I think an accurate upper is the collective result of several smaller issues.
I just finally got out to shoot here in AZ (after 110+ days of 100+ heat), and the AR Stoner sporter billet receiver with integrated handguard makes a difference. It's rigid, period. I lapped and glued the Barrel Extension and Upper together. The barrel is a heavyweight 24" fluted AR Stoner (Liberty?) barrel. There is no perceptible play between the AR Stoner Upper and PSA Classic receiver.
I also believe that a good trigger is a crucial plus over the standard AR trigger, and I installed a moderately economical PSA Nickel-Boron 2-Stage.
The chambering is 6.5 Grendel, brass is Starline, bullet is Speer 90gr TNT, primer is CCI BR-4, and powder is IMR-8208XBR. All of these components, except the BR-4's, are new to me. This is the same rifle that broke its extractor using 120gr SGD's and an excessive load of W-748. The extractor as replaced with the J-P Improved Grendel Extractor, and henceforth all loads will conform to published load limits.
The charge is rounded down to the next .5gr below max as I find that most of my accurate loads are just a few tenths below max. No report on MOA's, too few groups for a representative average, but the rifle does shoot small. Next few groups will be the same load to get a good statistical sample, and then testing of bracket loads will commence to see if any improvement is available. The resulting load is 31.5gr of IMR-8208XBR (31.7gr is max from the Speer Manual), with an estimated 3000ft/sec MV in the 24" barrel with a listed BC of.281. My range altitude is around 4500ft ASL.
In summary, I think the initial results are promising, especially so in the light of the rifle being built with economy in mind, and an effort was made to do the lapping and gluing to see if it is compatible with good performance.
So far, I think it is.
There is another rifle, built on a factory 20" AR Stoner 6.5 Grendel Upper and the same PSA Classic Lower with the same PSA Nickel-Boron 2-Stage Trigger. It comes without the lapping and gluing. Once a reliable benchmark level of accuracy is established, I will lap and glue it to see if that changes anything.
Greg