Attribute Inspection P Chart

HI , i need help
can someone help me to get p chart for variable sample?
(Attributs inspection)
thnx

aadem said:

HI , i need help
can someone help me to get p chart for variable sample?
(Attributs inspection)
thnx

Click to expand...

Here are instructions for making a p-chart:

http://www.efcog.org/wg/esh_es/Stat...Trending_Primer/daytwo/18 P Chart Day Two.pdf

And an example data set

http://www.efcog.org/wg/esh_es/Stat...ol/SPC_Trending_Primer/daytwo/18a P_Chart.xls

Thanks a lot Steve that's very helpful

The link is no longer available, please send another link....thanks

Due to that website shutting down, I loaded all the files here:

https://elsmar.com/Forums/showthread.php?t=69486

Hello,

Can someone enlighten me about the p-chart question. A medical device - monthly service/repair with monthly install base volume. Im not sure if I am doing the right thing but when I plugged in the data using p-chart. The total installed base volume is too large and not so real from p_bar formula = total defective/total subgroup. Please see attached.

Romvill2002 said:

Hello,

Can someone enlighten me about the p-chart question. A medical device - monthly service/repair with monthly install base volume. Im not sure if I am doing the right thing but when I plugged in the data using p-chart. The total installed base volume is too large and not so real from p_bar formula = total defective/total subgroup. Please see attached.

Click to expand...

Way to think about it is How Many Opportunities for Failure have you had. There apparently have been 49,000 opportunities for failure over the time interval.

It does look like you have an improving trend and should split the chart into two baselines.

However, be careful that each month, there are units that range from coming from the first month you sold these to present. You do need to watch out for burn-in and burn-out. Also, be sure to take units OUT of consideration for the chart as they complete their life span.

Thanks Steve, I thought 49000 is total volume out in the field wherein the actual is only 5000 devices in circulation. I would split the chart to a new baseline as the trend is going steady. The first 12 months were ramp up from product launch.

Soon, these install base volume would variety of major repaired and refurbished devices as they go 4-5 yrs in circulation.

So Binomial is better?

Binomial is Better IF:

1. It is truly Bernoulli trials. That is, each "trial" (a piece of equipment operating for a month) has an equally likely probability of failure as any other piece of equipment, and there are no dependencies between failed equipment (if one item fails, it doesn't knock out four others).

2. There are no real world issues like burn in, burn out (ie Non-Exponential failure rates). No aging of equipment. Preventive Maintenance is OKAY, if it is CONSISTENT across all items.

Thank you very much for the enlightenment.