Reloading Art vs Scientific Method

[clmayfield]

Re: Reloading Art vs Scientific Method

Well, I designed me an experiment using charge weight and seating depth as the two factors for a load using 155 Scenars and Varget.

I made the seating depth a quadratic function, which should work out pretty well. I made the charge weight a cubic function. If you think about the optimal barrel time theory, there should be a local max and a local min, so I think a cubic will work OK. In reality, a quartic function is probably more realistic, but that just increase my number of runs. I am also assuming an interaction between charge weight and seating depth.

I am going ten runs. My plan is to use CEP (circular error probable) as the response variable. I will do 7 shot groups. I really feel like you need that many shots to get an accurate estimate of CEP. Using the McMillan and McMillan chart, this is giving me the same error as using ten shot groups, which is pretty good.

We will see if I can come up with a load that yields better accuracy than old faithful (175grn SMK + 43.0grn Varget at .015" of jam). I have a lot of 75 pieces of brass, so I have 5 sighters / foulers. I will randomize the shot order much like the OCW method. 70 shots is quite a lot, but with good record keeping, I am hoping for a successful experiment.

When this is all done, I will check it against a load that Mike Miller was suggesting at one point, which is 45.0 gr Varget and 2.82" OAL (I will need to figure out how much jam/jump that is).

I am hoping this experiment will yield something fruitful. I will let y'all know the results. Here are the runs:
Charge Jam(-)/Jump(+)
46.735 -0.015
44.5 -0.015
44.5 0.1
45.125 -0.015
47 -0.015
45.75 0.0425
45.125 0.1
44.5 0.0425
47 0.1
46.375 0.1

 

[clmayfield]

Re: Reloading Art vs Scientific Method

Reporting back... I ended up testing 155 Scenar's with 8208xbr using JMP to design my experiment. As it turns out, I could have increased my powder at the high end. I was showing no pressure signs with 45.3 grains of 8208.

I used CEP as the stimulus variable and worked the loads up in an odd way... 1 shot groups. I randomized all of the cartridges fired and after the fact plotted X's and Y's using On Target's Aim function.

I shot 7 shot groups per run, which is quite a few. If I had to do it again, I would probably do 5 shot groups and test more levels.

After the fact, I reconstructed a 7 shot group to find that it was .654"... not bad.

I built a model using the results and found that the highest F Ratio was using a cubic polynomial for charge weight and a square for seating depth. The interaction between seating depth and charge weight (as well as interactions with the squares, etc.) decreased the F Ratio of my model indicating that the interaction between charge weight and seating depth is weak.

That being the case, it seems completely sensible to optimize charge weight, then optimize seating depth. Most people do it this way, but the statistics confirm that as a valid method.

My optimum charge weight is 44.9 grains and optimum seating depth is .005" jammed. That should give me a CEP of .0091". I'll do some confirmation testing to confirm this.

Interestingly, I also looked at SD of Velocity as a variable and found that the optimal charge weight for minimal SD is 44.6 grains. Looking at the difference in drops for the different velocities with JBM, I determined that I am better off at 1,000 yards with 44.6 grains, but just by a fraction. This also takes me deeper into the transonic region, so I think I will stick with the greater charge for now.

This is all really interesting stuff. I can't wait to confirm my results.

As a side note, as I was working with a new powder, I should have done some pressure and velocity testing prior to designing the experiment. I need to achieve at least 2,900 fps to remain safely above the transonic. I could have ranged my variables up quite a bit. I also should have chosen a tighter range of seating depths. Berger recommends testing at .120". I don't know anyone who runs loads with that type of jump. I should have ranged from .020" of jam to .050" of jump. Even that much jump seems a little excessive.