Practice Session

Qualitek-4 Practice Session


The first task to accomplish from the software is to design the experiment. But, before using the software to design the experiments, you must plan your experiment with utmost care. Follow the Recommended Experiment Planning Steps to conduct the planning session.

To start with, prepare a list of factors and levels for a practice experiment (say 5 factors at 2 levels each). Click on the DESIGN menu and select Manual Design. Check L-8 for your design. When in the design screen, describe factors and their levels by assigning them to any rows (which are columns of the array). Since you will be using five of the seven available columns, click on the UNUSED button to designate the two columns as unused. Click OK to move to the orthogonal array screen. Click OK to experiment file screen. Supply only the first eight characters of your file name. The file extension, Q4W, is automatically added.

After the design is completed, you may review the trial condition by clicking on the REVIEW menu and selecting Trial Condition. Should you want to carry out these experiments, you may print some or all the trial conditions.

Result Entry Once the experiments are carried out, you will need to enter the results by clicking on the RESULT menu and then, selecting Enter Result option. If your result includes multiple evaluation criteria, you will select the Multiple Criteria (OEC) option. Enter results/evaluation criteria as applicable. When done entering results, perform analysis following steps 3 – 8 as described above.

 To build familiarity with the software, follow the practice session as shown below:

  1. Run the Qualitek-4 (QT4) program from your computer and click OK past the Registration Screen to the experiment configuration screen.
  2. Notice that the experiment file PISTON.Q4W is already loaded in memory. If not, OPEN the file PISTON.Q4W. Since this experiment is already designed and carried out, it has results, and you are in a position to proceed with the analysis of results. (In case of a new experiment, like the one you may want to try, you will always start with the design option.)

Proceed to Review and Analyze the Results of PISTON and other example experiments. There are over 50 example experiments of all sizes included with the program.

  • Click the Analysis menu and select Standard Analysis. Since PISTON has three samples/results per trial, QT4 will remind you to perform S/N. Click OK to proceed. In the next screen check Bigger is Better Quality Characteristic and click OK.

4. Review results and click on Graph if you wish to view it. Click OK when done.

  • While in the Main Effect screen, click Plot to view plots of Main Effect. Once you are on the graphics screen, click the “<<” or “>>” button at the bottom of the graph to display other factor plots. Click OK to return to the Main Effect screen.
  • Click on the Interaction button, to display interaction plots between any two factors. Select the Automatic option if you want QT4 to calculate the interaction between all possible factors, even though you may not have thought about all the interactions or included them in your study. QT4 calculates Nx(N-1)/2 possible pairs of interactions

for N factors and ranks them by the Severity of their presence (Severity Index, 0 – 100%). Review the list to note the important pairs of interaction for future reference. Click OK or Cancel to return to the Main Effect screen and click OK to proceed to the ANOVA screen.

  • In the ANOVA screen, try POOLING the factor with the least amount of influence (last column). To POOL, double-click on the factor description. At the prompt, review % Confidence Level and click OK. Notice how the ANOVA table gets updated. You can also try one or more of the following tasks.
  • a) Pool another factor with the least influence
    b) Reset ANOVA
    c) Pool all factors with less than 98.6% confidence level.
    d) Plot bar graph, pie chart, and Reset ANOVA
    e) Review the Pie diagram

Click the OK or Optimum button to move to the optimum screen

8. At the Optimum screen:

  1. Read expected response
  2. Convert this result to a real number when S/N analysis is used.
  3. Plot graph of factor contribution
  4. Find response for all factors at level 2
  5. Go back to ANOVA and pool all factors below 10% influence
  6. Determine 90% confidence interval of the optimum response
  7. Go to the optimum screen

Click OK to return to the main screen. From the file menu select Open and select any other experiment file from the list of over 100 files. Analyze results by following steps 3 – 8. Manual Design. Check L-8 for your design. When in the design