I don’t turn necks unless I have to, I prefer to set my no turns up with .004 clearance.
That’s why I set my tension with a mandrel. I size down .004 below and then run a carbide mandrel that’s .001-.003 under, usually .002 unless it’s not satisfactory. This does a decent job of uniforming tension and it most definitely shows on paper.
When I do have to turn, I still turn for .004 clearance and size the same way, which makes tension as uniform as it’s ever going to get.
I have tested freshly annealed cases with and without turning and with and without the mandrel and the improvement is definitely linear in a positive direction the more steps I take.
I want good clearance, imperfections on the outside or nonexistent and low (but not too low) seating force.
People have been turning to the .0001 and annealing for decades. I think that’s many a thousand “ayes” for uniform neck tension.
In addition to that, check out some groups fired from custom muzzle loaders. They will all shoot in the .1s. Why? The absence of a case screwing things up is my opinion.
Bottom line, absolutely anything or everything you can do to make one case exactly like the other 4 in a string is going to show on paper. Uniform neck tension is near the top of that list.
The amount of the interference fit between the bullet and the case neck is commonly referred to as "neck tension". I do not think that is the best technical term, but I will use it here to help with communication. So, using this definition, If I seat a 0.224 diameter bullet in a case with an internal neck diameter of 0.221 the result is 0.003 neck tension.
I have measured the expander ball on my standard Remmington 223 case resizing die and it is 0.223 and this gives me 0.003 or more neck tension depending on the spring back of the brass. With neck tension at 0.003 or greater it takes what I would subjectively judge to be considerable force to seat the bullet in the case. I have concerns that this level of force may do some damage to the bullet and may be detrimental to accuracy.
I have found that I can control neck tension by the use of the proper diameter expander mandrel. There are now good sources of expander mandrels in measurement increments of 0.0005 and with a little experimentation I can set neck tension at the amount I choose within the limitations of the variation of the spring back of the brass. This helps reduce the seating force and gives a much more consistent feel during the bullet seating process.
For a semi auto like the AR it is not recommended to go too low due to the risk of the bullet moving in the case during the dynamics of recoil and of the cycling of the action. So, I set neck tension between 0.002 and 0.003 for that application. For the single shot bolt action, I set it at approximately 0.001.
My question is, has anyone done any accuracy testing to see if there is any effect on accuracy due to variations in neck tension?
Yes, Neck tension, makes a difference in accuracy. Some powerders are more forgiving than others. You have to start with quailty brass, like Lapua.