JP3678916B2 - Non-contact 3D measurement method - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention provides a non-contact image measurement function for measuring a contour shape or the like of an object to be measured from an image obtained by imaging a workpiece with an imaging means such as a CCD camera, and a distance between the measurement surface of the workpiece as a displacement amount. The present invention relates to a non-contact three-dimensional measurement method effective when measuring a precision component such as an IC package using a non-contact three-dimensional measurement apparatus having a non-contact displacement detection function for detecting contact.
[0002]
[Prior art]
Conventionally, an image measuring device has been used for measuring the contour shape of precision parts. The image measuring device images a workpiece to be measured at an arbitrary magnification using a CCD camera, detects an edge from the obtained two-dimensional image, and obtains a coordinate value of a necessary portion using various measurement tools. Is. When performing three-dimensional measurement including the height direction of the workpiece with this image measurement device, focus determination is performed from the contrast of the image on the measurement surface, and this focus position is set as a height direction position.
[0003]
[Problems to be solved by the invention]
The coplanarity (flatness) of the BGA (Ball Grid Array) of the LSI package is a major factor that determines the yield. For this reason, an apparatus capable of measuring the flatness of the BGA with high accuracy is desired. In the conventional image measuring apparatus, the position in the height direction (Z-axis direction) is measured by autofocus. Therefore, there is a restriction that the position to be measured must be parallel to the XY plane. Only the average value of the set area is measured by the focus. Therefore, contour shape measurement or surface roughness measurement that measures fine irregularities in the Z-axis direction cannot be performed.
[0004]
The present invention has been made in view of these points, and an object of the present invention is to provide a non-contact three-dimensional measurement method capable of measuring three-dimensionally the flatness of a BGA of a fine LSI package with high accuracy.
[0005]
[Means for Solving the Problems]
The non-contact three-dimensional measuring method according to the present invention includes an imaging unit that images a workpiece made of an IC package and outputs two-dimensional image information for image measurement, and a distance from a predetermined measurement point on the workpiece as a displacement amount. Three-dimensional point sequence data is obtained by moving an imaging unit in which a non-contact displacement meter that can be detected is maintained in a fixed positional relationship within the measurement three-dimensional space, and based on this three-dimensional point sequence data A non-contact three-dimensional measuring method for measuring the flatness of a ball grid array of the IC package, the step of measuring an image of the IC package using the imaging means, and based on a measurement result obtained by the image measurement along said specified measurement track non-contact displacement meter, and the vertex scanning pattern specified a predetermined range including the apex of each ball of the ball grid array A step of acquiring three-dimensional point sequence data by scanning measurement and a step of executing a flatness evaluation process on the point sequence data obtained in this step. 0006
According to the present invention, when the non-contact displacement meter measures the displacement amount of the workpiece along the measurement trajectory designated using the two-dimensional image information obtained by the imaging means, the non-contact displacement meter Since a predetermined range including the apex of each ball of the BGA is measured according to a specified apex pattern, the measurement trajectory of the non-contact displacement meter is shifted from the center position of the ball when the apex of the ball is not at the center of the circle Even in this case, by performing scanning over a predetermined range, vertex coordinates can be detected with extremely high probability, and flatness evaluation of BGA can be performed with high accuracy.
[0007]
DETAILED DESCRIPTION OF THE INVENTION
Preferred embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a perspective view showing the overall configuration of a non-contact three-dimensional measuring apparatus according to one embodiment of the present invention.
This apparatus includes a coordinate measuring machine 1 having a non-contact image measuring function and a non-contact displacement measuring function, and a computer system 2 that drives and controls the coordinate measuring machine 1 and executes necessary data processing. It is configured.
[0008]
The three-dimensional measuring machine 1 is configured as follows. That is, a measurement table 13 on which a workpiece 12 to be measured is placed is mounted on the gantry 11, and this measurement table 13 is driven in the Y-axis direction by a Y-axis drive mechanism (not shown). Support arms 14 and 15 extending upward are fixed to the center of both side edges of the gantry 11, and an X-axis guide 16 is fixed so as to connect both upper ends of the support arms 14 and 15. An imaging unit 17 is supported on the X-axis guide 16. The imaging unit 17 is driven along the X-axis guide 16 by an X-axis drive mechanism (not shown). The computer system 2 includes a computer 21 that performs measurement information processing and various controls, a keyboard 22 that inputs various instruction information, a joystick box 23 and a mouse 24, a CRT display 25 that displays a measurement screen, an instruction screen, and measurement results, And a printer 26 that prints out the measurement result.
[0009]
The inside of the imaging unit 17 is configured as shown in FIG. That is, the slider 31 is provided so as to be movable along the X-axis guide 16, and the Z-axis guide 32 is fixed to the slider 31 integrally. A support plate 33 is provided on the Z-axis guide 32 so as to be slidable in the Z-axis direction. The support plate 33 is provided with a CCD camera 34 as an imaging means for image measurement, and a laser probe as a non-contact displacement meter. 35. As a result, the CCD camera 34 and the laser probe 35 can move simultaneously in the three-axis directions of X, Y, and Z while maintaining a fixed positional relationship. An illumination device 36 for illuminating the imaging range is added to the CCD camera 34. In order to confirm the measurement position of the laser probe 35 by the laser beam, a CCD camera 38 that images the periphery of the measurement position and an illumination device 39 for illuminating the measurement position of the laser probe 35 are located near the laser probe 35. And are provided. The laser probe 35 is supported by a vertical movement mechanism 40 for retracting the laser probe 35 when the imaging unit 17 is moved, and a rotation mechanism 41 for adapting the directivity of the laser beam to an optimum direction. .
[0010]
FIG. 3 is a diagram showing details of the laser probe 35.
The light emitted from the semiconductor laser 51 passes through the beam splitter 52 and the quarter-wave plate 53 and is then converted into a parallel light beam by the collimator lens 54 and passes through the mirrors 55 and 56 and the objective lens 57 to measure the workpiece 12. To form a light spot. The light reflected from the measurement unit of the workpiece 12 follows the reverse path of the mirrors 56 and 55, the collimating lens 54 and the quarter wavelength plate 53, is reflected by the beam splitter 52, and is divided into two parts by the edge mirror 58. . The vertically divided light is detected by the two-divided light receiving elements 59 and 60 arranged above and below. The detection circuit 61 outputs a signal corresponding to the amount of deviation from the focal position of the objective lens 57 to the measurement surface 62 of the workpiece 12 based on the output signals from the two-divided light receiving elements 59 and 60. The servo circuit 63 outputs a drive signal for driving the objective lens 57 to the drive mechanism 64 based on the detection output of the detection circuit 61. When the objective lens 57 moves up and down, the movable member 67 of the displacement detector 66 moves with respect to the fixed member 68. This movement amount is output as a displacement amount.
[0011]
FIG. 4 is a block diagram of the entire apparatus showing the configurations of the coordinate measuring machine 1 and the computer system 2 in more detail.
In the CMM 1, image signals obtained by imaging the workpiece 12 with the CCD camera 34 for image measurement and the CCD camera 35 for confirming the measurement position of the laser probe 35 are A / D converters 71 and 72, respectively. Then, one of them is selected by the selection circuit 73 and supplied to the computer 21. Illumination light necessary for imaging by the CCD cameras 34 and 38 is given by the illumination controllers 74 and 75 controlling the illumination devices 36 and 39 based on the control of the computer 21, respectively. A displacement amount signal obtained from the laser probe 35 is supplied to the computer 21 via the A / D converter 76. Then, the imaging unit 17 including these is driven in the XYZ axis direction by an XYZ axis driving unit 77 that operates based on the control of the computer 21. The position of the imaging unit 17 in the XYZ axis direction is detected by the XYZ axis encoder 78 and supplied to the computer 21.
[0012]
On the other hand, the computer 21 is stored in the multi-value image memory 81, the CPU 81 that is the center of control, the multi-value image memory 82 connected to the CPU 81, the program storage unit 83, the work memory 84, and the interfaces 85 and 86. The display control unit 87 is configured to display multi-value image data on the CRT display 25. The CPU 81 switches the selection circuit 73 between the image measurement mode and the laser measurement mode. The multi-value image data for image measurement or the multi-value image data for laser measurement selected by the selection circuit 73 is stored in the multi-value image memory 82. The multi-value image data stored in the multi-value image memory 82 is displayed on the CRT display 25 by the display control operation of the display control unit 87. On the other hand, operator instruction information input from the keyboard 22, joystick 23, and mouse 24 is input to the CPU 81 via the interface 85. Further, the CPU 81 captures the displacement detected by the laser probe 35, XYZ coordinate information from the XYZ axis encoder 78, and the like. The CPU 81 executes various processes such as stage movement by the XYZ axis driving unit 77 and calculation processing of measured values based on the input information, the operator's instructions, and the program stored in the program storage unit 83. The work memory 84 provides a work area for various processes of the CPU 81. The measured value is output to the printer 26 via the interface 86.
[0013]
Next, measurement processing and data processing of the non-contact three-dimensional measuring apparatus according to the present embodiment configured as described above will be described. This apparatus has an image measurement mode and a laser measurement mode. In the image measurement mode, the same operation as that of the conventional image measurement apparatus is performed, so here the laser measurement mode will be described.
[0014]
FIG. 5 is a flowchart showing a procedure for scanning measurement in the laser measurement mode. First, the image measurement image and the laser probe 35 are calibrated (S1). That is, a jig 91 including two non-parallel linear components L1 and L2 that can be measured by the CCD camera 34 and the laser probe 35 as shown in FIG. For example, the jig 91 may be such that a trapezoidal pattern 93 having a predetermined width h is arranged on a substrate 92. By measuring the straight lines L1 and L2 in the projection plane in the Z-axis direction by the CCD camera 34 and the laser probe 35, respectively, the equations of these straight lines are obtained, and the obtained formulas are processed, whereby the CCD camera 34 and the laser probe An offset value between the 35 coordinate axes is obtained, and this offset value is used as position calibration data for the CCD camera 34 and the laser probe 35.
[0015]
When the calibration process is finished, next, the image of the workpiece 12 is measured to confirm the position of the workpiece 12, and the measurement point by the laser probe 35 is moved to the measurement start point (S2). During the image measurement, there is a possibility that the laser probe 35 interferes with the workpiece 12. Therefore, during the image measurement, the laser probe 35 is retracted upward by the vertical movement mechanism 40. The control is performed by, for example, an air cylinder. Next, when the laser measurement mode is selected (S3), the selection circuit 73 is switched, and the screen of the CRT display 25 changes from the CCD camera 34 to the screen of the CCD camera 38 for laser measurement. From this screen, the position (measurement position) of the laser beam spot from the laser probe 36 is confirmed (S4). Here, the position of the beam spot can be finely adjusted using the joystick 23, the mouse 24, or the like. The CCD camera 38 is for confirming whether or not the laser beam spot correctly hits the target position on the workpiece 12, and the image data is not used for measurement. For this reason, it is not necessary to have a high definition like the CCD camera 34 for image measurement. In addition, only the position of the laser spot appears bright with only the laser beam, and the surrounding area becomes dark and a beautiful image cannot be obtained. Of course, the CCD camera 38 and the illumination device 39 can also be used as the CCD camera 34 and the illumination device 36.
[0016]
Next, in order to provide a scanning measurement path, a measurement tool is selected and necessary parameters are set (S5). As a measurement tool, for example, the one shown in FIG. 7 can be considered.
(A) Point tool Measures the X, Y, and Z coordinate values of the current measurement point (black circle).
(B) A linear tool end point position Pe is given, and the measurement is performed by copying along the straight line from the current measurement point to the end point Pe.
(C) Given the width W, height H, and pitches PT1 and PT2 for the area tool area search, the scanning measurement is performed while reciprocating within the designated area at the designated pitch from the current measurement point.
(D) A circular tool radius R, a pitch PT, and a start angle θ are given, and a concentric circle is measured from the current measurement point.
(E) A rectangular tool width W and a height H are given and measured along the rectangle.
(F) Cross tools The lengths L1 and L2 of two line segments orthogonal to each other are given, and the cross is measured along the cross.
(G) Giving a spiral tool maximum radius R and a pitch RT (radius value increasing with one rotation), and measuring a spiral.
(H) Focus tool Simply focus at the current position.
[0017]
When a measurement tool is selected and necessary parameters are set, scanning measurement is executed (S6). The displacement detection accuracy of the laser probe 35 has a slight direction. For this reason, when measuring the contour and the surface roughness along the trajectory, the laser probe 35 is rotated by the rotation mechanism 41 so that the laser probe 35 is oriented in an optimum direction with respect to the traveling direction of the trajectory. When measuring along a circular or spiral trajectory, it is more effective to measure while rotating the laser probe 35.
[0018]
In the scanning measurement, the amount of displacement from the laser probe 35 is set so as to obtain a coordinate value in the Z-axis direction within a measurement range of the laser probe 35, for example, exceeding ± 0.5 mm, as shown in FIG. Based on this, the XYZ-axis drive unit 77 is driven to move the imaging unit 17 up and down in the Z-axis direction. Thereby, control is performed so that the in-focus position of the laser probe 35 is always at the center of the measurement range. In this case, the Z-axis coordinate value obtained by the XYZ-axis encoder 78 is the displacement amount in the Z-axis direction. In order to perform high-speed measurement so that the position control in the Z-axis direction is not in time, the correct displacement amount may be calculated by correcting the Z-axis coordinate value with the displacement amount of the laser probe 35. When measuring a minute surface roughness within the measurement range of the laser probe 35, as shown in FIG. 9, the Z-axis direction position of the laser probe 35 is fixed, and the objective lens 57 in the laser probe 35 is fixed. In this case, higher-speed processing is possible, and measurement with higher resolution (eg, 0.01 μm) than resolution by Z-axis drive (eg, 0.1 μm) is possible. Become. By such scanning measurement, coordinate values in the Z-axis direction are obtained as point sequence data along with the X and Y-axis coordinate values at predetermined intervals along the designated measurement trajectory, and stored in the work memory 84. When the point sequence data is obtained, the point sequence data is analyzed (S7).
[0019]
Next, analysis processing of point sequence data will be described. The conventional contour shape measuring machine and the surface roughness measuring machine are two-dimensional data, whereas the contour shape measuring data obtained by this non-contact three-dimensional measuring apparatus is three-dimensional data. In addition, since the scanning measurement is performed along the designated two-dimensional trajectory, the data processing becomes more complicated. Therefore, in order to simplify the data processing, the following point sequence data analysis processing is executed. Based on the flowchart of FIG. 10 and the waveform diagram of FIG. 11, this point sequence data analysis processing will be described.
[0020]
First, since the workpiece 12 itself may be inclined, an average surface (an average line in the case of a straight line) is obtained from the point sequence data, and data trend correction is executed on this surface (S11). As a result, the trend-corrected point sequence data shown in FIG. 11B is obtained from the tilted point sequence data shown in FIG. Next, the three-dimensional point sequence data is converted into two-dimensional point sequence data with the traveling direction along the measurement trajectory as the first axis direction and the normal direction of the average plane as the second axis direction (S12). Thereby, data as shown in FIG. 11C is obtained. Since this point sequence data is obtained by scanning with acceleration / deceleration along the measurement trajectory, it is not a constant pitch. If the pitch is not constant, the Gaussian filter processing normally used in FFT (Fast Fourier Transform), shape measuring machine, etc. cannot be executed, so here the constant pitching processing is executed (S13: FIG. 11 ( d)). Next, Gaussian filter processing is executed (S14: FIG. 11 (e)). Then, the fixed pitch data is returned to the original position (indefinite pitch position) (S15: FIG. 11 (f)). Next, two-dimensional → three-dimensional conversion is performed to return the two-dimensional data converted in step S12 to the original measurement trajectory position (XY position) (S16: FIG. 11G). The inclination correction processed by the data trend correction in S11 is returned to the original, and the data is converted in the direction in which the workpiece 12 is originally inclined (S17: FIG. 11 (h)).
[0021]
Through the above processing, it is possible to easily filter the three-dimensional point sequence data. Further, since the data after the processing in step S14 is two-dimensional data filtered at a constant pitch, various analysis processes such as those performed by a normal contour shape measuring machine or surface roughness measuring machine are possible. become.
[0022]
Next, a procedure for evaluating coplanarity (flatness) of an LGA (Land Grid Array) and a BGL (Ball Grid Array) of an LSI package as an example of the work 12 will be described.
FIG. 12 is a plan view and a longitudinal sectional view of the LSI packages 101 and 102 as the work 12. In the case of the LSI package 101 having the LGA of FIG. 10A, the height of the land 103 is used, and in the case of the LSI package 102 having the BGA of FIG. Using this method, point sequence data is obtained by reciprocating linear scanning. And based on the obtained point sequence data, as shown in FIG. 13, the coplanarity of the vertex part or the valley bottom part is evaluated.
[0023]
FIG. 14 shows a procedure for evaluating the coplanarity of LGA.
Here, since the process from the trend correction of the data in step S21 to the execution of the Gaussian filter in step S24 is the same as the process described in FIG. 10, the description is omitted here. The filter constants necessary for this processing are set using a parameter setting viewer such as shown in FIG. Presence / absence of trend correction, filter high frequency, low frequency cutoff frequency, calculation conditions, etc. can be set.
[0024]
FIG. 16 shows two-dimensional point sequence data (b) obtained by executing filter processing from two-dimensional point sequence data (a) obtained by scanning the LGA and having a constant pitch. From this point sequence data, the center position of each valley (Land) is extracted using the valley detection algorithm to obtain point cloud data (FIG. 16C) (S25). As shown in FIG. 17, the Yamatani algorithm gives a minimum depth h and a minimum pitch p from the reference position in advance, and the point sequence data existing in the range from the most valley portion to the minimum depth h is only the minimum pitch p. A valley portion is detected by judging whether or not it continues. This method has the advantage that valleys can be detected without having a design value file.
[0025]
As shown in FIG. 18, the number Nx of the land in the horizontal direction, the number Ny in the vertical direction, the scanning start positions Stx and Sty, the pitch Px in the horizontal direction, and the pitch Py in the vertical direction are prepared in advance as design data. If so, the valley point cloud data may be extracted using this design data.
[0026]
When point cloud data as shown in FIG. 16C is obtained by such processing, this is converted from 2D point cloud data to 3D point cloud data (S26), and finally the coplanarity is calculated (S26). S27).
[0027]
Similarly, when evaluating the coplanarity of BGA, the process shown in FIG. 19 may be executed. The process from the data trend correction (S31) to the Gaussian filter execution process (S34) is the same as the process from steps S21 to S24 in FIG. 14, and after the filter process, the apex of each ball is detected using the Yamatani detection algorithm. What is necessary is just to detect. In this case, the minimum height h and the minimum pitch p are given, and a mountain that meets the condition is detected. Subsequent processing (S36, S37) for the extracted point cloud data is the same as steps S26 and S27 in FIG.
[0028]
However, in this BGA detection, as shown in FIG. 20A, the apex (the highest portion) of the ball 104 may not be located at the center of the circle. Further, as shown in FIG. 5B, the laser probe 35 has an allowable tilt angle of, for example, ± 13 °, and the reflected light of the laser beam is not detected and correct measurement cannot be performed due to the shift of the scanning position. There is. With this in mind, for example, as shown in FIG. 21A, it is desirable to perform zigzag scanning on the upper surface of the ball 104, or spiral scanning as shown in FIG. Specifically, for example, XY coordinates of each ball 104 are prepared by design values or image measurement, and as shown in FIG. 22, the stage is stopped at the point a within the allowable tilt angle range (laser probe 35). ) And the pattern ab is measured with a vertex scanning pattern as shown in FIG. Between bc, it moves at high speed without performing measurement. FIG. 23 is a diagram showing a setting screen for setting each scanning parameter and scanning pattern of such a grid array.
[0029]
【The invention's effect】
As described above, according to the present invention, when the non-contact displacement meter measures the displacement amount of the workpiece along the measurement trajectory designated using the two-dimensional image information obtained by the imaging means, Since the non-contact displacement meter measures and measures a predetermined range including the vertex of each ball of the BGA according to the designated vertex pattern, the measurement trajectory of the non-contact displacement meter is not at the center of the circle, Even in such a case, it is possible to detect vertex coordinates with a very high probability by performing scanning over a predetermined range, and it is possible to evaluate the flatness of the BGA with high accuracy.
[Brief description of the drawings]
FIG. 1 is a perspective view of a non-contact three-dimensional image measuring apparatus according to an embodiment of the present invention.
FIG. 2 is a perspective view of the inside of an imaging unit in the apparatus.
FIG. 3 is a view showing a configuration of a laser probe in the apparatus.
FIG. 4 is an overall block diagram of the apparatus.
FIG. 5 is a flowchart showing a laser measurement procedure performed by the apparatus.
FIG. 6 is a diagram for explaining a calibration method of an image and a laser probe in the apparatus.
FIG. 7 is a diagram showing an example of a measurement tool used in the apparatus.
FIG. 8 is a diagram for explaining an example of scanning measurement of the apparatus.
FIG. 9 is a diagram for explaining another example of the scanning measurement of the apparatus.
FIG. 10 is a flowchart of point sequence data analysis processing of the apparatus.
FIG. 11 is a diagram for explaining the analysis processing;
FIG. 12 is a diagram showing an LSI package which is an example of a workpiece.
FIG. 13 is a diagram for explaining coplanarity measurement of BGA of the same package.
FIG. 14 is a flowchart showing a procedure of LGA coplanarity evaluation.
FIG. 15 is a diagram showing a parameter setting screen for performing the evaluation.
FIG. 16 is a diagram showing data obtained at each point of the evaluation process.
FIG. 17 is a diagram for explaining a Yamatani algorithm used in the evaluation process.
FIG. 18 is a diagram for explaining the same evaluation process using design data.
FIG. 19 is a flowchart showing a procedure of BGA coplanarity evaluation.
FIG. 20 is a diagram for explaining problems of BGA coplanarity evaluation;
FIG. 21 is a diagram showing an example of a vertex scanning pattern used in a scanning method for solving the same problem.
FIG. 22 is a diagram for explaining details of the scanning method.
FIG. 23 is a diagram showing a parameter setting screen for realizing the scanning method.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 1 ... CMM, 2 ... Computer system, 11 ... Mount, 12 ... Work, 13 ... Measurement table, 14, 15 ... Support arm, 16 ... X-axis guide, 17 ... Imaging unit, 21 ... Computer, 22 ... Keyboard 23 ... Joystick box, 24 ... Mouse, 25 ... CRT display, 26 ... Printer, 34, 38 ... CCD camera, 35 ... Laser probe, 36, 39 ... Illumination device.

Claims (3)

An image pickup means for picking up an image of a workpiece made of an IC package and outputting two-dimensional image information for image measurement, and a non-contact displacement meter capable of detecting a distance from a predetermined measurement point on the workpiece as a displacement amount with a certain positional relationship an imaging unit formed by arranging keeping to give a three-dimensional point sequence data by moving in the measuring three-dimensional space, measuring the flatness of a ball grid array of the IC package based on the three-dimensional point sequence data A non-contact three-dimensional measuring method,
Measuring the image of the IC package using the imaging means;
Based on the measurement result obtained by the image measurement, the non-contact displacement meter is copied along a designated measurement trajectory and in accordance with a designated vertex scanning pattern including a predetermined range including the vertex of each ball of the ball grid array. Obtaining three-dimensional point sequence data by measuring;
A non-contact three-dimensional measurement method comprising: a step of executing a flatness evaluation process on the point sequence data obtained in this step.   The non-contact three-dimensional measurement method according to claim 1, wherein the vertex scanning pattern is a zigzag pattern.   The non-contact three-dimensional measurement method according to claim 1, wherein the vertex scanning pattern is a spiral pattern.

Images