Test system and test method thereof
By using an artificial intelligence-generated model to analyze the degradation of probe card parameters, the problem of test quality degradation caused by wear of the probe card during CP testing was solved, and effective monitoring of the probe card status and timely shutdown were achieved.
Patent Information
- Application Number
- CN202511065974.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2025-06-23
- Filing Date
- 2025-07-31
- Publication Date
- 2025-10-10
AI Technical Summary
Existing probe cards suffer from poor test quality during CP testing due to foreign matter or wear on the probe tips. Therefore, effectively testing their normal operation has become an important issue.
An artificial intelligence algorithm is used to generate a model that performs degradation analysis on the probe card's parameters, determines its status, and generates test results, including abnormal warning signals and detection flags, to determine the degradation trend of the probe card.
This enables effective testing of probe card degradation trends, ensuring the normal operation of test sites and preventing degradation of test quality through timely shutdown.
Smart Images
Figure CN120761812A_ABST
Abstract
Description
Technical Field
[0001] Embodiments of the present invention relate to a testing system and a testing method, and more particularly to a testing system and a testing method for testing whether a probe card can operate normally. Background Art
[0002] After completing conventional semiconductor processes, a circuit probing (CP) test using a probe card is typically performed to verify the wafer's electrical functionality. Existing probe cards connect the wafer to the tester and perform CP testing. As the number of CP tests increases, probe card tip contamination and wear can affect CP test quality.
[0003] How to test whether the probe card can operate normally to avoid affecting the quality of CP testing is an important issue that those skilled in the art need to deal with. Summary of the Invention
[0004] The present invention provides a test system. The test system includes: an information collection unit and a processing unit. The information collection unit is used to receive multiple first parameters from a first test site having a first probe card. The first parameters are related to multiple test results of the first test site testing a wafer through the first probe card. The processing unit is coupled to the information collection unit and is used to: generate a first model corresponding to the first test site through an artificial intelligence algorithm and the first parameters; and perform degradation analysis on the first probe card based on the first model and the first parameters, and generate a first test result. The first model is used to predict the first probe card under the first test condition of the first test site.
[0005] In some embodiments, the degradation analysis is used to: convert the multiple first parameters into corresponding multiple test values based on the artificial intelligence algorithm; determine whether each of the multiple test values is an error value to generate the first detection result; and in response to the number of the multiple test values being error values being higher than a threshold, cause the processing unit test to issue an abnormal warning signal.
[0006] In some embodiments, the test system further includes a feedback unit coupled to the processing unit, configured to receive the abnormal warning signal and transmit the abnormal warning signal to the first test site to shut down the first test site.
[0007] In some embodiments, the feedback unit includes a notification unit, and the notification unit is configured to send the abnormality warning signal to an external device in response to the feedback unit receiving the abnormality warning signal.
[0008] In some embodiments, the testing system further includes a report generating unit, coupled to the processing unit, and configured to: receive the first test result; and add the first test result to a chart.
[0009] In some embodiments, the plurality of first parameters at least include an impedance value of the first probe card, a cumulative number of wafer tests of the first probe card, a cumulative number of probe insertions and cleanings of the first probe card, and a wafer yield rate.
[0010] In some embodiments, the information collection unit is further used to receive multiple second parameters from a second test site having a second probe card, wherein the multiple second parameters are related to multiple detection results of the second test site testing the wafer through the second probe card; and the processing unit is further used to: convert the first model into a second model, wherein the second model is used to predict the state of the second probe card at the second test site under second test conditions; and perform the degradation analysis on the second probe card through the second model and the multiple second parameters to generate a second detection result.
[0011] In some embodiments, the processing unit is further configured to perform the degradation analysis on the first probe card using the first model, the plurality of first parameters, the second model, and the plurality of second parameters to generate a detection flag.
[0012] In some embodiments, the first test condition is suitable for high temperature testing, and the second test condition is suitable for low temperature testing.
[0013] The present invention provides a testing method. The testing method includes: receiving a plurality of first parameters from a first test site having a first probe card, wherein the first parameters are related to a plurality of test results of a wafer tested by the first probe card at the first test site; generating a first model corresponding to the first test site using an artificial intelligence algorithm and the first parameters, wherein the first model is used to predict the first probe card under a first test condition at the first test site; and performing a degradation analysis on the first probe card using the first model and the first parameters to generate a first test result.
[0014] In some embodiments, the degradation analysis includes: converting the multiple first parameters into corresponding multiple test values based on the artificial intelligence algorithm; determining whether each of the multiple test values is an error value to generate the first detection result; and generating an abnormal warning signal in response to the number of error values of the multiple test values being higher than a threshold.
[0015] In some embodiments, the plurality of first parameters at least include an impedance value of the first probe card, a cumulative number of wafer tests of the first probe card, a cumulative number of probe insertions and cleanings of the first probe card, and a wafer yield rate.
[0016] In some embodiments, the testing method further includes: receiving a plurality of second parameters from a second site having a second probe card, wherein the plurality of second parameters are related to a plurality of detection results of the second test site testing the wafer through the second probe card; converting the first model into a second model, wherein the second model predicts the state of the second probe card at the second test site under second test conditions; and performing the degradation analysis on the second probe card using the second model and the plurality of second parameters to generate a second detection result.
[0017] In some embodiments, the testing method further includes: performing the degradation analysis on the first probe card using the first model, the first parameter, the second model, and the plurality of second parameters to generate a detection flag.
[0018] In some embodiments, the first test condition is suitable for high temperature testing, and the second test condition is suitable for low temperature testing.
[0019] In summary, the test system and test method of the present invention utilize artificial intelligence algorithms to perform degradation analysis on parameters obtained from the probe card to test the degradation trend of the probe card and further determine whether the test site can operate normally. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 FIG. 1 is a schematic diagram of a test system and a test station according to an embodiment of the present invention.
[0021] Figure 2 Based on Figure 1 Test chart of embodiment.
[0022] Figure 3 FIG. 1 is a schematic diagram of a test system and a test station according to an embodiment of the present invention.
[0023] Figure 4 Based on Figure 3 Test chart of embodiment.
[0024] Figure 5 FIG. 1 is a flow chart of a testing method according to an embodiment of the present invention. DETAILED DESCRIPTION
[0025] The following will illustrate embodiments of the present invention with reference to the accompanying drawings, in which the same reference numerals represent the same or similar elements or method flows.
[0026] Please refer to Figure 1 , Figure 1 FIG. 1 is a schematic diagram of a test system 100 and a test site 200 according to an embodiment of the present invention. Figure 1 In the embodiment of FIG. 1 , the test system 100 can be coupled to the test site 200 and the external device 150 . The test system 100 includes an information collecting unit 110 , a processing unit 120 , a feedback unit 130 , and a report generating unit 140 .
[0027] The information collecting unit 110 may be a data transmission interface (eg, a probe card interconnection socket, a USB interface, an SPI bus, or a telecommunication transceiver) for receiving the plurality of parameters PR1 from the test site 200 .
[0028] The processing unit 120 may be a single-chip processor, a microcontroller, or a processor, which is not limited in the embodiment of the present invention.
[0029] The feedback unit 130 can also be a single chip processor, a microcontroller or a processor. The report generation unit 140 can also be a single chip processor, a microcontroller or a processor.
[0030] In some embodiments, the feedback unit 130 and the report generating unit 140 may be disposed in the same circuit element as the processing unit 120 . Alternatively, the three may be separate circuit elements, which is not limited in the embodiment of the present invention.
[0031] The test site 200 includes a probe card 210 and a test head 220. The probe card 210 can be placed on the test site 200 to connect to the test head 220. The probe card 210 is used to provide a wafer to be tested (eg, Figure 1 The probe card 210 is connected to the test station 200, thereby enabling the test station 200 to complete the test on the wafer WF1 (for example, the CP test mentioned above). Such testing is typically performed before wafer dicing and chip packaging. The embodiments of the present invention do not limit the physical characteristics of the probe card 210 itself or the details of its test operation.
[0032] The probe card 210 can be used to form a good electrical contact with the wafer WF1 during the test. However, regardless of the material of the probe tip of the probe card 210, due to continuous contact events (i.e., physical contact between the probe card 210 and the wafer WF1), each probe of the probe card 210 will accumulate more or less contaminants, which will have an adverse effect on the probe card 210's measurement of the wafer WF1, thereby preventing the test site 200 from operating normally. Through the test system 100 of the embodiment of the present invention, the processing unit 120 equipped with an artificial intelligence algorithm can be used to perform degradation analysis on the probe card 210, determine the degradation trend of the probe card 210, and then determine whether the test site 200 can operate normally. In addition, the test system 100 can also determine the time point when the test site 200 changes from "can operate normally" to "cannot operate normally."
[0033] like Figure 1 As described above, the information collection unit 110 can receive multiple parameters PR1 from the test site 200. These parameters PR1 are multiple test results related to the probe card 210 testing the wafer WF1. For example, the parameters PR1 may include: the impedance value of the contact resistance between the probe card 210 and the wafer WF1, the cumulative number of tests performed by the probe card 210, the cumulative number of probe insertions performed by the probe card 210, the number of probe cleanings performed by the probe card 210, the yield of the wafer WF1 itself, and abnormality warning signals issued by the test site 200.
[0034] The needle cleaning may be a probe card laser cleaning or any probe card cleaning method, which is not limited in the embodiment of the present invention.
[0035] exist Figure 1 In an embodiment, the processing unit 120 may be coupled to the information collection unit 110, the feedback unit 130, and the report generation unit 140. The processing unit 120 is equipped with an artificial intelligence algorithm. The artificial intelligence algorithm may be a machine learning classification algorithm. The processing unit 120 may generate a first model corresponding to the test site 200 using the artificial intelligence algorithm and a plurality of parameters PR1. The first model is used to predict the state of the probe card 210 when the test site 200 performs a first test condition.
[0036] The first test condition may be a temperature condition for operating the test station 200. When the temperature condition changes, the parameter PR1 measured by the test station 200 from the wafer WF1 via the probe card 210 will vary, and the test condition of the test station 200 will be adjusted accordingly. The temperature condition of the first test condition in this embodiment is not limited.
[0037] The processing unit 120 may perform degradation analysis on the probe card 210 based on the first model and the parameter PR1 and generate a detection result DR1. The detection result DR1 may be a string or data in any format, indicating whether the probe card 210 is operating normally.
[0038] It's worth noting that in the present invention, the first model is a machine learning model. The term "prediction" refers to the first model's operation of testing and analyzing the probe card 210 to obtain the test result DR1. The test result (also referred to as the "prediction result") of the first model can be used to predict the state of the probe card 210. However, the test result is not identical to the actual operating state of the probe card 210.
[0039] When performing the degradation analysis, the processing unit 120 can convert the parameter PR1 (for example, the impedance value of the contact resistance when the probe card 210 contacts the wafer WF1, the cumulative number of tests of the probe card 210, the cumulative number of needle insertions of the probe card 210, the number of needle cleanings of the probe card 210, the yield of the wafer WF1 itself, and the abnormal warning signal issued by the test site 200) into corresponding multiple test values TV1 based on the first model. Specifically, the processing unit 120 can perform feature extraction and feature transformation on the parameter PR1 through the above-mentioned machine learning classification algorithm to reduce the feature dimension of the parameter PR1 and retain important information of the parameter PR1, and use the important information as the multiple test values TV1. The feature extraction method used in this embodiment can be principal component analysis (PCA), independent component analysis (ICA) or t-distributed stochastic neighbor embedding (t-SNE), and the embodiment of the present invention is not limited to this.
[0040] Furthermore, the processing unit 120 may determine whether each of the plurality of test values TV1 is an error value, thereby generating a corresponding detection result DR1. In some embodiments, the error value may be a standard value preset by the processing unit 120. If any of the test values TV1 does not meet the standard value (e.g., is greater than, less than, or exceeds a range defined by the standard value), the test value TV1 is classified as an error value.
[0041] The processing unit 120 can predict the status of the probe card 210 based on whether the number of test values TV1 classified as error values is higher than a threshold value. The threshold value can be a value preset by the processing unit 120. Among the multiple test values TV1, when the number of test values TV1 classified as error values is higher than the threshold value, the processing unit 120 can generate an abnormal warning signal EWS1. For example, the threshold value is set to "10" in one embodiment. If there are less than ten error values in the test values TV1 of the embodiment, it means that the status of the probe card 210 is still within an acceptable range; on the contrary, if there are more than or equal to ten error values in the test values TV1 of the embodiment, it means that the probe card 210 may be degraded, and the processing unit 120 can issue an abnormal warning signal EWS1 in this case.
[0042] The processing unit 120 can transmit an abnormality warning signal EWS1 to the feedback unit 130. When the feedback unit 130 receives the abnormality warning signal EWS1, it indicates that the probe card 210 has failed, preventing the test site 200 from operating normally. In this case, the feedback unit 130 can transmit the abnormality warning signal EWS1 to the test site 200, causing the test site 200 to shut down according to the abnormality warning signal EWS1.
[0043] Furthermore, the feedback unit 130 may include a notification unit 135. The notification unit 135 may send an abnormal warning signal EWS1 to the external device 150 to notify the engineering personnel that the probe card 210 placed on the test site 200 has deteriorated.
[0044] The notification unit 135 may transmit the signal via wireless communication or wired signal transmission, which is not limited in the present embodiment. The external device 150 is another device independent of the test system 100 and may be any computing device capable of receiving signals, such as a computer, server, or mobile communication device, which is not limited in the present embodiment.
[0045] In this embodiment, the report generating unit 140 may receive the test result DR1 and the test value TV1, and add the test result DR1 to a chart (eg, Figure 2 Test chart TAB1).
[0046] In addition to the report generation unit 140, the test system 100 may further include a visual abnormality trend graph generation unit. This unit can record multiple sets of parameters PR1 collected by the test system 100 over and over, along with the test values TV1 processed by the processing unit 120, the test results DR1, and the abnormality warning signal EWS1, in the form of a trend graph, thereby enabling engineers to clearly understand the test history of the probe card 210.
[0047] In summary, the test system 100 can utilize the artificial intelligence algorithm of the processing unit 120 to perform degradation analysis on the parameter PR1 obtained by the probe card 210 to test the degradation trend of the probe card 210 and further determine whether the test site 200 can operate normally.
[0048] Please refer to Figure 2 , Figure 2 Based on Figure 1 The test chart TAB1 of the embodiment of the present invention records a plurality of test results of the products PD1 - PD5 tested by the test stations S1 - LT. Figure 2 The test site S1-LT can be used for Figure 1 Test station 200. In this embodiment, products PD1-PD5 may be electronic products having wafers (e.g., wafers similar to wafer WF1). Test system 100 can determine whether probe card 210 is functioning properly by testing the wafers of products PD1-PD5. After a period of testing, test system 100 can record multiple test results into a table.
[0049] As shown in test chart TAB1, the test results of test site S1-LT for product PD1 showed that the error value was 0 out of 97 test values. This means that test system 100 performed 97 tests on the wafers of product PD1 and found no errors throughout the entire testing process. Furthermore, the test results of test site S1-LT for product PD2 showed that the error value was 0 out of 166 test values; the test results of test site S1-LT for product PD3 showed that the error value was 0 out of 100 test values; the test results of test site S1-LT for product PD4 showed that the error value was 0 out of 81 test values; and the test results of test site S1-LT for product PD5 showed that the error value was 0 out of 30 test values.
[0050] In this embodiment, the number of test values for each of the products PD1-PD5 is preset based on the products PD1-PD5. The embodiment of the present invention does not limit the number of test values for different products. The numbers recorded in the test chart TAB1 are only examples to facilitate those skilled in the art to understand the technology of the embodiment of the present invention.
[0051] Please refer to Figure 3 , Figure 3 FIG. 1 is a schematic diagram of a test system 300 and test sites 200 and 250 according to an embodiment of the present invention. Figure 3 The test system 300 may correspond to Figure 1 Test system 100. Compared to Figure 1 The test system 100, Figure 3The test system 300 can simultaneously perform tests from multiple test sites with different probe cards and simultaneously perform degradation analysis on the different probe cards.
[0052] In Figure 3 In an embodiment, the test site 250 includes a probe card 260 and a test head 270. The probe card 210 can be placed on the test site 200 to connect the test head 220. The probe card 260 can be placed on the test site 250 to connect the test head 270. The information collection unit 310 can receive a plurality of parameters PR1 from the test site 200 and a plurality of parameters PR2 from the test site 250. The plurality of parameters PR1 are detection results about testing a wafer (e.g., the wafer WFl) by the probe card 210, and the plurality of parameters PR2 are detection results about testing a wafer (e.g., the wafer WFl or another wafer) by the probe card 260. Figure 1 Figure 1
[0053] The processing unit 320 can be coupled to the information collection unit 310 and the feedback unit 330. The processing unit 320 is loaded with an artificial intelligence algorithm. The artificial intelligence algorithm can be a machine learning classification algorithm. Similar to the processing unit 120, Figure 1 the processing unit 320, Figure 3 the processing unit 320 can generate a first model corresponding to the test site 200 by the artificial intelligence algorithm and the plurality of parameters PR1. The first model is used to predict a state of the probe card 210 of the test site 200 under a first test condition.
[0054] In this embodiment, the artificial intelligence algorithm loaded by the processing unit 320 can further convert the first model suitable for the test site 200 to a second model suitable for the test site 250. The second model is used to predict a state of the probe card 260 of the test site 200 under a second test condition, and the second model is also a machine learning model.
[0055] Specifically, the first test condition can be a temperature condition for operating the test site 200, and the second test condition can be another temperature condition for operating the test site 250. For example, the first test condition is a test suitable for a first temperature, and the second test condition is a test suitable for a second temperature, and the first temperature is significantly higher than the second temperature. That is, the operating temperature of the test site 200 under the first test condition is greater than the operating temperature of the test site 250 under the second test condition. Therefore, the characteristics of the parameters PR1 tested by the test site 200 are different from the characteristics of the parameters PR2 tested by the test site 250.
[0056] The processing unit 320 can perform degradation analysis on the probe card 210 based on the first model and parameter PR1 and generate a detection result DR1 and an abnormality warning signal EWS1. The processing unit 320 can perform degradation analysis on the probe card 260 based on the second model and parameter PR2 and generate a detection result DR2 and an abnormality warning signal EWS2.
[0057] When performing the degradation analysis, the processing unit 320 may convert the parameter PR1 into corresponding test values TV1 based on the first model, and the processing unit 320 may convert the parameter PR2 into corresponding test values TV2 based on the second model.
[0058] The processing unit 320 can determine whether the number of test values TV1 classified as error values is higher than a threshold value, thereby predicting the status of the probe card 210. Among the plurality of test values TV1, when the number of test values TV1 classified as error values is higher than the threshold value, the processing unit 320 can generate an abnormality warning signal EWS1. Furthermore, the processing unit 320 can determine whether the number of test values TV2 classified as error values is higher than a threshold value (the threshold value corresponding to the test values TV2 may be the same as or different from the threshold value corresponding to the test values TV1), thereby predicting the status of the probe card 260. Among the plurality of test values TV2, when the number of test values TV2 classified as error values is higher than the threshold value, the processing unit 320 can generate an abnormality warning signal EWS2.
[0059] The processing unit 320 can transmit the detection results DR1 and DR2 to the feedback unit 330. When the feedback unit 330 receives the abnormal warning signal EWS1, it can transmit the abnormal warning signal EWS1 to the test site 200, causing the test site 200 to shut down according to the abnormal warning signal EWS1. When the feedback unit 330 receives the abnormal warning signal EWS2, it can transmit the abnormal warning signal EWS2 to the test site 250, causing the test site 250 to shut down according to the abnormal warning signal EWS2.
[0060] In some embodiments, the processing unit 320 can further perform a degradation analysis on the probe card 210 using the first model, the parameter PR1, the second model, and the parameter PR2 to generate a detection flag DF. Specifically, the processing unit 320 can compare the test value TV1 with the test value TV2 to determine whether the probe card 210 is more degraded than the probe card 260, and indicate this determination result using the detection flag DF. The detection flag DF can be transmitted to the feedback unit 330.
[0061] In the various embodiments described above, when the number of error values in the test value TV1 is significantly higher than the number of error values in the test value TV2, the processing unit 320 may determine that the probe card 210 is degraded. In this case, the test flag DF will include the abnormality warning signal EWS1. When the test flag DF includes the abnormality warning signal EWS1, the feedback unit 330 may transmit the test flag DF including the abnormality warning signal EWS1 to the test site 200, causing the test site 200 to shut down based on the abnormality warning signal EWS1.
[0062] The report generation unit 340 may receive the detection result DR1, the test value TV1, the detection result DR2, and the test value TV2, and add the detection result DR2 to a chart (eg, Figure 4 (see chart TAB2).
[0063] Please refer to Figure 4 , Figure 4 Based on Figure 3 The test chart TAB2 of the embodiment. Figure 4 The test sites S1-LT, S1-WLBT, S2-HT, and P1-HT are all available. Figure 3 Test sites 200, 250. For example, Figure 4 The test site S2-HT can correspond to Figure 3 The test site 200, Figure 4 The test site S1-WLBI or P1-HT can correspond to Figure 3 The test site 250. The test chart TAB2 records the test results of placing the test sites S1-LT, S1-WLBT, S2-HT, and P1-HT on multiple products PD1-PD5.
[0064] In the test chart TAB2, the test result of the test site S2-HT for product PD2 is that the error value is 4 out of 231 test values, and the test result for product PD5 is that the error value is 5 out of 338 test values.
[0065] Compared with other test sites, the test results of test site S1-WLBI for product PD2 were that the error value was 0 in 143 test values, and the test results for product PD5 were that the error value was 0 in 95 test values; the test results of test site P1-HT for product PD2 were that the error value was 0 in 225 test values, and the test results for product PD5 were that the error value was 0 in 230 test values.
[0066] From this, we can see that the test results of the test site S2-HT have more error values, which means that the probe card placed at the test site S2-HT should be relatively degraded. Figure 3The test system 300 may enable the detection flag DF to include the abnormal warning signal EWS1 and transmit the detection flag DF including the abnormal warning signal EWS1 to the test site S2-HT to shut down the test site S2-HT.
[0067] Please refer to Figure 5 . Figure 5 FIG. 5 is a flow chart of a testing method 500 according to an embodiment of the present invention. The testing method 500 is a method used by the testing system 100 to test whether the probe card 210 placed in the testing station 200 can function normally.
[0068] In step S510 , the test system 100 may receive a plurality of first parameters (ie, parameters PR1 ) from a first test site (ie, the test site 200 ) having a first probe card (ie, the probe card 210 ).
[0069] In step S520 , the testing system 100 may generate a first model corresponding to the first testing site using an artificial intelligence algorithm and a plurality of first parameters.
[0070] In step S530 , the testing system 100 may perform degradation analysis on the probe card 210 using the first model and the first parameters.
[0071] In step S540 , the testing system 100 may generate a first test result (ie, test result DR1 ).
[0072] In summary, the test system and test method of the present invention utilize artificial intelligence algorithms to perform degradation analysis on parameters obtained from the probe card to test the degradation trend of the probe card and further determine whether the test site can operate normally.
[0073] The above are only preferred embodiments of the present invention. Various modifications and equivalents may be made to the present invention without departing from the scope or spirit of the present invention. In summary, all modifications and equivalents made to the present invention within the scope of the following claims are within the scope of the present invention.
[0074]
Explanation of symbols
[0075] 100,300: Test system
[0076] 110,310: Information Collection Unit
[0077] 120,320: processing units
[0078] 130,330: Feedback unit
[0079] 135: Notification unit
[0080] 140,340: Report generation unit
[0081] 150: External device
[0082] 200,250: Test site
[0083] 210,260: Probe card
[0084] 220,270: Test head
[0085] WF1: Wafer
[0086] PR1, PR2: parameters
[0087] DR1, DR2: test results
[0088] TV1, TV2: test value
[0089] EWS1, EWS2: abnormal warning signal
[0090] TAB1, TAB2: test chart
[0091] PD1,PD2,PD3,PD4,PD5: Products
[0092] S1-LT, S1-WLBI, S2-HT, P1-HT: Test Site
[0093] 500: Test Method
[0094] S510, S520, S530, S540: steps.
Claims
1. A testing system, characterized in that: Include: an information collecting unit configured to receive a plurality of first parameters from a first test site having a first probe card, wherein the plurality of first parameters are related to a plurality of detection results of the first test site testing a wafer through the first probe card; as well as a processing unit, coupled to the information collecting unit, configured to: generating a first model corresponding to the first test site using an artificial intelligence algorithm and the plurality of first parameters, wherein the first model is used to predict a state of the first probe card when the first test site is subjected to a first test condition; as well as Based on the first model and the plurality of first parameters, a degradation analysis is performed on the first probe card, and a first detection result is generated.
2. The test system according to claim 1, wherein: The degradation analysis is used to: converting the plurality of first parameters into corresponding plurality of test values based on the artificial intelligence algorithm; determining whether each of the plurality of test values is an error value to generate the first detection result; as well as In response to the number of the plurality of test values being erroneous values being higher than a threshold, the processing unit test is caused to issue an abnormal warning signal.
3. The test system according to claim 2, wherein: The testing system further comprises: The feedback unit is coupled to the processing unit and is used for receiving the abnormal warning signal and transmitting the abnormal warning signal to the first test site to shut down the first test site.
4. The test system according to claim 3, characterized in that: The feedback unit includes a notification unit configured to send the abnormality warning signal to an external device in response to the feedback unit receiving the abnormality warning signal.
5. The test system according to claim 2, wherein: The test system further includes a report generating unit, coupled to the processing unit and configured to: receiving the first detection result; and The first detection result is added to the chart.
6. The test system according to claim 1, wherein: The plurality of first parameters at least include an impedance value of the first probe card, a cumulative number of wafer tests of the first probe card, a cumulative number of probe insertions and cleanings of the first probe card, and a wafer yield rate.
7. The test system according to claim 1, wherein: The information collecting unit is further configured to receive a plurality of second parameters from a second test site having a second probe card, wherein the plurality of second parameters are related to a plurality of inspection results of the second test site testing the wafer through the second probe card; and The processing unit is further configured to: converting the first model into a second model, wherein the second model is used to predict a state of the second probe card when the second test site performs a second test condition; as well as The degradation analysis is performed on the second probe card using the second model and the plurality of second parameters to generate a second detection result.
8. The test system according to claim 7, characterized in that: The processing unit is further configured to perform the degradation analysis on the first probe card using the first model, the plurality of first parameters, the second model, and the plurality of second parameters to generate a detection flag.
9. The test system according to claim 7, characterized in that: The first test condition is suitable for high temperature testing, and the second test condition is suitable for low temperature testing.
10. A testing method, characterized in that: Include: receiving a plurality of first parameters from a first test site having a first probe card, wherein the plurality of first parameters are related to a plurality of inspection results of the first test site testing a wafer through the first probe card; generating a first model corresponding to the first test site using an artificial intelligence algorithm and the plurality of first parameters, wherein the first model is used to predict a state of the first probe card when the first test site is subjected to a first test condition; as well as A degradation analysis is performed on the first probe card using the first model and the plurality of first parameters, and a first detection result is generated.
11. The testing method according to claim 10, characterized in that: The degradation analysis includes: converting the plurality of first parameters into corresponding plurality of test values based on the artificial intelligence algorithm; determining whether each of the plurality of test values is an error value to generate the first detection result; as well as In response to the number of the plurality of test values being erroneous values being higher than a threshold, an abnormality warning signal is generated.
12. The testing method according to claim 10, characterized in that: The plurality of first parameters at least include an impedance value of the first probe card, a cumulative number of wafer tests of the first probe card, a cumulative number of probe insertions and cleanings of the first probe card, and a wafer yield rate.
13. The testing method according to claim 10, characterized in that: The test method further comprises: receiving a plurality of second parameters from a second site having a second probe card, wherein the plurality of second parameters are related to a plurality of inspection results of the wafer tested by the second probe card at the second test site; converting the first model into a second model, wherein the second model predicts a state of the second probe card when the second test site performs a second test condition; as well as The degradation analysis is performed on the second probe card using the second model and the plurality of second parameters to generate a second detection result.
14. The testing method according to claim 13, characterized in that: The test method further comprises: The degradation analysis is performed on the first probe card using the first model, the first parameter, the second model, and the plurality of second parameters to generate a detection flag.
15. The testing method according to claim 13, characterized in that: The first test condition is suitable for high temperature testing, and the second test condition is suitable for low temperature testing.