AR-15 Lower: How a Midsize Manufacturer Does It

Most manufacturers guard their manufacturing processes about the way KFC guards Colonel Sanders’s 11 Herbs and Spices Coca-Cola guards the Coke recipe. They would no more show you a video of lower receiver machining than they would give you the product. Indeed, some mnufacturers not only won’t let us tour their plants, they won’t give us still photos, talk about processes, or even let their own workers bring smartphones or cameras in certain areas.

A time- and money-saving way to do a standard process can be that big of a competitive advantage.

But here’s midsize manufacturer Palmetto State Defense using a 5-axis machine to take AR-15 lower receiver forgings to 80%.

To us, it was interesting to watch how they rough milled the magazine well to approximate shape, and then CNC broached it to final shape (the rasp of the broach is unmistakeable). The end mill removes lots of metal fast, but the broach gets net shape and surface finish up to standard.

Note how much coolant/cutting fluid is used on these operations!

The link at the YouTube page no longer takes you to a buy page; PSD no longer sells the 80% lowers, apparently. They do make some attractive finished rifles, and one thing they’re selling is 200 rifles that were, supposedly, used in the Range 15 movie. (At press time, they still have 92 5.56 and 95 .300 Blackout rifles in stock). See ’em here.

33 thoughts on “AR-15 Lower: How a Midsize Manufacturer Does It

  1. whomever

    “Note how much coolant/cutting fluid is used on these operations!”

    That’s mostly about clearing the chips, I think.

    It was fun to watch. It sure is faster than me and my manual mill :-). For example, it only takes a few seconds to thread mill the buffer tube threads. For me, that means take it off the mill (losing positioning), move to the lathe, tediously single point the threads, move back to to the mill, reacquire positioning, …

    It was fun to see the single point broaching. Once upon a time that operation would have used a huge multi tooth broach a few feet long (that cost megabucks), on a dedicated broaching machine. Hobbyists like me have long done single point broaching on e.g. a Bridgeport by manually cycling the quill; I hadn’t realized the big boys had gone that way.

    Thanks for finding that!

    1. Claypigeonshooter

      I’m interested in seeing the set up you use to hold the lower in the lathe chuck if it is being machined from a raw forging. I could see how the buffer threads could easily be done in a lathe starting with just a square block of aluminum using a 4 jaw chuck. However just using a tap would seem to me to be easiest way to do it. Never seen a broach like that in a CNC as big square corners if needed at my work place are just cut on the EDM (precise) or water jet (not precise). The vast majority of time I work with manual machines so my CNC knowledge is limited to what I’ve learned in college.

      1. whomever

        Without looking up the dimensions, let’s assume that the centerline of the buffer tube threads is 0.750 above the top surface of the receiver. So, I clean up the top of the forging – that’s the reference surface. On the mill I drill/tap two holes – maybe 3/8-16 – into the metal where the magazine well and fire control pocket will be, as well as two 1/4 dowel pin holes for alignment. All these are on the centerline, at some convenient spacing.

        On the lathe I check a bar – let’s say 2 in dia by a foot long, and center it. I have machined a flat on one side, so the flat is 0.750 (the receiver top to buffer tube center dimension), and installed two dowel pins and two thru holes to match the threaded holes. Put the receiver blank onto the bar flat so the dowel pins match, then stick some 3/8 socket head screws in from the other side of the bar. So the dowel pins force alignment to the centerline, and the bolts putt the receiver down onto the flat on the bar. Now, as the spindle/chuck/bar rotates, the attached receiver rotates around the centerline of the lath, so you drill/bore/thread.

        Heh. After all that typing, I thought ‘if only I could post a pic’ – and I saw the ‘select an image’ thingy, so I’ll try to attach a picture.

        1. Claypigeonshooter

          Thanks, that is pretty creative set up. One of the things I still need to improve on is visualizing setups, but that will come as I get more experience.

      2. whomever

        Well, no joy on pics. FWIW, I see a ‘Select an image for your comment’ thing, with a ‘Browse’ button. I click that and find the appropriate pic, then ‘Post comment’. What should I be doing?

        1. Claypigeonshooter

          I go a pretty good mental image of your setup. Trying to post a pic myself.

        2. Claypigeonshooter

          Works for me. Is the file name showing up at the bottom near the choose file button after you select a picture?

          1. Claypigeonshooter

            I’ll have to keep that set up in mind not only if I machine a lower but for other parts I might make.

      3. Hognose Post author

        I’d assume (that word!) that the lower is fixed and the chuck moves. Otherwise you’d need some kind of a jig that fits in a 4-jaw chuck, wouldn’t you?

        I do have a tap for those threads. It’s the biggest tap I have by large numbers. It was fairly expensive as taps go, and I’ve never actually used it to cut threads, just to clean up buggered ones.

        (Uh, disregard. I was reading the comments alone and didn’t see that whomever had responded. his way’s very clever, and works with a common 3-jaw chuck. Ingenious!)

        1. whomever

          “I’d assume (that word!) that the lower is fixed and the chuck moves.”

          That would work, too, and would work on a lathe with smaller swing. Smaller lathes frequently come with T-slotted cross slides and milling attachments – sort of a milling vise that can be raised and lowered. Hobbyists who didn’t have a mill would put the end mill/boring bar/whatever in the lathe chuck(or collet) and use the lathe feeds plus the vertical axis provided by the attachment. Here’s a pic:

          shdesigns dot org /Craftsman-12×36/mill1.jpg

          I went the way I did because:
          -I have a big enough lathe to swing the whole contraption
          -I don’t have a milling attachment (because I have a mill :-))
          -my cross slide doesn’t have tee slots or any way to attach one anyway
          In a different shop, a milling attachment might be the best way to go. Some people make their own tap (making a tap that cuts AL is pretty easy). Hey, it’s definitely easier than making an airplane :-)

          As long as I’m being the ‘AR-15 from forgings’ pied piper, Ray Vin has a fantastic step by step tutorial:

          arlower dot ray-vin dot com /ar15/ray%20brandens%20complete%20ar%2015%20build.pdf

          if the link is mangled, hunt around on ray-vin dot com.

  2. whomever

    Here’s the old school way for an AR:

    youtube dot com /watch?v=Qib6okWndrU

    and a couple of Garand broaching ops:

    ww3 dot rediscov dot com /springar/full/10608-SA.A.1.jpg
    ww3 dot rediscov dot com /springar/full/10610-SA.1.jpg

    (had an uncle who made a lot of broaches for one of the Arsenals back in the day)

    1. Claypigeonshooter

      Thanks for the pics. I see a bunch of people with out safety glasses though. Could imagine working in a machine shop with out them.

      1. whomever

        I have a book with a pic somewhere – late 1800’s factory, overhead leather belts everywhere waiting to grab someone, and a guy (no safety glasses, of course, plus with the obligatory necktie) leaning over a workpiece in a lathe during a boring op. He’s trying to see up the bore by holding a candle. Wooden ships and iron men.

        My father, as a boy during the depression, would go along with his father to the railroad union hall to see if there was any work that day. Whoever ran things would come out periodically and say ‘need two engineers for 4 days’ or whatever. When he asked for experienced brakemen, he’d have them hold up their hand. If you had 5 digits, you weren’t experienced. This was before the safety couplers which latch when you back one rail car into another; for the old ones the brakeman had to trip something at the last minute; sooner or later you were a tiny bit slow and left a finger behind.

      2. Dyspeptic Gunsmith

        If you look at Gerstner machinists’ tool boxes, you see that there’s a mirror inset into the inside of the lid.

        Most people think that machinists must have been a very vain lot, to each have their own mirror in the lid of their toolboxes. Not so. That mirror was there in the toolbox to enable them to see chips in their eyes and try to pull them out by finger or magnet.

  3. Miles

    Someplace, far away now, I had access to one of these things, along with wire EDM and CNC lathes.
    I told my supervisor to never put let me be alone with them. The Temptation™ was just too great.

  4. DSM

    I think it’s fascinating just to watch a CNC machine do its work. All those tool paths and lines of code and then presto! Out pops an AR lower. Shoot it could’ve turned out a door knob mechanism and I’d still think it was cool as all get out. I really should cash in my GI Bill down to the community college on some machine courses if for no other reason than it’s something that interests me.

    1. Hognose Post author

      A few years ago we posted a 6-axis Swiss machining center (which is a type of tool, not necessarily from Der Schweiz any more) doing AR bolts automatically. In one shot — raw billets went in, finished bolts came out, with the machine both handling tool changes and fixturing changes as it went. Almost magic, but better, because it’s real.

      1. DSM

        I’ll look for that post on the AR bolts. Sounds great!

        The crew up at Long Rifles, Inc have turned to CNC for custom bolt rifle work. It may not be production to that scale but given the usual turnaround time for such from a traditional shop it’s practically light speed. My two are tack drivers. Only shop I’ll use now.

  5. Matt

    I’d always wondered what point on the forging they used as the reference for starting the program as every forging will be slightly different dimensionally. The Co-ordinate measuring attachment was something I didn’t know existed but makes perfect sense. Thanks for this!

    1. Claypigeonshooter

      From the video I am assuming that the forgings are close enough to a constant dimension that feature locations that are more critical are machined in the same setup because their location with certain other machined features are what is important. One of the datums in the video is the boss where the mag catch goes (which itself isn’t a critical dimension otherwise it wouldn’t have been forged). I’m guessing that one would not have to use that boss to find zero and could use other forged features (in the PDF posted by Whomever the author just eyeballs the center of the boss), as long as you keep that setup with machining operations that are critical with each other you are okay. An example from the video is the take down pin and top contour of the lower which are machined in the same setup so that they are as close to dead on with each other because if a pin hole is say 0.010in off location from the other hole the lower won’t work but if its off 0.010in from a forging feature it’s not a big deal because the critical locations are machined in relation to each other. If you were to for some reason change set up after one takedown pin hole you loose your datum so you would have to find your datum off of the pin hole that was already done instead of the forged feature that you used before doing the hole.

      1. whomever

        I wish I knew more of the details. The old school way to handle raw forgings or castings was to do a trial layout. You’d eyeball a start and start laying out all the features (with height gage and surface plate) If you were halfway through and found out the layout was extending outside the metal or whatever, you’d tweak your initial point and try again (or scrap the blank). That was for sand castings or open die forgings, which I think will vary quite a bit more than a closed die forging like an AR. But after they pick up the boss, it looks like it checks a couple of places on the deck, then one over by the grip. It would be interesting to know what it does with that collection of data? Does it do some kind of fitting among those 4 locations to best ‘center’ all the features in the forging? Just lock out and say ‘bad forging’ if it won’t clean up? I haven’t a clue.

        AR lowers may be consistent enough that you can just start from that boss and have an acceptable scrap rate, too.

        One thing that does vary, in my limited experience, is the width of the forging. I dunno if it would matter functionally, but the places with thin walls like the trigger well are going to look odd if one wall is .025 thicker than the other. I try to determine the true center individually for each forging.
        I’d love to know if the CNC is doing that.

        Another measurement they didn’t seem to take – for a couple of mine, centering on the boss would have left very little metal around the front takedown pin, so I adjusted the layout. That may be because people selling raw forgings cheap might not be pulling them from the top quality control band.

        1. Claypigeonshooter

          Looking at the video in full screen it looks like when they touch off the area near the grip it looks like they are checking the Z axis most likely to make sure everything is flat. So they can’t be checking the boss location as they skip the X axis. When they check the Deck of the lower I assume they are checking to make sure its in the fixture right or that there is clean up material.

          As for what I wrote above you I based that off of what I observed and what makes sense in terms of what dimensions are critical on a lower as my experience is limited to work where nearly everything is machined from a billet and the one 80% I have made so far; where I didn’t quite get the FCG pin holes centered in the pocket, and had to go back and make clearance for the hammer due to different Datum points. I didn’t have a dial indicator at home that would allow me to check the take down pin location for locating the FCG holes, so I used a drill bit instead if I remember right (yeah I know its the wrong way to go about it, but it worked); I also found the FCG pocket location by touching off the inside of the magwell with an edge finder where I should have stuck a pin in the hole to touch off of. After adding clearance to the front end of the pocket the hammer went in and everything functioned properly when I shot the rifle. Now my interest is piqued in machining raw forgings.

          1. Claypigeonshooter

            Only had to add about 0.03in clearance and it looks like I could have taken more out compared to my factory made lower.

    2. Dyspeptic Gunsmith

      There are very slick probing systems available now on CNC machines, mills especially.

      The company that makes quite a few of them found on CNC machines is here:–32083

      The probes get mounted in the machine spindle, as you see in the video. The probe communicates with the controller by use of the LED’s you see flashing on the probe. There will be an optical transceiver mounted on the inside of the mill enclosure, aimed at the spindle’s work envelope.

      I think one of the the slickest things that a probing system can do is probe the center of a bore hole. It used to be that finding center of a hole or outside round feature on a mill was a pain in the rump, but a Renishaw makes quick work of that problem with the probe and the software in the controller.

      Typically, you tell the probing program a “close enough” size (say, to within 1/4″”) of a feature, workpiece, hole or extrusion. You approximately center the probe by hand-jogging it over the part, about 1/4″ above the top of the part, then you run the little program. In a few moments, the dimensions are snapshotted into the controller.

      Now, if you want to see something really neat, go look up “CMM” in Youtube. That other thing you can look for is a Faro Arm. CMM’s are more precise, but Faro Arms are very quick at getting you some “good enough” dimensions to start with.

      1. whomever

        “The probe communicates with the controller by use of the LED’s you see flashing on the probe. ”

        Thanks! I’ve wondered how that works.

        At the state-of-the-art end of things, how much can the controller do with the position data.

        I.e., in ancient CNC, IIUC, the controller just played out moves – the operator might manually set zero, and past that the machine just moved per spec. In the modern world, can the controller receive a bunch of positions and then do a ‘best fit’, moving the to-be-machined features around to best fit the individual forging/casting? Can the controller say ‘oops, this one won’t clean up’ so the operator can scrap it before wasting machine time?

        AS you can tell, my understanding of CNC is back in the paper tape era. What other modern magic is out there?

        1. Dyspeptic Gunsmith

          OK, I’ll try to give you some idea of how the modern CNC controller can play magical games….

          Actually, allow me to back up and lay out the larger picture here:

          1. Castings come from the foundry oversize to the final desired result (obviously), but what most people don’t understand is how close to a specified size forgings and castings can be nowadays. Forgings of this size can be within 0.010″ of a specified size pretty reliably from a high-quality forge, with dies that pay attention to wear limits.

          2. That said, the point of forgings and precision investment casting (ala Ruger) isn’t to eliminate machining completely, but to remove the need for heavy material removal to get down to finish size. eg could you make a AR receiver set out of bar stock (not billet, dammit, billet aluminum are huge rods of aluminum you need a shop crane to lift off the truck). You could get a big slab of 7075 bar stock and just mill off everything that doesn’t look like an AR receiver.

          The trouble is, when you’re done, you’ve got a huge wad of money tied up in the bar stock material that has left the shop as chips to the recycler.

          Further, with forging (esp in steel), you can get a better metallurgical result in some aspects. With casting, you can form some really complicated shapes close to finish size far more easily than you can machine them from bar stock (eg, hammer on a 1911 or revolver).

          3. OK, back to CNC milling. On a CNC mill, there’s two issues involved in hitting your size:

          3a) Choosing and locating the origin of your part. All your positioning operations happen relative to the origin you choose for the part.

          3b) touching off all your tools in a uniform manner relative to your origin(s) on the part. The CNC machines needs to know both where your part is, and where the cutting edge of the tool is, so that it can bring the two together in a way that gets you the results you’re seeking. Typically, the most important thing that has to be set for each tool is the “Z offset,” ie, how far is the tool sticking out of the spindle/chuck? The diameter of the cutting tool can be assumed or programmed into the CNC machine.

          4) Something that most all people who haven’t really operated a CNC machine fail to understand is that an experienced CNC machinist can play lots of games with a properly written CNC program without re-writing the CNC program.

          There are two “big picture” ways to write a CNC program:

          4a) You can use CAM software to create a program that describes the toolpath to an absolute or relative position in the workpiece coordinate space. ie, you set up your forging/casting in the machine, you touch off on the workpiece, you see your X/Y/Z workpiece coordinate system (the G54… coordinate system) and then all your tool moves are described in absolute or relative movements inside that co-ordinate system. Some CAD/CAM systems do this. The CAM program has been told “We’re doing this milling operation with a 0.5″ end mill, and the center of the spindle (and endmill) traces a line along the part like this” in XYZ space, with an assumption that the centerline of the tool is 0.250″ from the line describing the boundary of your part. This is easily done because the CAM program is doing all the math in order to create a toolpath that is 0.250″ (half the diameter of a 0.5″ end mill) outside your part’s boundaries.

          If you want to take off more or less material in this sort of system, you play with the tool wear offsets in the CNC controller. You can tell the machine “OK, the program wants to cut with a 0.5″ end mill, which is called Tool 6… but I’m going to tell the CNC controller that the end mill has positive wear – ie, that the tool is actually 0.005″ larger than 0.5” and the tool moves away from the final toolpath that you programmed, leaving extra material behind on the workpiece.

          You then can take some measurements and determine “OK, now we’re 0.006 over final size – so we need to set the wear offset to -0.001 (from 0.005) and we should hit our final sizes.

          4b) You can write the program by hand, describing your toolpath as the boundary of your part (ie, reading you dimensions right off a print), and then you use the modern magic in CNC controllers of cutter compensation to get the correct toolpath for the cutters you have.

          “Cuttter compensation” means that you can tell the CNC controller that Tool (eg) 6 is a 0.5″ diameter end mill. The CNC controller will then do all the adjustment of the toolpath you’ve read off the prints to create a toolpath that is 0.25″ off the edge of your part. Let’s say you want a cylindrical pocket in your part 1″ in diameter. You can program a pocket milling canned cycle (G12 or G13), give it a 1.000″ diameter for the pocket diameter, if your tool selected is the 0.5″ diameter end mill, the controller will know it has to make the end mill run in a circle 0.5″ in diameter, with the center of the end mill tracing that 0.5″ diameter circle. You can get larger/smaller results by telling the controller that your end mill is larger or smaller – or you can use the wear offsets I mentioned above.

          Either way, you can tell the CNC controller to take more or leave more material, by playing with tool definitions on the controller.

          5) Here’s where a probing system gets really slick: You’re machining along, and you want to know “OK, so what resulting size did I get for this feature?” You can command the probe to take measurements and throw them up on the screen. You can ask for the distance between two points, you can ask for the centerline location of a symmetric feature, you can measure points off a centerline, etc. Once you have a macro programming package on a CNC controller, the only real limiting factors are your time and the memory capacity of your controller.


    I’m an engineer with a 25 year history in real time control systems, so I’d rate that as among the best 15 minutes of porn I’ve ever seen on on the internet…..ooh yeah…so much lubricant….:-)

  7. Dyspeptic Gunsmith

    One of the best ways to get a better surface finish when machining is to get the chips out of the cut. Otherwise, they get mashed back into the previously cut surface, and your surface finish looks like crap. That’s the point of that huge flow of coolant, aka “flood coolant” – push the chips out of the cut with sheer volume, rather than just high pressure, which tends to fling the chips all over the inside of the enclosure.

    On manual mills, you rarely see these “flood coolant” setups, because you really need an enclosure to keep the coolant on the mill. On CNC machines, the coolant is usually some synthetic, soluble/miscible oil/water solution, and it is mostly water.

    On manual mills, there are often provisions for dribbling cutting oil (typically high-sulphur cutting oil, or “dark” cutting oil) onto the material, keeping the cutting tool cool and lubed. High sulphur cutting oil is wonderful stuff on steel and steel alloys – on aluminum, the results look pretty mediocre, and high-sulphur oil is often slightly corrosive on brass and other yellow metals. When I’m cutting aluminum on my manual machines, the best coolant/chip remover is WD-40. No lie. I can make aluminum cut on a Bridgeport look 100% better with a can of liberally applied WD-40 while the mill is cutting. Brass, I leave dry.

    re: threading for the buffer tube on a mill. With a Blake co-ax indicator, or using a test indicator to sweep the hole, you can get the axis of the receiver’s buffer hole and the axis of the mill spindle within 0.0002 with some effort, and 0.0005 without much effort. Blakes are wonderful instruments for lining up work to a spindle co-axially, regardless whether it’s a lathe or mill. It’s faster and easier to use than a test indicator, but a test indicator will work.

    Once the axis of the mill is lined up with the axis of the hole, you can just run the tap into the hole at low speed until the tap “grabs up,” then let go of the sensitive feed handle until the tap is well into the hole. Stop the spindle, and then reverse it. Easy. I “power tap” on a Bridgeport-style mill all the time. (My mill is a Sharp, but it’s just a clone of the Bridgeports).

    Most of that toolpath will have been generated by a CAM program, which can take a CAD drawing and create a toolpath for making the stock into the final shape represented by the CAD drawing/model.

Comments are closed.