JP3511632B2 - Mounting process failure factor analysis method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a mounting process defect factor analysis method for pursuing quality defect factors in an electronic component mounting process for mounting electronic components on a printed circuit board.

[0002]

2. Description of the Related Art Up to now, quality inspection and control in the electronic component mounting process have been performed mainly by manual work, or due to the limit of manual work due to high density and high production volume, mainly automatic inspection mainly in the final process. I installed the device and used it for inspection and tabulation.

However, even if only the inspection after the soldering, which is the final step, is carried out by inserting an automatic inspection device, it is possible to find out in which step of printing, mounting and soldering the cause of the occurrence actually occurs. It was difficult to do so, and as a result, it was only possible to judge the quality of the final product.

Originally, in the electronic component mounting process, which is roughly divided into three processes of printing, mounting, and soldering, it is desirable to perform quality inspection and control for each process. Therefore, in recent years, automatic inspection devices have been developed not only for the soldering process but also for the printing / mounting process. Therefore, it has become relatively easy to obtain quality inspection results for each process. However, as shown in FIG. 6 , the conventional mounting process defect factor analysis method aggregates and analyzes the inspection results independently for each process.

[0005]

However, the quality defect found in each process is not always caused by the process. Due to the occurrence of defects in the previous process, even if the production work of the next process is correct, it may also result in defects.

For example, if a defect such as component displacement occurs due to the mounting process, even if the soldering process itself operates properly, an inspection result of component displacement may still appear during soldering inspection. Therefore, even if quality control is performed based on the inspection result only for each process, it is impossible to pursue the true cause of the defect.

On the other hand, for this reason, manually comparing inspection failure results between processes and pursuing the cause is a huge amount of work in view of the current state of the mounting process that produces more than 1000 printed circuit boards per day. Requested and virtually impossible.

In view of the above problems, the present invention compares the inspection results of the printing, mounting, and soldering processes for each printed circuit board to be mounted, and determines the influence of each process on the final defect. It is an object of the present invention to provide a mounting process defect factor analysis method that can be calculated and automatically predict the cause of a defect.

[0009]

The mounting process failure factor analysis method of the present invention is applied to printing, mounting, and mounting in mounting electronic components.
In each soldering process, a print result input process that inputs the print quality result of the printing process for each printed circuit board, a mounting result input process that inputs the mounting quality result of the mounting process for each printed circuit board, and a soldering process of the soldering process Attach quality results,
The soldering result input process to input for each printed circuit board,
Among the quality results between these processes, the process of registering the quality defect related rule indicating the possibility of occurring in relation to each other in advance in the quality defect related rule recording part, the print result input process, the mounting result input process, and the soldering result input A mounting process defect factor analysis method that consists of a process that compares the quality results of each process input through each process for each printed circuit board and refers to the quality defect related rules to calculate the degree of influence of each process on the final defect. is there.

Further, the mounting process defect factor analysis method of the present invention is based on the quality result of each process of printing, mounting, and soldering in the mounting of electronic parts and the actual occurrence frequency of the defective factor identified. The process includes the step of registering and correcting a quality defect related rule that indicates a possibility of occurrence related to the quality result between processes.

[0011]

According to the present invention, it is possible to automatically predict how much one of printing, mounting, and soldering processes has affected a defect in mounting an electronic component.

[0012]

EXAMPLES Hereinafter, with reference to FIG. 5 an embodiment of the present invention from FIG.

The mounting process defect factor analysis method of this embodiment is as follows:
The structure is as shown in FIG. The printed circuit board is manufactured by passing through a printing process 1, a mounting process 2, and a soldering process 3 in this order. The print inspection means 4 after the printing step obtains the quality inspection result of the printing step 1, and the print result input means 5
Through the printed circuit board. Similarly, the quality inspection result of the mounting step 2 is obtained by the mounting inspection means 6 after the mounting step, and is summed up for each printed circuit board through the mounting result input means 7. Further, the quality inspection result of the soldering step 3 is obtained by the soldering inspection means 8 and is summed up for each printed circuit board through the soldering result input means 9.

On the other hand, among the quality results between the respective processes, the quality defect related rules indicating the possibility of occurrence in relation to each other are recorded in the quality defect related rule recording unit 10 in advance. Then, the result comparison / analysis means 11 extracts the quality result of each process for each printed circuit board, and with respect to the result of each process,
The degree of influence of each process on the final defect is calculated with reference to the quality defect related rule.

The specific procedure of the mounting process failure factor analysis method in the above configuration will be described below. It is assumed that a plurality of printed circuit boards 12 as shown in FIG. 2 have gone through the mounting process. For the sake of explanation, each printed circuit board has 1001, 1002, ...
Number 14 such as. Each component on each board can be identified by the circuit number 13. Next, the inspection results in each step are shown in (Table 1) to (Table 3).

[0016]

[Table 1]

[0017]

[Table 2]

[0018]

[Table 3]

FIG. 3 shows the registered contents of the quality defect related rule in the quality defect related rule recording unit 10. 12
((A) (b) (c)), 13 ((a) (b)
(C)), 14 ((a) (b) (c)) are the inspection results, and 15 ((a) (b)), 16 ((a) (b)),
17 ((a) and (b)) are possible defect-causing events, which have the occurrence probability of each defect factor and the degree to which each factor affects the inspection result. All the straight lines in FIG. 3 are flows from the left to the right, and each straight line has a transition probability. For example, the mounting failure item 1 can occur at a certain transition probability from the printing failure items 1, 2 ... In the previous process and the mounting failure causes 1, 2 ... caused by the mounting process. On the other hand, in this state, the soldering inspection result is
From bad soldering 1, 2 ... to no bad soldering, each has a certain probability. Here, only the most prominent defect results are shown. Therefore, each inspection result item is independent of each other, and the sum of transition probabilities from defective soldering 1, 2, ...

Now, comparing the quality results for the circuit number R101 of the printed circuit board number 1001, no defect is found in the printing process, but there is a component misalignment defect in the mounting process and the soldering process. Here, the quality defect related rule is referred to.

An example of registration for this result will be shown in FIG. 4 following FIG. In FIG. 4, the numerical values on each straight line indicate the transition probabilities. However, the example of FIG. 4 is merely for the purpose of explanation, and although it differs from the actual numerical value, it does not deny this discussion. To simplify the explanation, FIG.
In, the transition probability 0 and some unnecessary parts are omitted. Furthermore, first, it is assumed that the probability of occurrence of each defect factor and each factor have the same effect on the inspection result.

First, "part shift" as a result of the soldering process
It can be said that there is a probability of 70% that "the soldering temperature condition is set incorrectly". In this figure, there are no other causes of soldering process defects that cause "component misalignment," so the remaining 3
With a probability of 0%, it can be said that the soldering process operation was correct. On the other hand, "component misalignment" has occurred in the mounting process of the previous process. Therefore, the probability that this defect is the cause is 80%.

Next, let us consider the "part shift" as a result of the mounting process. In the mounting process, there is a 50% chance of being caused by "suction displacement" and a 3 chance of being "wrong component dimensions".
It is 0%. On the other hand, in the printing process of the previous process, no printing defect has occurred, so the probability that the printing defect is the cause is 0%.

From the above, the probability that the cause of the defect in the mounting process affects the defect result "component deviation" in the soldering process which is the final process is that each defect cause is the "component deviation" in the mounting process.
It is sufficient to multiply the probability of causing the “component deviation” in the mounting process by the probability of causing the “component deviation” in the soldering process. (Table 4) shows the probability of failure cause of "component misalignment" in the soldering process.

[0025]

[Table 4]

Up to this point, it has been assumed that the probability of occurrence of each defect factor and the degree to which each factor affects the inspection result are equal, but in reality their values are different. Appropriate weighting is necessary for each failure factor in order to know which failure factor has the greatest effect on the final failure result. To take this into account, these parameters may be further multiplied. Table 5 shows an example of weighting of each factor with respect to the result of Table 4.

[0027]

[Table 5]

Here, the result of (Table 4) is multiplied by the occurrence frequency of each defect factor and the degree of influence of the defect factor on each inspection result. Since this value serves as a weight for each possible cause of failure, the expected cause of failure can be obtained with importance by calculating the ratio of the values. In this example, the possibility of “component misalignment” in the mounting process occupies a ratio of about 3/4, and is considered to be a probable cause. On the other hand, it can be seen that the possibility of "component size error" in the mounting process is considerably low. Actually, the degree of influence of the defect factor on each inspection result can be dynamically determined by a measurement value such as a deviation amount at the time of inspection.

In this example, as described above, it is assumed that a plurality of inspection result items do not occur at the same time and that the causes of each defect are independent of each other. In the actual mounting process, this assumption does not always hold. However, the possibility of interdependence can be reduced sufficiently if the items with high interdependence are combined into one, and each inspection result and itemization of the cause of failure are taken into consideration. Under such conditions, the possibility that a plurality of inspection result items will occur at the same time and the causes of each defect will be mutually related can be sufficiently ignored in the prediction calculation. On the other hand, it is considered that it is not practically useful to calculate the final result in consideration of the above conditions because the calculation complexity increases and a large load can be expected.

Next, a second embodiment will be described with reference to FIG. The difference from FIG. 1 is that the quality defect related rule input means 15 is provided in FIG. When the apparatus is actually operated for each inspection result, some expected causes of failure are obtained as in FIG. On the other hand, the operator confirms the actual cause of the failure based on this prediction and clarifies the true cause of the failure.

When the operation up to this point is repeated statistically a sufficient number of times, the occurrence frequency of the failure cause based on the actual result and the probability of occurrence of the inspection result from each failure cause are obtained from the history of occurrence of the actual failure cause and the inspection result. To be The actual occurrence frequency and the occurrence probability are registered in the quality defect related rule recording section 10 by the quality defect related rule input means 15 at any time. This allows
Even if the defect occurrence condition fluctuates over a long period due to the maintenance status of the mounting process, the value of the quality defect related rule recording unit 10 can be automatically maintained as a highly accurate value at any time without manually correcting it. Becomes

[0032]

According to the mounting process defect factor analysis method of the present invention, the inspection result of each step of printing, mounting and soldering in mounting of electronic parts is performed for each printed board to be mounted.
By comparing each other and calculating the degree of influence of each process on the final failure, and automatically predicting the cause of the failure, you can compare the true cause of failure to the conventional method without resorting to manual work. You can stop it faster and more accurately.

[Brief description of drawings]

FIG. 1 is a flowchart of a mounting process defect factor analysis method.

FIG. 2 is a front view of a numbered printed circuit board.

FIG. 3 is a related explanatory diagram showing registered contents of a quality defect related rule.

FIG. 4 is a related explanatory diagram showing an example of registration of quality defect related rules.

FIG. 5 is a flowchart of a mounting process defect factor analysis method according to the second embodiment.

FIG. 6 is a flow chart showing a conventional mounting process failure factor analysis method.

[Explanation of symbols]

4 Print result input means 5 Mounting result input means 6 Soldering result input means 7 Quality defect related rule recording section 8 Result comparison analysis means 15 Quality defect related rule input means

─────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-3-36800 (JP, A) JP-A-59-201500 (JP, A) JP-A-4-64286 (JP, A) (58) Field (Int.Cl. ⁷ , DB name) H05K 13/08

Claims (2)

(57) [Claims] 1. Performing a plurality of work steps and a plurality of inspection steps
After that, data that aggregates the results of the multiple inspection processes
And a plurality of causes of defects in the plurality of work processes
Relationship with the defective result in the inspection process of
The probability that the bad result will occur from a good cause.
And the defective result from the final inspection process.
A method of analyzing a cause of a process failure for identifying a cause, wherein a totaling process of collecting results of the plurality of inspection processes and the defective result in the final inspection process are generated.
At least one possible defect cause is
Extraction using the result of the aggregation process and the table
The process and the defective result are generated from the extracted defective cause.
The probability and the number of times the extracted cause of failure has occurred.
The occurrence frequency and the extracted failure cause are
Calculate the product with the influence degree for each extracted defect cause
And the product of the calculation step and the maximum value
At least one defect cause candidate including
A process failure cause analysis method comprising a specified process to be determined. 2. Data obtained by aggregating the results of a plurality of inspection processes
Based on the
Billing that includes a data change process that changes the rate or frequency of occurrence
Item 1. The method for analyzing the cause of a process defect according to Item 1.