I drew a magazine well cross-section and calculated tool paths which I hope to apply this evening. Electronically constructing everything was more work than machining the entire buffer end of the forging. I've attached my Google Sketchup file with dimensioned cuts to save other people the trouble. No warranty - you'll want to compare to the M16 mechanical drawing.
Most of the extra marks and guides show where the walls would be approaching the tightest and loosest in-spec magazine wells.
My motivation in adjusting the magazine well was to get more wall thickness at the rivet clearance cut due to Ray's deep lug cut start and my deeper screw-up. That was entirely straight forward - I got .013" of wall thickness by getting rid of Ray's .003" forward bias, shifting the magazine well to the aft end of its +/- .005" location specification, and moving the rivet clearance radius to the aft end of its +/- .005" tolerance.
While I had Ray's instructions and M16 lower blueprint jpeg (as I noted earlier in the thread it includes things omitted from D Snider's redraw and lacks some tolerance errors) in front of me I made some other comparisons which were out of spec
That is entirely reasonable -
The M16 mechanical drawings are not available via a Freedom Of Information Act inquiry because Colt considers them proprietary. You get them when you win a government contract and sign a non-disclosure agreement with Colt. Even if civilian market AR15 manufacturers were in a legal position to share there's no motivation. That means no canonical source of dimensions and understandable ambiguity on measurements.
but I wanted to do the right thing to avoid magazine fit problems and offending my sense of craftsmanship and did my own calculations + drawings regarding where to remove metal.
I don't currently have a CAD program on my laptop, so I used Google Sketchup with constructive geometry as a digital straight edge + compass. For example if an end-mill needs to touch a point (the apex of the outside corner transitioning between wide magazine rear and slot for its seam) and a line (the inside of that slot) you can put a circle with a matching radius centered on the point, line offset by the radius from the straight edge, and have the necessary center at the intersection. With Sketchup it's easiest to add the locating geometry, add and copy the feature or is control point, undo the locating bits (Sketchup sometimes exploding arcs into their component straight segments when one intersects something can make them hard to remove with the erase tool), and paste what's needed.
The M16 lower blueprint calls for a flat side bulge of .485" +/- .010" long terminating in .030" radii at its fore and aft end ends which would make it .475" wide at the radii ends and .535" at its wide part in the Maximum Material Condition producing the smallest hole and tightest in-spec magazine well.
Ray approximates that bulge with a 1/8" drill bit producing a .0625" radius, with holes centered at X 2.289 and 1.923 from the pivot pin axis. That's .366" (.109" narrower than spec at MMC) at top and .491" (.044" narrower). Presumably this comes from working off a redraw of a redraw - D Snider's omits the .030" radius.
I changed that without going outside the Least Material Condition producing the biggest hole and loosest in-spec magazine well.
The aft end of the magazine well is supposed to transition from its wide flat part to the narrow slot for the spot welded magazine rear rib with .062" radii. It'd be a lot of work to do that on a manual mill with no functional advantage so Ray approximates which is sound engineering. However doing the arithmetic on his plunge cuts at X 2.631 Y +/- 0.121 shows they're just knocking off the tip; and the 3/8 cut at X 2.712 between Y+/- 0.066 is leaving metal which should be missing. While probably OK, I made the rear cut .015" wider and moved the plunge cuts to knock off the other side of each corner so none of my approximated outside corners extend past where a "correct" cut would be.
The intersections with the specified .062" radii are visible below, but more so in the double sized drawings and scalable Sketchup file.
The magazine well sides are relieved .0175" +/- .0075" between .31" +/- .010" long pads at each end. Ray makes thin tails which intrude into that area; and while that definitely works for flat sided USGI magazines I adjusted by .020" to meet spec in the unlikely event some one has decided to use that space.
Here are my .0625" radius corners including the relocations necessary for the above geometry:
I calculated the rest of the tool paths needed and have a nice ordered set to reference when milling. Everything "should" match Ray's instructions apart from the second 3/4 roughing pass. My long 3/4 inch end-mill is actually .740" which I took into account for the second roughing pass with it. I left a few thousandths extra so accidentally re-using a .750" end-mill wouldn't touch the finished surfaces, although if I intended to use a 3/4" mill I'd adjust the coordinates.
