Rolled Throughput Yield and Confidence Interval

[fuadhisham]

Hi all,

I'm currently conducting a study on manufacturing yield performance from pilot stage until actual production built.

The performance measure that I used is Rolled Throughput Yield(RTY). RTY is calculated by multiplication of yield at every station.

Since I would like to do multiple comparison between each shop order, I need to know the confidence interval of RTY on each shop order.

Firstly, I thought I can use proportional test in order to determine the confidence interval but yet I'm not sure if applicable to RTY as it has confounding measurement from each stations.

Question: Anybody can help me on the confidence interval of RTY? Or is there any other ways to do multiple comparison on RTY ?

Thanks in advance for any help.

 

[Stijloor]

A Quick Bump!

Can someone help with this?

Thank you very much!

Stijloor.

 

[Bev D]

I've never needed to compare point estimates of RTY so am not aware of the formula if any. A quick internet search was not helpful.

Most of the time we are comparing RTY over time (weekly or monthly or by lot). When doing this confidence intervals are not appropriate, a simple control chart is best. Since RTY is a unique 'statistic' I use the I, MR chart and that has been sufficient for my uses....

 

[fuadhisham]

Thanks for the reply. Yes, I did quick internet searching but fail to get promising answer.

I can't use control chart such as I, MR etc, as I'm interested on comparing the performance during pilot vs ramp-up production vs production. Comparing the single number of RTY does not make sense as the sample size vary upon Shop Order(SO).

Using proportion test does not make sense either as each station has different sample size and source of variance is comfounding. Proportion test is valid for yeild but yet variation on sample size on each station, making overall yield calculation does not make sense.

Please help me if my statement below is correct...
We will be able to derive the confidence interval on each station. Since the confidence interval is actually probablity of combination from sample size, sample variability and confidence level. Then the confidence interval of RTY is the multiplication of the confidence interval of each station.
Eg:
Upper Bound of RTY = upper bound of station 1 x upper bound of station 2 x ........x upper bound of station n.
Same thing as Lower bound of RTY.

Thanks in advance