[Unclenick]

Too fast with a light bullet can spoil accuracy by causing core stripping, which is where the jacket is spun faster than the core in the bore. A bonded core bullet should withstand it better than a standard cup and core bullet. It does deteriorate accuracy. Harold Vaughn was able to use a magnetometer to measure it in bullets that were running around 3200 fps, IIRC? I think they were maybe 6 mm, too, but I'd have to double-check. May have been .270's.

In addition, when you spin a bullet too fast, every tiny imperfection in mass symmetry is exaggerated as wobble. It is the trade-off between that and fast settling that caused Vaughn to estimate a gyroscopic stability factor of 1.4 was about nominal, while Don Miller put it about 1.5. Close enough to each other that I usually put 1.45 into my calculator for nominal results.

Your 8" twist at 3800 fps with the 55 grain Sierra BlitzKing, for which I show a length of .747", would give a stability factor of 3.66. Sierra recommends not exceeding 3.00, generally, so I think your in potential problem territory. A 12.7" twist would give 1.45 with that bullet at that velocity.

I don't know if the bullets will actually fly apart or not? If they "disappear" on their way to the paper, then that's what happened. You'd have to try it to see? I would not expect great accuracy, even if everything holds together. Some of the bullets designed for extreme high velocity have thicker jackets and may do much better than expected. Jacket thickness being variable means there is no fixed spin limit for all bullets of the same caliber. Only those of the same caliber with the same jacket thickness.