A method for setting a fixed-time dose threshold for X-ray short-circuit detection on PCBA boards

Abstract

The present invention relates to the technical field of PCBA board detection and discloses a method for setting a dose threshold for X-ray short circuit detection of PCBA boards for a fixed time. The method comprises the following steps: taking U PCBA multilayer boards with the same substrate thickness and short circuit defects as a sample set and dividing the sample set into a training sample set m, a test sample set n, and a verification sample set q; performing short circuit detection at different doses on the PCBA multilayer boards in the training sample set m to obtain a lower dose prediction threshold and an upper dose prediction threshold; further obtaining a lower dose transition threshold, a lower dose final threshold, an upper dose transition threshold, and an upper dose final threshold through a machine learning system; and finally generating a lower dose threshold and an upper dose threshold respectively through weighted combination. The present invention generates the lower dose threshold and the upper dose threshold through weighted combination, which can reduce the influence of the specificity of the PCBA boards in the training sample set m, the test sample set n, and the verification sample set q on the final threshold, thereby making the final threshold more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of PCBA board detection, and in particular to a method for setting a dose threshold for a fixed duration for X-ray short circuit detection of a PCBA board. Background Art

[0002] The entire process of PCB blank board going through SMT mounting and DIP plug-in soldering is referred to as PCBA. After SMT and DIP plug-in soldering, the finished PCBA cannot guarantee that the electronic components soldered on the PCBA are soldered correctly. For example, there may be problems such as missing solder joints, cold solder joints, and short circuits. In order to sort out defective PCBAs from good ones and repair them, and reduce the probability of rework and rework in subsequent processes, inspection is required.

[0003] Existing PCBA inspection methods include thermal imaging, ultrasonic testing, short-circuit detector testing, and X-ray testing. For multi-layer or complex-structured PCBA boards, short circuits inside or hidden under components may be difficult to detect directly. Therefore, ultrasonic testing is used on the market to detect short circuits on multi-layer or complex-structured PCB boards, but X-ray testing is almost non-existent. This is because the radiation dose of X-ray testing may affect the normal use of the electronic components on the PCBA board. Therefore, it is necessary to confirm the X-ray dose threshold setting to ensure that the electronic components can operate normally under the premise that short-circuit defects can be detected. Therefore, a dose threshold setting method with a fixed duration for X-ray short-circuit detection on PCBA boards is proposed to solve the above problem. Summary of the Invention

[0004] (1) Technical problems solved

[0005] The purpose of this invention is to solve the problem that ultrasonic detection is used to detect short circuits on multi-layer or complex PCBs on the market, but there is almost no X-ray detection for PCBAs. A method for setting a fixed-time dose threshold for X-ray short-circuit detection on PCBAs is proposed.

[0006] (2) Technical solution

[0007] The technical solution of the present invention to solve the above technical problems is as follows:

[0008] A method for setting a fixed-time dose threshold for X-ray short circuit detection on a PCBA board comprises the following steps:

[0009] S1: Take U PCBA multilayer boards with the same substrate thickness and short circuit defects as the sample set and divide the sample set into a training sample set m, a test sample set n and a verification sample set q;

[0010] S2: randomly divide the PCBA multilayer boards in m into m igroup, use X-ray to sequentially scan m at the same time t i All PCBA multilayer boards in the group were subjected to short circuit testing at different doses;

[0011] S3: Preprocess the short circuit detection results, compare the processed data with the actual defects, and use the judgment accuracy formula to find the minimum dose when the judgment accuracy is 1 to establish it as the lower dose prediction threshold D min , for m i The electronic component E in the group is tested, and the maximum dose that can work normally in the electronic component E is used as the upper dose prediction threshold D max ;

[0012] S4: Set the lower dose prediction threshold D min and the upper dose prediction threshold D max Input into the machine learning system, take the test sample set n and put it into the machine learning system, use the lower dose prediction threshold D min and the upper dose prediction threshold D max Two doses of X-rays were used to perform short circuit detection on all PCBA multilayer boards in n twice. The lower dose prediction threshold D was calculated by machine learning system. min and the upper dose prediction threshold D max Correction processing is performed, and the lower dose transition threshold D is obtained after correction processing. min ' and upper dose transition threshold D max ';

[0013] S5: Set the lower dose transition threshold D min ' and upper dose transition threshold D max 'Replace the dose prediction threshold D min and the upper dose prediction threshold D max Input into the machine learning system and add the substrate material influence factor to confirm the final dose threshold D min ", reduce the upper dose transition threshold D max The value of ' is determined as the final upper dose threshold D max ", take the verification sample set q and put it into the machine learning system for verification;

[0014] S6: Set the lower dose prediction threshold D min , lower dose transition threshold D min ' and the final threshold value of the lower dose D min "Generate the lower dose threshold T by combining the weights min , the upper dose prediction threshold D max , upper dose transition threshold D max ' and the final threshold of upper dose D max "Generate the upper dose threshold T by combining the weights max .

[0015] On the basis of the above technical solution, the present invention can also be improved as follows.

[0016] Preferably, the S1 specifically includes:

[0017] Take U PCBA multilayer boards with the same substrate thickness and short circuit defects as the sample set, and randomly divide the sample set into training sample set m, test sample set n and verification sample set q in the ratio of 7:2:1.

[0018] Preferably, the m in S2 i The X-ray doses for the groups were:

[0019] d i =d0+Δd·(i-1)

[0020] Among them, d i is the X-ray dose of group i, d0 is the initial X-ray dose, and Δd is the incremental X-ray dose.

[0021] Preferably, the judgment accuracy formula in S3 is:

[0022]

[0023] where ρ d is the judgment accuracy, m d To determine the correct number, M d is the total number of short circuit defects.

[0024] Preferably, the electronic component E in S3 is a collection of several electronic components, and the electronic component E includes {M i , M s , M a , M m , P e}, M i Represents a collection of inorganic non-metallic material components, M s Represents a collection of semiconductor material components, M a Represents a collection of metal and alloy material components, M m Represents the magnetic component set, P e Represents a collection of piezoelectric ceramic elements;

[0025] The maximum dose for confirming that the electronic component E can work normally is the upper dose prediction threshold D max The specific formula is as follows:

[0026]

[0027] in, M i Corresponding to the dose prediction threshold, Ms Corresponding to the dose prediction threshold, M a Corresponding to the dose prediction threshold, M m Corresponding to the dose prediction threshold, P e Corresponding to the dose prediction threshold, The dose prediction threshold corresponding to the set of inorganic non-metallic material components, The dose prediction threshold corresponding to the semiconductor material component set, The dose prediction threshold corresponding to the metal and alloy material component set, The dose prediction threshold corresponding to the magnetic element set, The piezoelectric ceramic element set corresponds to an upper dose prediction threshold.

[0028] Preferably, the correction process in S4 obtains the lower dose transition threshold D min ' and upper dose transition threshold D max The specific process includes:

[0029] ΔD min The lower dose prediction threshold D min Perform incremental operation, use X-ray to perform short circuit detection on the test sample set, compare the test results with the actual defect situation of the test sample, and use the judgment accuracy formula to calculate whether the detection accuracy reaches the expected value of the lower dose test. If the detection accuracy does not reach the expected value of the lower dose test, continue to use ΔD min The lower dose prediction threshold is incremented until the detection accuracy reaches the expected value of the lower dose test. The lower dose prediction threshold after this increment is used as the lower dose transition threshold. The increment formula is:

[0030]

[0031] in, It is the d min The lower dose transition threshold of the incremental operation, D min is the prediction threshold without the increment operation, ΔD min is the standard deviation of the test sample set, d min is the number of increments;

[0032] ΔD max Upper dose prediction threshold D max Perform a decreasing operation, use X-ray to perform short circuit detection on the test sample set, compare the test results with the actual defect situation of the test sample, and use the judgment accuracy formula to calculate whether the detection accuracy reaches the expected value of the upper dose test. If the detection accuracy does not reach the expected value of the upper dose test, continue to use ΔD maxThe upper dose prediction threshold is incremented until the detection accuracy reaches the upper dose test expected value. The upper dose prediction threshold after this increment is used as the lower dose transition threshold. The increment formula is:

[0033]

[0034] in, It is the d max The upper dose transition threshold of the incremental operation, D max is the predicted threshold without the decrement operation, ΔD max is the standard deviation of the test sample set, d max Incremental times.

[0035] Preferably, the final threshold value D of the dose determined in S5 is min " and the upper dose final threshold D max The confirmation process includes:

[0036] Establish the final threshold of lower dose D min "Model, the final threshold of lower dose D min The model formula is as follows:

[0037] D min ”=ax1+b

[0038] Among them, D min " is the final threshold, x1 is the substrate material influence factor, a is the first fitting parameter, and b is the second fitting parameter;

[0039] Collect the verification accuracy ρ of the machine learning system under different substrate materials in the verification sample set v , the verification accuracy ρ in the verification sample set v and basic materials are recorded one by one, and the first fitting parameter and the second fitting parameter are estimated by the least square method;

[0040] Finally, the upper dose transition threshold D max 'Lower its threshold as the final upper dose threshold D max ", the specific reduction formula is:

[0041] D max =D max '-2ΔD

[0042] Among them, ΔD is the safety control parameter.

[0043] Preferably, the lower dose threshold T min The specific formula is:

[0044]

[0045] in, is the lower dose prediction threshold D min The weight factor, is the lower dose transition threshold D min 'The weight factor, is the final threshold value of the lower dose D min ” weight factor.

[0046] Preferably, the upper dose threshold T max The specific formula is:

[0047] T max =D max μ α +D max '·μ β +D max ”·μ γ , μ α +μ β +μ γ =1

[0048] Among them, μ α is the upper dose prediction threshold D max The weight factor, μ β is the upper dose transition threshold D max 'The weight factor, μ γ The final threshold value of upper dose D max ” weight factor.

[0049] (3) Beneficial effects

[0050] Compared with the prior art, the technical solution of this application has the following beneficial technical effects:

[0051] 1. The present invention generates the lower dose threshold T by combining weights min and upper dose threshold T max , which can reduce the impact of the specificity of the PCBA boards in the training sample set m, the test sample set n, and the verification sample set q on the final threshold, making the final threshold more accurate.

[0052] 2. The present invention can reduce the upper dose final threshold D by adding safety control parameters. max "The value of the upper dose final threshold D max The value of " is safer and more reasonable, reducing the possibility of damage to electronic components E. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0054] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0055] A method for setting a fixed-time dose threshold for X-ray short circuit detection on a PCBA board comprises the following steps:

[0056] S1: Take U PCBA multilayer boards with the same substrate thickness and short circuit defects as the sample set and divide the sample set into a training sample set m, a test sample set n and a verification sample set q;

[0057] S2: randomly divide the PCBA multilayer boards in m into m i group, use X-ray to scan m in sequence at the same time t i All PCBA multilayer boards in the group were subjected to short circuit testing at different doses;

[0058] S3: Preprocess the short circuit detection results, compare the processed data with the actual defects, and use the judgment accuracy formula to find the minimum dose when the judgment accuracy is 1 to establish it as the lower dose prediction threshold D min , for m i The electronic component E in the group is tested, and the maximum dose that can work normally in the electronic component E is used as the upper dose prediction threshold D max ;

[0059] S4: Set the lower dose prediction threshold D min and the upper dose prediction threshold D max Input into the machine learning system, take the test sample set n and put it into the machine learning system, use the lower dose prediction threshold D min and the upper dose prediction threshold D max Two doses of X-rays were used to perform short circuit detection on all PCBA multilayer boards in n twice. The lower dose prediction threshold D was calculated by machine learning system. min and the upper dose prediction threshold D max Correction processing is performed, and the lower dose transition threshold D is obtained after correction processing. min ' and upper dose transition threshold D max ';

[0060] S5: Set the lower dose transition threshold D min ' and upper dose transition threshold D max 'Replace the dose prediction threshold D min and the upper dose prediction threshold D maxInput into the machine learning system and add the substrate material influence factor to confirm the final dose threshold D min ", reduce the upper dose transition threshold D max The value of ' is determined as the final upper dose threshold D max ", take the verification sample set q and put it into the machine learning system for verification;

[0061] Considering that different substrate materials will affect the final threshold value D of the lower dose min Therefore, the concept of substrate material is introduced and the influence factor of substrate material is added to make the final threshold value of lower dose D min "More in line with actual needs;

[0062] S6: Set the lower dose prediction threshold D min , lower dose transition threshold D min ' and the final threshold value of the lower dose D min "Generate the lower dose threshold T by combining the weights min , the upper dose prediction threshold D max , upper dose transition threshold D max ' and the final threshold of upper dose D max "Generate the upper dose threshold T by combining the weights max .

[0063] The S1 specifically includes:

[0064] Take U PCBA multilayer boards with the same substrate thickness and short circuit defects as the sample set, and randomly divide the sample set into training sample set m, test sample set n and verification sample set q in the ratio of 7:2:1.

[0065] Preferably, the m in S2 i The X-ray doses for the groups were:

[0066] d i =d0+Δd·(i-1)

[0067] Among them, d i is the X-ray dose of group i, d0 is the initial X-ray dose, and Δd is the incremental X-ray dose.

[0068] The judgment accuracy formula in S3 is:

[0069]

[0070] where ρ d is the judgment accuracy, m d To determine the correct number, M d is the total number of short circuit defects.

[0071] The electronic component E in S3 is a collection of several electronic components, and the electronic component E includes {M i , M s , M a , M m , P e}, M i Represents a collection of inorganic non-metallic material components, M s Represents a collection of semiconductor material components, M a Represents a collection of metal and alloy material components, M m Represents the magnetic component set, P e Represents a collection of piezoelectric ceramic elements;

[0072] The maximum dose for confirming that the electronic component E can work normally is the upper dose prediction threshold D max The specific formula is as follows:

[0073]

[0074] in, M i Corresponding to the dose prediction threshold, M s Corresponding to the dose prediction threshold, M a Corresponding to the dose prediction threshold, M m Corresponding to the dose prediction threshold, P e Corresponding to the dose prediction threshold, The dose prediction threshold corresponding to the set of inorganic non-metallic material components, The dose prediction threshold corresponding to the semiconductor material component set, The dose prediction threshold corresponding to the metal and alloy material component set, The dose prediction threshold corresponding to the magnetic element set is The piezoelectric ceramic element set corresponds to an upper dose prediction threshold.

[0075] The correction process in S4 obtains the lower dose transition threshold D min ' and upper dose transition threshold D max The specific process includes:

[0076] ΔD min The lower dose prediction threshold D min Perform incremental operation, use X-ray to perform short circuit detection on the test sample set, compare the test results with the actual defect situation of the test sample, and use the judgment accuracy formula to calculate whether the detection accuracy reaches the expected value of the lower dose test. If the detection accuracy does not reach the expected value of the lower dose test, continue to use ΔD minThe lower dose prediction threshold is incremented until the detection accuracy reaches the expected value of the lower dose test. The lower dose prediction threshold after this increment is used as the lower dose transition threshold. The increment formula is:

[0077]

[0078] in, It is the d min The lower dose transition threshold of the incremental operation, D min is the prediction threshold without the increment operation, ΔD min is the standard deviation of the test sample set, d min is the number of increments;

[0079] ΔD max Upper dose prediction threshold D max Perform a decreasing operation, use X-ray to perform short circuit detection on the test sample set, compare the test results with the actual defect situation of the test sample, and use the judgment accuracy formula to calculate whether the detection accuracy reaches the expected value of the upper dose test. If the detection accuracy does not reach the expected value of the upper dose test, continue to use ΔD max The upper dose prediction threshold is incremented until the detection accuracy reaches the upper dose test expected value. The upper dose prediction threshold after this increment is used as the lower dose transition threshold. The increment formula is:

[0080]

[0081] in, It is the d max The upper dose transition threshold of the incremental operation, D max is the predicted threshold without the decrement operation, ΔD max is the standard deviation of the test sample set, d max Incremental times.

[0082] Confirm the final threshold value D of the lower dose as described in S5 min " and the upper dose final threshold D max The confirmation process includes:

[0083] Establish the final threshold of lower dose D min "Model, the final threshold of lower dose D min The model formula is as follows:

[0084] D min ”=ax1+b

[0085] Among them, D min " is the final threshold, x1 is the substrate material influence factor, a is the first fitting parameter, and b is the second fitting parameter;

[0086] Collect the verification accuracy ρ of the machine learning system under different substrate materials in the verification sample set v , the verification accuracy ρ in the verification sample set v and basic materials are recorded one by one, and the first fitting parameter and the second fitting parameter are estimated by the least square method;

[0087] Finally, the upper dose transition threshold D max 'Lower its threshold as the final upper dose threshold D max ", the specific reduction formula is:

[0088] D max =D max '-2ΔD

[0089] Among them, ΔD is the safety control parameter.

[0090] The method for establishing the safety control parameter ΔD is as follows:

[0091] First, refer to the X-ray machine manual or technical specifications to understand the manufacturer's recommended irradiation time range and safety operating specifications, preliminarily establish the value range of the safety control parameter ΔD, take the middle value of the safety control parameter ΔD range as the safety control parameter ΔD, and then use the upper dose final threshold D max "Directly use the maximum input of the X-ray machine to test the verification sample set q, and determine the safety control parameter based on the damage status of the electronic components E in the verification sample set q. If the electronic components E are damaged, increase the value of the safety control parameter ΔD by 0.1 times ΔD until the electronic components E are no longer damaged during the X-ray testing of the verification sample set q. If the electronic components E are not damaged, directly use the middle value of the safety control parameter ΔD value range as the safety control parameter ΔD.

[0092] Since different PCBA multilayer boards have certain specificity, the upper dose final threshold D max " and the lower dose final threshold D min "It is generally not possible to satisfy all samples in the verification sample set q. In order to minimize the final upper dose threshold D max " and the lower dose final threshold D min "The impact on the threshold setting, so the upper dose transition threshold D max ' and the final threshold of upper dose D max "Generate the upper dose threshold T by combining the weights max .

[0093] The lower dose threshold T min The specific formula is:

[0094]

[0095] in, is the lower dose prediction threshold D min The weight factor, is the lower dose transition threshold D min 'The weight factor, is the final threshold value of the lower dose D min ” weight factor.

[0096] Generally, 0.05-0.15 is used. Generally, 0.15-0.25 is used. Generally, it is taken as 0.65-0.75.

[0097] The upper dose threshold T max The specific formula is:

[0098] T max =D max μ α +D max '·μ β +D max ”·μ γ , μ α +μ β +μ γ =1

[0099] Among them, μ α is the upper dose prediction threshold D max The weight factor, μ β is the upper dose transition threshold D max 'The weight factor, μ γ The final threshold value of upper dose D max ” weight factor.

[0100] μ α Generally, 0.05-0.1 μ β Generally, 0.05-0.2 μ γ Generally, it is taken as 0.7-0.9.

[0101] Lower dose threshold T min and upper dose threshold T max constitute the lower and upper limits of the threshold, so the reasonable detection dose is [T min , T max ], the test dose is generally taken in the test It's important to note that X-rays are harmful to the human body. Regulations must be followed: the average annual effective dose for any worker must not exceed 20mSv for five consecutive years, and the effective dose in any given year must not exceed 50mSv. The annual equivalent dose to the limbs (hands and feet) or skin must not exceed 500mSv. Furthermore, there are strict limits on the radiation leakage of X-ray inspection equipment during operation, typically no more than 1μSv / h (microsievert per hour).

[0102] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0103] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for setting a fixed-time dose threshold for X-ray short circuit detection on PCBA boards, characterized in that: The following steps are involved: S1: Take U PCBA multilayer boards with the same substrate thickness and short circuit defects as the sample set and divide the sample set into a training sample set m, a test sample set n and a verification sample set q; S2: randomly divide the PCBA multilayer boards in m into m i group, use X-ray to sequentially scan m at the same time t i All PCBA multilayer boards in the group were subjected to short circuit testing at different doses; S3: Preprocess the short circuit detection results, compare the processed data with the actual defects, and use the judgment accuracy formula to find the minimum dose when the judgment accuracy is 1 to establish it as the lower dose prediction threshold D min , for m i The electronic component E in the group is tested, and the maximum dose that can work normally in the electronic component E is used as the upper dose prediction threshold D max ; S4: Set the lower dose prediction threshold D min and the upper dose prediction threshold D max Input into the machine learning system, take the test sample set n and put it into the machine learning system, use the lower dose prediction threshold D min and the upper dose prediction threshold D max Two doses of X-rays were used to perform short circuit detection on all PCBA multilayer boards in n twice. The lower dose prediction threshold D was calculated by machine learning system. min and the upper dose prediction threshold D max Correction processing is performed, and the lower dose transition threshold D is obtained after correction processing. min ' and upper dose transition threshold D max '; S5: Set the lower dose transition threshold D min ' and upper dose transition threshold D max 'Replace the dose prediction threshold D min and the upper dose prediction threshold D max Input into the machine learning system and add the substrate material influence factor to confirm the final dose threshold D min ", reduce the upper dose transition threshold D max The value of ' is determined as the final upper dose threshold D max ", take the verification sample set q and put it into the machine learning system for verification; S6: Set the lower dose prediction threshold D min , lower dose transition threshold D min ' and the final threshold value of the lower dose D min "Generate the lower dose threshold T by combining the weights min , the upper dose prediction threshold D max , upper dose transition threshold D max ' and the final threshold of upper dose D max "Generate the upper dose threshold T by combining the weights max .

2. A method for setting a fixed-duration dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, characterized in that , the S1 specifically includes: Take U PCBA multilayer boards with the same substrate thickness and short circuit defects as the sample set, and randomly divide the sample set into training sample set m, test sample set n and verification sample set q in the ratio of 7:2:

1.

3. The method for setting a fixed-time dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, characterized in that: S2 i The X-ray doses for the groups were: d i =d0+Δd·(i-1) Among them, d i is the X-ray dose of group i, d0 is the initial X-ray dose, and Δd is the incremental X-ray dose.

4. The method for setting a fixed-duration dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, characterized in that: The judgment accuracy formula in S3 is: where ρ d is the judgment accuracy, m d To determine the correct number, M d is the total number of short circuit defects.

5. The method for setting a fixed-duration dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, characterized in that: The electronic component E in S3 is a collection of several electronic components, and the electronic component E includes {M i , M s , M a , M m , P e }, M i Represents a collection of inorganic non-metallic material components, M s Represents a collection of semiconductor material components, M a Represents a collection of metal and alloy material components, M m Represents the magnetic component set, P e Represents a collection of piezoelectric ceramic elements; The maximum dose for confirming that the electronic component E can work normally is the upper dose prediction threshold D max The specific formula is as follows: in, M i Corresponding to the dose prediction threshold, M s Corresponding to the dose prediction threshold, M a Corresponding to the dose prediction threshold, M m Corresponding to the dose prediction threshold, P e Corresponding to the dose prediction threshold, The dose prediction threshold corresponding to the set of inorganic non-metallic material components, The dose prediction threshold corresponding to the semiconductor material component set, The dose prediction threshold corresponding to the metal and alloy material component set, The dose prediction threshold corresponding to the magnetic element set is The piezoelectric ceramic element set corresponds to an upper dose prediction threshold.

6. The method for setting a fixed-duration dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, characterized in that: The correction process in S4 obtains the lower dose transition threshold D min ' and upper dose transition threshold D max The specific process includes: ΔD min The lower dose prediction threshold D min Perform incremental operation, use X-ray to perform short circuit detection on the test sample set, compare the test results with the actual defect situation of the test sample, and use the judgment accuracy formula to calculate whether the detection accuracy reaches the expected value of the lower dose test. If the detection accuracy does not reach the expected value of the lower dose test, continue to use ΔD min The lower dose prediction threshold is incremented until the detection accuracy reaches the expected value of the lower dose test. The lower dose prediction threshold after this increment is used as the lower dose transition threshold. The increment formula is: in, It is the d min The lower dose transition threshold of the incremental operation, D min is the prediction threshold without the increment operation, ΔD min is the standard deviation of the test sample set, d min is the number of increments; ΔD max Upper dose prediction threshold D max Perform a decreasing operation, use X-ray to perform short circuit detection on the test sample set, compare the test results with the actual defect situation of the test sample, and use the judgment accuracy formula to calculate whether the detection accuracy reaches the expected value of the upper dose test. If the detection accuracy does not reach the expected value of the upper dose test, continue to use ΔD max The upper dose prediction threshold is incremented until the detection accuracy reaches the upper dose test expected value. The upper dose prediction threshold after this increment is used as the lower dose transition threshold. The increment formula is: in, It is the d max The upper dose transition threshold of the incremental operation, D max is the predicted threshold without the decrement operation, ΔD max is the standard deviation of the test sample set, d max Incremental times.

7. The method for setting a fixed-time dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, characterized in that: Confirm the final threshold value D of the lower dose as described in S5 min " and the upper dose final threshold D max The confirmation process includes: Establish the final threshold of lower dose D min "Model, the final threshold of lower dose D min The model formula is as follows: <h2 style=";text-align:left;direction:ltr">D<h2 style=";text-align:left;direction:ltr"> min <h2 style=";text-align:left;direction:ltr"> “=ax1+b Among them, D min " is the final threshold, x1 is the substrate material influence factor, a is the first fitting parameter, and b is the second fitting parameter; Collect the verification accuracy ρ of the machine learning system under different substrate materials in the verification sample set v , the verification accuracy ρ in the verification sample set v and basic materials are recorded one by one, and the first fitting parameter and the second fitting parameter are estimated by the least square method; Finally, the upper dose transition threshold D max 'Lower its threshold as the final upper dose threshold D max ", the specific reduction formula is: IS max ”=D max '-2ΔD Among them, ΔD is the safety control parameter.

8. The method for setting a fixed-duration dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, wherein: The lower dose threshold T min The specific formula is: in, is the lower dose prediction threshold D min The weight factor, is the lower dose transition threshold D min 'The weight factor, is the final threshold value of the lower dose D min ” weight factor.

9. The method for setting a fixed-duration dose threshold for X-ray short circuit detection on a PCBA board according to claim 1, wherein: The upper dose threshold T max The specific formula is: T max =D max ·m α +D max '·m β +D max "m" γ ,m α +m β +m γ =1 Among them, μ α is the upper dose prediction threshold D max The weight factor, μ β is the upper dose transition threshold D max 'The weight factor, μ γ The final threshold value of upper dose D max ” weight factor.

Images