 118 is connected to counter 112. An amplifier circuit 118 is connected to the inverter circuit 114, and amplifies the amplitude of the pulses supplied from the circuit 114 to an amplitude capable of driving the stepping motor 29 to apply the amplified pulses to the stepping motor 29. The stepping motor 29 performs the normal or reverse rotation in synchronism with the rotation of the bobbin 15 in accordance with the number and polarity of pulses supplied from the amplifier circuit 118 to move the rollers 20a and 20b of the feed means forward or backward in the direction parallel to the axis of the bobbin 15. An arithmetic circuit 121 is connected to the position detecting circuit 102, the diameter-of-wire measuring circuit 105 and the origin detecting circuit 107, and calculates a distance from the origin to such an initial position of the roller 20a that the roller 20a can supply the wire 19 to a position on the bobbin at which the wire 19 is first wound, in the direction perpendicular to the axis of the bobbin 15, on the basis of the position of the flange 15a of the bobbin 15 supplied from the circuit 102, the diamer of the wire 19 supplied from the circuit 105 and the position of the roller 20a (or 20b) supplied from the circuit 109, in order to calculate and deliver the number of pulses corresponding to the above-mentioned distance. A counter 123 is connected to the arithmetic circuit 121 and the pulse generating circuit 122, sets therein the number of pulses supplied from the circuit 121, counts up pulses supplied from the circuit 122, and delivers a single signal when the set number of pulses coincides with the number of pulses supplied from the circuit 122. A gate circuit 126 connected to the arithmetic circuit 121, the pulse generating circuit 122 and the counter 123, is enabled by the pulses supplied from the circuit 121 to pass therethrough pulses supplied from the circuit 122, and is disabled by the signal delivered from the counter 123 to stop the passage of the pulses from the circuit 122. The pulses having passed through the gate circuit 126 are applied to and amplified by the amplifier circuit 118 to apply the amplified pulses to the stepping motor 29. Then, the stepping motor 29 turns in correspondence with the applied number of pulses to move the roller 20a so as to place the contact between the roller 20a and the wire 19 in a plane containing therein the inner side face of the flange 15a. Thus, the wire 19 can be supplied to the position on the bobbin 15 at which the wire 19 is first wound, in the direction perpendicular to the axis of the bobbin 15. A change-over switch 119 is connected to the counter 112 and the gate circuit 126, is put in the on-state by the pulses supplied from the circuit 126, and is put in the off-state by the pulse signal delivered from the counter 112. A solenoid valve 120 is connected to the change-over switch 119, and opens or closes the path of the compressed air supplied to the cylinder 17 of the push means according as the switch 119 is put in the on- or off-state.

With the circuit arrangement as above, the following operation is performed. Firstly, a total number of turns desired is set in the number-of-turn setting circuit 101, and the number of pulses corresponding to that degree of rotation of the bobbin 15 which is equivalent to the delay of a predetermined amount between the supply position of wire and the winding position on the bobbin, is set in the counter 112. The bobbin 15 is interposed between the shaft 3 and the tail stock shaft 9. Then, the stepping motor 14 is turned on to move the tail stock 10 forward and to push the bobbin 15 with such a pressure as not to change the width of the bobbin 15. Thus, the flanges 15a and 15b are kept in close contact with the shaft 3 and the tail stock shaft 9, respectively. Next, the initial end of the wire 19 which is held between the rollers 20a and 20b, is fastened at a predetermined position on the flange 15a. In this state, the pneumatic cylinder 49 is actuated to move the detectors 46a and 46b. Thus, the contacts of the detectors are put in contact with the facing surfaces of the flanges 15a and 15b to measure the position of each of the flanges 15a and 15b. Then, the position detecting circuits 102 and 103 detect the coordinates Wa and Wb of the flanges 15a and 15b in the direction parallel to the axis of the bobbin 15, and the arithmetic circuit 104 calculates the width W (W=Wa-Wb) of the bobbin 15. While, the diameter D of the wire 19 which is placed between the reference roller 34 and the measuring roller 36, is measured by the differential transformer or potentiometer 39, and is detected by the diameter-of-wire measuring circuit 105. The arithmetic circuit 108 calculates the number n of turns per layer from the width W of the bobbin 15 and the diameter D of the wire 19. (In this case, however, the number n is an integer indicating a quotient in the division W/D and is so determined as to make a residue in the division W/D less than D2.) Subsequently, the arithmetic circuit 109 calculates the winding width w(w=(n+1/2)D) of the bobbin 15 required to conduct the perfect layer coil winding from the number n of turns per layer and the diameter D of the wire 19. The comparator circuit 115 compares the width W and the winding width w of the bobbin 15 which are supplied from the arithmetic circuits 104 and 109 respectively, to deliver the amount A.sub.1 (A.sub.1 =W-w) of adjustment by which the width of the bobbin 15 is varied. The pulse generating circui 116 produces pulses to the number corresponding to the amount A.sub.1 of adjustment to apply the pulses to the stepping motor 14 through the amplifier circuit 117. The stepping motor 14 turns in correspondence with the number of pulses applied. Thus, the tail stock 10 is moved toward the shaft 3 with the rotation of the screw 13, and the flange 15b is pushed by the tail stock shaft 9 to adjust the winding width W of the bobbin 15. The origin detecting circuit 107 detects the origin S or standby position of the roller 20a (or 20b) of the feed means. The arithmetic circuit 121 calculates, from the coordinate Wa of the flange 15a supplied from the position detecting circuit 102, the diameter D of the wire 19 supplied from the diameter-of-wire measuring circuit 105 and the origin S of the roller 20a (or 20b) supplied from the origin detecting circuit 107, an amount ##EQU1## of movement of the roller 20a required to move the roller 20a to the supply position shown in FIG. 4 and capable of supplying the wire 19, in the direction perpendicular to the axis of the bobbin 15, to that position s on the bobbin 15 at which the wire 19 is first wound. The arithmetic circuit 121 delivers pulses in the number corresponding to the amount A.sub.2 of movement. The number of pulses delivered from the circuit 121 is set in the counter 123. Simultaneously, the gate circuit 126 is enabled, namely, put in the onstate to pass the pulses from the pulse generating circuit 122 and to apply the pulses to the stepping motor 29 through the amplifier circuit 118. The stepping motor 29 turns in correspondence with the number of pulses applied thereto. Thus, the table 24 is moved with the rotation of the screw 27 to move the rollers 20a and 20b to the supply position s. When the number of pulses set in the counter 123 becomes coincident with the number of pulses which are supplied from the pulse generating circuit 122 and are counted up by the counter 123, the counter 123 delivers a signal to disable the gate circuit 126. The pulses having passed through the gate circuit 126 are applied to the change-over switch 119 to close the switch 119, and an electric current is thereby applied to the solenoid valve 120. Thus, the solenoid valve 120 opens to supply the compressed air to the cylinder 17. The introduction of the compressed air into the cylinder 17 rotates the shaft of the cylinder 17, and therefore the push plate 16 turns round. Thus, the push plate 16 is interposed between the flanges 15a and 15b to push the wire 19 against the flange 15a, as shown in FIG. 4. The arithmetic circuit 110 calculates the rotational frequency dividing ratio on the basis of the diameter D of the wire 19 supplied from the diameter-of-wire measuring circuit 105 to set the ratio in the rotational-frequency-dividing-ratio counter 111. In this state, the motor 8 is turned on to commence the winding operation. At the beginning of the winding operation, the wire 19 is wound along the flange 15a, since the winding position is defined by the push plate 16. The rotation of the bobbin 15 is detected by the rotary encoder 5, and the degree-of-rotation detecting circuit 106 delivers pulses to the number corresponding to the degree of rotation of the bobbin 15. The above pulses are counted up by the counter 112, and the counter 112 ceases to count up the pulses and delivers a single signal when the number of counted pulses becomes equal to the number previously set in the counter 112. The signal from the counter 112 is applied to the gate circuit 113, the inventer circuit 114 and the changeover switch 119. The switch 119 opens upon the application of the above signal to interrupt the electric current for the valve 120. Thus, the solenoid valve 120 shuts to interrupt the supply of compressed air into the cylinder 17 and to exhaust the compressed air within the cylinder 17 to the atmosphere. Thus, the cylinder turns round and the push plate 16 is rotated to be spaced apart from the bobbin 15. The gate circuit 113 is enabled by the signal from the counter 112 to apply the pulses from the rotational-frequency-dividing-ratio counter 111 to the inverter circuit 114 through the circuit 113. The inverter circuit 114 defines the polarity of the pulses from the gate circuit 113 by the signal from the counter 112, and the pulses thus defined are applied to the stepping motor 29 through the amplifier circuit 118. The stepping motor 29 turns with a predetermined sense of rotation in correspondence with the polarity and number of pulses delivered from the amplifier circuit 118, to move the slide table 24 with the rotation of the screw 27. Since the rollers 20a and 20b move together with the slide table 24, the supply position of the wire 19 moves in the direction parallel to the axis of the bobbin 15 with the delay of a predetermined amount for the winding position on the bobbin 15, as shown in FIG. 5.

For example, in the case that the counter 112 delivers the signal when the bobbin 15 has rotated by a 180-degree angle, during the first one half turn of the bobbin 15, the winding operation is controlled by the push plate 16 and the wire 19 is wound in close contact with the flange 15a. When the bobbin 15 has just rotated by a 180-degree angle, the push plate 16 is spaced apart from the bobbin 15, the rollers 20a and 20b begin to move, and the wire 19 is wound round the bobbin 15 spaced from the flange 15a. When the degree of rotation of the bobbin 15 reaches 300 to 360 degrees, the wire 19 is placed on that initial portion of the wire 19 which passes through the flange 15a and corresponds to the starting position of the winding operation. At this time, the supply position of the wire 19 has been already moved, and therefore the supply position is shifted from the center of the above-mentioned initial portion of the wire 19 toward the flange 15b. Thus, the wire 19 slides down the initial portion to be placed parallel with the initial portion. Simultaneously with the above slide, the portion of the wire 19 which has been wound round the bobbin 15 during the rotation of the bobbin 15 from 180 to 360 degrees, is corrected in position by the tension of the wire 19 to take the shortest path. The subsequent turns of a coil are wound parallel with the first turn to form a perfect layer coil, and turns 1-a, 1-b, 1-c, 1-d, and so on are wound in this order, as shown in FIG. 7. The supply position of the wire 19 follows the winding position with the delay of one half of the diameter D of the wire 19 during the winding operation for the turns 1-b, 1-c, 1-d, and so on. Thus, the angle .theta. of inclination of the wire 19 resulting from the slide of the wire 19 indicates the so-called angle of lag.

When the wire 19 is wound round the bobbin 15 to the number n of turns per layer which is calculated by the arithmetic circuit 108 and is set in the counter 124, the reset signal from the counter 124 is applied to the counter 112. Thus, the counter 112 is reset, and commences again to count up the pulses delivered from the degree-of-rotation detecting circuit 106. Further the counter 112 delivers a single signal which is applied to the gate circuit 113 and the inverter circuit 114, and ceases to count up the pulses, when the number of pulses supplied from the circuit 106 becomes coincident with the number set in the counter 112. At this time, the rollers 20a and 20b have been moved, as shown in FIG. 6, to the position capable of supplying the wire 19 to the bobbin 15 along the flange 15b in the direction perpendicular to the axis of the bobbin 15. The inverter circuit 114 reverses the polarity of the pulses from the gate circuit 113 when applied with the signal from the counter 112, to apply the pulses of reverse polarity to the stepping motor 29 through the amplifier circuit 118. Thus, the stepping motor 29 turns in the reverse sense of rotation, and therefore the rollers 20a and 20b move in the opposite direction. At this time, the delay of the predetermined amount is given by the counter 112.

In more detail, as shown in FIG. 7, the winding operation for the first layer is conducted in the order of turns 1-a, 1-b, . . . , 1-k, 1-l, 1-m, 1-n, and 2-a, and is completed at the turn 2-a. The end of the first layer is identical with the beginning of the second layer, and is indicated in FIG. 7 by the upper one of circles containing therein reference symbol 2-a. When the winding operation for the first layer is completed, the delay between the supply position of the wire 19 and the winding position on the bobbin 15 is eliminated, and the wire 19 is supplied to the bobbin 15 along the flange 15b in the direction perpendicular to the axis of the bobbin 15. Further, when the wire 19 is shifted from the turn 1-n to the turn 2-a, the wire 19 is wound along the preceding turn while being kept in contact with the bobbin 15 during the period of the winding operation corresponding to the first one quarter turn of the bobbin 15. The wire 19 is placed on the preceding turn 1-n for more than one quarter turn of the bobbin 15, and is corrected in position by the tension of the wire 19 to take the shortest path. Thus, the portion of the wire which has been wound in contact with the bobbin 15, is also lifted out of the surface of the bobbin, and a gap is formed between the surface of the bobbin 15 and the end portion of the turn 1-n in such a manner as being gradually enlarged. At this time, the rollers 20a and 20b stop, and are left as they are till the signal is delivered from the counter 112 by more one half turn of the bobbin 15. During the period when the rollers 20a and 20b stop, the wire 19 is wound along the flange 15b with one half of the cross section of the wire 19 placed on the turn 1-n. When the signal is delivered from the counter 112, the rollers 20a and 20b begin to move in the direction opposite to the direction for the first layer. When the bobbin 15 further effects one half turn, the supply position of the wire 19 is moved to a position 2-b'. Thus, the wire 19 slides down the turn 2-a, and is wound along the turn 2-a to form the turn 2-b. When the wire 19 is shifted from the turn 2-a to the turn 2-b, the wire 19 pushes the portion of the wire at which the wire 19 is shifted from the turn 1-n to the turn 2-a, reduces the previously-mentioned gap, and takes the shortest path. Subsequently, the wire 19 is wound in the same manner as the first layer to form the second layer. It should be noted that the portion of each turn of the second layer which is parallel to the flanges 15a and 15b is placed in valleys formed by respective portions of all turns of the first layer which are parallel to the flanges 15a and 15b.

The above-mentioned winding operation is performed till the predetermined number of turns is obtained. When the number of turns reaches the number N set in the number-of-turn setting circuit 101, the number-of-turn counter 125 delivers the signal indicating that the winding operation is completed, and the signal is applied to the gate circuit 113. Simultaneously, the motor 8 is turned off, and therefore the rotation of the bobbin 15 is stopped. The gate circuit 113 is disabled upon the application of the signal from the counter 125 to prevent the pulses from the rotational-frequency-dividing-ratio counter 111 from passing through the gate circuit 113. Thus, the stepping motor 29 is turned off, and the rollers 20a and 20b stop.

In the above-mentioned manner, the wire 19 is wound round the bobbin 15 in a desired number of turns to form a perfect layer coil.

According to the present invention, a perfect layer coil can be formed irrespective of the non-uniformity in diameter of the wire 19 and in width of the bobbin 15. Further, the winding operation can be automatically performed from the initial stage thereof and therefore is high in speed, and moreover the coil formed has the highest possible density of wire.