A fast testing and correction method for a thread needle testing machine

Abstract

The present invention discloses a method for fast testing and correcting a wire needle tester, including: step one: obtaining the test result of the ABCD point area; step two: according to the distribution of bad graphics in the ABCD area, calculating the direction and displacement of the correction movement, and performing the correction test; and also including the following steps: first: using the golden section ratio algorithm to correct the deviation, setting a cycle number, completing the correction within the cycle number means completing the test, and running the Archimedes screw algorithm if the test is not completed within the cycle number; second: using the Archimedes screw algorithm, setting a cycle number, completing the correction within the cycle number means completing the test, and if the test is not completed within the cycle number, the circuit board is judged to be unqualified and the test is terminated. It avoids the misjudgment of the test result caused by the manufacturing error of the circuit board, and at the same time adopts a combined algorithm, taking into account the characteristics of point search efficiency and completeness, and preferentially searching for points in the most likely direction, saving the test time and avoiding misjudgment.

Description

Technical Field

[0001] The invention relates to the technical field of needle testing machines, and in particular to a fast testing and deviation correction method for a wire needle testing machine. Background Art

[0002] The high-density wire probe tester uses a wire probe fixture to test the circuit board. During the test, the wire probes on the wire probe fixture need to form good contact with the tested points on the circuit board to ensure the true test results without misjudgment due to poor contact.

[0003] However, due to the many variables in the manufacturing process of the circuit board itself, such as repeated alignment, long process flow, alternating hot and cold during the production process, the circuit board will have errors such as relative benchmark offset and expansion and contraction deformation. During the test process, it will show a large number of defects in a certain area of ​​the circuit board, and the bad test results are highly correlated with the position information.

[0004] In this case, the equipment needs to fine-tune the test position to repeat the test on the circuit board and eliminate the influence caused by the above factors. How to effectively fine-tune the position and save the cycle time (equipment single cycle operation time) is a problem to be solved. Summary of the invention

[0005] The purpose of the present invention is to provide a fast test and correction method for a wire needle tester, which can effectively avoid the misjudgment of test results caused by circuit board manufacturing errors, and adopts a combined algorithm, taking into account the characteristics of point finding efficiency and completeness, and can prioritize point finding in the most likely direction, saving test time and avoiding misjudgment. It solves the problem that due to many variables in the manufacturing process of the circuit board itself, such as repeated alignment, long process flow, and alternating hot and cold changes in the production process, the circuit board will have errors such as relative reference offset and expansion and contraction deformation, which manifests in a large number of defects in a certain area of ​​the circuit board during the test process, and the bad test results are highly correlated with the position information.

[0006] To achieve the above-mentioned object, the present invention provides a method for quickly testing and correcting a thread needle tester as follows: the method for quickly testing and correcting a thread needle tester comprises:

[0007] Step 1: Get the test results of the ABCD point area to determine whether it is an occasional defect or a regional large-area defect. If it is an occasional defect, it is determined to be a circuit board defect; if it is a regional large-area defect, it is determined to be a misalignment defect and needs to be corrected;

[0008] Step 2: According to the distribution of defective patterns in the ABCD area, the direction and displacement of the correction movement are calculated according to the algorithm designed by the present invention, and the correction test is performed; the following steps are included;

[0009] First: Use the golden ratio algorithm to correct the deviation, set a cycle number, if the correction is completed within the cycle number, the test is completed, if the test is not completed within the cycle number, run the Archimedean screw algorithm;

[0010] Second: Use the Archimedean screw algorithm and set a cycle number. If the correction is completed within the cycle number, the test is completed. If the test is not completed within the cycle number, the circuit board is judged to be unqualified and the test is ended.

[0011] Step 3: Correction control parameters k1, k2, k1 is the number of repetitions of the golden section algorithm, k2 is the number of repetitions of the Archimedean spiral algorithm, and are initially set to a constant. The specific correction algorithm is as follows:

[0012] First: When an offset occurs, calculate the position of the graphic first, find the diagonal line, and perform an offset test on the most likely offset direction. The offset is 1 / n of the pad point size. If the test passes, it is OK and the subsequent process is carried out; if the test is NG, there are two situations:

[0013] (1) If a point that was originally OK becomes NG, it can be judged that the offset is excessive and needs to be offset back. The amount of the offset back is used to find the point using the golden ratio algorithm to improve the probability of finding the point. Specifically, the amount of the offset back is (1-0.618) times the previous offset;

[0014] (2) If the original OK points are still OK, and the original NG points are still NG, it can be determined that the offset is insufficient, and the offset is continued until it is OK or the offset is excessive. Each time the golden ratio algorithm is run, the k1 value is reduced by 1 until it reaches zero;

[0015] Second: When the k1 value is zero, run the Archimedean spiral algorithm, and use the original NG center as the center to perform Archimedean spiral point search. If the test result is OK, proceed according to the OK result; if NG, continue to search for points. Each time a point is searched, the k2 value is reduced by 1 until it is zero. Specifically, the parameters a and b are selected according to the empirical parameters.

[0016] If both K1 and K2 are zero, the correction algorithm ends and the product is judged to be NG.

[0017] Optionally, the adjustment process of the offset direction in step 2 is as follows:

[0018] When there are four test points ABCD in the figure, ABC is OK, and D is NG, because AC are all OK points, the most likely offset direction is the direction of the line connecting BD, so the BD line is taken as the offset direction.

[0019] Optionally, the ABCD test point is described above as a test point. In actual testing, a bad test result will appear in the form of an area. ABCD can also be an area. The position of ABCD in the middle of the area is calculated according to an algorithm, such as the geometric center of the area.

[0020] Optionally, the golden ratio algorithm in step three is specifically y=0.618x.

[0021] Optionally, the Archimedean spiral algorithm in step three is specifically y=a+bθ.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] The present invention can effectively avoid misjudgment of test results caused by circuit board manufacturing errors, and at the same time adopts a combined algorithm, taking into account the characteristics of point finding efficiency and completeness, and can preferentially find points in the most likely direction, saving test time and avoiding misjudgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a test flow chart of the present invention;

[0025] Figure 2 This is the NG point diagram for testing the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] Example 1

[0028] See also Figure 1 and Figure 2 The present invention provides a fast test and correction method for a thread needle testing machine: the fast test and correction method for a thread needle testing machine comprises:

[0029] Step 1: Get the test results of the ABCD point area to determine whether it is an occasional defect or a regional large-area defect. If it is an occasional defect, it is determined to be a circuit board defect; if it is a regional large-area defect, it is determined to be a misalignment defect and needs to be corrected;

[0030] Step 2:

[0031] According to the distribution of bad patterns in the ABCD area, the direction and displacement of the correction movement are calculated according to the algorithm designed by the present invention, and the correction test work is performed; the following steps are included:

[0032] First: Use the golden ratio algorithm to correct the deviation, set a cycle number, if the correction is completed within the cycle number, the test is completed, if the test is not completed within the cycle number, run the Archimedean screw algorithm;

[0033] Second: Use the Archimedean screw algorithm and set a cycle number. If the correction is completed within the cycle number, the test is completed. If the test is not completed within the cycle number, the circuit board is judged to be unqualified and the test is ended.

[0034] It is worth noting that the principle of the diagonal calculation method is that a large area of ​​defect is found in one of the four areas ABCD, while the other three areas are normal. For example, assuming that ABC is normal and D is defective, AC is diagonal and BD is diagonal in the graphic distribution. At this time, the diagonal direction of BD is determined to be the most likely deviation direction, and the diagonal direction of BD is taken as the approaching direction of movement.

[0035] Step 3:

[0036] The correction control parameters k1 and k2, k1 is the number of repetitions of the golden section algorithm, and k2 is the number of repetitions of the Archimedean spiral algorithm. They are initially set to a constant. The specific correction algorithm is as follows;

[0037] First: When an offset occurs, calculate the position of the graphic first, find the diagonal line, and perform an offset test on the most likely offset direction. The offset is 1 / n of the pad point size. If the test passes, it is OK and the subsequent process is carried out; if the test is NG, there are two situations:

[0038] (1) If a point that was originally OK becomes NG, it can be judged that the offset is excessive and needs to be offset back. The amount of the offset back is used to find the point using the golden ratio algorithm to improve the probability of finding the point. Specifically, the amount of the offset back is (1-0.618) times the previous offset;

[0039] (2) If the original OK points are still OK, and the original NG points are still NG, it can be determined that the offset is insufficient, and the offset is continued until it is OK or the offset is excessive. Each time the golden ratio algorithm is run, the k1 value is reduced by 1 until it reaches zero;

[0040] Second: When the k1 value is zero, run the Archimedean spiral algorithm, and use the original NG center as the center to perform Archimedean spiral point search. If the test result is OK, proceed according to the OK result; if NG, continue to search for points. Each time a point is searched, the k2 value is reduced by 1 until it is zero. Specifically, the parameters a and b are selected according to the empirical parameters.

[0041] If both K1 and K2 are zero, the correction algorithm ends and the product is judged to be NG.

[0042] See also Figure 1 to Figure 2 The adjustment process of the offset direction in step 2 is as follows:

[0043] When there are four test points ABCD in the figure, ABC is OK, and D is NG, because AC are all OK points, the most likely offset direction is the direction of the line connecting BD, so the BD line is taken as the offset direction.

[0044] See also Figure 1 to Figure 2 , ABCD test points are described above as test points. In actual tests, bad test results will appear in the form of areas. ABCD can also be an area. The position of ABCD in the middle of the area is calculated according to the algorithm, such as the geometric center of the area.

[0045] See also Figure 1 to Figure 2 , The golden ratio algorithm in step three is specifically y=0.618x.

[0046] See also Figure 1 to Figure 2 , The Archimedean spiral algorithm in step three is specifically y=a+bθ.

[0047] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A fast test and correction method for a thread needle testing machine, characterized in that: The fast testing and correction method of the wire needle testing machine comprises: Step 1: Get the actual test results of the ABCD test points in the region, and determine whether the test results are occasional defects in the region or large-area defects in the region; if it is an occasional defect, the circuit board is determined to be unqualified, and if it is a large-area defect in the region, it is determined to be poor contact, and the correction algorithm needs to be run; Step 2: According to the position characteristics of the bad graphics, the most likely offset direction is calculated by diagonal calculation, and the golden ratio algorithm is used to approximate the position; The diagonal calculation method is: in the four areas of ABCD, assume that ABC is normal and D is bad. In the graphic distribution, AC is the diagonal and BD is the diagonal. At this time, the diagonal direction of BD is determined to be the most likely deviation direction, and the diagonal direction of BD is taken as the approaching direction of movement; Step 3: Correction control parameters k1, k2; k1 is the number of repetitions of the golden ratio algorithm, k2 is the number of repetitions of the Archimedean spiral algorithm, and they are initially set to a constant. The specific correction algorithm is as follows: First: When an offset occurs, calculate the position of the graphic first, find the diagonal line, and perform an offset test on the most likely offset direction. The offset is 1 / n of the pad point size. If the test passes, it is OK and the subsequent process is carried out; if the test is NG, there are two situations: (1) If a point that was originally OK becomes NG, it can be judged that the deviation is excessive and needs to be offset back. The golden ratio algorithm is used to find the point based on the offset amount to increase the probability of finding the point; (2) If the points that were OK are still OK and the points that were NG are still NG, it can be determined that the offset is insufficient. Then continue to offset until it is OK or the offset is excessive. Each time the golden ratio algorithm is run, the k1 value is reduced by 1 until it reaches zero. Second: When the k1 value is zero, run the Archimedean spiral algorithm, and use the original NG center as the center to search for points. If the test result is OK, proceed according to the OK result; if NG, continue to search for points. Each time a point is searched, the k2 value is reduced by 1 until it reaches zero. If both K1 and K2 are zero, the correction algorithm ends and the product is judged to be NG.

2. The method for rapid testing and correcting a thread needle testing machine according to claim 1, characterized in that: The following steps are also included between step 1 and step 2: First, the image is fed into a preset model, which usually comes from a data graphics file; Then, the preset model is used to extract features from the image and compare the actual test results of the current ABCD test points; Finally, the preset model performs correction calculation on the position recognition results of the ABCD test points based on the ABCD mapping relationship.

3. The method for rapid testing and correcting a thread needle testing machine according to claim 1, characterized in that: The adjustment process of the offset direction in step 2 is specifically as follows: When there are four test points ABCD in the figure, ABC is OK, and D is NG, because AC are all OK points, the most likely offset direction is the direction of the line connecting BD, so the BD line is taken as the offset direction.

4. A method for rapid testing and correcting a thread needle testing machine according to any one of claims 1 to 3, characterized in that: The ABCD test point is described above as a test point. In actual testing, a bad test result will appear in the form of an area. ABCD is the area, and the position of ABCD in the middle of the area is calculated according to an algorithm.

5. The method for rapid testing and correcting a thread needle testing machine according to claim 1, characterized in that: The golden ratio algorithm in step three is specifically y=0.618x.

6. The method for rapid testing and correcting a thread needle testing machine according to claim 1, characterized in that: The Archimedean spiral algorithm in step three is specifically y=a+bθ.

Images