I've been seeing a lot of threads on round ball weight variation and its effect on repeatable accuracy. Most of the threads people stated their opinions and how they liked to sort their round balls. Some of the threads contained some data on the actual observed weights of the balls, but usually there was not a high sample quantity. So I decided to perform a small Gage R&R study on the round balls. I started out to perform a fairly robust study, but after getting into it and seeing some of the results I backed off to a more simple analysis. Plus my license of minitab apparently is expired and I am too lazy to do the analysis by hand. So here are some of my observations.
Hornady 6090 0.490 Round Balls
Lot #: 2192135
Sample Size: 30
Repetitions: 2
Scale: AWS-100
Scale was calibrated with a 100 gram weight prior to use. The rubber bung was used to hold the round ball in place. The scale was tared with the bung on the scale.
Sample Mean (Average Weight): 177.565 gn
Sample Standard Deviation: 0.192 gn
Observation #1: As shown in the Xbar chart, the process for round balls is stable and predictable.
Observation #2: The data is chunky. If you draw a horizontal line through the data you can see that several of the data points are in a row. This indicates that the scale does not have the resolution to properly measure the variation present in the round balls. This is supported by the histogram. There should not be gaps in the bins.
Conclusion #1: Will not proceed further with the Gage R&R due to the chunky data.
Conclusion #2: Variation of the weight of round balls is not as much as I was expecting based on previous threads.
Conclusion #3: Effect of round ball weight on repeatability is going to be insignificant compared to the effect of the powder charge on repeatability. This is an assumption right now and I can't prove it yet. I suspect that the variation of the powder charge is going to be much higher due to it being a manual process that the operator is going to have a huge impact on.
Anyway, this thread is not intended to change anyone's opinion or change the way they sort their round balls. It is for informational purposes only. I happen to be an engineer and find statistical analysis very interesting. So I wasted several hours on it this morning when I should be working on other things including my Kibler Kit. The kids were helping me with keeping track of samples and measuring. So it at least was a family activity.
My Kibler SMR is a 32 caliber, so once I track down a better scale, I'll perform a better study using hornady 310 balls.
Hornady 6090 0.490 Round Balls
Lot #: 2192135
Sample Size: 30
Repetitions: 2
Scale: AWS-100
Scale was calibrated with a 100 gram weight prior to use. The rubber bung was used to hold the round ball in place. The scale was tared with the bung on the scale.
Sample Mean (Average Weight): 177.565 gn
Sample Standard Deviation: 0.192 gn
Observation #1: As shown in the Xbar chart, the process for round balls is stable and predictable.
Observation #2: The data is chunky. If you draw a horizontal line through the data you can see that several of the data points are in a row. This indicates that the scale does not have the resolution to properly measure the variation present in the round balls. This is supported by the histogram. There should not be gaps in the bins.
Conclusion #1: Will not proceed further with the Gage R&R due to the chunky data.
Conclusion #2: Variation of the weight of round balls is not as much as I was expecting based on previous threads.
Conclusion #3: Effect of round ball weight on repeatability is going to be insignificant compared to the effect of the powder charge on repeatability. This is an assumption right now and I can't prove it yet. I suspect that the variation of the powder charge is going to be much higher due to it being a manual process that the operator is going to have a huge impact on.
Anyway, this thread is not intended to change anyone's opinion or change the way they sort their round balls. It is for informational purposes only. I happen to be an engineer and find statistical analysis very interesting. So I wasted several hours on it this morning when I should be working on other things including my Kibler Kit. The kids were helping me with keeping track of samples and measuring. So it at least was a family activity.
My Kibler SMR is a 32 caliber, so once I track down a better scale, I'll perform a better study using hornady 310 balls.
Attachments
Last edited:
