## Tuesday, November 8, 2016

### A Machining Puzzle

I like to say that every part is a puzzle, and one of the puzzles that interests me the most is rifling a barrel.  Especially a rifle barrel.  I'll be the first to say that creating a match grade rifle barrel is a work of art on a par with any sculpture or high art you can imagine and I don't see any way that's in my future.  But what about applications where the demands aren't quite as severe, like a pistol used out to a few yards or a battle rifle used to a hundred or two?

This is one of those "how would you cope if the system collapsed" questions: barrels wear out.  What can we do?  Stockpile barrels?  Hat tip to Weapons Man today for getting me thinking about this, after posting a couple of interesting videos of rifle barrel manufacture, although thinking of how I might rifle the barrel of the GB-22 started this thought process days ago.

A few days ago, I ran across this interesting video of a guy who's making a .380 pocket pistol, sort of based on a Colt Mustang (I gather) that he calls the Kolt380. Take a look at this:

Now think about this: how is rifling different from an cutting an internal screw thread?  The main thing I see is that instead of "turns per inch" you're interested in "inches per turn".  Rifle twist rates are described as "1 turn in 16 inches" or 16 IPT.  Cutting threads on the lathe is a fundamental operation, with the thread TPI set by changing gear sets to fix the ratio of turns of the work to the advance of the cutting bit.  In this video, the barrel is clamped motionless and the cutter is spun into the work.  In internal threading, the cutter wouldn't rotate, the barrel would.  As a practical consideration, the number of TPI that can be cut with the typical lathe's set of gears is going to be much larger than you'd want for a gun barrel; say a minimum of 4.  A .223 barrel might be 1 turn in 7 inches, or .1429 TPI, while 1 in 10 inches, 0.1 TPI, is common for .308.  A typical .22LR is 1 in 16" or .0625 TPI.  Handguns run in the same general ranges.

How do we get those very small TPI values?  Extremely wide range gears?  I'm guessing the gear ranges are limited by what fits in the machine.  What if we didn't use gears? What if we drove the chuck holding the barrel with a stepper motor and advanced the cutting tool with another stepper?  There could still be gears or something to slow down either rotation, if necessary.  The most common stepper motors have 200 steps per turn.  That could be geared to 400 or more, if necessary.  The ratio of speeds would be set with CNC controller.  It would need feedback so it knows the motors are in the position they're supposed to be, but rotational encoders are pretty common in CNC threading.

Saying "let the CNC do it" doesn't let one off the hook!  There are still going to be ratios of speeds that have to be considered.  I haven't looked into this, so I can't say for sure, but I can imagine needing to run a range of speeds that might be impractical.  I might have enough numbers to start looking at potential solutions, though.

Well, this is kind of rambling, "what if?" pondering.  Something to think about as we wait to find if our long national nightmare is ending or just beginning.  Just kidding. The Deep State always wins.

1. Rifling is just a really deep pitch square thread. Ayup.

2. Stepper motors controlling pre-loaded reduction gears is pobably the answer. Many years back I used Compumotor controllers on standard 1-8 motors (200 steps/rev = 1.8 degrees per step) to make them into 5K steps/rev motors (using Compumotor's own motors with those controllers produced a 25K steps/rev combination). That was mid-1980s so I'm betting the tech has gotten much better since, and even back then there were other outfits beside Compumotor doing that sort of thing, probably bunches more now (I used CM controllers because up to 8 controllers could go on a common buss and be easily managed by addressing move commands to a particular controller with the controller's assigned address on the buss; easy-peasy for our application.)

A key component to the system would be the feedback system - even 5K steps/rev is pretty precise (especially on fine threaded pre-loaded leadscrews) and it's useful to confirm - if not actually manage - an exact number of steps executed per move command.

3. It's no problem at all to set up however you want to do these ratios. As Nosmo says, microstepping controllers have been around for ages. The thing that kills me about rifling a barrel is how the heck do you keep such a deep bore hole straight? It would seem to me that this is the bigger problem.

On the other hand, perhaps a wander in the barrel isn't so important, since the last inch of the barrel is what determines its direction, and in any case whatever that direction is it can be sighted in.

1. I don't think rifling will do that (wander in the barrel), the don't have enough cutting surface to wander, but drilling the barrel can. Drilling a centered hole through a 30" barrel (or even a 16" barrel!) is tough and I don't pretend to know the details. I know they use special drill bits that don't have flutes and multiple cutting surfaces. They're more like an arrowhead, and called D drills. See the second post here. Apparently general production barrels are made by a forging process onto orange hot steel. I'm pretty sure the really high end barrels are closer to hand made using other methods.

2. Thanks. It sounds pretty hairy, actually. I would consider these "drill bits" more like specialized boring bars. Perhaps the hobbyist should stick to 2" barrels!

3. With those long D bits, they pump cutting fluid down the bit at fairly high volumes. You'll see they pump fluid down the cutting bar in those rifling videos on Weapons Man.

I gather from reading around that Pratt&Whitney doesn't make those barrel rifling machines anymore and used ones will cost you about \$80,000 to bring up to production ready.

I think we could do a 2" barrel with a plain old drill bit followed by boring bar. Maybe even a 4" barrel. Fixturing is everything.