Designing A Rifle From Scratch(ish)

[Sturgeon]

Significant BCG redesign today. Initially I was going to have the firing pin ride with the bolt, like an AK. The benefits of this are that the bolt stem retracts into the bolt carrier when the bolt unlocks, saving some overall length with the bolt group and allowing additional overtravel. Here's the original design with the bolt unlocked (notice the retracted firing pin and bolt stem):

 

This has a couple disadvantages, the chief one being that it fucks you coming and going on your mass ratio. Initially, I thought I would have plenty of excess mass in the bolt carrier for this to give me a ratio of 5.6:1 or above, but once the design was complete it turned out I only had a ratio of 4.3:1, which is not terrible but not what I'd hoped for. A few days earlier @Collimatrix had insisted that, in fact, I would get whupped in mass ratio by this so 10 internet points to him for getting that right.

 

I switched over to a firing-pin-in-carrier arrangement, which you can see below. The mass ratio jumped up to 6.3:1, and all is well in the world.

 

 

The carrier has an extension on the back of it which houses the firing pin. It could probably be made a little more manufacturable but for right now it's fine. During the recoil stroke the extension just telescopes into the recoil buffer. Not a big deal.

[Collimatrix]

2 hours ago, Lord_James said:

 

This might be stupid, but couldn’t you make that 300mm long handguard in 3x 100mm long pieces and fuse them together? It might not look pretty, but it should get the job done without having to worry to much about warping. 

 

You could fuse multiple pieces of plastic together, and in fact this is done for the Steyr AUG stock.  But I'm not sure how strong two pieces of plastic welded together are at the seam.

 

Polymer molecules are these long, noodly things comprised of a bunch of monomers linked end to end to end.  Prior to injection molding, the plastic gets churned up in a barrel with a helical screw.  Imagine a big bowl of spaghetti that gets stirred up and then left out to cool and then dry out.

 

I'm not sure how well plastic welding techniques stir the polymer molecules from the two parts together.  That might mean you end up with something like two bowls of spaghetti, and one was poured on top of the other, and then left out to cool and then dry out.

[Sturgeon]

[Sturgeon]

Bolt group is complete:

 

 

 

[Toxn]

9 hours ago, Collimatrix said:

 

You could fuse multiple pieces of plastic together, and in fact this is done for the Steyr AUG stock.  But I'm not sure how strong two pieces of plastic welded together are at the seam.

 

Polymer molecules are these long, noodly things comprised of a bunch of monomers linked end to end to end.  Prior to injection molding, the plastic gets churned up in a barrel with a helical screw.  Imagine a big bowl of spaghetti that gets stirred up and then left out to cool and then dry out.

 

I'm not sure how well plastic welding techniques stir the polymer molecules from the two parts together.  That might mean you end up with something like two bowls of spaghetti, and one was poured on top of the other, and then left out to cool and then dry out.

https://en.wikipedia.org/wiki/Plastic_welding

[Xoon]

19 hours ago, Collimatrix said:

 

That handguard is rather long.  It could be made of injection-molded plastic, but injection molding something that long while still holding good dimensional tolerances is challenging.  The mold halves have to close exactly on the part.  Those mold halves would be something on the order of 300mm long.  At that length, slight differences in temperature between the mold halves are enough to cause distortion and misalignment due to differential thermal expansion.

 

I think there are plastics that could be rigid enough and stable enough to be used in a handguard.  PEEK, for example, has excellent mechanical properties and a melting point that's 100 degrees or so higher than nylon 6.  A polysulfone might work, as might PPS.  But these are fairly expensive plastics that have additional challenges in processing that would slow down production.

Coming back to this, wouldn't it be possible to monitor the temperature of the mold and keep it heated at a constant temperature? 

Can't simple thicker plastic be used to make it more rigid? 

 

18 hours ago, Collimatrix said:

 

There are plastic handguards, they are almost all injection molded.  Injection molding has potentially amazing cycle times.  We're talking less than a minute per part in a process that can be run nearly continuously.

3D printed plastic parts are weaker than injection molded ones, even if they're made from the same plastic.  3D printing is also a lot slower.  It's really not cost-competitive for mass-manufacture.

I was more thinking of low volume runs or something like that.  3D printing has come a long way, I remember a 3D printer that uses light to harden a resin into a shape, which is then cured, this avoids the layered approach which makes the plastic much stronger. 

 

18 hours ago, Sturgeon said:

 

Because allen screws suck dick.

I like hex keys, probably because every fastener in my workplace is either hex key or a bolt. Though, I think the torx is superior, except in price. 

 

17 hours ago, Lord_James said:

 

This might be stupid, but couldn’t you make that 300mm long handguard in 3x 100mm long pieces and fuse them together? It might not look pretty, but it should get the job done without having to worry to much about warping. 

You could make it a little thicker than needed, and smooth out the welds, to remove the weld mark. 

 

15 hours ago, Collimatrix said:

 

You could fuse multiple pieces of plastic together, and in fact this is done for the Steyr AUG stock.  But I'm not sure how strong two pieces of plastic welded together are at the seam.

 

Polymer molecules are these long, noodly things comprised of a bunch of monomers linked end to end to end.  Prior to injection molding, the plastic gets churned up in a barrel with a helical screw.  Imagine a big bowl of spaghetti that gets stirred up and then left out to cool and then dry out.

 

I'm not sure how well plastic welding techniques stir the polymer molecules from the two parts together.  That might mean you end up with something like two bowls of spaghetti, and one was poured on top of the other, and then left out to cool and then dry out.

Wouldn't stir welding accomplish this? 

 

13 hours ago, Sturgeon said:

The rifle is looking nice. 

[Xlucine]

53 minutes ago, Xoon said:

I was more thinking of low volume runs or something like that.  3D printing has come a long way, I remember a 3D printer that uses light to harden a resin into a shape, which is then cured, this avoids the layered approach which makes the plastic much stronger. 

 

Stereolithography is great if UV-cured resins are good enough for what you're doing, which is rare (short of prototypes that you're testing for fit&feel). The final products tend to embrittle over time (especially if exposed to sunlight, as the UV continues crosslinking), which is not ideal for a firearm. It would be an elegant solution to getting a long handguard, as the final part can be much bigger than the original tank (if you're using the bottom-up build process, with UV projected on the bottom of the resin tank)

[Xoon]

1 hour ago, Xlucine said:

 

Stereolithography is great if UV-cured resins are good enough for what you're doing, which is rare (short of prototypes that you're testing for fit&feel). The final products tend to embrittle over time (especially if exposed to sunlight, as the UV continues crosslinking), which is not ideal for a firearm. It would be an elegant solution to getting a long handguard, as the final part can be much bigger than the original tank (if you're using the bottom-up build process, with UV projected on the bottom of the resin tank)

Good point. 

 

Though, it made me think, what about 3D printed aluminum? 
You could accomplish new types of geometry with it.
Crazy shapes like these which are pretty much impossible to cast and very expensive to machine:

 

Another thing would be carbon fiber construction. 

[Sturgeon]

1 minute ago, Xoon said:

 

Though, it made me think, what about 3D printed aluminum? 

Not anywhere close to as strong as forged or extruded aluminum, and the cycle times would be ridiculous. Look at how fast extrusions can be made (starting at about 1:00):

 

 

That would produce almost fully formed receivers that just need to be cut to length, sorted for straightness, and very minor finishing operations (like drilling holes).

 

[Xoon]

3 minutes ago, Sturgeon said:

Not anywhere close to as strong as forged or extruded aluminum, and the cycle times would be ridiculous. Look at how fast extrusions can be made (starting at about 1:00):

 

 

That would produce almost fully formed receivers that just need to be cut to length, sorted for straightness, and very minor finishing operations (like drilling holes).

 

How long does it take to mill the rest?

[Sturgeon]

3 minutes ago, Xoon said:

How long does it take to mill the rest?

 

It depends on your design, but in the case of the FN SCAR, it doesn't take hardly any milling:
 

 

The biggest bit of milling is for the pic rail slots.

[Sturgeon]

For comparison, here's how long it takes to print an engine block from plastic:

 

 

 

If you imagine that makes, oh, let's say 20 receivers in a go, then you're talking about over four hours of cycle time per machine per unfinished receiver. That's way too long.

[Xoon]

20 minutes ago, Sturgeon said:

For comparison, here's how long it takes to print an engine block from plastic:

 

 

 

If you imagine that makes, oh, let's say 20 receivers in a go, then you're talking about over four hours of cycle time per machine per unfinished receiver. That's way too long.

I am not trying to dispute that 3D printers are slow, I am just brainstorming. In theory, you could mill your handguard in 5 seconds if you wanted. 

[Sturgeon]

16 minutes ago, Xoon said:

I am not trying to dispute that 3D printers are slow, I am just brainstorming. In theory, you could mill your handguard in 5 seconds if you wanted. 

 

I think 5-axis machines enable you to do much of what you could do with a 3D printer, as far as small arms applications go. Now for aerospace, that's a very different story.

[Collimatrix]

3 hours ago, Xoon said:

Coming back to this, wouldn't it be possible to monitor the temperature of the mold and keep it heated at a constant temperature? 

 

 

Not really.

 

The molten plastic is being shot into the mold, and since we're talking thermoplastics, the mold is what is removing the heat from the plastic via conduction.  There are cooling lines running through the mold to remove this heat, but the more complex the mold geometry, the harder it is for the cooling lines to remove the heat evenly.

[Xoon]

32 minutes ago, Sturgeon said:

 

I think 5-axis machines enable you to do much of what you could do with a 3D printer, as far as small arms applications go. Now for aerospace, that's a very different story.

Slapping a mill bit on a industrial robot? 

 

25 minutes ago, Collimatrix said:

 

Not really.

 

The molten plastic is being shot into the mold, and since we're talking thermoplastics, the mold is what is removing the heat from the plastic via conduction.  There are cooling lines running through the mold to remove this heat, but the more complex the mold geometry, the harder it is for the cooling lines to remove the heat evenly.

Hmm, thats too bad. 

[Xlucine]

3 hours ago, Xoon said:

Good point. 

 

Though, it made me think, what about 3D printed aluminum? 
You could accomplish new types of geometry with it.
Crazy shapes like these which are pretty much impossible to cast and very expensive to machine:

 

Another thing would be carbon fiber construction. 

 

Materially it's one big weld (thinking of SLM, which is laser welding aluminium powder into the desired shape - in an inert atmosphere ofc). Not great for fatigue (e.g. you're very likely to end up with porosity), which might not be ideal for a receiver (I don't want to say it can't be done, but it'd need testing before saying it can be done). If this was really getting made then it'd be a much better option for the first half dozen prototypes, before the geometry was finalised enough for the big spend on the die for the extrusion.

 

Upper receivers are hollow bodies of almost constant cross section by definition (as they're enclosing the reciprocating moving parts), so it's a natural match for extrusion for full production

[OnlySlightlyCrazy]

I promised I would provide feedback at some point, but my brain is slow and timid. My chief thoughts are as follows.

1. The arrangement of the bolt is one of the most peculiar aspects of the weapon. Your design certainly appears supported by sound technical reasoning regarding interplay with the rounds in the magazine. My concern is that a reduction in the number of bolt lugs will correspondingly degrade the precision of the weapon. Of course, this is in the context of military ammunition, and yet more worrisome,  military shooters. Further, it is certainly reasonable to accept some loss of precision to gain increase reliability of the weapon during feeding, given your starting premise. I would, in short, like to see the data.

 

2. The retention of the AR-15 FCG components is an interesting move.  I always thought it odd that fewer companies exploited reusing existing components in this fashion. If I'm understanding correctly, it would also be easy to move towards a "cassette" style of trigger pack, which is of growing popularity with the AR-15 market. 

3. Your decision to devise your own magazine is admirable, and you certainly seem to have succeeded. My friend is curious if you can handle longer ogive, possibly ~66gr range EPR style 5.56 projectiles in the magazines. I'm not sold on drums, as their weight and bulk are disconcerting, but the ability to have them is never a bad thing.

4. I have no idea how you're going to solve the action-spring issue. I'm eager to see to what degree you can balance the forces acting on the bolt. Given you're trying to be unique, I wonder to what degree you'll end up matching the SCAR or other AR-18 descendants. Given the shape of the space you're working with, I do have my predicitions. (They mostly involve it being real goose hours, so I shan't share.)

[Sturgeon]

29 minutes ago, OnlySlightlyCrazy said:

My concern is that a reduction in the number of bolt lugs will correspondingly degrade the precision of the weapon.

 

wat

 

29 minutes ago, OnlySlightlyCrazy said:

My friend

 

lol

 

30 minutes ago, OnlySlightlyCrazy said:

curious if you can handle longer ogive, possibly ~66gr range EPR style 5.56 projectiles in the magazines.

 

Realistically, it's designed for 2.26" OAL rounds. However, the action can handle rounds much, much longer as its overtravel is significant.

 

31 minutes ago, OnlySlightlyCrazy said:

I have no idea how you're going to solve the action-spring issue.

 

Come again?

 

[OnlySlightlyCrazy]

Huh. Yeah perhaps I shouldn't have written my thoughts right before going to bed.

 

8 hours ago, Sturgeon said:

 

wat

My understanding of the appeal of multi-lug rotating bolt systems was that they offer a much more consistent lockup due primarily to the number of lugs involved, so that in theory an AR-10 would have a more consistent lockup than a Garand, or possibly in some odd world we could compare the descendants of those weapons. Thus, to my inexperienced eye, moving to fewer larger locking lugs seems like it could introduce inconsistency into the lockup, leading to a reduction in the precision of the weapon.

 

 

8 hours ago, Sturgeon said:

Realistically, it's designed for 2.26" OAL rounds. However, the action can handle rounds much, much longer as its overtravel is significant.

 

That would be a limitation of the magazine, correct? 

 

 

8 hours ago, Sturgeon said:

Come again?

The SCAR has it's action spring (or maybe it's called the return spring? I don't study bad rifles other than the AR-15) hanging out behind the bolt carrier group and occupying that space in the receiver, with a single spring as shown. I'm curious how you will design your action spring, whether you'll use a single spring in line with the bolt ala SCAR, an AR-18 two-spring assembly, or something novel. 

 

[Sturgeon]

3 hours ago, OnlySlightlyCrazy said:

My understanding of the appeal of multi-lug rotating bolt systems was that they offer a much more consistent lockup due primarily to the number of lugs involved, so that in theory an AR-10 would have a more consistent lockup than a Garand, or possibly in some odd world we could compare the descendants of those weapons. Thus, to my inexperienced eye, moving to fewer larger locking lugs seems like it could introduce inconsistency into the lockup, leading to a reduction in the precision of the weapon.

 

Oh I see what you are talking about. So my understanding of that is that it has less to do with how many lugs and more to do with how those lugs are spaced. So you want lugs that cover both axes rather than just two opposed lugs. I have four which more or less cover the two axes, and that arrangement was designed to do that (among other things - it was a compromise of course).

 

3 hours ago, OnlySlightlyCrazy said:

That would be a limitation of the magazine, correct? 

 

Yes the magazine is designed for a 2.26" OAL plus 0.065" of tolerance. This is similar tolerance to an AK-74 magazine, and does mean you could stuff slightly longer rounds (as long as 2.32") into the mag, potentially. However while this should be fine for the casual handloader, I wouldn't recommend doing it for a full military small arms system. Easy enough to just lengthen the mag, instead. 

 

3 hours ago, OnlySlightlyCrazy said:

The SCAR has it's action spring (or maybe it's called the return spring? I don't study bad rifles other than the AR-15) hanging out behind the bolt carrier group and occupying that space in the receiver, with a single spring as shown. I'm curious how you will design your action spring, whether you'll use a single spring in line with the bolt ala SCAR, an AR-18 two-spring assembly, or something novel. 

 

Currently the design is intended to use one action spring, nested in the bolt carrier coaxial to the piston, but I may have to move it slightly lower.

[OnlySlightlyCrazy]

1 minute ago, Sturgeon said:

So you want lugs that cover both axes rather than just two opposed lugs.

I'm rather perplexed that I hadn't noticed that before. Ah well, it is a good day to learn!

[Oedipus Wreckx-n-Effect]

Updates! Feed us, leader!

[Sturgeon]

I am away from my drafting station for a week or so, but after that I'll get back to it.

[Oedipus Wreckx-n-Effect]

9 hours ago, Sturgeon said:

I am away from my drafting station for a week or so, but after that I'll get back to it.

 

Quitter.