Method for improving a manufacturing process

Abstract

A factorial experiment is conducted on a manufacturing process to generate a response matrix. The responses are used to calculate individual contrasts in a document as well as replicates effects. The contrast sums are also calculated and displayed in the document. The largest of the contrast sums are identified, and effects associated with those contrast sums are tested for significance using an end count method. The information from the process transformed into “significant effects” information is used to adjust process variables to improve the manufacturing process by avoiding the effect or imparting it to a measurable response of the process.

1 . A method for improving a manufacturing process wherein there are a plurality of process variables and a value of a measurable response of the manufacturing process is indicative of an improvement to the process, the method comprising: 
 conducting a full factorial experiment by setting a plurality of the process variables at a plurality of settings in a plurality of combinations of settings and receiving at least one measurement of the response of the process for each combination of level settings;    calculating individual contrasts for each process variable and each interaction among the process variables using the received responses of the full factorial experiment and displaying the individual contrasts and the sums of the contrasts for each variable and each interaction;    verifying that both variables of an interaction contrast must be set at the levels of the interaction to impart an effect substantially equal to the effect of the interaction by evaluating the variance of the contrasts displayed; and    setting the process variables as a function of the effect of the interaction on the measurable response of the process.    
   
   
       2 . The method of  claim 1 , further comprising conducting repeat tests for the full factorial experiment and calculating and displaying individual replicate effects, wherein only contrast sums greater than the sums of any set of replicate effects are identified.  
   
   
       3 . A method of improving a manufacturing process wherein a target is determined for a measurable response, the target being indicative of an improvement in the process, the method comprising: 
 conducting a full factorial experiment with at least two process variables being adjusted between at least two level settings with output responses being measurements of the response for which the target is determined;    receiving the responses of the full factorial experiment and using the responses to calculate individual contrasts for each process variable and each interaction among the process variables and displaying each of the contrasts in a document;    adding the individual contrasts of each process variable and each interaction to generate separate contrast sums;    selecting at least one of the contrast sums when it is greater than at least one of the other contrast sums by a predefined factor; and    adjusting the level settings of the process variables as a function of an estimated effect associated with the selected contrast sum.    
   
   
       4 . The method  claim 3 , wherein the level settings of the process are adjusted as a function of the estimated effect only if the estimated effect is determined to be significant.  
   
   
       5 . The method of  claim 4 , wherein when the estimated effect is an interaction effect, determining the significance of the effect comprises removing lower order effects from the responses of the full factorial experiment then testing the significance of the effect.  
   
   
       6 . The method of  claim 3 , wherein there are a plurality of contrast sums greater than at least one of the other contrast sums by a predefined factor and a plurality of contrast sums are selected, and each of the effects of the contrast sums are tested for significance with each significant effect being removed before testing the significance of another effect.  
   
   
       7 . A method of manufacturing wherein there is a required target for a measurable response of the process, the method comprising: 
 conducting a full factorial experiment by setting a plurality of process variables of the process at a plurality of settings, in a plurality of combinations of settings, and receiving at least one measurement of the response of the process for each combination of level settings;    receiving the response of the full factorial experiment and using the responses to calculate individual contrasts to calculate individual contrasts for each process variable and each interaction among the process variables and displaying each of the contrasts in a display at a particular location of the display corresponding to a notation, the notation indicating the level settings of the other of the process variables not involved in the particular contrasts;    adding the individual contrasts of each process variable and each interaction to generate separate contrast sums;    calculating effects estimates for each of the contrast sums;    identifying contrast sums that are greater than at least one of the other contrast sums;    verifying that both variables of an estimated interaction effect must be set at the levels of the interaction to impart an effect substantially equal to the effect of the interaction by evaluating the variance of the contrasts displayed when the sum of contrasts for that interaction has been identified; and    adjusting the level settings of the process variables as a function of at least one of the estimated effects.    
   
   
       8 . A method for improving a manufacturing process wherein there are a plurality of process variables and a value of a measurable response of the manufacturing process is indicative of an improvement to the process, the method comprising: 
 conducting a full factorial experiment by setting a plurality of process variables of the process at a plurality of settings, in a plurality of combinations of settings, and receiving at least one measurement of the measurable response of the process for each combination of level settings;    receiving the responses of the full factorial experiment and using the responses to calculate individual contrasts for each process variable and each interaction among the process variables;    testing the significance of effects associated with the contrasts, wherein before each effect is tested, the previously tested effect is removed from the responses if found significant, and when the effect to be removed is an interaction effect it is removed to achieve the smallest remaining estimates for the lower order effects; and    adjusting the level settings of the process variables as a function of a least one of the significant effects.    
   
   
       9 . The method of  claim 8 , wherein when the effect to be removed is a spike interaction effect, it is removed only from a response in which it was observed.  
   
   
       10 . The method of  claim 8 , wherein a graphical representation of the responses associated with an interaction effect is used to assist in removal of the interaction effect to achieve the smallest remaining estimates for the lower order effects.  
   
   
       11 . The method of  claim 9 , wherein each of the remaining lower order effects of an interaction are estimated with the other of the process variables of the interaction effect set to a level not associated with the spike interaction removed.  
   
   
       12 . A computer implemented method for use in improving a manufacturing process wherein the transformed information from implementation of the method on the computer is used to adjust the process, the method comprising: 
 receiving a plurality of level settings for a plurality of process variables and a plurality of responses from a full factorial experiment;    calculating individual contrasts for each process variable and each interaction among the process variables;    testing the significance of effects associated with the contrasts, wherein when an effect is found to be significant and is an interaction effect, it is removed before testing the significance of another affect, the removal being done to achieve the smallest remaining estimates for the lower order effects; and    displaying the significant effects of the process and the associated process variables as well as the settings of the variables not associated with the effect.    
   
   
       13 . A computer implemented method for use in improving a manufacturing process comprising: 
 receiving level settings for a plurality of process variables and a plurality of responses of the process from a full factorial experiment;    calculating individual contrasts for process variables; and    testing the significance of effects associated with the contrasts, wherein when an effect is found to be significant and is an interaction effect, it is removed before testing the significance of another affect, the removal being done to achieve the smallest remaining estimates for the lower order effects.    
   
   
       14 . The method of  claim 12  further comprising: 
 receiving the target response; and    setting the process variables as a function of the significant effects and the target response.    
   
   
       15 . A computer readable medium for instructing a computer to perform a method for improving a manufacturing process, comprising: 
 receiving level settings and responses for a factorial experiment;    calculating individual contrasts for each process variable and each interaction among the process variables; and    testing the significance of effects associated with the contrasts, wherein when an effect is found to be significant and is an interaction effect, it is removed before testing the significance of another affect, the removal being done to achieve the smallest remaining estimates for the lower order effects.    
   
   
       16 . The computer readable medium of  claim 15  wherein the significant interaction effect removed is removed from only the response in which it was observed.