Probe card stroke compensation system and method

Abstract

The present invention relates to the field of probe card technology, and specifically discloses a probe card stroke compensation system and method, including: analyzing the distance value of each probe in the probe card to obtain a distance characterization value WB, judging whether an adjustment signal is generated based on the distance characterization value WB due to the abnormality of the probe card, and based on the adjustment signal, applying test pressure to a single probe at a fixed rate, analyzing the change trend of the distance value as the test pressure increases, judging whether the change of the probe distance value as the test pressure increases is linear, and selecting a suitable model as a prediction model based on the judgment result of whether the change of the probe distance value as the test pressure increases is linear. The present invention can timely adjust the distance standard value according to the actual monitoring situation, and adjust the pressure during the probe stroke compensation accordingly, effectively solving the problem caused by the large difference between the actual distance value and the predicted distance value, and ensuring the normal operation of the probe card and the accuracy of the test effect.

Description

Technical Field

[0001] The present invention relates to the technical field of probe cards, and in particular to a probe card stroke compensation system and method. Background Art

[0002] Wafer testing is a key link in the chip manufacturing process. By testing the electrical properties of the chips on the wafer with probe cards, unqualified chips can be screened out in a timely manner, reducing production costs. In order to ensure the accuracy of wafer testing, the probe card needs to be able to accurately connect to the chip pins, and the stroke compensation technology can compensate for the influence of factors such as the height difference of the chip pins and the deformation of the probes themselves, ensuring that each probe can have good contact with the pins. The worn probe tip may become blunter, and the contact area with the chip pins will increase under the same pressure, and the contact distance will also change accordingly.

[0003] The invention patent with publication number CN115113011B records a probe card stroke compensation system and method, including a measuring unit for testing the stroke of a single probe and the flatness of all probes; and calculating the total elastic force of the probe card based on the values ​​of the stroke and flatness; a pressure sensor unit, wherein the pressure sensor and the probe head are interchangeable, and the stroke compensation value of the probe card is obtained based on the relationship between pressure and stroke measured by the pressure sensor.

[0004] The above-mentioned prior art can compensate for the stroke of the probe card probe in wafer testing, ensure effective contact between the probe and the wafer, and thus ensure the wafer test yield and reliability. However, as the number of times the probe card is used increases, the probe wears out. The above-mentioned prior art lacks adjustment of the distance standard value. In order to ensure accurate testing and good contact, the distance standard value needs to be adjusted according to the actual usage of the probe. Summary of the invention

[0005] The object of the present invention is to provide a probe card stroke compensation system and method to solve the technical problems existing in the above background.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] In a first aspect, the present invention provides a probe card stroke compensation method, comprising:

[0008] Step 1: Analyze the distance value of each probe in the probe card to obtain the distance characterization value WB, and determine whether the state of the probe card is abnormal based on the distance characterization value WB. If abnormal, generate an adjustment signal;

[0009] Step 2: Based on the adjustment signal, by applying the test pressure to a single probe at a fixed rate, the change trend of the distance value with the increase of the test pressure is analyzed to obtain the linear coincidence value FX. Based on the linear coincidence value FX, it is determined whether the change of the probe distance value with the increase of the test pressure is linear. The determination results include linear relationship and nonlinear relationship.

[0010] Step 3: Based on the judgment result of whether the change of the probe distance value with the increase of the test pressure is linear, a suitable model is selected as the prediction model, and the adjustment pressure for the probe to reach the distance standard value is obtained based on the prediction model analysis, and the adjustment pressure segment is obtained based on the adjustment pressure, and the predicted distance values ​​of different pressure points in the adjustment pressure segment are obtained through the prediction model;

[0011] Step 4: In the pressure adjustment section, the distance value of the probe is monitored in real time, and the probe distance value monitored in real time in the adjustment period is compared and analyzed with the predicted distance value corresponding to different pressures to obtain the change similarity value QJ. Based on the change similarity value QJ, the degree of convergence between the probe distance value monitored in real time in the adjustment period and the predicted distance value is determined;

[0012] Step 5: Based on the inaccurate signal, the distance values ​​of different pressure points are analyzed to obtain the non-converging similarity value XS. Based on the non-converging similarity value XS, determine whether the distance standard value needs to be adjusted. If adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-entered to re-obtain the adjusted pressure. Based on the re-obtained adjusted pressure, the pressure during probe stroke compensation is adjusted.

[0013] In a second aspect, the present invention provides a probe card stroke compensation system, comprising:

[0014] Abnormal state judgment module: analyzes the distance value of each probe in the probe card to obtain the distance characterization value WB, and judges whether the state of the probe card is abnormal based on the distance characterization value WB. If abnormal, generates an adjustment signal;

[0015] Linear relationship judgment module: Based on the adjustment signal, by applying the test pressure to a single probe at a fixed rate, the change trend of the distance value with the increase of the test pressure is analyzed to obtain the linear coincidence value FX. Based on the linear coincidence value FX, it is judged whether the change of the probe distance value with the increase of the test pressure is linear. The judgment results include linear relationship and nonlinear relationship;

[0016] Model building module: Based on the judgment result of whether the change of the probe distance value with the increase of the test pressure is linear, a suitable model is selected as the prediction model, and the adjustment pressure for the probe to reach the distance standard value is obtained based on the prediction model analysis, and the adjustment pressure segment is obtained based on the adjustment pressure. The predicted distance values ​​of different pressure points in the adjustment pressure segment are obtained through the prediction model;

[0017] Approach degree judgment module: In the pressure adjustment section, the distance value of the probe is monitored in real time, and the probe distance value monitored in real time during the adjustment period is compared and analyzed with the predicted distance value corresponding to different pressures to obtain the change similarity value QJ. Based on the change similarity value QJ, the approach degree of the probe distance value monitored in real time during the adjustment period and the predicted distance value is judged;

[0018] Adjustment analysis module: Based on the inaccurate signal, the distance values ​​of different pressure points are analyzed to obtain the non-converging similarity value XS. Based on the non-converging similarity value XS, it is determined whether the distance standard value needs to be adjusted. If adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-input to re-obtain the adjustment pressure. Based on the re-obtained adjustment pressure, the pressure during probe stroke compensation is adjusted.

[0019] Beneficial effects of the present invention:

[0020] (1) The present invention analyzes the distance value of each probe in the probe card to obtain the distance characterization value WB, and determines whether to generate an adjustment signal based on the distance characterization value WB. By analyzing the distance data of each probe in the probe card, the present invention can timely and accurately determine the abnormality of the probe card and generate an adjustment signal to prompt the relevant system to take measures, which is beneficial to improving the working reliability of the probe card and the accuracy of chip testing;

[0021] (2) The present invention is based on an adjustment signal. By applying a test pressure to a single probe at a fixed rate, the change trend of the distance value as the test pressure increases is analyzed to determine whether the change of the probe distance value as the test pressure increases is linear. Based on the determination result of whether the change of the probe distance value as the test pressure increases is linear, a suitable model is selected as a prediction model. The predicted distance values ​​of different pressure points within the adjustment pressure section are obtained through the prediction model. Within the adjustment pressure section, the distance value of the probe is monitored in real time. The probe distance value monitored in real time during the adjustment period is compared and analyzed with the predicted distance values ​​corresponding to different pressures to determine the degree of convergence between the probe distance value monitored in real time during the adjustment period and the predicted distance value. The present invention can more accurately understand the change law of the distance value of the probe under different pressures, and then accurately determine and optimize the adjustment pressure of the probe according to the prediction model, thereby improving the working performance and test accuracy of the probe card.

[0022] (3) The present invention analyzes the distance values ​​of different pressure points based on inaccurate signals to determine whether the distance standard value needs to be adjusted. If adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-input to re-obtain the adjusted pressure. The pressure during probe stroke compensation is adjusted based on the re-obtained adjusted pressure. The present invention can timely adjust the distance standard value according to the actual monitoring situation and adjust the pressure during probe stroke compensation accordingly, effectively solving the problem caused by the large difference between the actual distance value and the predicted distance value, thereby ensuring the normal operation of the probe card and the accuracy of the test results. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The present invention will be further described below in conjunction with the accompanying drawings.

[0024] Figure 1 It is a flowchart of a probe card stroke compensation method according to an embodiment of the present invention;

[0025] Figure 2 The system block diagram of a probe card stroke compensation system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0027] Embodiment 1:

[0028] like Figure 1 As shown, a probe card stroke compensation method according to an embodiment of the present invention includes the following steps:

[0029] Step 1: obtaining distance data of each probe in the probe card, wherein the distance data includes a distance value, analyzing the distance data of each probe in the probe card to obtain a distance characterization value WB, and determining whether to generate an adjustment signal based on the abnormality of the probe card;

[0030] It should be noted that the distance value represents the movement distance of the probe relative to the initial position;

[0031] In some embodiments, a distance value of each probe in the probe card is subjected to difference processing with a distance standard value, and the difference is taken as an absolute value to obtain a distance deviation value;

[0032] Compare the distance deviation value to the distance deviation threshold:

[0033] If the distance deviation value is greater than or equal to the distance deviation threshold, a distance abnormality signal is generated;

[0034] If the distance deviation value is less than the distance deviation threshold, a distance normal signal is generated;

[0035] Count the number of abnormal distance signals in the detection area, and compare it with the total number of probes in the detection area to obtain the abnormal number ratio YC;

[0036] Obtain the distance deviation value corresponding to the distance abnormal signal in the detection area, perform subtraction processing on it and the distance deviation threshold to obtain the distance relative deviation value, sum and average all the distance relative deviation values ​​in the detection area to obtain the distance relative deviation mean, perform ratio processing on the distance relative deviation mean and the distance deviation threshold to obtain the distance deviation degree ratio PC;

[0037] The abnormal number ratio YC and the distance deviation ratio PC are processed by the formula: The distance characterization value WB is calculated, where a1 and a2 are both preset proportional coefficients, a1 is 1.832, and a2 is 2.052;

[0038] Compare the distance characterization value WB with the distance characterization threshold:

[0039] If the distance characterization value WB is greater than or equal to the distance characterization threshold, it indicates that the abnormality of the probe card is serious, and an adjustment signal is generated;

[0040] If the distance characterization value WB is less than the distance characterization threshold, it indicates that the abnormality of the probe card is minor and a non-adjustment signal is generated;

[0041] The technical solution of an embodiment of the present invention is mainly as follows: obtaining distance data of each probe in a probe card, wherein the distance data includes a distance value, analyzing the distance data of each probe in the probe card to obtain a distance characterization value WB, and judging the degree of abnormality of the probe card based on the distance characterization value WB to generate an adjustment signal. By analyzing the distance data of each probe in the probe card, the present invention can timely and accurately judge the degree of abnormality of the probe card, and generate an adjustment signal to prompt the relevant system to take measures, which is beneficial to improving the working reliability of the probe card and the accuracy of chip testing.

[0042] Embodiment 2:

[0043] On the basis of Example 1, Figure 1 As shown, a probe card stroke compensation method according to an embodiment of the present invention further includes the following steps:

[0044] Step 2: Based on the adjustment signal, by applying the test pressure to a single probe at a fixed rate, the change trend of the distance value with the increase of the test pressure is analyzed to obtain the linear coincidence value FX. Based on the linear coincidence value FX, it is determined whether the change of the probe distance value with the increase of the test pressure is linear;

[0045] Apply test pressure to a single probe at a fixed rate (increasing the pressure by 0.1N per second), obtain the distance value corresponding to the probe test pressure, and mark and connect it in the XY two-dimensional coordinate system to obtain a distance value change curve, where the X-axis represents the test pressure and the Y-axis represents the distance value;

[0046] Divide the distance value change curve into a number of curve segments, measure the slope of the curve segments, connect the two end points of the distance value change curve to obtain a linear reference line, and measure the slope of the linear reference line;

[0047] Compare the slope of the curve segment to the slope of the linear reference line:

[0048] If the slope value of the curve segment is the same as the slope value of the linear reference line, the curve segment is marked as a curve segment with the same direction;

[0049] If the slope value of the curve segment is different from the slope value of the linear reference line in sign, the curve segment is marked as a non-co-directional curve segment;

[0050] It should be noted that the same-direction curve segment means that the slope of the curve segment is the same as the slope of the reference line, and the non-same-direction curve segment means that the slope of the curve segment is different from the slope of the reference line, or one is 0 and the other is not 0.

[0051] Count the number of curve segments in the same direction among all curve segments, and perform ratio processing on it with the total number of curve segments to obtain the ratio of the number of curve segments in the same direction JL;

[0052] Integrate the slope values ​​of all the curve segments in the same direction into a slope data group, and obtain the variance value CD of the slope data group;

[0053] The proportion of the number of curve segments in the same direction JL and the variance value CD of the slope data group are processed by the formula: The linear overlap value FX is calculated, where s1 and s2 are preset proportional coefficients, s1 is 1.555, s2 is 1.653, and s3 is 1.668;

[0054] Compare the linear coincidence value FX with the linear coincidence threshold:

[0055] If the linear coincidence value FX is greater than or equal to the linear coincidence threshold, it means that there is a linear relationship between the probe distance value and the test pressure;

[0056] If the linear coincidence value FX is less than the linear coincidence threshold, it means that there is a nonlinear relationship between the probe distance value and the test pressure;

[0057] Step 3: Based on the judgment result of whether the change of the probe distance value with the increase of the test pressure is linear, a suitable model is selected as the prediction model, and the adjustment pressure for the probe to reach the distance standard value is obtained based on the prediction model analysis, and the adjustment pressure segment is obtained based on the adjustment pressure, and the predicted distance values ​​of different pressure points in the adjustment pressure segment are obtained through the prediction model;

[0058] The distance values ​​of all probes are integrated into a data set, which is used as a training data set. If there is a linear relationship between the distance value and the test pressure, a linear regression model is selected as the prediction model. If there is a nonlinear relationship between the distance value and the test pressure, a random forest network model is selected as the prediction model. The prediction model is trained using the training data set to obtain a trained prediction model.

[0059] The standard value of the distance is input into the prediction model to obtain the pressure when the probe distance reaches the standard value of the distance, and is marked as the arrival pressure. The arrival pressure is processed with the current pressure to obtain the adjustment pressure when the probe distance reaches the standard value of the distance, and the interval between the adjustment pressures is marked as the adjustment pressure segment;

[0060] Input different pressures in the regulated pressure section into the prediction model to obtain predicted distance values ​​of different pressures in the regulated pressure section;

[0061] Step 4: In the pressure adjustment section, the distance value of the probe is monitored in real time, and the probe distance value monitored in real time in the adjustment period is compared and analyzed with the predicted distance value corresponding to different pressures to obtain the change similarity value QJ. Based on the change similarity value QJ, the degree of convergence between the probe distance value monitored in real time in the adjustment period and the predicted distance value is determined;

[0062] In the regulating pressure section, a monitoring pressure section is preset, wherein the interval length corresponding to the preset monitoring pressure section is smaller than the interval length corresponding to the regulating pressure section;

[0063] It should be noted that the purpose of presetting the monitoring pressure section within the adjustment pressure section is to: monitor in real time whether the distance value within the preset monitoring pressure section is close to the predicted distance value, and then analyze the accuracy of the prediction;

[0064] Get the distance values ​​of different pressures in the preset monitoring pressure range and compare them with the predicted distance values ​​of different pressures predicted by the model. Specifically:

[0065] Based on any pressure within the preset monitoring pressure range;

[0066] If the distance value corresponding to the pressure in the preset monitoring pressure section is not equal to the predicted distance value, the corresponding pressure is marked as a different pressure point;

[0067] If the distance value corresponding to the pressure in the preset monitoring pressure section is equal to the predicted distance value, the corresponding pressure is marked as the same pressure point;

[0068] Count the total number of pressure points and the number of different pressure points in the preset monitoring pressure section, perform ratio processing on the number of different pressure points and the total number of pressure points in the preset monitoring pressure section, and obtain the change deviation number ratio BT;

[0069] Obtain the distance values ​​of different pressure points, perform difference processing on them and the predicted distance values ​​of the corresponding pressures, and take the absolute value of the difference to obtain the distance deviation values ​​of different pressure points, perform ratio processing on the distance deviation values ​​of different pressure points and the predicted distance values ​​to obtain the distance deviation ratios of different pressure points, sum and average the distance deviation ratios of all different pressure points to obtain the change deviation degree value BJ;

[0070] The change deviation quantity ratio BT and the change deviation degree value BJ are processed by the formula: The change similarity value QJ is obtained, where b1 and b2 are both preset proportional coefficients, b1 is 1.323, and b2 is 1.556;

[0071] In some embodiments, the change similarity value QJ is compared with a change approach threshold;

[0072] If the change similarity value QJ is greater than the change similarity threshold, it means that the distance change trend of the probe at different pressure points is close to the predicted distance value change trend, and an accurate signal is generated;

[0073] If the change similarity value QJ is less than or equal to the change similarity threshold, the distance change trend of the probe at different pressure points is close to the predicted distance value change trend, and an inaccurate signal is generated;

[0074] The technical solution of the embodiment of the present invention is mainly as follows: based on the adjustment signal, by applying the test pressure to a single probe at a fixed rate, the change trend of the distance value with the increase of the test pressure is analyzed to obtain the linear overlap value FX, based on the linear overlap value FX, it is judged whether the change of the probe distance value with the increase of the test pressure is linear, based on the judgment result of whether the change of the probe distance value with the increase of the test pressure is linear, a suitable model is selected as the prediction model, based on the prediction model analysis, the adjustment pressure for the probe to reach the distance standard value is obtained, based on the adjustment pressure, the adjustment pressure segment is obtained, and the predicted distance values ​​of different pressure points in the adjustment pressure segment are obtained through the prediction model, and in the adjustment pressure segment, the distance value of the probe is monitored in real time, and the probe distance value monitored in real time during the adjustment period is compared and analyzed with the predicted distance values ​​corresponding to different pressures to obtain the change similarity value QJ, and based on the change similarity value QJ, the degree of convergence between the probe distance value monitored in real time during the adjustment period and the predicted distance value is judged. The present invention can more accurately understand the change law of the distance value of the probe under different pressures, and then accurately judge and optimize the adjustment pressure of the probe according to the prediction model, thereby improving the working performance and test accuracy of the probe card.

[0075] Embodiment 3:

[0076] On the basis of Example 1 and Example 2, Figure 1 As shown, a probe card stroke compensation method according to an embodiment of the present invention includes:

[0077] Step 5: Based on the inaccurate signal, the distance values ​​of different pressure points are analyzed to obtain the non-converging similarity value XS. Based on the non-converging similarity value XS, it is determined whether the distance standard value needs to be adjusted. If adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-entered to re-obtain the adjustment pressure. Based on the re-obtained adjustment pressure, the pressure during the probe stroke compensation is adjusted;

[0078] Obtain the distance values ​​of different pressure points, perform difference processing on them and the predicted distance value, and take the absolute value of the difference to obtain the relative distance difference;

[0079] Integrate the relative distance differences of all different pressure points into a relative distance difference data group, and obtain the variance value FC of the relative distance difference data group;

[0080] The variance value FC and the change deviation quantity ratio BT of the distance relative difference data group are processed by the formula: A non-approaching similarity value XS is obtained, where m1 and m3 are both preset proportional coefficients;

[0081] Compare the non-converging similarity value XS with the non-converging similarity threshold;

[0082] If the non-converging similarity value XS is greater than the non-converging similarity threshold, the relative distance differences of all different pressure points are obtained and summed up to obtain the target distance difference mean, the distance standard value and the target distance difference mean are subjected to difference processing to obtain the adjusted distance standard value, and the value prediction model is re-input to re-obtain the adjusted pressure, and the pressure during the probe stroke compensation is adjusted based on the re-obtained adjusted pressure;

[0083] If the non-approaching similarity value XS is less than or equal to the non-approaching similarity threshold, no operation is performed;

[0084] The technical solution of the embodiment of the present invention is mainly as follows: based on the inaccurate signal, the distance values ​​of different pressure points are analyzed to obtain the non-converging similarity value XS; based on the non-converging similarity value XS, it is determined whether the distance standard value needs to be adjusted; if adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-input to obtain the adjusted pressure; based on the re-obtained adjusted pressure, the pressure during probe stroke compensation is adjusted; the present invention can timely adjust the distance standard value according to the actual monitoring situation, and adjust the pressure during probe stroke compensation accordingly, which effectively solves the problem caused by the large difference between the actual distance value and the predicted distance value, thereby ensuring the normal operation of the probe card and the accuracy of the test effect.

[0085] Embodiment 4:

[0086] On the basis of Example 1 and Example 2, Figure 1 , Figure 2 As shown, a probe card stroke compensation system according to an embodiment of the present invention includes:

[0087] Abnormal state judgment module: analyzes the distance value of each probe in the probe card to obtain the distance characterization value WB, and judges whether the state of the probe card is abnormal based on the distance characterization value WB. If abnormal, generates an adjustment signal;

[0088] Linear relationship judgment module: Based on the adjustment signal, by applying the test pressure to a single probe at a fixed rate, the change trend of the distance value with the increase of the test pressure is analyzed to obtain the linear coincidence value FX. Based on the linear coincidence value FX, it is judged whether the change of the probe distance value with the increase of the test pressure is linear. The judgment results include linear relationship and nonlinear relationship;

[0089] Model building module: Based on the judgment result of whether the change of the probe distance value with the increase of the test pressure is linear, a suitable model is selected as the prediction model, and the adjustment pressure for the probe to reach the distance standard value is obtained based on the prediction model analysis, and the adjustment pressure segment is obtained based on the adjustment pressure. The predicted distance values ​​of different pressure points in the adjustment pressure segment are obtained through the prediction model;

[0090] Approach degree judgment module: In the pressure adjustment section, the distance value of the probe is monitored in real time, and the probe distance value monitored in real time during the adjustment period is compared and analyzed with the predicted distance value corresponding to different pressures to obtain the change similarity value QJ. Based on the change similarity value QJ, the approach degree of the probe distance value monitored in real time during the adjustment period and the predicted distance value is judged;

[0091] Adjustment analysis module: Based on the inaccurate signal, the distance values ​​of different pressure points are analyzed to obtain the non-converging similarity value XS. Based on the non-converging similarity value XS, it is determined whether the distance standard value needs to be adjusted. If adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-input to re-obtain the adjustment pressure. Based on the re-obtained adjustment pressure, the pressure during probe stroke compensation is adjusted.

[0092] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments, and that various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention, and these changes and improvements all fall within the scope of the present invention claimed for protection.

Claims

1. A probe card stroke compensation method, characterized in that: include: Step 1: Analyze the distance value of each probe in the probe card to obtain the distance characterization value WB, and determine whether the state of the probe card is abnormal based on the distance characterization value WB. If abnormal, generate an adjustment signal; Step 2: Based on the adjustment signal, by applying the test pressure to a single probe at a fixed rate, the change trend of the distance value with the increase of the test pressure is analyzed to obtain the linear coincidence value FX. Based on the linear coincidence value FX, it is determined whether the change of the probe distance value with the increase of the test pressure is linear. The determination results include linear relationship and nonlinear relationship. Step 3: Based on the judgment result of whether the change of the probe distance value with the increase of the test pressure is linear, a suitable model is selected as the prediction model, and the adjustment pressure for the probe to reach the distance standard value is obtained based on the prediction model analysis, and the adjustment pressure segment is obtained based on the adjustment pressure, and the predicted distance values ​​of different pressure points in the adjustment pressure segment are obtained through the prediction model; Step 4: In the pressure adjustment section, the distance value of the probe is monitored in real time, and the probe distance value monitored in real time in the adjustment period is compared and analyzed with the predicted distance value corresponding to different pressures to obtain the change similarity value QJ. Based on the change similarity value QJ, the degree of convergence between the probe distance value monitored in real time in the adjustment period and the predicted distance value is determined; Step 5: Based on the inaccurate signal, the distance values ​​of different pressure points are analyzed to obtain the non-converging similarity value XS. Based on the non-converging similarity value XS, determine whether the distance standard value needs to be adjusted. If adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-entered to re-obtain the adjusted pressure. Based on the re-obtained adjusted pressure, the pressure during probe stroke compensation is adjusted.

2. A probe card stroke compensation method according to claim 1, characterized in that: The distance characterization value WB is obtained as follows: Obtain the distance value of each probe in the probe card, and analyze to obtain the abnormal number ratio YC and the distance deviation ratio PC; The abnormal number ratio YC and the distance deviation ratio PC are processed by the formula: The distance characterization value WB is calculated, where a1 and a2 are both preset proportional coefficients, a1 takes a value of 1.832, and a2 takes a value of 2.

052.

3. A probe card stroke compensation method according to claim 2, characterized in that: The abnormality ratio YC and the distance deviation ratio PC are obtained as follows: Count the number of abnormal distance signals in the detection area, and compare it with the total number of probes in the detection area to obtain the abnormal number ratio YC; Obtain the distance deviation value corresponding to the distance abnormal signal in the detection area, perform subtraction processing on it and the distance deviation threshold to obtain the distance relative deviation value, sum and average all the distance relative deviation values ​​in the detection area to obtain the distance relative deviation mean, perform ratio processing on the distance relative deviation mean and the distance deviation threshold to obtain the distance deviation degree ratio PC.

4. The probe card stroke compensation method according to claim 1, characterized in that: The linear overlap value FX is obtained by: Obtain the distance value corresponding to the probe test pressure, construct a distance value change curve, and analyze the proportion of the number of curve segments in the same direction JL and the variance value CD of the slope data group based on the distance value change curve; The proportion of the number of curve segments in the same direction JL and the variance value CD of the slope data group are processed by the formula: The linear overlap value FX is calculated, where s1 and s2 are both preset proportional coefficients, s1 takes a value of 1.555, s2 takes a value of 1.653, and s3 takes a value of 1.

668.

5. A probe card stroke compensation method according to claim 4, characterized in that: The method for obtaining the proportion of the number of curve segments in the same direction JL and the variance value CD of the slope data group is as follows: Apply test pressure to a single probe at a fixed rate, obtain the distance value corresponding to the probe test pressure, and mark and connect it in the XY two-dimensional coordinate system to obtain a distance value change curve, where the X-axis represents the test pressure and the Y-axis represents the distance value; Divide the distance value change curve into a number of curve segments, measure the slope of the curve segments, connect the two end points of the distance value change curve to obtain a linear reference line, and measure the slope of the linear reference line; Compare the slope of the curve segment to the slope of the linear reference line: If the slope value of the curve segment is the same as the slope value of the linear reference line, the curve segment is marked as a curve segment with the same direction; If the slope value of the curve segment is different from the slope value of the linear reference line in sign, the curve segment is marked as a non-co-directional curve segment; Count the number of curve segments in the same direction among all curve segments, and perform ratio processing on it with the total number of curve segments to obtain the ratio of the number of curve segments in the same direction JL; The slope values ​​of all the curve segments in the same direction are integrated into a slope data set, and the variance value CD of the slope data set is obtained.

6. A probe card stroke compensation method according to claim 1, characterized in that: The predicted distance value is obtained as follows: The distance values ​​of all probes are integrated into a data set, which is used as a training data set. If there is a linear relationship between the distance value and the test pressure, a linear regression model is selected as the prediction model. If there is a nonlinear relationship between the distance value and the test pressure, a random forest network model is selected as the prediction model. The prediction model is trained using the training data set to obtain a trained prediction model. The standard value of the distance is input into the prediction model to obtain the pressure when the probe distance reaches the standard value of the distance, and is marked as the arrival pressure. The arrival pressure is processed with the current pressure to obtain the adjustment pressure when the probe distance reaches the standard value of the distance, and the interval between the adjustment pressures is marked as the adjustment pressure segment; Different pressures within the regulated pressure section are input into the prediction model to obtain the predicted distance values ​​of different pressures within the regulated pressure section.

7. The probe card stroke compensation method according to claim 1, characterized in that: The method for obtaining the change similarity value QJ is as follows: In the adjustment pressure section, a monitoring pressure section is preset, and the distance values ​​of different pressures in the preset monitoring pressure section are obtained and analyzed and processed to obtain the change deviation quantity ratio BT and the change deviation degree value BJ; The change deviation quantity ratio BT and the change deviation degree value BJ are processed by the formula: The change similarity value QJ is obtained, where b1 and b2 are both preset proportional coefficients, b1 takes a value of 1.323, and b2 takes a value of 1.

556.

8. A probe card stroke compensation method according to claim 7, characterized in that: The change deviation quantity ratio BT and the change deviation degree value BJ are obtained as follows: Get the distance values ​​of different pressures in the preset monitoring pressure range and compare them with the predicted distance values ​​of different pressures predicted by the model. Specifically: Based on any pressure within the preset monitoring pressure range; If the distance value corresponding to the pressure in the preset monitoring pressure section is not equal to the predicted distance value, the corresponding pressure is marked as a different pressure point; If the distance value corresponding to the pressure in the preset monitoring pressure section is equal to the predicted distance value, the corresponding pressure is marked as the same pressure point; Count the total number of pressure points and the number of different pressure points in the preset monitoring pressure section, perform ratio processing on the number of different pressure points and the total number of pressure points in the preset monitoring pressure section, and obtain the change deviation number ratio BT; Get the distance value of different pressure points, perform difference processing on it and the predicted distance value of the corresponding pressure, and take the absolute value of the difference to obtain the distance deviation value of different pressure points, perform ratio processing on the distance deviation value of different pressure points and the predicted distance value to obtain the distance deviation ratio of different pressure points, sum and average the distance deviation ratios of all different pressure points to obtain the change deviation degree value BJ.

9. The probe card stroke compensation method according to claim 1, characterized in that: The non-approaching similarity value XS is obtained as follows: Obtain the distance values ​​of different pressure points, perform difference processing on them and the predicted distance value, and take the absolute value of the difference to obtain the relative distance difference; Integrate the relative distance differences of all different pressure points into a relative distance difference data group, and obtain the variance value FC of the relative distance difference data group; The variance value FC and the change deviation quantity ratio BT of the distance relative difference data group are processed by the formula: A non-approaching similarity value XS is obtained, where m1 and m3 are both preset proportional coefficients; Compare the non-converging similarity value XS with the non-converging similarity threshold; If the non-converging similarity value XS is greater than the non-converging similarity threshold, the relative distance differences of all different pressure points are obtained and summed up to obtain the target distance difference mean, the distance standard value and the target distance difference mean are subjected to difference processing to obtain the adjusted distance standard value, and the value prediction model is re-input to re-obtain the adjusted pressure, and the pressure during the probe stroke compensation is adjusted based on the re-obtained adjusted pressure; If the non-converging similarity value XS is less than or equal to the non-converging similarity threshold, no operation is performed.

10. A probe card stroke compensation system, characterized in that: The system is used to execute the method described in any one of claims 1 to 9, and the system comprises: Abnormal state judgment module: analyzes the distance value of each probe in the probe card to obtain the distance characterization value WB, and judges whether the state of the probe card is abnormal based on the distance characterization value WB. If abnormal, generates an adjustment signal; Linear relationship judgment module: Based on the adjustment signal, by applying the test pressure to a single probe at a fixed rate, the change trend of the distance value with the increase of the test pressure is analyzed to obtain the linear coincidence value FX. Based on the linear coincidence value FX, it is judged whether the change of the probe distance value with the increase of the test pressure is linear. The judgment results include linear relationship and nonlinear relationship; Model building module: Based on the judgment result of whether the change of the probe distance value with the increase of the test pressure is linear, a suitable model is selected as the prediction model, and the adjustment pressure for the probe to reach the distance standard value is obtained based on the prediction model analysis, and the adjustment pressure segment is obtained based on the adjustment pressure. The predicted distance values ​​of different pressure points in the adjustment pressure segment are obtained through the prediction model; Approach degree judgment module: In the pressure adjustment section, the distance value of the probe is monitored in real time, and the probe distance value monitored in real time during the adjustment period is compared and analyzed with the predicted distance value corresponding to different pressures to obtain the change similarity value QJ. Based on the change similarity value QJ, the approach degree of the probe distance value monitored in real time during the adjustment period and the predicted distance value is judged; Adjustment analysis module: Based on the inaccurate signal, the distance values ​​of different pressure points are analyzed to obtain the non-converging similarity value XS. Based on the non-converging similarity value XS, it is determined whether the distance standard value needs to be adjusted. If adjustment is required, the relative difference of the distances of all different pressure points is processed and analyzed to obtain the adjusted distance standard value, and the value prediction model is re-input to re-obtain the adjustment pressure. Based on the re-obtained adjustment pressure, the pressure during probe stroke compensation is adjusted.

Images