EP0510027A1 - System for mounting components on circuit boards - Google Patents

Abstract

An apparatus for installing the finest components on printed circuit boards includes a component gripping stage where a gripping head (13) grasps a component in a component gripping area, a second stage which includes the case if necessary a video inspection camera (154) which examines the component in order to check whether it can be positioned, and a positioning stage with a positioning head (20) which is used to install the component on a printed circuit board. The three floors operate concomitantly. The positioning stage comprises a mechanism (202) for transporting and clamping the pads with adjustable width movable along a longitudinal horizontal axis and a positioning head mounted on a hollow shaft laterally and horizontally movable along a carriage (21), so as to move the positioning head vertically and rotatably about a vertical axis. The positioning head carries the component over an upward facing camera (208) so as to obtain positioning information. A support (12) underlying the movable plate under the positioning head supports the printed circuit board directly below the installation position of the component. A conveyor (6) transports plates through the system and tabs (200) lift the plates to the transport and clamping mechanism. Other plates can pass through the system, passing under the plate on which the components are installed.

Description

SYSTEM FOR MOUNTING COMPONENTS ON CIRCUIT BOARDS

TECHNICAL FIELD

This invention relates to apparatus for surface mounting components onto printed circuit boards (PCBs). More specifically, the invention relates to a system which combines component acquisition, inspection and placement in a highly efficient and effective manner.

The system is especially suitable for use with fine pitch components. Fine pitch components are those electrical components typically intended for surface mounting on printed circuit boards and having very small spacing . between leads. Surface mount technology has recently become well known, and typically involves reflow soldering of leads to the surface of the printed circuit board, rather than passing the leads through holes in the board.

Certain terms, well understood in the industry, are used in the following description. Nevertheless, the following definitions are provided for possibly greater clarity. ^■■Excising" refers to the process of stripping the component from its tape backing or other mounting means. "Forming" refers to forming the leads to the right shape to contact the printed circuit board. "Placement" refers to positioning of the component at the right place on the circuit board and with the correct orientation. "Soldering" of course refers to the actual fastening of the component to the circuit board, by reflow soldering.

BACKGROUND ART

Systems which combine various ones of the above functions, i.e. component acquisition, inspection and placement, are known in the prior art. However, the invention offers various advantages over the prior art systems in terms of speed and accuracy. DISCLOSURE OF INVENTION

It is an object of the invention to provide improved apparatus for the surface mount installation of components, especially fine- pitch components, on printed circuit boards.

In the invention, the apparatus includes a component acquisition stage, an optional inspection stage, and a placement stage. The stages operate concurrently such that while the placement stage is installing a component, the inspection stage (if applicable) is inspecting the next component to be installed, and while the inspection stage is inspecting that next component, the component acquisition stage is acquiring yet another component. Thus, system cycle time is substantially reduced and is dependent only upon the longest step in the process. The inspection stage tends to be the slowest, and can be bypassed or operated on an intermittent sampling basis if desired, to achieve greater speed at the expense of some certainty of accuracy. The placement stage uses a "split axis" concept involving a three-axis manipulator, and a single-axis board positioning subsystem. Motions of the board carriage are independent of the placement head positions thus allowing the placement head and the board carriage to be moved simultaneously and reducing the time required for precise positioning of the board. The main axis uses a precision lead screw and linear encoder arrangement. The circuit board is held in a rigid supporting frame and moves in one axis under the placement head through a similar lead screw and linear encoder arrangement.

The system was developed to help ease the transition from today's mixed technology boards to the highly populated low real estate surface.mount boards that are becoming increasingly prevalent in the electronics industry. The system concept was developed with the - 3 - primary focus of meeting the objectives of high precisi movements for placement of fine pitch leaded and/ leadless components, and high quality workmanship.

Due to its combination of economy and speed, t system serves as an excellent platform for both entry lev and advanced surface mount placement of leaded and leadle components.

The system provides the high precision movemen required for placement of fine pitch components. A wi variety of component feeders can be accessed due to t flexibility of the component acquisition stage.

Further features of the invention will described or will become apparent in the course of t following detailed description.

BRIEF DESCRIPTION OF DRAWINGS

In order that the invention may be more clear understood, the preferred embodiment thereof will now described in detail by way of example, with reference the accompanying drawings, in which: Fig. 1 is a perspective view of the overa system;

Fig. 2 is another perspective view of the syste in less detail, showing the external guarding; Fig. 3 is a plan view of the system; Fig. 4 is another plan view of the system;

Fig. 5 is an elevation view of the system; Figs. 6 through 9 are elevation views of t conveyor sub-system;

Fig. 10 is an elevation view showing t mechanism for adjusting the width of the conveyor;

Fig. 11 is a plan view corresponding to Fig. 9 Fig. 12 is an elevation view showing t component acquisition stage; Fig. 13 is another elevation view of the component acquisition stage, at 90 degrees to Fig. 12;

Fig. 14 is a perspective view of the placement stage; Fig. 15 is a top view of the placement head area;

Fig. 16 is a sectioned first side view of a first portion of the quill assembly;

Fig. 17 is a sectioned second side view of the first portion of the quill assembly, the view extending normal to the view of Fig. 16;

Fig. 18 is a sectioned side view of that part of the first portion of the quill assembly that is adapted to secure an article to the arm mechanism;

Fig. 19 is a cross-sectional view of the part of the first portion of the quill assembly shown in Fig. 18, the view being taken on the line 19-19 in Fig. 18;

Fig. 20 is a first cross-sectional view of the first portion of the quill assembly, the view being taken on the line 20-20 in Fig. 16; Fig. 21 is a second cross-sectional view of the first portion of the quill assembly, the view being taken on the line 21-21 in Fig. 16;

Fig. 22 is a third cross-sectional view of the first portion of the quill assembly, the view being taken on the line 22-22 in Fig. 16;

Fig. 23 is an elevation view of the placement stage, showing an underboard support;

Fig. 24 is a side elevation view corresponding to Fig. 23; and - Fig. 25 is a plan view of the underboard support.

BEST MODE FOR CARRYING OUT THE INVENTION Overview

The preferred embodiment of the invention provides a system which is capable of excising, forming, inspecting, placing and soldering fine pitch components onto printed circuit boards.

The system has a base 1 to which all cell tooling and the placement robot framework is mounted. The base 1 has a fabricated welded steel frame 2 on which is mounted a top plate 3 with an antistatic surface such as MicaStat (trademark) laminate. The top plate is drilled to accept the mounting of all component feeders. All pneumatic and electrical connections are enclosed under the table top and are accessible via hinged access doors.

As shown in Fig. 2, the entire system with the exception of the entry and exit conveyor, the electrical panels and the CRT terminals are enclosed in safety interlocked bronzed guarding 4, made from Lexan (trademark) for example. An emergency stop palm button is located outside of the cell guarding as well as a keylock switch to allow maintenance personnel to bypass the cell guarding when required.

Physical configuration of the system is such that it presents a relatively small system footprint of approximately 1.8 x 1.8 meters (6x6 feet) exclusive of the space required for feeders. The system can be linked to existing lines in a fully pass-through configuration allowing one board to be processed while other boards pass through the system.

The system has the following main components or subsystems, all directed and sequenced by a personal computer 5 such as an AT-compatible controller:

(1) PCB conveyor: A dual belt PCB conveyor, which moves boards through the placement stage and allows pass-through board handling.

(2) Component installation system: A four-axis placement platform is provided for placement and "hot bar" reflow soldering of fine pitch components onto the surface of the circuit board. The system includes a component acquisition stage, an inspection stage, and a placement stage.

The subsystems and components will now be described in turn.

Conveyor System

As mentioned above, the system has been configured as a pass-through system which connects directly into the customer's existing PCB transport system. A dual belt PCB conveyor 6 provided with a manual handcrank 8 permits manual width adjustment to accommodate board sizes ranging from about 5 cm. (2 inches) up to about 61 cm. (24 inches) wide. Where required the handscrew can be replaced with an optional servo motor 9 to allow automatic programmable adjustment of the conveyor width.

The circuit boards 7 are supported by the edges on thin conveyor belts 6, which are independently driven section to bring boards into the system working envelope. The conveyor passes through the system between a single axis board carriage 10 and an underboard support 12. The board carriage includes a board clamping arrangement 202, as described in greater detail below. Boards are lifted from the conveyor to the clamping mechanism by tongues 200, as also described in greater detail below. The conveyor belts are routed along the inside of machined anodized aluminum rails 14. The rails function as rigid stress supporting members as well as a mounting rail for associated accessories such as board sensors and soft stops. The conveyor can be quickly modified or reconfigured with an entirely new accessory complement in a very short time simply by unfastening the mounting bolts and sliding the conveyor accessories to their new location.

Boards are queued up on the conveyor, and one board at a time is released into the system. The conveyor is provided with all required queue stops and board locating tooling required to ensure that circuit boards are held in queue until required and then released one at a time into the system work envelope, to be processed. As seen in Figs. 5-9, the board to be processed is lifted from the conveyor by tongues 200 into a rigid board carriage which holds the board by its edges using a pneumatic clamp arrangement 202.

As seen in Figs. 7 and 11, the width is adjustable to accommodate boards of different widths. Figs. 8 and 9 showing the lifting and clamping operation, with the tongues 200 being lifted by cylinders 203 against the upper reference stops 205. Actuation of the cylinders 207 rotates the clamping mechanism 202 as shown in Fig. 9. Since the circuit board is top referenced against the stops 205, board thickness is not a consideration and does not have a detrimental effect upon placement.

Component Acquisition Stage

As seen in Figs. 12 and 13, the component acquisition stage has a two-axis servo-driven pick and place unit 121, which travels the length of a hardened steel track 122, guided by cam followers 125 on rail 127. The pick and place unit picks up the component from the feeder and deposits it on an inspection platform, as described later. The length is about 1.2 meters (48 inches). Additional lengths may be added according to the number of feeders that must be accessed. The pick and place unit is equipped with a pneumatically-actuated Z-axis pickup head in the form of a quill 130 including a vacuum gripper 124 at the bottom thereof, on a vertically compliant slide 126, constituting the third axis. The pick and place unit is driven into position by DC servo motors which are controlled by their own servo controller cards. Each axis is driven by a DC servo motor. The motors provide high accuracy in a small lightweight package, thus minimizing inertia and allowing the system to operate at high speeds.

The quill 130 is moved to position its vacuum gripper 124 in line with the component pickup or dropoff point. The gripper is then extended and lowered into the feeder for component pickup. The gripper is provided with vacuum parts presence sensors indicating to the system controller that a part has been correctly picked from the tooling.

The component acquisition stage as illustrated in the Fig. 1 system configuration is capable of accessing the following: a. Lead form and excise tool 132 for preparation and presentation of parts packaged in coinstack carriers 134. b. Reject tray 136. c. Vision inspection pedestal 138 of the inspection stage. The Fig. 1 embodiment has a coinstack carrier feeder 134 which incorporates lead excise and forming tooling 132. A component is excised, the leads are formed, and the component is then picked up by the gripper for transfer to the vision inspection pedestal 138 of the inspection stage.

Other component feeding and presentation alternatives are available according to the particular components that must be handled. That is, the system accommodates a variety of different types of feeders positionable for access by the quill 130. For example, Fig. 3 shows a large number of reel-type component feeders 123 and a tray feeder 119.

Component Feed_f Excise and Form Tool _ 9 _

For fine pitch components, it is desirable t perform lead excise and forming operations directly fro the carrier in which the part was presented. By performin lead excising, lead forming and acquisition in one syste and indeed in the same stage, component handling an resulting lead damage potential is greatly reduced.

A tool 132 which combines the lead excise forming and feeding operations on one common platform i provided in the Fig. 1 embodiment. One tool is require for each different device window size. Each die set i dedicated to the feeding of one window size only Different lead pitches and pin counts within tha particular window can be accommodated using the same di tooling. In one type of feeder, components are presente to the component acquisition stage stacked vertically in a aluminum extrusion. The operator loads the extrusion ont the feeder base and locks it into position. The feede base is provided with low level sensors to indicate to th operator when replenishment of the parts stock is required.

The excise/form tool 132 uses a pneumaticall powered miniature C-frame press 140. Forming operation are performed by the upper die set 142. One carrier at time is moved from the bottom of the stack out into th excise/form position 144. The component and carrie assembly are accurately located by tooling pins in th lower die set. The press lowers to excise the componen leads from the carrier. Further downward movement of th press forms the leads as required. As the press retracts to the home position, th vacuum quill 130 removes the component from the lower di 143 for transfer to the vision inspection pedestal 138 o the inspection stage. After removal of the component fro the tooling, the scrap carrier is expelled into a bin. Inspection Stage

From the component acquisition stage, the component is deposited onto the vision inspection pedestal

138, which has upwardly facing vacuum quills 139 to hold the component in place. Which quill is to be used is selected depending on the size of the component.

The pedestal 138 is mounted on a programmable micro positioning X-Y table 152 which precisely moves the component past two downward-looking area cameras 154 mounted on an overhead rotary turntable 156. The X-Y table has two programmable axes each with an approximate 25 cm. (ten inch) programmable travel and extremely high repeatabi1i y.

The optics incorporated in the vision system allow each camera to see a top view of the component as well as a side elevation view of the lead tips in space. Illumination is provided by LED light sources arrayed in such a manner that various rows and columns of LEDs can be programmably switched on and off according to the illumination pattern required for each particular component. By using red LEDs and red filters on all cameras, the system's susceptibility to ambient lighting conditions is eliminated as ambient lighting normally has only a Very small component of red energy. A precision camera reticule is utilized for camera calibration. The reticule incorporates a precision grid of squares photo-etched onto a glass plate. By viewing the reticule with the camera, any optical distortions in the camera lens are evident in the displacement of the grid lines with respect to the vision system coordinates. A set of translating coordinates are developed to transform the image seen by the vision system to real world coordinates based on the known spacing of the reticule pattern. This mapping increases the accuracy of the vision system. After components are loaded onto the pedestal, the components are moved to the location of the downward- looking cameras. A "snapshot" is taken of the area of the leads on one side of the component, the camera rotates 90 degrees, the component is moved to position another side under the cameras, and another snapshot is taken. The process is repeated until all four sides have been viewed. This inspection may take place on every component, or on an intermittent sampling basis. A separate 386 compatible computer is provided for control of the vision system and motion control of the X-Y table 152. Camera data is collected and assembled for display on a CRT monitor by vision imaging boards installed into the computer backplane. Camera data is analyzed by high level routines in the C programming language.

The vision routines perform lead inspection including lead splay, lead toe in, toe out, lead count, coplanarity and lead presence. Based on the data received as a result of these inspections, "pass/fail" decisions are made, and component offsets for pickup are then passed to the IC controller.

Component inspection algorithms are data driven such that new components can be accommodated simply by making a local data file entry specifying major component attributes such as lead count, lead spacing, and toe-to-toe dimensions. Parts are not taught as such, but rather a generic software model of the components to be processed is provided allowing new components to be learned from design data alone. Component inspection and positioning algorithms are implemented in a layered fashion such that some parts of the inspection process can be disabled, or executed on a sampling basis only.

Components rejected by the vision inspection stage are picked by the placement head and deposited onto a reject shuttle which moves the component back within reach of the component acquisition stage. When the acquisition stage has an empty quill and there is sufficient time available, the rejected part is transferred to a reject tray 136. By operating in this manner, operator attention is not required to remove a rejected part from the placement quill and the reject cycle does not affect the system throughput.

The component position within the reject tray is recorded and the failure data corresponding to each component is written to disk for later retrieval.

Upon completion of the inspection process, components are positioned by the vision pedestal under the carriage 21 for pickup by the placement head of the quill assembly 20 for placement. If no vision inspection is to be performed on the component, or if the system is configured without vision inspection, the component is still transferred to the pedestal 138, for subsequent pickup by the placement head.

Placement Stage The placement stage uses a "split axis" concept involving a three-axis manipulator 20 and a single-axis board positioning subsystem, all directed and sequenced by a personal computer such as an AT-σompatible controller. Incoming boards are positioned in the X axis while the placement head simultaneously moves overhead along the Y axis. The placement head also contains a Z and theta axis resulting in a total of four degrees of freedom for the system. A swivelling ball and socket head at the end of the placement head actually provides even further freedom of movement, and is lockable in the desired orientation for correct placement.

The swivelling ball and socket head has a compliant vacuum cup to automatically compensate for parts that are not planar. The cup is relaxed such that it is free to self-compensate for any top surface planarity errors as the component is held stationary by the vision inspection pedestal. Upon lowering for pickup, the cup is locked into position to bottom reference the component for all subsequent operations including transport and placement.

Since the bottom surface of the component is used as the reference surface for lead forming and excise operations, picking components in this manner allows an implied reference to be made with the component leads without actually touching or damaging the leads. This compliant technique also can be used to handle components with heat sinks mounted to the top surface.

Components are picked from the pedestal by the quill assembly 20 and are then moved in a straight line past an upward looking camera 208. As the camera field of view is filled strobed "snapshots" are taken of the component to corner register the component leads.

"Snapshot" data is collected by an imaging board which is installed into the system controller backplane. Camera data is analyzed by high level software routines.

"Snapshots" are taken of the component held in the placement quill while it is in movement without slowing the motions of the placement head. The placement head has previously examined local placement fiducials with its camera and the board positioning system has been moved to the proper position for placement. The component is then lowered to the surface of the board.

Extremely high precision motions are critical to the design of the X, Y and theta axes. However, the same design constraints do not exist for the vertical motions of the placement head. Typical placement of fine pitch and . TAB devices requires the component to be lowered until contact with the board is indicated by tactile sensors rather than being lowered to a preprogrammed height. Although the same degree of precision is not required for this axis, economies of scale can be achieved by using the same motors, encoders, and drivetrain as used for the other linear axes of the machine. A system of this nature is only as good as the components that make up the whole. Thus top quality lead screws and linear motion guides must be used, all machined to very fine tolerances. The system preferably uses preloaded antifriction linear motion guides between all critical moving surfaces. These bearings use four rows of recirculating balls which provide high payload capacities and smooth rolling properties. All bearing surfaces and impact points should be hardened and heat treated to achieve precision, straightness and flatness. Differences in ambient as well as operation- induced temperatures can potentially cause calibration errors. By mounting the linear encoders away from the motors, heat generated by the motors does not affect the encoder readings. Software enhancements may be used to actively measure the temperature of the linear encoder scales and to compensate for their expansion or contraction accordingly.

Linear expansion along major machines axes has minimal effect on performance, as the system uses machine vision to seek out the final position of local placement fiducials. The only distance over which the vision system is unable to compensate for thermal expansion is the distance between the placement quill and the downwards looking camera 204 mounted adjacent thereto. Since this distance is only about 10 cm. (four inches), thermal expansion over that length has a minimal effect on placement accuracy.

The system is equipped with two ; independent downwards force mechanisms including a light placement force for attachment of components onto the surface of the board and a stronger force used to drive the heater bars against the board.

The placement head incorporates a load cell which bears against the vacuum cup, this load cell providing control of the vacuum cup placement pressure from 0 to 500 grams in 5 gram increments.

The placement head is equipped with heater bars used to hold component leads onto the surface of the board and to heat the leads to the reflow temperature. These heater bars can apply downwards pressure onto the surface of the board of from 0 to 10 lbs in .25 lb steps.

The placement head will now be described in greater detail. A quill assembly generally designated as 20 is mounted for horizontal movement on a supporting carriage generally designated as 21. Quill assembly 20 is moved on carriage 21 by rotation of a ball screw 22 which passes through a complementary thread within quill assembly 20. A motor 19 within an end housing 23 rotates ball screw 22. A set of guide members 24 and 25, each extending parallel to ball screw 22, are used to stabilize the movement of quill assembly 20.

The quill assembly 20 is movable between the pick-up position at the right end of carriage 21 in Fig. 14, i.e. from the inspection stage, and the placement position at which quill assembly 20 sits in Fig. 14. Assembly 20 picks up a component 27 at the pick-up position by means of a swivelling suction cup member 29, as shown in outline in Fig. 16 and more fully described subsequently, and then carries that package to the deposition position. A support platform 30, movable in a direction normal to carriage 21, is adapted to carry a printed circuit board on which component 27 is to be mounted.

The ball screw 22 extends through a frame member

31 of quill assembly 20 and engages a complementary thread within that member, as earlier mentioned. The frame member 31 has a profile that is complementary to the profile on the guide members 24 and 25. Mounted on frame member 31 by a band 32 is a motor 33 which has its rotor connected in¬ line with a ball screw 34 which extends vertically. With the exception of frame member 31, motor 33, and ball screw 34, the remaining parts of quill assembly 20 move vertically as a single unit; that unit has a pair of guide members 36 and 37 each fitted to slide within a respective one of a pair of vertical grooves 38 and 39 in frame member 31.

An outer cylindrical housing 42 of quill assembly 20 has an inner cylinder 43 mounted concentrically within it. A motor 44 is mounted to an upper portion of outer housing 42 such that the rotor of motor 44 extends parallel to the axis of inner cylinder 43. A gear wheel 46 connected to the rotor of motor 44 is positioned such that its teeth mesh with the teeth of a gear wheel 47 connected to inner cylinder 43. Inner cylinder 43 also has connected to it a disc 48 with radial calibrations markings, as shown in Fig. 14. An optical scanning unit 49 is mounted to housing 41 to sense the angular position of inner cylinder 43; that angular position can then be modified by actuation of motor 44.

Inner cylinder 43 of quill assembly 20 is shown enlarged in Figs. 16 and 17; those figures do not show the gear 47 or disc 48 that are connected to the upper end of cylinder 43. Connected to the lower end of cylinder 43 is a heater bar assembly 55 including floating heater bars 56, similar to the type which has been generally described in U.S. Patent No. 4,894,506 entitled "Method and Apparatus for Reflow Soldering of Electrical Component Leads, including Floating Heater Bar", or in British patent application ser. no. 90.04247, entitled "Floating Heater Bar and Guide". The lower end of heater bar assembly comprises four spring-mounted heater bars 56 arranged such that each bar 56 forms one of the sides of a rectangle. A series of heater bar assemblies of differing sizes may be used, the particular assembly in use at any given time depending on the outer dimensions of the component 27 being carried on quill assembly 20. Each assembly 55 is sized such that each of its four heater bars 56 is adapted to extend across all of the gullwing-shaped leads 58 that extend from a respective one of the four sides of the corresponding component 27. As shown in Figs. 16 and 17, a tube 60 runs along the inner surface of inner cylinder 43. A motor 61 is supported on a cross-member 62 connected to one end of tube 60. A coupling 63 connects the rotor 64 of motor 61 to an in-line sleeve shaft 65 which extends symmetrically from one end of a cylindrical sleeve 66, the sleeve 66 thus being rotatable with rotor 64 by motor 61. Sleeve 66 is positioned within tube 60 by a pair of cross-members 67, each having a bronze ring 68 sitting between it and sleeve 66. The other end of sleeve 66 is fitted with a ball nut 69 through which extends a complementary ball screw 70. Rotation of sleeve 66 causes ball screw 70 to move in or out of sleeve 66.

The ball screw 70 has a threaded end portion (not shown) which extends through a hole centrally positioned in a plate 71, a nut 72 fitted on the threaded end portion securing plate 71 to ball screw 70. A pair of guide rods 75 are press fitted into a yoke 76. Each guide rod 75 extends through a hole in a respective opposite end of plate 71. A nut 77 is secured to a threaded end portion of each guide rod 75 to limit movement of plate 71 relative to yoke 76. The upper end of a spring 81 is fitted over the threaded end portion of ball screw 70. The lower end of spring 81 is fitted over a threaded rod (not shown) which extends out of the top side of yoke 76, the bottom end of spring 81 abutting a nut 82 on the threaded rod. Slide member 85 is fitted into a complementary aperture in the lower side of yoke 76. A load cell 86 is fitted to the upper end of slide member 85 to sense pressure exerted by the one on the other. The nut 82 secures a plate 87 to the top of the yoke 76. A pair of springs 88 each extend between a respective opposite end of the plate 87 and a respective opposite end of a rod 89 extending centrally through the slide member 85; the springs 88 are utilized for maintaining a positive pressure on load cell 86. Four guide supports 90 extend in parallel spaced relation to each other in a generally square configuration for restricting yoke 76 to movement parallel to tube 60. The four supports 90 are integrally connected to one end of a rectangular beam 91 to form a support structure, that structure being positioned within tube 60 by a pair of cross-members 92 and 93.

The slide member 85 is slidably attached to a first portion 94 of a pair of slide bearings by a first series of bolts 95. The first portion 94 of each slide bearing and a second portion 96 of the bearing each have a facing elongated channel within which are mounted a series of roller bearings (not shown) . The second portion 96 of each slide bearing is attached to the beam 91 by a second series of bolts 97. The series of circles 98 shown in outline adjacent to the perimeter of cross-member 67 in Fig. 21 represent wires that extend between inner cylinder 43 and tube 60 for supplying electrical current to the heater bars 56.

The slide member 85 is connected at its lower end to extend in-line with a cylindrical element 100, the connection being by means of a threaded rod 101. A bore 102 extends centrally through the lower portion of element 100, the upper end of bore 102 being connected through a fixture 103 to a flexible vacuum hose 104. Fitted to the slide member over element 100 is a hollow cylinder 106. The one end of cylinder 106 has a integral annular flange 107 which is secured to an annular ring 108 by a series of screws 109. An annular piston 110, which is press fitted to cylindrical element 100,. is sandwiched between flange 107 and ring 108. The cylinder 106 moves relative to cylindrical element 100 whenever a pressure differential exists between the annular chambers 111 and 112. Annular sealing rings 113, 114, and 115 provide seals for air entering chambers 111 and 112 through the fixtures 116 and 117.

The lower ends of cylindrical element 100 and cylinder 106 are contoured to form a spherical chamber within which sits a spherical member 120. Fitted into spherical member 120 is a suction cup member 29. The bore 102 is in flow communication with a flow channel and vacuum chamber within cup member 29. When cylinder 106 assumes a first position relative to cylindrical element 100, spherical member 120 and cup member 29 are free to rotate; when cylinder 106 assumes a second position, spherical member 120 is prevented from rotating. By applying a differential air pressure to the fixtures 116 and 117 the suction cup member 29 can be maintained in a fixed orientation.

The package 27 sometimes has a top surface that is not parallel to its bottom surface. For instance, if a heat sink defines the top of the package the epoxy may not be applied evenly between the heat sink and the remainder of the package. When quill assembly 20 is lowered onto package 27 in the pick-up position described earlier, cylinder 106 is positioned relative to cylindrical element 100 such that cup member 29 is free to rotate to orient its suction surface to extent parallel with the top surface of package 27. Variation of the air pressure entering one or both of fixtures 116 and 117 then results in cylinder 106 moving relative to cylindrical element 100 to prevent spherical member 120 and cup member 29 from any further rotation. When the package 27 is subsequently deposited on a printed circuit board resting on the support platform 30, the bottom surface of the package will extend parallel to the surface of that board.

The sequence of operations of the placement stage is as follows. The quill assembly 20 is moved to the pick¬ up position, such that it sits directly above the package 27 that is being picked up. The pressure into apertures 116 and 117 is such that spherical member 120 and cup member 29 are free to rotate. The motor 33 is then actuated to move the vertically-movable portion of quill assembly 20 toward package 27. The suction surface of cup member 29 contacts the top of package 27, and cup member 29 rotates until the suction surface is flush with the top surface of package 27. As housing 41 continues downward, the spring 81 becomes compressed and exerts increasing pressure on load cell 86. Load cell 86 transmits its signal to the control system of the apparatus, which in turn actuates motor 61 to rotate sleeve 66. Screw ball 70 then moves into sleeve 66 to relieve the pressure on load cell 86. This arrangement is calibrated such that only a limited force of typically approximately four ounces is experienced by package 27. It is also calibrated such that motor 33 stops within a small distance of the expected contact with package 27. Each of the heater bars 56 is then adjacent one of the sides of package 27.

Then, the pressure into apertures 116 and 117 is modified such that cup member 29 is prevented from rotating, and suction is applied to hose 104 for holding package 27 on cup member 29. The motor 33 is then actuated to raise the vertically-movable portion of quill assembly 20 and package 27, after which ball screw 22 is rotated to move quill assembly 20 such that it sits above the position on the printed circuit board at which package 27 is to be deposited.

The motor 33 is then actuated to lower the vertically-movable portion of quill assembly 20 toward the printed circuit board. The package 27 comes into contact with the printed circuit board, and spring 81 begins to compress. Load cell 86 senses the increased pressure and sends a signal to the control system of the apparatus. Motor 61 is then actuated to maintain the limited force on package 27. The heater bars 56 continue to move downward around the periphery of the package 27. Prior calibration of the apparatus results in the motor 33 stopping as soon as the heater bars 56 begin to press the leads 58 on package 27 against the circuit board. The spring-mounting of each of the heater bars 56 prevents damage to the leads and circuit board. The heater bars 56 are then activated to momentarily reflow solder the leads 58, each lead 58 being thereby soldered to a respective metal trace on the circuit board. The suction is then removed from hose 104, and the motor 33 is actuated to raise the vertically- movable portion of quill assembly 20; as it rises, the pressure on load cell 86 is reduced and motor 61 is automatically actuated to move screw ball 70 out of sleeve 66 to return the pressure on load cell 86 to its equilibrium value. The pressure into apertures 116 and 117 is then adjusted to allow cup member 29 to again freely rotate. The quill assembly is then in a state ready for fetching another package 27.

The stage is equipped with flux and adhesive dispensers 35. A dispensing nozzle (not shown) is used for dispensing of flux onto the surface of the board. Flux is dispensed by a programmable dispenser. This tool . is provided with a height sensing pneumatically actuated probe on a compliant vertical slide. The probe is lowered until contact with the board is made and sensed by a sensor incorporated into the vertical slide. The placement head then raises by a known height offset from the top of the board. In this manner the dispense height is always at a known height offset from the height of the board surface. Where circuit boards are subject to warpage this procedure is performed repeatedly.

By splitting the axes of the gantry such that both the board and the placement head move independently to achieve relative X-Y motion, both major axes (X and Y) can be mounted to the system base. Thus, the moving mass and motor sizes are kept to a minimum and positioning errors are reduced. With the lower inertia that can be obtained with this arrangement, the speed of movement is correspondingly enhanced. Unlike typical prior art systems where rotary encoders may be mounted to the drive shaft of the motor, and errors can be introduced in the drivetrain linkage between the motor and the driven mechanism, direct reading linear encoders are used on the X and Y axes. (A rotary encoder is used for the Z axis, on the motor shaft.) By using linear encoders mounted at the point of movement rather than rotary encoders mounted directly to the driving motor, the system can compensate for flexibility and backlash errors, and resolutions that allow a final placement accuracy of typically ± .01 mm. (approximately 0.005 inches) are probably attainable.

The accuracy of the placement quill rotational axis is critical to the overall placement accuracy of the system. This axis is provided with a direct reading glass scale encoder with 18,000 lines per revolution. By recording the position and directions of the encoders output zero crossing points, mathematical interpolations can be performed to achieve a fourfold increase in resolution. When bearing eccentricity is taken into account, calculations yield a conservative resolution of .004.

Each axis is driven by DC servomotors. The particular motors chosen provide high accuracy in a small lightweight package thus minimizing inertia and allowing the system to operate at high speeds. The main axis uses a precision lead screw and linear encoder arrangement. The circuit card is held in a rigid supporting frame and moves in one axis under the placement head through a similar lead screw and linear encoder arrangement.

As the placement accuracy is a function of working area size, the system design minimizes the envelope of each stage. The X axis has a 61 cm. (24 inch) range of travel, the Y axis is 122 cm. (48 inches) and the Z axis is limited to a 10 cm. (4 inch) stroke while the rotate or wrist axis is 370 degrees. This working envelope allows board sizes up to 61 cm. x 61 cm. (24 inches x 24 inches) to be easily accommodated. The placement head can move its rated payload of 12 lbs at speeds in excess of 150 cm. (about 60 inches) per second.

Underboard support

Referring particularly to Figs. 23 - 25, mounted to the system base, under the board carriage, is a servo driven programmable underboard support 12 which minimizes board deflection during the placement of components. The support consists of a support tool 12 mounted on a single axis table 162 driven by a servo motor 163. The support is raised until contact is made with the bottom of the board and is then locked into position. This prevents possible flexing of the board as the heater bars are lowered onto the component leads and pressure is applied.

After the board is clamped into the carriage, other circuit boards can pass through the system while the underboard support is retracted, thus providing a parallel material handling capacity without the use of diverting conveyor spurs.

Hot Bar Reflow Soldering Tooling

When the component has been placed onto the surface of the board and is held in position by the placement quill, four heater bars (or fewer, depending on the component being installed) simultaneously lower onto the leads on all four (or fewer) sides of the component to press and hold the leads on the board during the solder reflow process.

This method of reflow soldering has been used for several years now and provides a proven technology to locally heat the component leads to the point of solder reflow without detrimentally affecting nearby components. The required items to perform hot bar reflow soldering consist of: a soldering head with four independent heater bars specifically configured to suit the size of components to be reflow soldered, a four channel reflow power supply to provide the current required to bring the heater bars to the reflow temperature, and a tool rack for exchange of component specific tools as required.

The soldering head is a high precision instrument that allows automatic end effector removal by utilizing two locating pins and a locking mechanism. Electrical and pneumatic connections are incorporated for the vacuum quill gripper and each individual heater bar. Connections are automatically made with each end effector exchange.

Each of the four individual bars incorporates its own thermocouple for temperature measurement and a spring loaded compliant pivot to allow compensation for warped circuit boards. As contact is made with the board, each heater bar complies against a light spring force allowing the heaters bars to conform to the actual plane of the board. Downwards pressure of the soldering head is automatically programmable from 0 to 10 lbs in .25 lb increments.

A further downward movement of the soldering head applies a predetermined load on the heater bars. The power supplies are energized, thus heating the bars and reflowing the solder pads to complete the solder joint.

Co-planarity correction is inherent as all leads are forced to the board plane, by virtue of the rocking feature of the heater bar or heater bar support. The amount of time and soldering temperature are programmably set. Typically the correct settings are determined by exhaustive trials and tests. After the programmed amount of time has elapsed and the heater bar temperature has dropped below the solder freezing point, the placement head is lifted from the device and cleared off the board.

Reflow Power Supply

The power supply has four pulse width modulated outputs for controlling four current transformers. Temperature is accurately controlled using a feedback loop that monitors the heater bar temperature using a thermocouple mounted to the heater bar.

Unique circuitry within the power supply is able to automatically detect thermocouple discontinuities. Temperature and power output are continuously indicated on the front panel via a quad 20 segment bar graph. A serial interface is provided for remote control.

A lock-out feature is incorporated to prevent remote firing of the controller when being manually operated. The reflow power supply is packaged in a 48 cm.

(19 inch) rack mount enclosure with individual removable modules that make both installation and servicing easy.

The channel (transformer control) module in the system is self contained in its operation, and from one to four of these modules can be installed into the rack.

The controller can be set and operated in three different manners: manually from the front panel, remotely from a host computer (using a serial communications line), or remotely using discrete I/O (such as a programmable controller or robot controller) . All three methods of operation are available concurrently.

Increasingly important to manufacturers is the ability to record all process variables sent to the power supply and the capability of measuring actual values attained to quickly determine problem areas of the process.

The basic reflow power supply does not gather data on the actual temperatures or applied power as a function of time. An optional add-on board can be provided to allow SPC tracking of process variables which can be stored offline or downloaded to a nearby printer for later analysis. This card when installed can sample, store, and download temperature and power information from up to four control channels. The period of time over which data is taken can be programmed.

Heater Bar Cleaning Station

Although the heater bars are made from a material that prevents solder wicking, the bars may still require cleaning at periodic intervals to remove flux buildup. An automatic heater bar cleaning station can be used which consists of an abrasive disk to remove contaminants as the stage moves the bars across the surface. Typically one or two passes over the disk removes the buildup of contaminants from the bar surface. Cleaning intervals are programmable and can be altered at any time. Alternatively, ultrasonic cleaning or any other suitable means may be employed. Controls Description System Controls

Control of the system and peripheral associated devices is directed by two AT-compatible computers. Each computer performs a specific task including axis motions and sequencing, vision stage control and data storage and messaging. Two CRT monitors are provided, one to display vision system data and the other to display operating system menus. Menus and choices are available to allow the download, upload, storage and editing of board data. Additional menus are provided to display and control the status of cell devices and to control the start of assembly sequences. Menus are provided for display and operator access to the operating system. The system accepts operator commands and displays or processed them to perform such functions as editing, storing, loading, downloading and debugging. System control is distributed among two 20 MHz

80386 AT computers. The cell control computer is equipped with a 1.2 MB microfloppy removable drive, 2 MB of memory, and a 60 MB hard drive. The system is equipped with a EGA video adapter card and touch screen for display of system operating data and relevant operating parameters.

The operating system may be, for example, QNX version 3.13 which provides a real-time multi-tasking, multi-user environment with which to provide dynamic editing of data set files and production data files. The system is entirely menu driven with various screens that are configured to supply the required system information. Fields within each menu can be easily changed to update such information as number of leads and toe to toe distance in a parts data file. Menus include a package definition menu, board configuration menu, assembly data menu, and a menu of process parameters such as part placement pressure etc. The menu editor alerts the operator if out of range or invalid data has been entered. Once all required changes to this information have been made the file is saved and becomes a new placement program. In this way all normal maintenance, product changeovers and control of the system are performed without the need for programming knowledge.

During normal operations of the system minor faults such as low parts levels, or failure of the system to pick a part from a feeder may occur. Sensors are provided throughout the system to detect faults of this nature. When a fault is sensed a Help menu appears which explains the fault and displays a message listing possible recovery methods. The number of possible error messages is limited by the number and placement of sensors.

The system is provided with vision data frame grabbers and a distributed control network (DCN) for control of the four axes of the system, the two axes of the component acquisition stage and the single axis of the underboard support. Sensors are provided for all critical functions or areas where maintenance is normally required.

There are a number of significant benefits to using distributed control over centralized control. A dedicated processor for a small number of I/O points ensures high performance. Secondly, with many control functions taken care of at the local level, the central computer has additional processing power that can be applied to arithmetic calculations. The data set for the system is stored on the hard disk media in flat-file format. The data set is loaded at the start of a given board. Data is subdivided as required by individual units of the system. The data set resides in memory and is acted upon by the system. The disk resident data set can then be edited as production is run. The data set relates the following information for the system:

Part form factor: allows the system to determine the suitable processing requirements for a part. Part type: source feeder number, board place location referenced from tooling hole location or local fiducial location.

Fiducial data: fiducial form factor, fiducial dimensions, fiducial relative location to lead pads.

Dispense data: patterns, dispense volumes Place data: z-force requirements, offset data for part location relative to either local fiducial, or board tooling holes. Test data: accept/reject criteria for all passive components

Vision data: lead inspection criteria. A screen is provided for monitoring process variables. A file is maintained to reflect the status of the system and is updated as changes occur. Production summary data is displayed to the system monitor during a production run and is also written to disk for later access. This file is updated once each board.

System Accuracy Overall practical component placement accuracy depends on a number of significant factors as follows (in order of magnitude) : a. Predictability of component slip at the PCB surface while complying to the seating plane b. Manufacturing accuracy of PCB fiducials c. Artwork plotting errors and photo-reproduction distortions d. Design data rounding errors e. Variation in physical shape and symmetry of similar components f. Thermal variation of fiducial items

Since the majority of cumulative placement error is attributable to process items as opposed to system technology, the vision system resolution and system accuracy can be taken to be sufficient.

All bearing surfaces and impact points are hardened and heat treated to achieve precision, straightness, and flatness.

General

It will be appreciated that the above description relates to the preferred embodiment by way of example only. Many variations on the invention will be obvious to those knowledgeable in the field, and such obvious variations are within the scope of the invention as described and claimed, whether or not expressly described.

INDUSTRIAL APPLICABILITY

As disclosed above, the invention has utility in the field of mounting components onto printed circuit boards.

Claims

CLAIMS :

1. In apparatus for the installation of fine pitch components onto printed circuit boards, comprising a component acquisition stage in which a pickup head (130) picks a component to be installed from a component acquisition area, a second stage which receives said component from said component acquisition stage, and a placement stage having a placement head (20) which picks up said component from said second stage for installing said component on a printed circuit board (7), the improvement characterized by said stages operating concurrently such that while said placement stage is operating to install a first component, said second stage is holding or if applicable operating on a second component for installation after said first component, while said component acquisition stage is acquiring a third component for installation after said first and second components, and so on for successive components intended to be installed on said board.

2. Apparatus as recited in claim 1, further characterized by said second stage comprising means (138) for receiving said component from said component acquisition stage at a first location and delivering said component to a second location for pickup by said placement head of said placement stage.

3. Apparatus as recited in claim 1, further characterized by said second stage comprising vision inspection means (154) for determining the suitability of said component for installation.

4. Apparatus as recited in claim 1, further characterized by said placement stage comprising circuit board clamping and carriage means (202) moveable along a longitudinal horizontal axis, and said placement head being mounted on a quill moveable laterally and horizontally along a carriage (21), said placement head being moveable vertically and rotatable about a vertical axis by said quill.

5. Apparatus as recited in claim 4, further characterized by an upward-looking camera (208) between said second stage and the circuit board location of said component installation stage, in the same vertical plane as the plane of movement of said placement head, such that said placement head passes said component picked up from said second stage over said camera, whereby said camera may record positioning information with respect to said component to ensure accurate placement on said circuit board.

6. Apparatus as recited in claim 4, further characterized by an underboard support (12) positioned beneath said circuit board clamping and carriage means and moveable along a lateral axis beneath said placement head between the edges of said clamping and carriage means, and moveable vertically upwardly and downwardly to and from the plane of the underside of the circuit board, whereby the circuit board may be supported directly beneath the component installation location during said installation.

7. Apparatus as recited in claim 4, further characterized by said board clamping and carriage means comprising two rails (14) spaced laterally apart from each other, each supporting clamping means, the spacing between said rails being adjustable to accommodate different circuit board widths.

8. Apparatus as recited in claim 4, further characterized by conveying means (6) for conveying circuit boards through said apparatus, and lifting means (200) for lifting selected ones of said circuit boards to said board clamping and carriage means, said lifting means being retractable to permit said conveying means to convey other boards through said apparatus below a circuit board held in said clamping and carriage means, if so desired.

9. In apparatus for the installation of fine pitch components onto printed circuit boards, comprising a component acquisition stage in which a pickup head (130) picks a component from a component acquisition area, a second stage optionally having automated vision equipment (154) for inspecting said component to ensure placeability, and a placement stage having a placement head (20) for installing said component on a printed circuit board (7), the improvement characterized by said stages operating concurrently such that while said placement stage is operating to install a first component, said second stage ^' is holding or if applicable is inspecting a second component for installation after said first component, while said component acquisition stage is acquiring a third component for installation after said first and second components, and so on for successive components intended to be installed on said board, said component acquisition stage passing acquired components to said second stage, and said placement head picking up components from said second stage.