From: Stan Hilliard
Date: 1/24/2006
Time: 11:13:37 PM
Hi Stan,
I was able to calculate and graph the oc curve for the lot using TP105 V3.0 and Excel in combination.
First I calculated the oc-curve of the containers: n=31, C=0. That defines a "defective" container. With tab3, I specified the table of p' vs Pa to cover the range p'=0 to 0.01.
I cut and pasted the p' and Pa columns into Excel using the "Copy Columns" command in TP105.
In Excel, I calculated the column Pr=1-Pa, (the fraction of "defective" containers). This column becomes the "p'" for the oc curve of the lot. I will call it p'c so as not to get the fraction defective containers mixed up with p' fraction defective individuals.
For each p'c on the Excel sheet, I used TP105 to calculate the Pa of the lot. This by recognizing that the beta risk is the Pa at RQL. So with n=50, Ac=1 on Tab2 of TP105, I entered p'c for RQL and calculated Beta. (neglecting AQL and Alpha, which are not needed for this.)
Now each row of my Excel sheet has a value for:
p'(individual), Pa(container), Pr=p'c(containers), Pa(Lot).
I copied the columns p'(individual) and Pa(Lot) to new columns on a second Excel sheet. This is the composit oc curve -- the oc curve of the lot. I used the chart wizard of Excel to embedd a graph of the oc curve. Chart Type=XY(scatter). Chart sub-type= "Chart with data points connected by smoothed lines without markers".
I can't paste the graph here but here are some of the points on the new composite oc curve:
p' Pa
0.0000 1.0000
0.0005 0.9351
0.0010 0.8273
0.0015 0.7098
0.0020 0.6004
0.0025 0.4990
0.0030 0.4116
0.0035 0.3357
0.0040 0.2722
0.0045 0.2197
0.0050 0.1761
0.0055 0.1403
0.0060 0.1116
0.0065 0.0882
0.0070 0.0698
0.0075 0.0549
0.0080 0.0431
0.0085 0.0336
0.0090 0.0263
0.0095 0.0205
0.0100 0.0159