Method of inspecting an array of solder ball connections of an integrated circuit module

• US 5,574,801 A
• Filed: 08/12/1994
• Issued: 11/12/1996
• Est. Priority Date: 08/12/1993
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Collet-Beillon, Olivier
• Primary Examiner(s): RAZAVI, MICHAEL
• Assistant Examiner(s): Patel, Jayanti K.
• Application Number:
  US08/289,972
• Time in Patent Office: 823 Days
• Field of Search: 382/150, 382/256, 382/144, 382/145, 382/146, 382/147, 382/148, 382/149, 382/141
• US Class Current: 382/150
• CPC Class Codes: G01N 21/95684 Patterns showing highly reflecting parts, e.g. metallic elements
• CPC Class Codes: G01R 31/311 of integrated circuits G01R31/31728 takes precedence

Abstract

A method of inspecting an array of balls used as connections in integrated circuit modules such as Solder Ball Connection modules, by means of an inspection apparatus that includes a microprocessor (34), a support (32) for holding in place the modules to be inspected (30), the microprocessor controlling a vertical camera (38) and a tilted camera (40) for obtaining images of the balls, wherein said method includes measuring the X and Y coordinates of each ball of the array to determine a best fitting grid of the balls, detecting the Z coordinate of each ball to determine the best fitting plane for the array of balls, offsetting the best fitting plane such that the offset plane also includes the lowest ball of the array, computing the deviation between each ball and the offset plane, and comparing the computed deviations with predetermined specifications to ascertain whether the inspected module is in compliance with the specifications.

Claims

1. A method of inspecting balls within a ball grid array, the balls used as connections of integrated circuit modules, using an inspection apparatus having a microprocessor, a support for holding in place the modules to be inspected, wherein positioning of the support is controlled by the microprocessor, at least one camera connected to the microprocessor, the camera providing images of the balls, the balls being illuminated by a light source, said method comprising the steps of:

measuring a centrality of the grid array by taking views of successive adjoining sub-arrays of said grid array and combining said views to form a composite image of the module;

from said composite image, measuring the X and Y coordinates of each of the balls forming said grid array to determine a best fitting grid of said balls;

determining the Z coordinate measured from the top of each of said balls to define a best fitting plane applicable to said array, based on said best fitting grid;

offsetting said best fitting plane so that the plane includes the lowest ball of said array of balls;

computing a deviation in the Z direction between each of said balls of said array and said offset plane; and

comparing the computed deviation against a predetermined specification and discarding as defective any said modules exceeding said first deviation and said deviation in the Z direction.

2. The method as recited in claim 1, wherein the step of determining the best fitting grid of said balls further comprises the steps of:

measuring the position of each of the balls of said array for each of said modules;

determining from the position of each of the balls the best fitting grid to best match all the balls of said array of balls; and

computing the deviation in the x and y directions between each of said balls and its position within said best fitting grid.

3. The method as recited in claim 2, wherein said step of measuring the position of each of said balls in said array further comprises taking successive juxtaposed views with the camera of sections of said array of balls, each section comprising a sub-array of m by n of said balls.

4. The method as recited in claim 1, wherein the step of determining the Z coordinate of each of said balls of said array further comprises taking views with a tilted camera of successive juxtaposed rows of said array of balls, wherein said balls are illuminated by the light source placed opposite to said camera, and wherein the image of said balls taken by said camera corresponds to a light reflection at the top of said ball.

5. The method as recited in claim 1, wherein said step of determining the best fitting plane for said array of balls further comprises applying a Least Square Regression technique to the computation of parameters that define said plane so as to minimize the sum of squared distances between the plane and the balls.

6. The method as recited in claim 4, wherein the Z coordinate measured at the top of each of said balls and detected by the tilted camera is corrected by determining a deviation of the position of each of said balls with respect to a predetermined position in said best fitting grid.

7. The method as recited in claim 6, wherein prior to the step of detecting the Z coordinate of each of said balls, the relative position of the module with respect to the vertical camera is adjusted to account for the thickness of the module.

8. The method as recited in claim 1, further comprising a calibration step that includes a coarse and a fine calibration, said calibration being performed prior to inspecting the module, wherein a scale and a drift required by the inspection are defined.

9. The method as recited in claim 8, wherein said coarse calibration further comprises obtaining an image of said array of balls and computing an average of the deviation between said balls, wherein said balls are combined two by two to define a pixel scale.

10. The method as recited in claim 9, wherein said fine calibration further comprises:

determining the drift resulting from a misalignment between an axis of said camera and an axis of the support by computing the difference between the variation between two successive views of said balls taken by the camera and the slope of the module with respect to the support;

determining the X and Y scales by computing the ratio of the distance between said balls captured by said camera in one of said views to the distance between balls in two of said views, and

determining the Z scale by using the angle 0 between the vertical plane and the image plane when a view is taken with a tilted camera.

11. A method of inspecting balls within a ball grid array, the balls used as connections of integrated circuit modules, using an inspection apparatus having a microprocessor, a support movable in the X and Y directions controlled by said microprocessor, first and second cameras coupled to said microprocessor for providing images of the balls, said method comprising the steps of:

a) measuring a centrality of the grid array by taking views of successive adjoining sub-arrays of said grid array and combining said views to form a composite image of the module;

b) combining said views to form an image of said array of balls;

c) from said image, measuring the position of each said balls, computing a best fitting grid and a first deviation in the X and Y directions between each of said balls and said best fitting grid;

d) determining the height of said balls with said second camera tilted with respect with said module, taking views of successive rows, while said balls are illuminated by a source of light placed opposite to said wherein an image of said balls in each said views corresponds to light reflections at the top of said balls;

e) computing a best fitting plane formed by said tops of said balls;

f) offsetting said best fitting plane by an amount determined by the top of the lowest of said balls in said array of balls to form an offset plane;

g) computing a deviation in the Z direction between said best fitting plane and said offset plane; and

h) comparing said first deviation and said deviation in the Z direction with predetermined specifications, and discarding as defective any said modules exceeding said first deviation and said deviation in the Z direction.

12. The method as recited in step d) of claim 11, further comprising the step of correcting said height of said tops of said balls by adjusting said height in accordance with said first deviation.

13. The method as recited in step d) of claim 11, wherein said first camera is adjusted to account for the thickness of said module.

14. The method as recited in claim 11, wherein calibration is performed prior to said inspection to quantify a drift that occurs while taking said views.

15. The method as recited in claim 14, wherein said calibration includes a coarse and a fine calibration.

16. The method as recited in claim 15, wherein said coarse calibration includes creating an image of said array of balls and subsequently computing a deviation in the X and Y directions between two of said balls by combining said balls two by two to determine an approximate pixel scale.

17. The method as recited in claim 15, wherein said fine calibration further comprises the steps of:

determining said drift resulting from a misalignment between said first camera and said support by computing a difference between a variation from one of said views to the next of said views taken with said first camera and the tilt of said module with respect to said support;

determining X and Y scales by computing a ratio of the distance between balls of one of said views over the distance between balls of two of said views; and

determining a Z scale by using an angle formed by a vertical plane and an image plane determined when a view is taken with said second camera.