JP2781694B2 - Semiconductor device manufacturing management 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 method of managing the production of semiconductor devices, and more particularly to a method of managing the status of a work-in-progress, a work period, and a planning support for an input / output plan.

[0002]

2. Description of the Related Art Generally, a variety of semiconductor devices are diversified even in one semiconductor device. In addition, even if one kind of semiconductor device is used, the number of manufacturing processes is 50 to 500. Furthermore, various conditions are different in one of the manufacturing processes, and equipment used is different. Also, the equipment to be used is not a single function at all, but entrusts various manufacturing processes. For example, in a normal facility, 2 to 10 manufacturing processes are assigned. In the manufacture of semiconductor devices, several operations are performed in the same facility under different processes (different conditions).

[0003] Therefore, in manufacturing such a variety of semiconductor devices, it is important to objectively grasp the status of the work in progress, or to estimate the work period (progress) and the receipt of goods in progress.

[0004] Conventionally, in a support device for assisting in grasping the in-process status, the in-process status has only been displayed as a graph or output as a list. Then, based on the graphs and the list, the human was predicting the construction period and warehousing by experience.

[0005]

In this conventional method, since a person estimates the construction period / reception from a graph or a list of the work-in-progress status, the difference between the estimated value and the actual value is very large depending on the person. Further, when a plurality of people make predictions at the same time, there is a problem that the prediction results are different in some cases.

[0006]

According to the present invention, there is provided a semiconductor device manufacturing management method, comprising: a terminal for transferring information to a host computer in each facility; Exchanging work information between the terminal and the host computer, the host computer adding / updating the information to the database as needed, and the host computer inputs the process procedure data. by-receipts and equipment operation plan data and data of the work history, chromatic status of progress, construction period, and predicting supporting incoming quantity, and displaying outputs various data processing results and the database by terminal Based on the data
In the above steps to assist in predicting the situation, construction period, and warehousing amount
So, is a certain variety on average close to the input?
Or a work-in-progress index indicating whether the work is in progress near the warehouse,
At that time, the progress of the lot of that type in the line
Average progress construction period, which averages the degree of progress converted to construction period, and constant
Estimate indicating the total number of work processes in which the type is operated during the period
Calculate the number of regular work and calculate the in-process index and the average progress
Period and the estimated number of work
It is characterized by obtaining appropriate prediction results.

[0007]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings. FIG. 1 is a schematic block diagram when an embodiment of the present invention is applied to a semiconductor device manufacturing line.

As an outline of the apparatus, a host computer 1 for integrally managing and controlling various data, a terminal 2 for reporting information from equipment to the host, and a display terminal 3 for displaying and outputting information in the host computer 1 Consists of

As an outline of the host computer 1, a procedure database unit (hereinafter referred to as DB) for managing process procedure information
4, a history DB 5 that manages work history information,
Planning DB6 that manages input, storage, and working day plans,
It calculates the work-in-process index from the data of the reporting unit 7 that controls and controls the data from the terminal device 2 and the procedure DB 4, the history DB 5, and the planning DB 6, estimates the construction period of the flowing lot, and predicts the receipt of goods. The support unit 8 performs calculations for various kinds of support.

Next, the internal functions of the host computer 1 will be described. FIG. 2 is a structural diagram showing a tree structure of the procedure DB 4. The procedure name 9 of the process procedure is set as a parent, and the process name 10 is arranged in the lower level according to the work process flow. In addition, a work equipment name and condition data are included as conditions 11 for each process at a lower level. The record of the procedure name 9 includes the procedure name and a plurality of product names.

FIG. 3 is a structural diagram showing a tree structure of the history DB 5. Product Lot No. 12 is the parent, and the lower level contains the procedure name or product name of the lot as lot information 13 and the work process name and work date and time of the lot as history 14.

FIGS. 4 to 6 show the internal structure of the plan DB 6, each of which has a relational type DB structure.
FIG. 4 is a DB relating to the input plan. The input amount for each type 16 is managed for each month 16. FIG. 5 is a DB relating to a storage plan. The storage quantity for each month 18 is managed for each product type 17. Fig. 6 shows the DB related to line operation days
It is. An operating day 19 is input for each month.

FIG. 7 is a flowchart showing the function of the report unit 7. When the processing of the lot in the equipment is completed 20, the operator 2 or the automatic transfer by the on-line equipment in step 21 causes the terminal 2 to report the name of the worked equipment and the lot No. to the report unit 7 of the host computer 1. Is transferred. From the data transferred in step 22, the equipment name and lot No. To get. Step 23
Then, in the lot No. obtained in step 22, More history D
B4 lot No. The same thing as 12 is searched, and the procedure name and the product name are obtained from the lot information 13 located below.
In step 24, as in step 23, the process name of the latest history is obtained. The procedure name 9 and the process name 10 of the procedure DB 4 are searched from the procedure name and the process name obtained in the steps 23 and 24, and the process name (process in process name) of the next process is obtained (the process name 10 is Searching is easy because they are arranged in order according to the manufacturing flow). The work equipment name is obtained in step 26 from the condition 10 below the process name (work process name). In step 27, the equipment name obtained in step 22 and the equipment name obtained in step 26 are compared. If the result of step 27 is equal, then in step 28 the history DB
The record at that time is added to the record of the process name (process in process name) obtained in step 25 as the latest history at the history 14 of the fifth item. When step 28 is completed, a message “send to the next process” is transferred to the terminal 2. If the result of step 28 is not equal, in step 30 an error message is sent to terminal 2 stating that "reporting facilities are different".

Next, the function of the support unit 8 will be described. This support unit performs a numerical operation with four mathematical expressions based on the data of the procedure DB 4, the history DB 5, and the plan DB 6.

These four formulas are formula (1) which is a formula for calculating a work-in-progress index, formula (2) which is a formula for calculating an average progress period, formula (3) which is a formula for calculating an estimated number of work, and a relational formula of a work-in-progress state. There is a certain equation (4).

Formula (1) of the work-in-progress index calculation formula is as follows:

[0017]

In this formula, for an arbitrary type A, the number of in-process lots of the type A is L, the number of process procedures of the type A (the number of processes) is M, and Serial numbers are assigned in order from 1 until entry. At this time, the serial number of the process i of the latest history of a certain lot N is Ni. The cumulative value of Ni of all lots in process for the type A is obtained, and the numerical value S defined by the equation (1) is called a process index. This in-process index S is
It is in the range of 0 ≦ S ≦ 1 and is an index indicating whether the type A is on average near the loading (near 0) or near the warehousing (near 1).

The formula (2) for calculating the average progress period is as follows:

[0020]

In this formula, for an arbitrary product type A, the number of work lots of product type A is L, the number of process procedures (product number) of product type A is M, and a certain lot G The number j of processes performed (passed) on that day is defined as Gj. Varieties A
And the numerical value T defined by the equation (2) is calculated as the average progress period. The average progress period is an average of the degree of progress of the lot in progress in the line at that time, which is converted into the period.

Formula (3) of the formula for calculating the estimated number of work is

[0023]

In this formula, the number of process steps of the type A (the number of processes) is M for an arbitrary type A, and the average progress period during an arbitrary period B (for example, one month) is shown. T, period B
Let K be the number of working days of the line, and L _M be the number of lots that are in process for the lot of the type A on average during the period B.
1) Defined by equation.

[0025]

The numerical value H defined by the equation (3) is called an estimated number of operations. The estimated number of work H indicates the total number of work processes performed by the type during the period B.

Equation (4) of the relational expression of the in-process state is as follows:

[0028]

This equation is expressed as follows: For an arbitrary type A and an arbitrary period B (for example, one month), the number of in-process lots at the end of the period B of the type A is L, and the period B of the type A is L ₁ widgets number of lots at the beginning of the time of, S ₁ widgets index at the beginning of the time period B widget exponent S, varieties a at the time the end of the period varieties a B, duration varieties a B M is the number of process steps (at the end) (the number of processes), H is the estimated number of operations during the period B of the type A, and O is the number of lots received during the period B of the type A.
As shown. By substituting equation (3) for H in equation (4), the input, storage, work-in-progress state (work-in-progress index), average progress period, and estimated number of work in period B are related.

FIG. 8 shows an example of an embodiment in which the data is output to the display terminal 3 using the support unit 8 or the respective mathematical expressions. In this example, the input, storage,
The in-process index is used as a parameter to calculate the number of in-process lots at the end of the month, the average progress of construction, and the estimated number of work. Here, what is used as a parameter and what is used as a calculation result are determined by equations (1) to (4) according to the situation to be simulated at that time.
Expressions are combined and transformed.

The table may be realized directly by the support unit 8, or necessary data may be transferred to the display terminal 3, where it is read by a commercially available spreadsheet S / W and read as shown in FIG. May be deployed.

Here, the former example will be described. In this example, the operation is performed on a monthly basis or on an arbitrary period such as a day or a quarter. The number of repetition days in an arbitrary period is also arbitrary (in this example, two days in November and December).
Times).

The table shown in FIG. 8 is composed of a total of 39 rows including a product name A31, a product name B32, and a product name C33, and columns of data on the names of October, November, and December. In this case, there are only three months, but the same applies to more than three months. The October month section includes a column of the item finger 40 of the work-in-process index at the end of October and a column of the work-in-process lot number item mouth 41 at that time.
Hereinafter, “item” is abbreviated, and for example, item connection 41 → connection 41
That. 100 under 35 processes with product name A31
Indicates the number of steps of the product name A. The number of steps 35 is
The product name at the procedure name 9 in the procedure DB is searched, and the number of lower steps of the procedure name 9 at that time is searched and obtained. The number of processes may be set in advance in a record together with the product name of the same level as the procedure name 9 and referred to. Similarly, 36 steps and 37 steps are obtained. Product name A31
"0.5" in the field of the finger 40 of the
The process of the lot as of the end of October (October 31, 23:00:59) is checked, and the lot is calculated by equation (1). Product name B32, Product name C3
The same applies to “0.7” and “0.3” of the third finger 40.

The work-in-progress index at the end of October is not calculated later based on the history, but at 23:59 on October 31.
There is also a method of checking the in-process index, storing it somewhere else, and taking it out as needed.

"30" in the column of connection 41 of the product name A31 is also
The number of in-process lots of the product name A at the end of October is counted from the history DB5. Here, the number of in-process lots can be calculated by counting lots that have a history of input and no history of storage. "40" of the connection 41 of the product name B32 and the product name C33,
The same applies to “50”. "12" of a total of 34 connections 41
"0" is the total number of lots for each product name. This column is recalculated and displayed each time the number of work-in-process lots changes.

"30" of the working day 38 indicates the number of working days of the line in November, which is transcribed with reference to the DB relating to the working days in the plan DB 6. The same applies to "25" of the operating day 39. The input 42 in November, the input 43, the input 48 in December, and the input 49 are also transcribed from the DB relating to input and input in the plan DB 6. November finger 44,
The initial value of the December finger 50 is the same as that of the October finger 40 when the spreadsheet is first expanded. The user of this spreadsheet changes this in-process index while referring to the average progress period and the estimated number of operations. When this is changed, all relevant values are recalculated each time. In FIG. 8, the product name A31 is not changed at 0.5, but the product name B32 is changed from 0.7 to 0.6,
It has been changed to 0.6. The product name C33 is also 0.3 to 0.1.
It has been changed to 4,0.5.

The work lot number port 45 for November is calculated by the following formula for each product name.

Terminal 45 = Terminal 41 + Input 42-Receiving 43 The number of in-process lots at the end of December The terminal 51 is also calculated for each product name by the following formula.

Port 51 = Port 45 + Input 48-Store 49 The estimated number of work 47 in November is calculated by transforming equation (4) to the following equation (4-1).

[0040]

Work 47 = connection 45 × finger 44 × number of processes 3
5 + warehousing 43 x number of processes 35-port 41 x finger 40 x number of processes 35 (This formula is for product name A, and for product name B, the number of processes is 3
6 is used in place of 35 steps, and when the product name is C,
(The process number 37 is used in place of the process number 35.) Similarly, the estimated work number work 53 for December is expressed by the following equation. In the case of the product name A, the work 53 = the connection 51 × the finger 50 × the number of processes 35 + the storage 49
× number of processes 35-connection 45 × fingers 44 × number of processes 35 The construction period 46 of the average progress construction period in November is obtained by transforming the formula (3) into the following formula (3-2) for each product name.

[0042]

In the case of the product name A, the construction period 46 = the number of processes 35 ×
It is calculated by working day 38 {work item 47 x {connection 41 + (input 42-storage 43)} 2}. If the product name changes, the calculation is performed by changing the number of processes 35.

Similarly, the construction period 5 of the average progress construction period in December
2 is the construction period 52 = the number of processes 35 x the working day 39 / the work 53 x
It is calculated by {port 45 + (input 48-storage 49)} 2 {for product name A).

The terminal 41, the input 42, the receiving 43, the terminal 45, the operation 47, the input 48, the receiving 49, the terminal 51, and the operation 53 of a total of 34 rows are the accumulated values of the respective columns.

If a parameter in this spreadsheet is changed, all values associated with it are recalculated. For example, if the number of input lots of the product name A in November changes, the number of in-process lots, the average progress period, the estimated number of work in November and December, and the accumulation related thereto are recalculated and updated.

[0047]

As described above, according to the present invention, a uniform and reasonable prediction result can be obtained by systematizing the in-process, storage, and construction period prediction of a semiconductor device based on intuition and experience of a person.
This has the effect of improving productivity.

[Brief description of the drawings]

FIG. 1 is a block diagram of a semiconductor device line according to an embodiment of the present invention.

FIG. 2 is a structural diagram of a database of process procedure data.

FIG. 3 is a structural diagram of a database of history data.

FIG. 4 is a structural diagram of a database of input data.

FIG. 5 is a structural diagram of a database of incoming data.

FIG. 6 is a structural diagram of a database of operating day data.

FIG. 7 is a flowchart for explaining the operation of the report unit 7 in FIG. 1;

FIG. 8 is a schematic diagram of a table output to the display terminal 3 of FIG. 1;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Host computer 2 Terminal 3 Display terminal 4-6 Database part 7 Reporting part 8 Supporting part 9-19 Database item name 20-30 Processing step 31-53 Item name

Claims (1)

(57) [Claims] 1. A method of manufacturing and managing a semiconductor device, comprising: a terminal for transferring information to a host computer in each facility;
Exchanging work information between the terminal and the host computer; the host computer adding and updating the information to the database as needed; and It has a step of predicting and supporting the status of the work in progress, the construction period, and the amount of storage based on the data of the receiving facility operation plan and the data of the work history, and a step of displaying and outputting the processing results and various data in the database on the terminal. the data based on the widget situation, construction period, predicting the incoming quantity supported
In the steps to assist, certain varieties are put on average
Is it set up nearby or is set up near the warehouse
In-progress index that indicates whether
Average the degree of progress of the lot of that type
The type of work that has been performed during the average progress
And the estimated number of work indicating the total number of work processes,
The in-process index, the average progress period and the estimated number of work
A method of manufacturing and managing a semiconductor device, wherein an appropriate prediction result is obtained by associating with each other .