Nelson's specification management method and system
Technical field
The application relates to the statistical Process Control specification trend management technology of semicon industry, particularly a kind of Nelson's specification management method and system.
Background technology
At semicon industry, product processing technique is very complicated, cost is high, therefore whole process needs constantly goods in process to be carried out quality monitoring, obtain Reliable guarantee with the quality that guarantees the per pass processes, and machine table is effectively available, thereby produces work of high quality.
Development along with semiconductor technology, the lifting of technique, the measurement data of product required collection in whole process is more and more, it is numerous that each product machines the rear measurement data of uploading, and therefore a large amount of measurement data is carried out express-analysis calculates to judge whether corresponding product data exceed specification and just become the problem of a key.
To the monitoring of Product Process processing, be all to adopt statistical process control method usually, namely to measurement collection in product processing to data carry out statistically analysis and necessary specification, trend management.Common statistical Process Control (Statistical Process Control-SPC) system adopts Nelson's rule (Nelson Rule) to carry out specification and trend management.
Nelson's rule (Nelson Rule) has 8 rules and forms, and the sampled point scope that every rule uses is all different, and wherein the part rule also needs historical data as sampled point, concludes as table 1; Nelson's rule (Nelson Rule) needed term Zone A (regional A), Zone B (regional B), Zone C (regional C), Upper Control Limit (UCL, the upper limit), Target (Center Line, center line), definition such as Fig. 1 of Lower Control Limit (LCL, lower limit);
Table 1
Common Nelson's specification management system, when the measurement data of product enters this system, system can obtain desired data one by one according to Nelson's rule (Nelson Rule), requirement sets according to product quality specification and needed data statistics sampled point, in conjunction with measurement data of the same type in historical data, the judgement data have or not and exceed the control specification, thereby whether the trend development that judges this technique is abnormal.The data processing algorithm of common Nelson's specification management system, although very directly perceived in logic, length consuming time, high to using the system resource load request, speed is slow.
For example, for the processing of one technique of product, set up the trend management figure of a machined parameters-average thickness.in average thickness trend management figure, parameter UCL (upper limit), Target (center line), LCL (lower limit) are respectively 9,5,1, and this parameter has been collected five product data of Lot1~Lot5, as table 2, algorithm according to Sigma, drawing corresponding Sigma is 1.294, the average thickness parameter value of the 6th product Lot6 is 5.5, do you how to judge that the 6th product Lot6 data point violate Nelson's rule (Nelson Rule)? common Nelson's specification management system, for each Nelson's rule (Nelson Rule), equal 4 as example take N, first obtain previous 3 nearest data points in this trend map, add new data point namely the data of the 6th product compare accordingly, thereby confirm that the 6th data point is normal, there is no OOC (exceed and control specification) phenomenon, see Table 3:
Table 2
Data point |
Product |
Parameter value: average thickness |
1 |
Lot1 |
4 |
2 |
Lot2 |
6 |
3 |
Lot3 |
6.5 |
4 |
Lot4 |
3.5 |
5 |
Lot5 |
4.5 |
Table 3
Summary of the invention
The technical matters that the application will solve is to improve the computing velocity of Nelson's specification management, reduces the resource load of data computing.
For solving the problems of the technologies described above, Nelson's specification management method that the application provides comprises the following steps:
One. according to following table, determine the binary value of 8 attributes of last N product data of collecting;
Attribute |
Attribute description |
Has this attribute |
Do not have this attribute |
1 |
Drop on beyond the A district |
1 |
0 |
2 |
Drop on the A district |
1 |
0 |
3 |
Drop on the B district |
1 |
0 |
4 |
Drop on the C district |
1 |
0 |
5 |
Drop on the center line upside |
1 |
0 |
6 |
Drop on the center line downside |
1 |
0 |
7 |
Value greater than previous point |
1 |
0 |
8 |
Value less than previous point |
1 |
0 |
A district in table, B district, C district, center line are the corresponding definition of Nelson's rule, and N is positive integer;
Two. according to the binary value of 8 attributes of N the product data of collecting, determine whether N product data of collection exceed the control specification, specific as follows:
(1) if the value of the attribute 1 of last product data of collecting is 0, do not violate article one Nelson rule;
(2) if the value of the attribute 5 of last N product data of collecting is not all 1, perhaps the value of the attribute 6 of last N product data of collecting is not all 1, does not violate second Nelson rule;
(3) if the value of the attribute 7 of last N product data of collecting is not all 1, perhaps the value of the attribute 8 of last N product data of collecting is not all 1, does not violate the 3rd Nelson's rule;
(4) if the value of the attribute 7 of last N product data of collecting is not 0,1 staggered, and the value of the attribute 8 of last N the product data of collecting of person is not 0,1 to interlock, and do not violate the 4th Nelson's rule;
(5) if having in the value of the attribute 1 of last 3 product data of collecting have at least two be 0 or the value of the attribute 2 of last 3 product data of collecting in have have at least two be 0 or the value of the attribute 1 of last 3 product data of collecting in have have at least 1 be 0 and the value of the attribute 2 of last 3 product data of collecting in have and have at least 1 to be 0, perhaps the value of the attribute 5 of last 3 product data of collecting be not all 1 and the value of the attribute 6 of last 3 product data of collecting be not all 1, do not violate the 5th Nelson's rule;
(6) have at least two to be 1 if having in the value of the attribute 4 of last 5 product data of collecting, and the value of the attribute 5 of last 5 product data of collecting is not all 1, the value of the attribute 6 of 5 product data of collection is not all 1 at last, does not violate the 6th Nelson's rule;
(7) if the value of the attribute 4 of last N product data of collecting is not all 1, do not violate the 7th Nelson's rule;
(8) if the value of the attribute 4 of last N product data of collecting is not all 0, do not violate the 8th Nelson's rule.
For solving the problems of the technologies described above, Nelson's specification management system that the application provides comprises product data attribute modular converter, controls the specification processing module;
Described product data attribute modular converter is used for according to following table, determines the binary value of 8 attributes of last N product data of collecting;
Attribute |
Attribute description |
Has this attribute |
Do not have this attribute |
1 |
Drop on beyond the A district |
1 |
0 |
2 |
Drop on the A district |
1 |
0 |
3 |
Drop on the B district |
1 |
0 |
4 |
Drop on the C district |
1 |
0 |
5 |
Drop on the center line upside |
1 |
0 |
6 |
Drop on the center line downside |
1 |
0 |
7 |
Value greater than previous point |
1 |
0 |
8 |
Value less than previous point |
1 |
0 |
A district in table, B district, C district, center line are the corresponding definition of Nelson's rule, and N is positive integer;
Described control specification processing module is used for the binary value according to 8 attributes of N the product data of collecting, and determines whether N the product data of collecting exceed the control specification, and is specific as follows:
(1) if the value of the attribute 1 of last product data of collecting is 0, do not violate article one Nelson rule;
(2) if the value of the attribute 5 of last N product data of collecting is not all 1, perhaps the value of the attribute 6 of last N product data of collecting is not all 1, does not violate second Nelson rule;
(3) if the value of the attribute 7 of last N product data of collecting is not all 1, perhaps the value of the attribute 8 of last N product data of collecting is not all 1, does not violate the 3rd Nelson's rule;
(4) if the value of the attribute 7 of last N product data of collecting is not 0,1 staggered, and the value of the attribute 8 of last N the product data of collecting of person is not 0,1 to interlock, and do not violate the 4th Nelson's rule;
(5) if having in the value of the attribute 1 of last 3 product data of collecting have at least two be 0 or the value of the attribute 2 of last 3 product data of collecting in have have at least two be 0 or the value of the attribute 1 of last 3 product data of collecting in have have at least 1 be 0 and the value of the attribute 2 of last 3 product data of collecting in have and have at least 1 to be 0, perhaps the value of the attribute 5 of last 3 product data of collecting be not all 1 and the value of the attribute 6 of last 3 product data of collecting be not all 1, do not violate the 5th Nelson's rule;
(6) have at least two to be 1 if having in the value of the attribute 4 of last 5 product data of collecting, and the value of the attribute 5 of last 5 product data of collecting is not all 1, the value of the attribute 6 of 5 product data of collection is not all 1 at last, does not violate the 6th Nelson's rule;
(7) if the value of the attribute 4 of last N product data of collecting is not all 1, do not violate the 7th Nelson's rule;
(8) if the value of the attribute 4 of last N product data of collecting is not all 0, do not violate the 8th Nelson's rule.
Nelson's specification management method and system of the application, characteristics according to each rule in the Nelson rule, each product data is carried out the feature analysis and summary, and it is transformed the binary value of 8 attributes, then according to the incidence relation of each rule in the Nelson rule and 8 attributes, adopt binary bit-based algorithm to judge whether corresponding product data exceed specification.Nelson's specification management method and system of the application are owing to adopting the binary number value-based algorithm, and not only speed is very quick, and the findings data form is simple, the needs of historical data have been avoided reviewing, improve computing velocity, simplified the data storage, reduced the resource load of data computings.
Description of drawings
In order to be illustrated more clearly in the application's technical scheme, the below does simple the introduction to the accompanying drawing of the required use of the application, apparently, accompanying drawing in the following describes is only some embodiment of the application, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain according to these accompanying drawings other accompanying drawing.
Fig. 1 is the needed term definition schematic diagram of Nelson's rule;
Fig. 2 is Nelson's specification management method schematic diagram of the application.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present application, the technical scheme in the embodiment of the present application is clearly and completely described, obviously, described embodiment is only the application's part embodiment, rather than whole embodiment.Should be appreciated that preferred embodiment described herein only is used for description and interpretation the application, and be not used in restriction the application.And in the situation that do not conflict, embodiment and the feature in embodiment in the application can make up mutually.Based on the embodiment in the application, the every other embodiment that those of ordinary skills obtain belongs to the scope that the application protects.
Below from Nelson rule (Nelson Rule) signature analysis, this case is elaborated.
Definition from Nelson's rule (Nelson Rule), 8 attributes that we can be defined as follows each data point (each measured data values of product), and represent with 1 or 0 whether this data point has corresponding attribute, as shown in table 4;
After each data point had had these 8 attributes, for continuous N point, each attribute was to be represented by continuous N individual 0 or 1;
After the attribute of N data point shows with 0,1 such mode, for each Nelson's rule (Nelson Rule), can carry out with the attribute of correspondence bit-based algorithm and process, as shown in table 5;
Table 4
Attribute |
Attribute description |
Has this attribute |
Do not have this attribute |
1 |
Drop on beyond the A district |
1 |
0 |
2 |
Drop on the A district |
1 |
0 |
3 |
Drop on the B district |
1 |
0 |
4 |
Drop on the C district |
1 |
0 |
5 |
Drop on the center line upside |
1 |
0 |
6 |
Drop on the center line downside |
1 |
0 |
7 |
Value greater than previous point |
1 |
0 |
8 |
Value less than previous point |
1 |
0 |
Table 5
Nelson Rule N Nelson rule elson Rule |
Required consideration attribute |
Rule one: a bit at regional A (3Sigma) in addition |
1 |
Rule two: a N continuous point is in center line the same side |
5||6 |
Rule three: a N continuous point, increasing or decreasing |
7||8 |
Rule four: a N continuous point is staggered up and down |
7&8 |
Rule five: have 2 points to drop on beyond B district, center line the same side in continuous 3 points |
5&1&2||6&1&2 |
Rule six: have in continuous 5 beyond 4 C districts that drop on center line the same side |
5&4||6&4 |
Rule seven: a N continuous point drops on the C district |
4 |
Rule eight: N continuous is put and is dropped on beyond the C district |
4 |
Embodiment one
As shown in Figure 2, Nelson's specification management method comprises the following steps:
One. according to table 4, determine the binary value of 8 attributes of last N product data of collecting;
A district in table 4, B district, C district, center line are the corresponding definition of Nelson's rule, and N is positive integer;
Two. according to the binary value of 8 attributes of N the product data of collecting, determine whether N product data of collection exceed the control specification, specific as follows:
(1) if the value of the attribute 1 of last product data of collecting is 0, do not violate article one Nelson rule;
(2) if the value of the attribute 5 of last N product data of collecting is not all 1, perhaps the value of the attribute 6 of last N product data of collecting is not all 1, does not violate second Nelson rule;
(3) if the value of the attribute 7 of last N product data of collecting is not all 1, perhaps the value of the attribute 8 of last N product data of collecting is not all 1, does not violate the 3rd Nelson's rule;
(4) if the value of the attribute 7 of last N product data of collecting is not 0,1 staggered, and the value of the attribute 8 of last N the product data of collecting of person is not 0,1 to interlock, and do not violate the 4th Nelson's rule;
(5) if having in the value of the attribute 1 of last 3 product data of collecting have at least two be 0 or the value of the attribute 2 of last 3 product data of collecting in have have at least two be 0 or the value of the attribute 1 of last 3 product data of collecting in have have at least 1 be 0 and the value of the attribute 2 of last 3 product data of collecting in have and have at least 1 to be 0, perhaps the value of the attribute 5 of last 3 product data of collecting be not all 1 and the value of the attribute 6 of last 3 product data of collecting be not all 1, do not violate the 5th Nelson's rule;
(6) have at least two to be 1 if having in the value of the attribute 4 of last 5 product data of collecting, and the value of the attribute 5 of last 5 product data of collecting is not all 1, the value of the attribute 6 of 5 product data of collection is not all 1 at last, does not violate the 6th Nelson's rule;
(7) if the value of the attribute 4 of last N product data of collecting is not all 1, do not violate the 7th Nelson's rule;
(8) if the value of the attribute 4 of last N product data of collecting is not all 0, do not violate the 8th Nelson's rule.
Embodiment two
Based on Nelson's specification management system of Nelson's specification management method of embodiment one, comprise product data attribute modular converter, control the specification processing module;
Described product data attribute modular converter is used for according to table 4, determines the binary value of 8 attributes of last N product data of collecting;
A district in table 4, B district, C district, center line are the corresponding definition of Nelson's rule, and N is positive integer;
Described control specification processing module is used for the binary value according to 8 attributes of N the product data of collecting, and determines whether N the product data of collecting exceed the control specification, and is specific as follows:
(1) if the value of the attribute 1 of last product data of collecting is 0, do not violate article one Nelson rule;
(2) if the value of the attribute 5 of last N product data of collecting is not all 1, perhaps the value of the attribute 6 of last N product data of collecting is not all 1, does not violate second Nelson rule;
(3) if the value of the attribute 7 of last N product data of collecting is not all 1, perhaps the value of the attribute 8 of last N product data of collecting is not all 1, does not violate the 3rd Nelson's rule;
(4) if the value of the attribute 7 of last N product data of collecting is not 0,1 staggered, and the value of the attribute 8 of last N the product data of collecting of person is not 0,1 to interlock, and do not violate the 4th Nelson's rule;
(5) if having in the value of the attribute 1 of last 3 product data of collecting have at least two be 0 or the value of the attribute 2 of last 3 product data of collecting in have have at least two be 0 or the value of the attribute 1 of last 3 product data of collecting in have have at least 1 be 0 and the value of the attribute 2 of last 3 product data of collecting in have and have at least 1 to be 0, perhaps the value of the attribute 5 of last 3 product data of collecting be not all 1 and the value of the attribute 6 of last 3 product data of collecting be not all 1, do not violate the 5th Nelson's rule;
(6) have at least two to be 1 if having in the value of the attribute 4 of last 5 product data of collecting, and the value of the attribute 5 of last 5 product data of collecting is not all 1, the value of the attribute 6 of 5 product data of collection is not all 1 at last, does not violate the 6th Nelson's rule;
(7) if the value of the attribute 4 of last N product data of collecting is not all 1, do not violate the 7th Nelson's rule;
(8) if the value of the attribute 4 of last N product data of collecting is not all 0, do not violate the 8th Nelson's rule.
Embodiment three
For the processing of one technique of product, set up the trend management figure of a machined parameters-average thickness.In average thickness trend management figure, parameter UCL, Target, LCL are respectively 9,5,1, and this parameter has been collected five product data of Lot1~Lot5, as table 2; According to the algorithm of Sigma, the average thickness parameter value that draws corresponding Sigma and be 1.294, the six product Lot6 is 5.5.
Nelson's specification management system by product data attribute modular converter, according to table 4, determines the binary value of 8 attributes of five product data of Lot1~Lot5 of previous collection as shown in table 2, as table 6; When the 6th data value is after 5.5 product data gather, according to the value 4.5 of last data point Lot5 in current trend map, to draw the binary value of 8 attributes of the 6th product data, as table 7;
Table 6
Data point |
Attribute 1 |
Attribute 2 |
Attribute 3 |
Attribute 4 |
Attribute 5 |
Attribute 6 |
Attribute 7 |
Attribute 8 |
1 |
?0 |
?0 |
?0 |
?1 |
?0 |
?1 |
?N/A |
?N/A |
2 |
?0 |
?0 |
?0 |
?1 |
?1 |
?0 |
?1 |
?0 |
3 |
?0 |
?0 |
?1 |
?0 |
?1 |
?0 |
?1 |
?0 |
4 |
?0 |
?0 |
?1 |
?0 |
?0 |
?1 |
?0 |
?1 |
5 |
?0 |
?0 |
?0 |
?1 |
?0 |
?1 |
?1 |
?0 |
Table 7
Data point |
Attribute 1 |
Attribute 2 |
Attribute 3 |
Attribute 4 |
Attribute 5 |
Attribute 6 |
Attribute 7 |
Attribute 8 |
6 |
?0 |
?0 |
?0 |
?1 |
?1 |
?0 |
?1 |
?0 |
According to pressing the relation of the Nelson's rule of each shown in table 5 (Nelson Rule) with 8 attributes shown in table 4, the attribute at respective counts strong point is carried out step-by-step to be calculated, thereby confirm whether these the 6th product data be normal, there is no OOC (exceed and control specification) phenomenon, see Table 8;
Table 8
Nelson's specification management method and system of the application, characteristics according to each rule in the Nelson rule, each product data is carried out the feature analysis and summary, and it is transformed the binary value of 8 attributes, then according to the incidence relation of each rule in the Nelson rule and 8 attributes, adopt binary bit-based algorithm to judge whether corresponding product data exceed specification.Nelson's specification management method and system of the application are owing to adopting the binary number value-based algorithm, and not only speed is very quick, and the findings data form is simple, the needs of historical data have been avoided reviewing, improve computing velocity, simplified the data storage, reduced the resource load of data computings.
The above is only the application's preferred embodiment, and is in order to limit the application, not all within the application's spirit and principle, any modification of making, is equal to replacement, improvement etc., within all should being included in the scope of the application's protection.