More PunkWorks Testing and Conclusions

[cockerpunk]

COMPRESSION TESTING -


video - http://www.youtube.com/watch?v=dbZEutYU8No

data sheet - http://spreadsheets.google.com/ccc?key=pym...HdSUomcM7hnHeeA

this is the first release of data, as the conversion continues we will release more of the data collected. me and bryce will try to stay out of the discussion unless you want clarifying questions answered.

ACCURACY CONCLUSIONS -

mine - http://www.youtube.com/watch?v=rDxWq...e=channel_page

bryces - http://www.youtube.com/watch?v=ru-IS...e=channel_page

[Chrishew09]

not trying to be a bone head but can you answer a question for me. what is the purpose of proving that a thicker shell wall takes more pressure to burst? just don't understand what the purpose of this extensive testing you guys are doing.

[cockerpunk]

not trying to be a bone head but can you answer a question for me. what is the purpose of proving that a thicker shell wall takes more pressure to burst? just don't understand what the purpose of this extensive testing you guys are doing.

this is the first baseline test to get a bearing on what is happening when you break a paintball. to see if orientation of the seam has an effect, baselines for the different types of paint and such. then we can change the application of the force, and compare it to something, or we can do different things to the paintballs and see what happens. the second page is with our distributed load or undeboring simulator.

[Hilltop Customs]

You dont have to go longer ranges on your "accuracy" testing.....increase # of tests and lower the alpha value will produce more accurate data. Trajectory calculations dont change with distance...prove the barrels are statistically the same at 50 feet and they will be the same at 100 feet.(barring spin inducing barrels and assuming the projectiles are the same between tests) If your looking to prove there is a difference you could just raise your % confidence on your current data if and it would narrow your ranges to a point the would be a "difference". Like every statistics teacher says, you can make statistics say anything you want them to.

How did you find the center of the shot grouping to take your measurements from, minimizing the average radii?

btw I think cockerpunk messed up in his explaination of the determination of the accuracy vector. The paper shows vectors and the math makes more sense to use x and y vectors to determine the radius, then use the radii to calc the std dev. Then use the std dev of the radii to calc the corresponding % of balls falling within a given target size. Had to point it out cause the way it was explainted in cockerpunks video(calculating the individual x and y std devs and using pythag) made me consider if it would work with that method or not.

cockerpunk, or anyone for that matter, look up "six sigma" if your interested in how the std dev relates to the % deffective(or off target in this case). Six sigma has more to do with industry, but most of the time it's explanation describes the relationship well.


Just an idea for further compression testing.....you guys did vertical, horizontal and 45* testing of the seams....orientation of the seam produced some pretty significant changes. How about breaking it down into finer * increments to see how close to vertical you need to be for the seam to cause lower pressure breakage. You guys do any measuring of or around the paintball seams while you were measuring shell thickness? Observe the verical breaks following the seam or area around the seam?(hehe its fun to try and see that with paint everywhere)



Neat test, makes me wonder how big of a role the fluid dynamics and inertia of the internal liquid of the paintball play in breakage. Then I guess you also have to consider the deformation of the target....blah blah blah, I could go on for days. thanks for the test...I'm going to have to venture over to punkworks sometime

[cockerpunk]

if we are open to the idea of the barrel changing accruracy, then we must be open to it changing accuracy at longer ranges. your argument is like saying we shot a target 3 inches away, and they all shot the same, therefore they should shoot the same at all ranges.

center of the grouping was simply the mean of all the shots, but standard deviation automatically takes that into account anyway.

as far as compression testing, there is alot more i have done, but nothing like what you said. after discovering what relationship we had there, i stuck to vertical and horizontal loading.

You dont have to go longer ranges on your "accuracy" testing.....increase # of tests and lower the alpha value will produce more accurate data. Trajectory calculations dont change with distance...prove the barrels are statistically the same at 50 feet and they will be the same at 100 feet.(barring spin inducing barrels and assuming the projectiles are the same between tests) If your looking to prove there is a difference you could just raise your % confidence on your current data if and it would narrow your ranges to a point the would be a "difference". Like every statistics teacher says, you can make statistics say anything you want them to.

How did you find the center of the shot grouping to take your measurements from, minimizing the average radii?

btw I think cockerpunk messed up in his explaination of the determination of the accuracy vector. The paper shows vectors and the math makes more sense to use x and y vectors to determine the radius, then use the radii to calc the std dev. Then use the std dev of the radii to calc the corresponding % of balls falling within a given target size. Had to point it out cause the way it was explainted in cockerpunks video(calculating the individual x and y std devs and using pythag) made me consider if it would work with that method or not.

cockerpunk, or anyone for that matter, look up "six sigma" if your interested in how the std dev relates to the % deffective(or off target in this case). Six sigma has more to do with industry, but most of the time it's explanation describes the relationship well.


Just an idea for further compression testing.....you guys did vertical, horizontal and 45* testing of the seams....orientation of the seam produced some pretty significant changes. How about breaking it down into finer * increments to see how close to vertical you need to be for the seam to cause lower pressure breakage. You guys do any measuring of or around the paintball seams while you were measuring shell thickness? Observe the verical breaks following the seam or area around the seam?(hehe its fun to try and see that with paint everywhere)



Neat test, makes me wonder how big of a role the fluid dynamics and inertia of the internal liquid of the paintball play in breakage. Then I guess you also have to consider the deformation of the target....blah blah blah, I could go on for days. thanks for the test...I'm going to have to venture over to punkworks sometime