Thanks all for replies. I believe I haven't been very clear what I'm trying to accomplish here. My primary purpose for starting the thread was just to learn. While I totally appreciate that it's difficult to build a load for a rifle to meet certain goals without doing lots and lots of testing, I find that understanding the theory behind what happens as a result of changing various parameters is really useful for me to make educated decisions on what to test next. In other words, I like to know the "why" and not just the "what."
I think one of the reasons things haven't been clear is that I'm asking some specific questions that relate to various ongoing projects of mine, so it looks like I'm just wandering around aimlessly here
Let me try to be a bit more specific.
Project #1 right now is finding a reasonably accurate cast .223 load for my AR-15 (aiming for 2 MOA), one which cycles with 100% reliability. The rifle is as basic as they come: 16" barrel, 1:9 twist, carbine-length gas system without adjustable gas block, standard buffer spring and weight. Basically made to sling cheap 55gr surplus ammo. The mold I have is the Lee 55gr FP (.225"). It's not that great of a mold and my static fit is pretty bad. I realize that the very best thing I could do to improve accuracy is to buy a better mold, but that's not in the cards right now, so I'm trying to use this as a learning opportunity, and see what I can achieve by experimenting with other changes that I CAN make right now. Specifically, bullet hardness as measured with BHN, powder type, and powder charge.
I have done lots and lots of testing. Without going into too much detail, I have been juggling various combinations of the three factors listed above. Namely, bullets ranging from BHN 12 to BHN 35, different powders (specifically IMR 4895 and H335), and loads everywhere from "so low it barely cycles" to "hot jacketed loads." So far, my best and most accurate results have inarguably come from the BHN 35 bullets. There has been no load in my testing using BHN 12, 18, or 26 bullets that was more accurate than the same load with BHN 35 bullets. To me, this makes sense. Yes, my bullet fit is working against me, and perhaps nothing else I can do will fully overcome that problem. But it seems like the higher the BHN, the more resistant the bullet will be to deformation under pressure. I have heard numerous people say that after a certain point, increasing BHN actually leads to worse accuracy, but nobody has actually explained why yet. So not only has this not matched my experience, it also seems to contradict my best attempts to understand WHY extremely high BHN bullets have performed best for me.
Interestingly, my best results have been with these 35 BHN bullets and also my highest velocity loads tested to date, using IMR 4895. I got about 4 MOA with a 10-shot group of these hard bullets over 20.5 grains of IMR 4895, which came out just over 2700 FPS. Oh, and no loads of mine have ever leaded my AR bore, no matter how hard or soft the alloy. I would think you'd need to mess up pretty badly to get leading with powder coated, gas checked bullets.
So I have just finished loading up what will probably be my final and definitive set of tests to see what my 35 BHN bullets are capable of. I have loads ranging from 18 grains of H335 to 25 grains (approaching book max for 55gr bullets), and from 18 grains of IMR 4895 to 24 grains. If none of these loads get me to 2 MOA as I'm hoping, I'm having a hard time believing that a high BHN is contributing to any inaccuracy, or that lowering the BHN might tighten up my groups. Perhaps I'm wrong, and if so I hope someone will explain to my why that is. If all of these loads fail, then the only thing I can think to try is a slower powder than 4895 (maybe RL 12 or 4064?), or playing with different additions to my alloy. I have a small amount of high-antimony lead which is a blend of linotype, electrotype, and monotype which I can mix in with my COWW, and I think that would probably help.
So bottom line is, I know the "right answer" is to get a better mold, but since that is not doable right now I'm trying to learn what else I can tweak to improve my results within the limitations presented by the poor fit of this mold. For the purposes of this load I actually don't care about velocity at all. It could be a 1000 FPS load or 3000 FPS, and as long as it hit 2 MOA and cycled 100% I'd be happy. Just so happens that my best results have been at quite high velocities with very hard bullets. Within my working hypothesis for what was mainly causing my inaccuracy (deformation upon firing), it kinda seems like the higher the BHN the better in my specific scenario, and my results bear that out thus far.
So as far as theoretical questions I have posed here, I think it boils down to:
1. What causes higher BHNs to be detrimental to accuracy? Is it failure to obturate? Something else? And why have my results been best with 35 BHN bullets if 35 is apparently way higher than I should be using?
2. If I wanted to use a lighter load (~2100 FPS) in an attempt to keep chamber pressure down and thus reduce bullet deformation, would increasing neck tension help me get a better burn? The answer here seems to be "probably, but at the expense of causing different and likely worse problems."
3. Is there some way to get a good consistent burn in a reduced load with either stick powder (4895) or ball powder (335)? The answer here seems to be that consistent burn simply requires higher pressure. It doesn't matter which method is used to get higher pressure (increased charge, magnum primer, filler, heavier bullet, what have you), it's the higher pressure itself that is going to deform the bullet anyway. Meaning that the best I can hope for with a reduced load is that maybe it's low enough pressure to allow my bullet to stay straight from chamber to muzzle, but the burn and therefore velocity will be inconsistent and I'll always have vertical stringing.
The verdict appears to be that "reduced loads" and "consistent burn" are diametrically opposed concepts. I'm beginning to lose hope that there might be a "sweet spot" window where pressure is high enough for a good consistent burn, but low enough not to screw me over when my static fit is as bad as it is. If this "sweet spot" exists at all, it seems to me that the harder bullet I can get, the higher pressures I can get away with before the bullet accuracy is destroyed. And if it exists with H335 or IMR 4895 and my current alloy, I suspect my next set of tests will find it. If not... maybe a different alloy, maybe a different powder. Maybe both. Maybe there really is no way to get 2 MOA at with any cycling load using this bullet in my rifle. I'd like to believe there is... but maybe not.