A sampling method for improving the detection efficiency of ceramic substrate performance

Abstract

A sampling method for improving the detection efficiency of ceramic substrate performance. Step 1: Collect customer quality requirements, including the total number of contacts on the ceramic substrate, the acceptable quality level, and the consumer risk. Step 2: The manufacturer sets the detection quality requirements, including the acceptable quality level and the producer risk. Step 3: Develop a sampling plan based on the customer quality requirements and the manufacturer's detection quality requirements, and determine the required sample contact number and the non-conductive contact number threshold in the sampling plan. Step 4: Use a flying probe machine to conduct a contact continuity test to obtain the number of non-conductive contacts. A sampling detection plan can be quickly developed according to customer requirements. While maintaining the defective rate of the finished products delivered to customers to meet the requirements, the number of contacts inspected is greatly reduced, which can effectively save detection time, improve production efficiency, and has high economic benefits.

Description

Technical Field

[0001] The present application relates to the detection technology of directly plated copper substrates (DPC substrates), and particularly to a sampling method for improving the detection efficiency of the performance of ceramic substrates. Background Art

[0002] The directly plated copper ceramic substrate (DPC substrate) is a commonly used packaging substrate in the electronics industry. Its main technical features are as follows: on the surface of the cleaned ceramic substrate, a metal seed layer is deposited using vacuum sputtering technology, and then the circuit is fabricated through photolithography, development, and etching processes. Finally, copper is plated by electroplating / electroless plating, etc., and the circuit fabrication is completed after removing the photoresist. The DPC substrate has the advantages of good thermal conductivity, heat resistance, good insulation performance, and high circuit accuracy, and is commonly used in the packaging of power devices such as light-emitting diodes (LEDs) and lasers (LDs).

[0003] When evaluating the quality of the DPC substrate, an important indicator is the continuity of the contacts (vertical vias) on the substrate. Due to the requirements of packaging technology, the contacts on the DPC substrate need to enable the front and back sides of the substrate to conduct with each other. This is mainly achieved by drilling vias on the DPC substrate and then electroplating copper to fill the vias. During the electroplating process, due to process limitations, defects may occur during the via filling process, resulting in the inability of the contacts to conduct (open circuit). In order to improve the subsequent packaging quality and efficiency, it is required that the number of non-conductive contacts on the substrate does not exceed a certain proportion, otherwise the entire substrate should be judged as a defective product.

[0004] Currently, DPC substrate manufacturers use a full inspection method for detecting the continuity of the substrate contacts, that is, using a flying probe machine to individually detect the continuity of each contact on the substrate (100% detection). If the proportion of non-powered contacts on the substrate reaches a certain level, the substrate is judged as a defective product. This method has a long detection time and low detection efficiency, which limits the production efficiency.

[0005] The above-mentioned prior art has the following defects: The traditional full inspection mode for detecting the continuity of the contacts of the DPC substrate takes too long, which limits the production efficiency of the production line. Summary of the Invention

[0006] In order to improve the problem of low detection efficiency caused by the full inspection mode for detecting the continuity of the contacts of the existing DPC substrate, the present application provides a sampling method for improving the detection efficiency of the performance of ceramic substrates. The following technical solutions are adopted:

[0007] A sampling method for improving the detection efficiency of the performance of ceramic substrates includes the following specific steps:

[0008] Step 1: Collect customer quality requirements, including the total number of contacts on the ceramic substrate, the acceptable quality level, and the consumer risk;

[0009] Step 2: The manufacturer sets the inspection quality requirements, including the acceptable quality level and the producer's risk;

[0010] Step 3: Develop a sampling plan based on the customer's quality requirements and the manufacturer's inspection quality requirements, and determine the required number of sample contacts and the threshold of non-conductive contacts in the sampling plan;

[0011] Step 4: Use a flying probe machine to perform a contact continuity test to obtain the number of non-conductive contacts;

[0012] Step 5: Compare the number of non-conductive contacts obtained in the test in Step 4 with the threshold of non-conductive contacts in Step 3 to obtain the sampling result of whether the ceramic substrate is qualified.

[0013] The acceptable quality level in Step 1 refers to the customer's judgment criterion for whether the ceramic substrate is a defective product. The consumer's risk refers to the maximum batch defective rate allowed among all products accepted by the customer.

[0014] The acceptable quality level in Step 2 refers to the actual non-conductive rate of the contacts of the ceramic substrate during the manufacturer's trial production. The producer's risk refers to the maximum proportion of qualified finished products that the manufacturer can accept as being misjudged as defective according to cost requirements;

[0015] Based on the customer's quality requirements and the manufacturer's inspection quality requirement data, determine the specific required number of sample contacts and the threshold of non-conductive contacts in the sampling plan. The size of the sample contacts refers to the number of samples required for a single sampling inspection during sampling inspection. If the number of non-conductive contacts among the sampled contacts is higher than the threshold of non-conductive contacts, this substrate is determined to be defective and repaired or scrapped. Otherwise, further determination is required.

[0016] Optionally, in Step 2, use Minitab software to generate the required number of sample contacts and the threshold of non-conductive contacts for the sampling plan. The specific steps are as follows: Run the attribute sampling acceptance of Minitab software, select Create Sampling Plan, input the total number of contacts in the customer's quality requirements in Step 1 into Lot Size, input the acceptable quality level into Lot Tolerance Percent Defective (LTPD), input the consumer's risk into Consumer's Risk (Beta), input the acceptable quality level in the inspection quality requirements in Step 2 into Acceptable Quality Level (AQL), input the producer's risk into Producer's Risk (Alpha), and run the software to obtain the sample quantity and acceptance number of the sampling plan. Among them, the sample quantity is the size of the sample contacts, and the acceptance number is the threshold of non-conductive contacts.

[0017] By adopting the above technical solution, specifically input the corresponding customer quality requirements and manufacturer's inspection quality requirement data into the relevant positions in the sampling plan creation interface of Minitab software for acceptance sampling by attributes, and running the software can obtain the sample size and acceptance number of the sampling plan, where the sample size is the size of the sample contact points, and the acceptance number is the threshold of non-conductive contact points.

[0018] Optionally, the threshold of non-conductive contact points refers to the maximum number of non-conductive contact points allowed for the contact continuity test using a flying probe machine among the sampled contact points. If the actually tested number of non-conductive contact points exceeds the threshold of non-conductive contact points, the ceramic substrate is determined to be a defective product.

[0019] By adopting the above technical solution, set the threshold of non-conductive contact points. When the number of non-conductive contact points obtained from the contact continuity test using a flying probe machine is greater than the threshold of non-conductive contact points, determine this substrate as a defective product. If the number of non-conductive contact points obtained from the test is not greater than the threshold of non-conductive contact points, further determination is required.

[0020] Optionally, when running Minitab software to generate a sampling plan, an operating characteristic (OC) curve graph and an average total inspection (ATI) comparison table are also generated.

[0021] By adopting the above technical solution, evaluate the sampling plan in step 2 through the operating characteristic (OC) curve graph to determine the probability that the batch of products can pass customer acceptance under the condition of corresponding different actual batch defective rates of the products when adopting this sampling plan.

[0022] The generation of the average total inspection (ATI) comparison table provides a reference for the standard of the re-inspection plan.

[0023] When the tested defective rate of the product is the acceptable quality level AQL, if the defective products detected according to the sampling plan definition are re-inspected and the re-inspection method is 100% inspection, then after inspecting all products according to this sampling plan, the value of the number of contact points that need to be detected for each product on average is the average total inspection (ATI) value, simply referred to as the ATI value. The increase of the ATI value means the extension of the detection time. Without affecting the detection accuracy, reduce the detection time as much as possible to reduce the detection cost. Therefore, selecting a suitable re-inspection threshold is an important factor in measuring whether the sampling method is suitable.

[0024] Optionally, the following method is specifically adopted in step 5 to determine the sampling result: if the number of non-conductive contact points obtained from the flying probe machine test is greater than the threshold of non-conductive contact points, determine the ceramic substrate as a defective product; if the number of non-conductive contact points obtained from the test is less than or equal to the threshold of non-conductive contact points, it is necessary to further determine whether re-inspection is required according to the average total inspection (ATI) comparison table.

[0025] By adopting the above technical solution, since the defect points are evenly distributed on the ceramic substrate, there is no need to consider the interval and distribution factors when extracting the detection points, and the points can be randomly selected. To simplify the programming of the flying probe machine, a continuous point selection method is adopted. Specifically, starting from any one of the four corners of the substrate, the same number of contacts as the sample contact number are sequentially selected for continuity testing. If the number of non-conductive contacts among these contacts is greater than the non-conductive contact number threshold, it can be determined that the substrate is a defective product and is scrapped or repaired. If the number of non-conductive contacts obtained from the test is less than or equal to the non-conductive contact number threshold, it is necessary to further determine whether a re-inspection is required according to the average total inspection number (ATI) comparison table.

[0026] Optionally, the specific method for determining whether a re-inspection is required is: define the total number of contacts as n, the average total inspection number (ATI) as ATI, the number of non-conductive contacts in step 4 as e, the non-conductive contact number threshold as x, set 35%n ≤ ATI ≤ 45%n, read the average total inspection number (ATI) comparison table, find the average total inspection number (ATI) that meets the requirements, and the non-conductive contact number corresponding to this average total inspection number (ATI) is set as the re-inspection threshold k. When k < e ≤ x, it is necessary to re-inspect the ceramic substrate.

[0027] By adopting the above technical solution, due to the inevitable possibility of missed inspections in sampling inspection, when the number of non-conductive detected by the flying probe machine is close to the non-conductive point number threshold, if there is a missed inspection, there is a risk of misjudging a defective product as a qualified product. Therefore, the sampling plan needs to set a re-inspection threshold k. However, the smaller the re-inspection threshold k, the larger the proportion of contacts that need to be re-inspected, which leads to a larger corresponding ATI value. An increase in the ATI value means an extension of the detection time. Without affecting the detection accuracy, the original intention of formulating the sampling plan is to reduce the detection time as much as possible and lower the detection cost. Therefore, after comprehensive consideration, it is set that when the average total inspection number (ATI) is in the range of 35% to 45% of the total number of contacts n, this plan is the best. By reading the average total inspection number (ATI) comparison table, find the average total inspection number (ATI) that meets the requirements, and the non-conductive contact number corresponding to this average total inspection number (ATI) is set as the re-inspection threshold k. When k < e ≤ x, it is necessary to re-inspect the ceramic substrate.

[0028] Optionally, the re-inspection method is to detect all the contacts of the ceramic substrate. When the defective rate of the re-inspected contacts is less than the acceptable quality level, it is determined that the ceramic substrate is a qualified product, otherwise it is determined that the ceramic substrate is a defective product.

[0029] By adopting the above technical solution, the full inspection method is used to continue the re-inspection of the ceramic substrate that meets the re-inspection conditions. In this way, the ratio of the number of non-conductive contacts in the re-inspection to the total number of contacts is the defective rate of the contacts of the ceramic substrate. When the defective rate of the contacts is less than the acceptable quality level, it is determined that the ceramic substrate is a qualified product, otherwise it is determined that the ceramic substrate is a defective product.

[0030] In summary, the present application includes at least one of the following beneficial technical effects:

[0031] The present invention provides a sampling method for improving the detection efficiency of ceramic substrate performance, which can quickly formulate a sampling detection plan according to customer requirements. While maintaining the defective rate of the finished products delivered to customers to meet the requirements, the number of contacts to be inspected is greatly reduced, the possibility of missed inspection and false inspection is minimized, the detection time can be effectively saved, the production efficiency can be improved, and it has high economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a schematic flow chart of the present invention.

[0033] Figure 2 is a schematic diagram of the software parameter setting interface for generating a sampling plan using Minitab software in step 2 of the present invention;

[0034] Figure 3 is an operating characteristic (OC) curve of a sampling inspection feature of an embodiment of the present invention;

[0035] Figure 4 is a sampling schematic diagram for testing the continuity of contacts by a flying probe machine of the present invention;

[0036] Figure 5 is a schematic diagram of the software parameter setting interface of the comparative experience plan of the present invention;

[0037] Figure 6 is an operating characteristic (OC) curve of the comparative experience plan of the present invention;

[0038] Figure 7 is a schematic diagram of the software parameter setting interface of another comparative experience plan of the present invention;

[0039] Figure 8 is an operating characteristic (OC) curve of another comparative experience plan of the present invention; DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] The following further elaborates on the present application in conjunction with the attached Figure 1-8 for a more detailed description.

[0041] The embodiments of the present application disclose a sampling method for improving the detection efficiency of ceramic substrate performance.

[0042] Referring to Figure 1-8 , a sampling method for improving the detection efficiency of ceramic substrate performance includes the following specific steps:

[0043] Step 1: Collect customer quality requirements, including the total number of contacts on the ceramic substrate, the rejectable acceptable quality level, and the consumer risk;

[0044] Step 2: The manufacturer sets the inspection quality requirements, including the acceptable quality level and the producer risk;

[0045] Step 3: Develop a sampling plan based on the customer quality requirements and the manufacturer's inspection quality requirements, and determine the required number of sample contacts and the threshold of non-conductive contact numbers in the sampling plan;

[0046] Step 4: Use a flying probe machine to conduct a continuity test on the contacts to obtain the number of non-conductive contacts;

[0047] Step 5: Compare the number of non-conductive contacts obtained in the test in Step 4 with the threshold of non-conductive contact numbers in Step 3 to obtain the sampling result of whether the ceramic substrate is qualified.

[0048] The rejectable acceptable quality level in Step 1 refers to the customer's judgment criterion for whether the ceramic substrate is a defective product. The consumer risk refers to the maximum batch defective rate allowed among all products accepted by the customer.

[0049] The acceptable quality level in Step 2 refers to the actual non-conductive rate of the contacts of the ceramic substrate during the manufacturer's trial production process. The producer risk refers to the maximum proportion of qualified finished products that the manufacturer can accept as being misjudged as defective according to its cost requirements;

[0050] Based on the customer quality requirements and the manufacturer's inspection quality requirement data, determine the specific required number of sample contacts and the threshold of non-conductive contact numbers in the sampling plan. The size of the sample contacts refers to the number of samples required for a single sampling inspection during sampling inspection. If the number of non-conductive contacts among the sampled contacts is higher than the threshold of non-conductive contact numbers, this substrate is judged as defective and should be repaired or scrapped. Otherwise, further judgment is required.

[0051] In Step 2, use Minitab software to generate the required number of sample contacts and the threshold of non-conductive contact numbers for the sampling plan. The specific steps are as follows: Run the attribute sampling acceptance of Minitab software, select Create Sampling Plan, input the total number of contacts in the customer quality requirements in Step 1 into Lot Size, input the rejectable acceptable quality level into Lot Tolerance Percent Defective (LTPD), input the consumer risk into Consumer's Risk Beta, input the acceptable quality level in the inspection quality requirements in Step 2 into Acceptable Quality Level (AQL), input the producer risk into Producer's Risk Alpha, and run the software to obtain the sample size and the acceptance number of the sampling plan. Among them, the sample size is the size of the sample contacts, and the acceptance number is the threshold of non-conductive contact numbers.

[0052] Specifically, input the corresponding customer quality requirements and manufacturer's inspection quality requirement data into the relevant positions in the sampling plan creation interface of Minitab software for attribute sampling acceptance. Running the software can obtain the sample size and acceptance number of the sampling plan, where the sample size is the size of the sample contact points, and the acceptance number is the threshold of non-conductive contact points.

[0053] The threshold of non-conductive contact points refers to the maximum number of non-conductive contact points allowed for the contact continuity test using a flying probe machine among the sample contact points drawn. If the actually tested number of non-conductive contact points exceeds the threshold of non-conductive contact points, the ceramic substrate is determined to be a defective product.

[0054] Set the threshold of non-conductive contact points. When the number of non-conductive contact points obtained from the contact continuity test using a flying probe machine is greater than this threshold of electrical contact points, determine this substrate as a defective product. If the tested number of non-conductive contact points is not greater than this threshold of electrical contact points, further determination is required.

[0055] When running Minitab software to generate a sampling plan, an operating characteristic (OC) curve graph and an average total inspection (ATI) comparison table are also generated.

[0056] Evaluate the sampling plan in step 2 through the operating characteristic (OC) curve graph to determine the probability that the batch of products can pass customer acceptance under the corresponding different actual batch defective rates of the products when adopting this sampling plan.

[0057] The generation of the average total inspection (ATI) comparison table provides a reference for the standard of the re-inspection plan.

[0058] When the tested defective rate of the product is the acceptable quality level AQL, if the defective products detected according to the sampling plan definition are re-inspected and the re-inspection method is 100% inspection, then after inspecting all products according to this sampling plan, the value of the number of contact points that need to be detected for each product on average is the average total inspection (ATI) value, simply referred to as the ATI value. The increase in the ATI value means the extension of the detection time. Without affecting the detection accuracy, reducing the detection time as much as possible to reduce the detection cost. Therefore, selecting a suitable re-inspection threshold is an important factor in measuring whether the sampling method is suitable.

[0059] Specifically, the following method is adopted in step 5 to determine the sampling result: If the number of non-conductive contact points obtained from the flying probe machine test is greater than the threshold of non-conductive contact points, then determine the ceramic substrate as a defective product; if the tested number of non-conductive contact points is less than or equal to the threshold of non-conductive contact points, it is necessary to further determine whether re-inspection is required according to the average total inspection (ATI) comparison table.

[0060] Since the defective points are evenly distributed on the ceramic substrate, there is no need to consider the interval and distribution factors when extracting the detection points, and the points can be randomly selected. To simplify the programming of the flying probe machine, a continuous point-taking method is adopted. Specifically, starting from any one of the four corners of the substrate, the same number of contacts as the number of sample contacts are taken in sequence for continuity testing. If the number of non-conductive contacts among these contacts is greater than the non-conductive contact number threshold, it can be determined that the substrate is a defective product and is scrapped or repaired. If the number of non-conductive contacts obtained from the test is less than or equal to the non-conductive contact number threshold, it is necessary to further determine whether a re-inspection is required according to the average total inspection number (ATI) comparison table.

[0061] The specific method for determining whether a re-inspection is required is as follows: Define the total number of contacts as n, the average total inspection number (ATI) as ATI, the number of non-conductive contacts in step 4 as e, the non-conductive contact number threshold as x, set 35%n ≤ ATI ≤ 45%n, read the average total inspection number (ATI) comparison table, find the average total inspection number (ATI) that meets the requirements, and the non-conductive contact number corresponding to this average total inspection number (ATI) is set as the re-inspection threshold k. When k < e ≤ x, the ceramic substrate needs to be re-inspected.

[0062] Since there is an inevitable possibility of missed inspections in sampling inspection, when the number of non-conductive detections by the flying probe machine is close to the non-conductive point number threshold, if there is a missed inspection, there is a risk of misclassifying defective products as qualified products. Therefore, the sampling plan needs to set a re-inspection threshold k. However, the smaller the re-inspection threshold k, the larger the proportion of contacts that need to be re-inspected, which leads to a larger corresponding ATI value. An increase in the ATI value means an extension of the detection time. The original intention of formulating the sampling plan is to reduce the detection time as much as possible and lower the detection cost without affecting the detection accuracy. Therefore, after comprehensive consideration, it is set that when the average total inspection number (ATI) is in the range of 35% to 45% of the total number of contacts n, this plan is the optimal. By reading the average total inspection number (ATI) comparison table, find the average total inspection number (ATI) that meets the requirements, and the non-conductive contact number corresponding to this average total inspection number (ATI) is set as the re-inspection threshold k. When k < e ≤ x, the ceramic substrate needs to be re-inspected.

[0063] The re-inspection method is to detect all the contacts of the ceramic substrate. When the defective rate of the re-inspected contacts is less than the acceptable quality level, the ceramic substrate is determined to be a qualified product; otherwise, the ceramic substrate is determined to be a defective product.

[0064] The full inspection method is used to continue the re-inspection of the ceramic substrate that meets the re-inspection conditions. In this way, the ratio of the number of non-conductive contacts in the re-inspection to the total number of contacts is the defective rate of the contacts of the ceramic substrate. When the defective rate of the contacts is less than the acceptable quality level, the ceramic substrate is determined to be a qualified product; otherwise, the ceramic substrate is determined to be a defective product.

[0065] The implementation principle of a sampling method for improving the detection efficiency of ceramic substrate performance in an embodiment of the present application is as follows:

[0066] The customer requirements collected are as follows: the total number of contacts on the substrate is 728, the consumer risk is 5%, the rejectable quality level is 5%, the acceptable quality level set by the manufacturer is 1%, and the producer risk is 2.5%.

[0067] Run the acceptance sampling by attributes in Minitab software, select to create a sampling plan, input the acceptable quality level of 1% into the acceptable quality level AQL (unit: %) in the acceptance sampling by attributes, input the rejectable quality level of 5% into the rejectable quality level LTPD (unit: %), input the total number of contacts on the ceramic substrate, 728, into the lot size n, input the producer risk of 2.5% into the producer risk Alpha, and input the consumer risk of 5% into the consumer risk Beta. Running the Minitab software can export the required sample size of 208 and the acceptance number of 5. Among them, the sample size of 208 is the size of the sample contact points to be sampled, and the acceptance number of 5 is the threshold of non-conductive contact points. The specific input interface is as Figure 2 shown, and the generated sampling plan and operating characteristic (OC) curve are as Figure 3 shown.

[0068] The generated average total inspection number (ATI) comparison table is shown in Table 1:

[0069] Table 1

[0070] Number of non-conductive contacts Average Total Inspection ATI 2 387.4 3 289.5 4 238.8 5 217.9

[0071] Set 35%n ≤ ATI ≤ 45%n, substitute n = 728, and get 254.8 ≤ ATI ≤ 327.6. Read the average total inspection number (ATI) comparison table, and find that the average total inspection number (ATI) that meets the requirements is 289.5. The non-conductive contact number corresponding to this average total inspection number (ATI) of 289.5 is 3, that is, set the re-inspection threshold k = 3. When 3 < e ≤ 5, the ceramic substrate needs to be re-inspected, that is, when the non-conductive contact numbers are 4 and 5, the ceramic substrate needs to be re-inspected, and the re-inspection method is full inspection.

[0072] After obtaining the sample contact point size of 208, the non-conductive contact threshold of 5, and the re-inspection threshold of 3 for the sampling plan, start to detect this batch of ceramic substrates. Clamp the single ceramic substrate to be inspected on the flying probe machine in turn, and set the test trajectory of the flying probe machine to start from any one of the four corners of the ceramic substrate to be tested, and perform tests in turn. The detection trajectory is as Figure 4 shown. Set the number of test contact points to the sample size of 208, and then the flying probe machine can be started to complete the test.

[0073] Randomly inspect 208 contacts. If the number of defective points among them is greater than 5, the substrate can be determined to be a defective product.

[0074] After the test, the number of non-conductive contacts is obtained. If the number of non-conductive contacts is 0 - 3, it can be directly determined that this ceramic substrate is a qualified product;

[0075] If the number of non-conductive contacts is 4 or 5, the ceramic substrate needs to be re-inspected. During the re-inspection, when the ratio of the number of re-inspected non-conductive contacts to the total number of contacts is the contact defective rate of this ceramic substrate, when the contact defective rate is less than the acceptable quality level of 5%, it is determined that this ceramic substrate is a qualified product; otherwise, it is determined that this ceramic substrate is a defective product.

[0076] After the inspection of this batch of ceramic substrates is completed, the total number of defective products of this batch of ceramic substrates will be obtained. The total number of defective products divided by the total number of this batch of ceramic substrates is the defective rate of this batch of products. The customer determines whether this batch of products is qualified based on the defective rate of this batch of products.

[0077] According to the sampling inspection characteristic (OC) curve, it can be seen that by using this sampling method, when the actual defective rate of this batch of ceramic substrates reaches 1%, the probability of passing the quality inspection and acceptance exceeds 95%. When the defective rate reaches 2%, the probability of passing the quality inspection and acceptance can also reach 75%. It is determined that the formulation of this sampling plan has high feasibility and the expectation of passing the quality inspection and acceptance.

[0078] Next, compare the feasibility of the sampling plan by modifying the plan:

[0079] Evaluate the sampling plan through the calculation of Minitab software and compare it with the empirical plan:

[0080] Empirical plan 1: Strict sampling to improve quality requirements. Since in production practice, a test defective rate of 1% can be achieved in terms of the on-off property of the contacts, the sampling plan can be designed according to experience as follows: when the number of non-conductive contacts (i.e., the defective rate) in the sampled sample is higher than 1%, it is determined to be a defective product.

[0081] Run the acceptance sampling by attributes of Minitab software, select to compare the user-defined sampling plan, input the acceptable contact defective rate of 1% into the acceptable quality level in the acceptance sampling by attributes, input the rejectable batch defective rate of 5% into the rejectable quality level, input the sample quantity into the value 208 generated by the previous sampling plan, adjust the acceptable number from 5 to 2, and the batch size is still the total number of contacts of this ceramic substrate, which is 728. Its specific input interface is as Figure 5 shown, and the generated sampling plan and sampling inspection characteristic (OC) curve are as Figure 6 shown.

[0082] According to the results of the operating characteristic (OC) curve, even if the defective rate of the test actually reaches 1%, the probability that a non-defective product is misjudged as a defective product is as high as 34.5%. That is, adopting such a strict plan will cause a large number of non-defective products to be misjudged as defective products, resulting in waste.

[0083] Empirical solution 2: Loose sampling. When the number of non-conductive contacts (i.e., the defective rate) in the sampled sample is higher than 5%, it is determined as a defective product.

[0084] Run the acceptance sampling by attributes in Minitab software, select to compare user-defined sampling plans, input the acceptable defective rate of contacts 1% into the acceptable quality level in the acceptance sampling by attributes, input the rejectable defective rate of batches 5% into the rejectable quality level, input the sample size into the value 208 generated by the previous sampling plan, adjust the acceptable number from 5 to 10, and the batch size is still the total number of contacts of this ceramic substrate 728. The specific input interface is as Figure 7 shown, and the generated sampling plan and operating characteristic (OC) curve are as Figure 8 shown.

[0085] It can be seen that when the defective rate of the test is 5%, this sampling plan will cause 53.2% of the defective products to be detected as non-defective products, far exceeding the range allowed by the customer, which will cause serious quality problems.

[0086] Therefore, the sampling plan deduced by the demand parameters of the OC curve has significant advantages compared with the sampling plan obtained by the traditional empirical method.

[0087] The above are all preferred embodiments of this application. The protection scope of this application is not limited by this. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. A sampling method for improving the detection efficiency of ceramic substrate performance, characterized in that: It includes the following specific steps: Step 1: Collect customer quality requirements, including the total number of contacts on the ceramic substrate, the rejectable acceptable quality level, and the consumer risk; Step 2: The manufacturer sets the inspection quality requirements, including the acceptable quality level and the producer risk; Step 3: Develop a sampling plan based on the customer quality requirements and the manufacturer's inspection quality requirements, and determine the required sample contact number and the non-conductive contact number threshold in the sampling plan; Step 4: Use a flying probe machine to conduct contact continuity tests to obtain the number of non-conductive contacts; Step 5: Compare the number of non-conductive contacts obtained in the test in Step 4 with the non-conductive contact number threshold in Step 3 to obtain the sampling result of whether the ceramic substrate is qualified; In Step 3, the Minitab software is used to generate the required sample contact number size and the non-conductive contact number threshold of the sampling plan. The specific steps are as follows: Run the attribute sampling acceptance of the Minitab software, select to create a sampling plan, input the total contact number of the customer quality requirements in Step 1 into the batch size, the rejectable acceptable quality level into the rejectable quality level LTPD, the consumer risk into the consumer risk Beta, input the acceptable quality level of the inspection quality requirements in Step 2 into the acceptable quality level AQL, and the producer risk into the producer risk Alpha. Running the software can obtain the sample quantity and the acceptance number of the sampling plan. Among them, the sample quantity is the sample contact number size, and the acceptance number is the non-conductive contact number threshold; when running the Minitab software to generate the sampling plan, an operating characteristic (OC) curve graph and an average total inspection (ATI) comparison table are also generated; In Step 5, the following method is specifically used to determine the sampling result: If the number of non-conductive contacts obtained by the flying probe machine test is greater than the non-conductive contact number threshold, it is determined that the ceramic substrate is a defective product; if the number of non-conductive contacts obtained by the test is less than or equal to the non-conductive contact number threshold, it is necessary to further determine whether a re-inspection is required according to the average total inspection (ATI) comparison table; The specific method for determining whether a re-inspection is required is as follows: Define the total contact number as n, the average total inspection (ATI) as ATI, the number of non-conductive contacts in Step 4 as e, the non-conductive contact number threshold as x, set 35%n ≤ ATI ≤ 45%n, read the average total inspection (ATI) comparison table, find the average total inspection number ATI that meets the requirements, and the non-conductive contact number corresponding to this average total inspection number ATI is set as the re-inspection threshold k. When k < e ≤ x, it is necessary to conduct a re-inspection on the ceramic substrate.

2. The sampling method for improving the detection efficiency of the performance of a ceramic substrate according to claim 1, wherein: The non-conductive contact number threshold refers to the maximum allowable number of non-conductive contacts in the sampled sample contacts when using a flying probe machine to conduct contact continuity tests. If the actually measured number of non-conductive contacts exceeds the non-conductive contact number threshold, the ceramic substrate is determined to be a defective product.

3. A sampling method for improving the detection efficiency of the performance of a ceramic substrate according to claim 1, characterized in that: The re-inspection method is to detect all the contacts of the ceramic substrate. When the defective product rate of the re-inspected contacts is less than the rejectable acceptable quality level, it is determined that the ceramic substrate is a qualified product; otherwise, it is determined that the ceramic substrate is a defective product.

Images