I'd be irresponsible if I tried to cite you a single numerical figure.
<br>
<br>Whether the difference between two diameters applies to belted, rimmed, straight, or necked cases, the minimum effective difference depends entirely on the fit of the loaded case inside the chamber.
<br>
<br>To illustrate the point, consider the force required to overcome the resistance of a sizing die, to make it "accept" a fired case. In this extremely snug "fit," the difference between the diameters of (say) fired .38 Special brass and a .38 Special case that can freely enter a .38 Special sizing die is very small. The headspace ledge in the mouth of a .30 Carbine, 9mm Parabellum, .423 OKH, .444 Marlin Rimless, or .45 ACP chamber isn't very wide, and the sized case has to fit into the chamber pretty snugly.
<br>
<br>The shoulder on a fire-formed .400 Whelen is adequate. The shoulder on another .400 Whelen case (NOT fire-formed) can be inadequate. Nominally, the difference between shoulder and neck diameters is only 0.024 inch (0.454 - 0.430 inch). The nominal shoulder angle becomes irrelevant in such a case, because the convex fillet at the base of the shoulder and the concave fillet at the base of the neck leave no straight line there to form a true angle.
<br>
<br>As a practical figure, the difference between the neck and shoulder diameters of UNfire-formed brass in an anonymous or typical chamber would have to be a bit more than the 0.024 inch that headspaces the fire-formed .400 Whelen case quite adequately.
<br>
<br>The nominal difference between the belt and body diameters of the H&H Magnum cases isn't much, either -- less than 0.020 inch (0.532 - 0.513 = 0.019 inch). If the shoulder of a SIZED necked case fits its chamber as closely as the base of a magnum case fits its recess, the same slight difference in diameters should be adequate to hold the case against the impact of the striker -- which is, after all, cushioned somewhat by the relative softness of the primer cup as it indents under that very localized impact.
<br>
<br>