Well, I found out why Barnes Bullets "like jump" ...

I measured 15 bullets (with a caliper and Hornady gauge) that were machined with a modern CNC machine. As you can see, despite the type of tooling enhancement, there is still a .005" spread. I plugged the numbers into an SD calculator (https://www.calculator.net/standard-deviation-calculator.html.), and below is the finding.

View attachment 408431
Which would insinuate the entire body of the bullet varies by .005.
In that sample. A disaster for the reloader and rifle.
Sounds like a Gage Repeatability Study ought to be done first with multiple operators in play.
 
Which would insinuate the entire body of the bullet varies by .005.
In that sample. A disaster for the reloader and rifle.
Sounds like a Gage Repeatability Study ought to be done first with multiple operators in play.
If I remember right Feenix uses Fowler-Sylvac 54-110-560-0 caliper so I'll bet his repeatability is pretty good?
Repeatability: 20µm/0.0008″
 
My simple bullet stuff number fixer came up with:

Screenshot (513).png


A variance of 2.0622222......E-6 is real tiny & with sensible views would be 0.000. Using a T Val of 2.145 for a 95% confidence level with 15 samples, (15 less 1), I produced an estimated high mean of .800 & low mean of .798. For a 95% confidence level it could be expected that the mean difference of bullet ogive lengths would be 0.002 or 2 thou. Replacing the 95% confidence T values with 99% confidence T Values would make little difference.

An insight to this stuff:


The TVal of 2.145 can be found in the table, row 14., column .975. or 1.0 - .975 = .025 on each tail of the 95% probability normal distribution curve.

Or - given a bucket of bullets from the same lot & measuring 15 bullets, assuming a normal distribution it can be estimated that the hi mean would be .800 & the low mean would be .798. The extreme spread of the data set would be .005. The variance & SD would be essentially zippo and the entire process would not provide any great surprise given the tiny differences of usually 1-2 thou. The .005 extreme spread would be dwarfed or minimized by a whole bunch of other measured bullets usually having length differences of 0-2 thou.

Running this stuff would be of interest to produce ammo for an army of 1,000's of shooters up against hordes of destructive rodents requiring .25 MOA accuracy. Diligent ammo production and quality control using statistical analyses would assure optimum results. Great economic benefits would be realized by many attendant gunsmiths in providing & replacing barrels having uniform chamber & bore dimensions.
 
It is because their tooling is not accurate.

Same box, getting 4-5 thou variance on base-to-ogive numbers.

When you jump them a lot, 4-5 thou variance is lost is in the statistical noise. (5 thou on 50 thou jump is a 10% deviation, which is hard to detect. 5 thou on a 10 jump is a 50% deviation and it shows up and you notice it more.)

Are hammer bullets any better?
I'm aware of that. Got box of 264 TTTX that shot terrible. They were off diameter wise. Some were .258. That's bad.
 
Which would insinuate the entire body of the bullet varies by .005.
In that sample. A disaster for the reloader and rifle.
Sounds like a Gage Repeatability Study ought to be done first with multiple operators in play.
The body diameter of the bullet does not vary by 0.005". The length of the bullet base to ogive varies by 0.005" in FEENIX's sample.

So, someone seating those bullets into the lands might have a problem.
 
That's then going to muck with the internal case volume as they will be seated shallower/deeper based on the bullet variance. In my case I'm already at slight compression (N570) so that will mean seating the short ones longer, or reducing the powder load.
The majority of bullets have variance. I've seen worse variance with cup and core ones than with copper ones.
I have yet to see the variance in bullet length affect the accuracy as you suggest, even with a slight compression. The typical variance in internal case volume between cases is likely more than any variance induced by bullet lengths affecting the capacity.
 
The "parting operation" of cutting the bullet from the copper rod stock is probably the largest variation in bullet geometry by a factor of 10. I really don't care if the base has .004 variation to itself. The meplat to boat tail contour on CNC machined bullets will be very repeatable. These machines replaced many secondary grinding operations because they are that good. And, it's machining copper.
Cold working like a Barnes is a tough thing to do. Ask them as I did.
 
Well your statement is quite interesting. I worried about bullet length differences enough that I opened all of my Barnes bullet boxes in 338 to make sure I was getting everything I could out of my (to me custom) 338 Edge. After measuring over three boxes I got tired of the boring consistency and decided that if the gun didn't shoot, it was the nut behind the scope and had nothing to do with the bullets I was loading. The Barnes bullets that I measured, base to ogive, varied less than one one thousandth of an inch. I don't worry about the Barnes "tooling". By the way, my 338 Edge consistently shoots five shots under a .5 moa and typically will shot five shots into less than .375 if I haven't had too many cups of coffee.
"I" always consider the "NUT" behind the trigger as the biggest factor, as shown in my sig line.
 

Recent Posts

Top