Category Archives: GunTech

More on Whitworth and his Rifle

Yesterday’s post on Civil War Sharpshooters was meant to be a shallow overview, but one thing leads to another, and so this morning’s scheduled post was thrown back into the sea so we can have some further discussion of this subject, and especially of the English rifle used to such great effect by the Confederates, the Whitworth.

Fred Ray, whom we cited in that post, commented and included a couple of links that may be of interest. First, he has traced the English word “sharpshooter” back to 1795.

I think the question is settled, at least as far as any reference to Christian Sharps is concerned.

Also check out my articles on the web site for my book, Shock Troops of the Confederacy, which deals with Confederate sharpshooters. There’s much more to them than Pat Cleburne! Couple of articles there on the killing of Gen. John ” Couldn’t hit an elephant at this distance” Sedgwick and one on Joseph Whitworth and his rifles.

Fred’s Whitworth article is this one. It’s well researched and was a good read.

Among the things we didn’t know about Whitworth were that he invented scraping for a flat surface; invented an early caliber, and was instrumental in standardization of threads and fasteners in Britain.

Because one good article deserves another, here’s the section on the Whitworth from Fuller & Steuart’s Firearms of the Confederacy. The Whitworth was not the only English rifle used by sharpshooters in gray, as they also made good use of Enfield and Kerr rifles, but the Whitworth is the one that has captured the imagination of historians, collectors and reenactors.

This rifle is of particular interest to students of Confederate arms, as it is believed to be the only one of the imported arms that was used exclusively by the South who used them in small numbers for arming sharpshooters. They were an accurate and powerful weapon — good for a range of half a mile and were responsible for the taking off of many a federal officer.


Note: all of these illustrations are pulled from the web, not from the book quoted. This Confederate Whitworth has a replaced lock, but is for sale for a LOT of money.


The specimen shown is marked on the lock plate Whitworth Rifle Company, Manchester and on top of the barrel Whitworth Patent.

This is a Whitworth lock; it’s not from a sharpshooter’s rifle, but from a target rifle sold by James Julia. Hammer price was $10,300 in 2014,

This is the barrel marking referenced in the text, also from the Julia target rifle.

Length of barrel, 33 inches. Total length of arm, 49 inches. The bore is hexagonal. Caliber .45, using an elongated bullet weighing 530 grains.  the twist is one in twenty inches. The arm is an exceptionally well-made piece — iron mounted throughout and besides the regular site equipment, is provided with attachments for a telescope site to be mounted on the left side of the gun. The stock is nicely checkered and the arm has all of the characteristics of the highest type sporting piece. All parts bear the serial number 554.

From the Julia target gun, period “globe” or aperture sights.

In the year 1852 when the British ordnance department conducted extensive experiments to test the comparative merits of various rifles submitted to the government they found a wide variation in the accuracy updatable. Whitworth, one of the leading technicians of the day, was commissioned to make exhaustive experiments at the cost of the Government in order to discover the best form of rifling.

This gentleman had devoted a great deal of time and study to the design and manufacture of cannon and had adopted the polygonal bore as giving the best results and decided to use this type of rifling for his small arms. The advantage of the elongated bullet had long been demonstrated but in attempting to use it in connection with the polygonal bore considerable trouble was experienced from the ball “capsizing” or “turning over”.  He became convinced that this action was due to the slow spiral and eventually after testing every graduation from one turn in seventy-eight inches to one turn in five inches found the necessary rotation to impart the required steadiness to the ball and cause it to maintain a flight parallel to its axis was best obtained at a pitch of one turn in twenty inches.

On tests before the Minister of War and many distinguished officers the Whitworth rifle of .45 caliber beat the Enfield of government factories by three to one. The mean deviation at 500 yards was four and one-half while the recorded best of any rifle previously tried was twenty-seven.

The rifle was never adopted into the Government service but 40 of them were made for the competitive shoot of 1860 for the Queen’s prize at the meeting of the [British] National Rifle Association. Plate XXIII shows an enlarged view of the bore of this arm and the machine made bullet used with it.

While the original bullet for the Whitworth rifle was hexagonal to fit the rifle bore, those used by the Confederates were for the most part cylindrical.

This is characteristic Whitworth hexagonal rifling. Also from the Julia gun.

He notes that “20 or 30” of these rifles were slipped through the blockade, and divided equally between Eastern and Western rebel forces, but Bilby thinks that a much greater number of the Whitworths must have been on hand — probably hundreds.

Fuller & Steuart also reproduce this period article:

The Richmond Daily Examiner of November 10, 1863, says:

We have a wonderful gun in our army, the Whitworth rifle. it kills it 2000 yards, more than a mile. It is no bigger than the Mississippi rifle. [US Rifle M1841 -Ed.] With a few of these rifles Longstreet shot across the Tennessee River, killing the Yankees and completely blocking the river road.

They go on to reproduce some combat tales of the rifle in action.

Sergeant Grace of the Fourth Georgia killed General Sedgwick of the Union Army with a Whitworth rifle at a range of 800 yards.

Sergeant Grace used a globe sight. Most of the Whitworths were equipped with telescope sights, but these were easily lost.

Whitworth rifles are said to have done terrible execution at Fort Wagner, Charleston.

General Cleburne, writing in 1863, said: “The fire of five Whitworth rifles appeared to do good service. Mounted men were struck it distances ranging from 700 to 1300 yards.”

Twenty men of Company F, Eighth North Carolina Regiment, were armed with Whitworth rifles with globe sights at Morris Island. South Carolina sharpshooters also had Whitworths and General Lytle is said to have been killed at Chickamauga with a bullet from a Whitworth rifle.

This is one type of False Muzzle, from another target gun (this one a Maine gun for sale by Joe Salter). It was used to ensure the bullet was started right, aligned with the bore. A False Muzzle was usually part of a target gun’s standard accessoried.

One of the great Whitworth mysteries remaining is: why did the Union never buy any? Both armies bought plenty of Enfields from Britain. But if the Union bought even a single Whitworth, no trace of the transaction has been found.


We’d be remiss if we didn’t point out Fred Ray’s book on Confederate States sharpshooters. You can probably find it at your favorite online bookstore, but the promotional website has more information…

…and you can buy it there. (We did, hard and soft cover editions).  There seems to be more useful information on the website in terms of an errata page, and more of Fred’s articles, that extend his work, and are free to read for those of who who can’t explain to Mama $30 on another book. (“What’s wrong with reading the books you already have?” <– one downside to making an important choice based on pulchritude alone, in the bloom of youth).

Update II

We let this go live without images, in the interests of speed, but Holy-Wall-o-Text, Batman, so we’ve added some illustrations to ease the barrage on your eyes. Apologies to early readers.

Homesmithing News: Printed Revolver, Easy Home Rifling

Here’s the Imura Works’ latest Zig-Zag Revolver, the Flying Swallow. Imura Works is named for the Japanese martyr to the cause of home gunsmithing.

It has quick-change cylinders for rapid reloads.

Of course, this wouldn’t be that interesting if files weren’t available. But files are available — how ’bout that. The file pack includes not only the .stls that you need to print it out, but also assembly drawings that show how it all goes together.

Happy printing!

Beyond Printing: The Steel Rifled Barrel

Of course, a printed pistol has its limits; we’ve seen a lot of better examples of homemade guns made from steel. In an austere or denied environment, raw steel will always be available, but one stumbling block many small and home machinists have encountered has been rifling a barrel. But it turns out, there is a high-throughput, benchtop way to do it: electro-chemical machining (ECM). You don’t need a big ECM machine: you can improvise with innocuous parts and chemicals (table salt!). This is the whole setup:

Yep, the el cheapo battery charger is all the power supply you need. They used 33 AWG copper wire.

Here’s some results from an early series of experiments.

HERE it is! Rifling a barrel using the ECM (electrochemical machining) process by the one and only Jeffrod. This fosscad project is still in its early infancy so expect more to come! ECM is like reverse-electroplating where you are removing material instead of adding it. ECM is a very precise method and is more suitable for mass production. The ECM process can be used on hard materials that cannot be machined by other more traditional processes. Unlike the EDM (Electrical discharge machining) process, no sparks are generated with ECM between the cathode and anode.

That particular barrel used a pentagonal 3D-printed mandrel, copper wires, and three rounds of 5 minutes in a saline solution with 6 volts and 6 amps running through the wires. The rifling is 5-thousandths deep. It’s not a target barrel, but it’s a process that will produce a legal and likely functioning rifled barrel. And you can experiment with it with just some steel tubing, a bucket, wires and a mandrel. This is the mandrel, with the copper machining wires attached.

The whole process is recounted on this imgur thread, including several earlier experiments before this last barrel was produced. These barrels are a proof of concept, not in any particular caliber or chambering at this time. So we’re a long way from making, say, AR barrels, but a Sten barrel is close.

 You can’t ban guns. Make one problem difficult, and the community reacts to the damage by detouring around it. Short of lobotomizing the whole community, this is only going to grow from here.

Have You Gymnasticated Your Gun Today?

You can, perhaps, Gymnasticating the M1 Garand via the op-rod and follower rod. Image via CMP.

Wait, what? “Gymnasticate?” What’s that?

Well, that was exactly our reaction to seeing this word in a very interesting book, Roy F. Dunlap’s Ordnance Went Up Front. Dunlap is an interesting character, about whom little is known apart from his two books — this history of his work as a small arms maintenance NCO in World War II, and a comprehensive gunsmithing manual that remains in print. Rereading Ordnance recently, these lines jumped out at us, right off of Page 94.

In the spring of 1943 an officer approached me with the idea of finding out what this M38 [German Anti-Tank rifle, per Dunlap – Ed.] could do, as he had a gun in perfect condition. I scratched my head, gymnasticated the rifle, trying to look intelligent, and finally gave my opinion it would penetrate 1/2 inch armor at 100 yards, but not much more.

Now, Dunlap went on to discover that the gun had considerably better performance than that, but what struck us funny was the word, “gymnasticated.” And Dunlap must have expected it to, for he provided a footnote:

That word “gymnasticate” may have a few of you on the ropes, but is simply an ordnance term meaning the artificial operation of the recoil mechanism of a weapon. Usually it is applied only to artillery, but is perfectly proper for any weapon operated by or having a recoil system. When you push back on the barrel of an auto loading shotgun or a Colt.45 pistol, you are gymnasticating the arm.

Auto loading shotgun? Remember, he wrote this book in 1948; in this passage he’s talking about experiences he had in North Africa in 1942 (he would later go to the Pacific Theater, putting him in position to be able to make comparisons of Allied and Axis weapons from most nations, although he only encountered those Russian small arms that were recycled for use by Germany early in the war). The Remington 58, first ancestor of the 1100, had yet to demonstrate that an effective gas-operated shotgun was a possibility. Almost all semiautomatic shotguns of the time were based on Browning long-recoil designs, and the conventional wisdom was that a gas system would not work with low-pressure and highly variable shotgun loads.

But seriously, has anyone encountered this term, “gymnasticate,” before? We haven’t seen it, but we just have books on artillery, which is not like having experience on artillery. We know we have cannon cockers in the audience. What say ye?

Meanwhile, let’s make 2017 the year in which we spread the word “gymnasticate” far and wide, in memory of Roy Dunlap, who sharpens swords in Valhalla lo these many years. And let’s all gymnasticate a gun today. The guns seem to enjoy it.

The KM17s Follows the Leader

Australia, which now prides itself on being semi-auto-rifle and homicide-free, once not only designed and manufactured such rifles, but exported them. The Australian Leader rifle in 5.56 never won any contracts we’re aware of. But a quantity of them sold here, and they have won some converts. It shares some mechanical features with the Armalite AR-18 series, and seems like it could work well.

Ian did a very good video on the Leader rifle at Forgotten Weapons.

Various versions including 7.62mm and bullpup versions were tried, and a version of the bullpup went into production by Bushmaster, ending when Cerberus yanked Bushmaster’s chain. A review in a recent American Rifleman brought to our attention the re-launch of a much improved version of the bullpup as the KM17S556. Like Charles St. George of Leader nearly 40 years ago, the head of K&M Arms, Ken McAllister, envisions a whole line of these rifles; success of the 5.56 model that has been making the review circuit will enable all that.

Bullpups have been in the news a lot, thanks largely to the Israeli adoption of the modern Tavor, which has also been an export success for Israel Weapons Industries, with about 20,000 Tavors a year finding new homes in America, despite a boycott maintained by dozens of smelly campus hippies as part of the Boycott, Divestment, Stink movement. Perhaps due to rapid modification of the issue gun and the slower pace of ATF approvals, the design of the civil and military Tavors has diverged, but the rifle remains popular despite a high price. (Almost $2k before optics). Other bullpups come and go, including AUGs, MSARs (an American AUG knock-off), and two models from Kel-Tec; the compactness of the bullpup system always intrigues buyers. The original AUG and a Croatian design, as well as FN’s 5.56 and PDW bullpups, continue in military service here and there, even though Steyr seems to have thrown in the towel on bullpup designs, and is returning to a conventional layout.

The Leader of 1978 is unrecognizable in the KM17S, apart from its internals. It has a three-lug bolt, a gas tappet design (like Tokarev, Simonov, and Saive FN designs) and is designed to quickly field-strip without tools into three large assemblies.

(NRA photo)

The Leader and all its successors are designed for easy manufacture. On the Leader, the upper receiver was pressed steel. On its bullpup derivations, it’s a machined aluminum extrusion — very fast and inexpensive to produce, in quantity.

Its path from Australia to Maine to Arizona has been long and has seen many minds and hands work on it; if McAllister can produce and distribute them to the legions of bullpup fans out there, the Leader may have a new lease on life. If so, it will be an overnight success — a generation in the making.

The original inventor of the Leader was an Aussie named Charles St. George. He developed the Leader T2 in the 1970s and modified it many times with the hopes of selling it to the Australian defence (as they spell it!) forces. Many improvements were developed for the military M18, such as ambidextrous dual non-reciprocating charging handles, and persist in its descendants today. The military bought the AUG instead — Australia still manufactures a version locally —  and St. George’s bullpup version, which he called the T18, never was produced as such in quantity. Instead, he licensed it to Bushmaster, which produced it in small quantity from about 1992 to 2005. At some point in time, Leader went paws up and a small quantity of continuation T2s were made by a firm called Australian Automatic Arms; at least some of these had wood stocks in a “postban configuration” per the US 1994 gun ban law, as seen here on Forgotten Weapons. Per Ian, importation to the USA of the Leader was only about 2,000 units.

Meanwhile, McAllister had a machine shop in Chandler, Arizona, called K&M Aerospace. Apart from the custom aerospace and automotive manufacturing he did, he began to make parts to customize and improve bullpup rifles after getting hooked on the platform by a Bushmaster M17S that he picked up. His upgrades included trigger improvements, weight reductions, rails systems, even cheekpieces, and he began to manufacture replacement parts for the M17 community. From there, it seemed a short step to manufacture, although the KM17S’s path from decision to market — if it is, really, indeed on the market now that reviews are showing up — was over five years long.

The bullpup design appeals to owners who would like to have a modern rifle with a different style or flare. It has proven to be a very hard sell to armies, but there’s certainly demand for bullpup rifles out there, and K&M’s design is unique in the market in not being largely polymer. (It’s mostly steel and aluminum). The NRA found it shot 2-inch groups.

For More Information:

When The Army Resisted the M16A2, Part 3 of 3

The previous two stories set the stage, for a look at a report drafted for the Army Research Institute for the Behavioral and Social Sciences the Army was still pursuing the “best” (an upgraded M16 meeting all Army objectives) instead of the “good” (the M16A2, which was developed and revised to meet Marine objectives). Of course, we all know the spoiler aleady: the Army accepted the Marine M16A2 as is, leaving the report as an orphaned artifact. The report is here:

Colt factory shot of the M16A2. The A2 was developed by the USMC, but was manufactured by Colt and FNMI.

This is the third of a three part series. In the first part, Thursday on, the Army contractors noted the specific solutions implemented on the A2 and the problems the Marines solved thereby, but complained that the problems and solutions were too USMC-specific. In the second part, posted yesterday, we discussed just what they thought was wrong about the Marines’ product. In this, third, part, we’ll list the modifications that they suggested in lieu of or in addition to the A2 mods.

Most of the Army’s problems with the A2 related to the burst mechanism, and the sights, especially the complicated rear sight. (This is actually an A3/A4 or M4: note the knobs, left, for removing the carrying handle. The A2 handle was forged as part of the upper receiver.


We should note that the Marines’ tests, as reported in this document (p,7), demonstrated significantly lower reliability, and increased fouling in the A2 compared to its older brother. These tests are suspect because the early lot of XM855 used was considered bad ammo, but the M16A1 did outperform the A2.

Thirty Ml6A1 rifles firing 26,010 rounds of M193

Failures to fire – none
Failures to feed – 3 (Not locking magazine in place)

Thirty M16A2 rifles firing 26,010 rounds of XM855

Failures to fire – 52 (27 – bad ammunition) (25 – mechnanical [sic] malfunctions)
Failures to feed – 3 (Improperly loaded magazines)

Those failures to fire that were not attributed to bad ammo were thought to be caused by the A2 trigger system’s Achilles’s heel, the burst trigger mechanism. The A2 performed even worse in a cold weather test, but again, it was with the questionable ammunition, and many of the failures to fire were also laid at the feet of the burst mechanism.

The report has an interesting discussion of the burst mechanism and its rationale in Marine, but not Army, small arms doctrine:

The M16A2 has less combat capability due to the elimination of full automatic fire. Full automatic fire enhances the ability of Army units to clear and defend buildings, to conduct final assaults on enemy positions, to defend against an enemy final assault, to conduct an ambush, to react to an enemy ambush, to engage an enemy helicopter or fast moving vehicle, etc.

While the Marines claim greater accuracy and conservation of ammunition for the 3-round burst control, no data were generated during the test to support these contentions and no supportative [sic] data are known to exist.

Also, it should be noted that room-to-room fighting was conducted with blanks, no close-in firing was conducted, no firing with short time limits was conducted, no firing at aircraft was conducted, etc. In other words, for all of the automatic/burst firing conducted during the test, a semi-automatic mode of fire would have probably resulted in a greater number of target hits.

Finally, to be given very serious consideration, is the fact that the burst control requires nine (9) new parts in the lower receiver, evidently contributing to the large number of weapon malfunctions during testing of the M16A2.

They also took issue with the heavy barrel (“heavy in the wrong place”), the twist rate (preferred 1:9), stock length increased when even the A1 stock was too long for small soldiers, and the fast twist’s incompatibility with the .22 subcaliber system. 

The article includes an extensive comparison of the pros and cons of Marine KD vs. Army Trainfire marksmanship modalities. These training differences result from the different combat envelopes for the rifleman: the Marines need to engage with rifles in the 300-to-800 meter space, because they don’t have the supporting arms that the Army can count on, at least, not in the same quantity. A unit that must fight with just its organic weapons needs to get the very most out of these weapons. The Army of 1986 did not consider a 500 or 600 meter target a primary rifle target, but a crew-served-weapons target.

In the end, the recommendations the contractors made were mostly about the sights. They put their recommendations in a table with the M16A1 and M16A2 stats. Since the latter are probably familiar to most readers, we omit them now to save time, and just show the contract recommendations.

Item Recommended
Front sight (day) Fixed blade, 0.090″
Front sight (night) Luminous dot on each sightguard
Rear Sight (day) single 2mm peep. A single elevation knob marked for 200, 250, 390, 25, 400, 500, 15, 600, 700, and 800 meters. Windage knob at rear. Each click equal to 1 MOA
Rear Sight (night) Two luminous dots on upper portion of receiver (or a single flip- up luminous dot located forward of the carrying handle) are aligned with front dots for shooting at night
Zero Recording Yes
Zero Inspection Yes
25m setting (day and night sights) Yes
Mechanical Zero Yes
250-m battlesight Yes
Firing mode Semi and Auto
Barrel 20″. Slightly heavier than A1 at receiver and mid-barrel. 1:9″ twist
Handguard Same as M16A2 except held in place with a securely fastened ring nut to provide rigidity.
Buttstock Same material as M16A2. Same length as M16A1. Option for adjustable length.

There are several interesting observations to make here. First, the contractors recommended that the Army make changes that would decrease the mechanical accuracy of the proposed M16Ax relative to the Marines’ A2. Specifically, these changes included the wider fixed front sight blade, the 1-MOA adjustments on the rear sight (A2 offers ½-MOA), and arguably the simplification of the rear sight. The trade-off was simplicity and ease of training, instead of superior bullseye performance.

Second, some of the proposals would definitely improve the utility of the firearm, including restoring the short stock, or replacing it with an adjustable one; increasing the barrel diameter towards the chamber rather than the muzzle, thus improving sustained fire accuracy and reliability; reverting to automatic fire from the burst mechanism (which also has side benefits, in improving the trigger’s feel and consistency). The night-sight proposal was truly ingenious.

Third, in some of these road-not-taken proposals, the Army was reverting to the original AR-10 design and rejecting changes that were largely imposed on the AR design by the Army in the previous decade. These include the rigid fastening of the handguard, and the fixed front sight blade.

Finally, these proposals were almost the last gasp of the iron-sighted military rifle. As this  document passed from the contracting officer to file cabinets across the service, without action, special operators were already wringing out scopes and single-point sights, and a few visionaries were already arguing that the day of the iron sight had run its three centuries, and was now at an end. A new generation of optical technology was eliminating the two objections that had kept optics off the rifles of most soldiers: less durability than irons, and slower target acquisition. Many men’s efforts went into winning over the Voices of Experience who still said “no” to anything with a lens, thanks to memories of Uncle Joe’s elk lost because his scope fogged up, or the VC that got away because somebody attached an unauthorized 4×32 Colt scope to the carrying handle of his M16.

How Many Johnsons Does One Man Need?

New Market Arms has a range of Johnson M1941 rifles including a rare tolroom prototype. A couple of them are for sale on GunBroker — thanks to commenter Josey Wales for tipping us off — and more are on the website. We want them all, but we’ve already got a couple of Johnsons, and how many does one need?

The right answer, of course, is “need doesn’t enter into it” with these rare and historic firearms.

The rarest of them is this tool room prototype, numbered S-3. (And see the GunBroker auction here).

This is a very rare, one of only six manufactured, Johnson Automatics Model 1941 Tool Room Sample Rifle in .30-06, that is still in its military configuration.

An amazing set of almost 200 photos of this rifle is available. It includes comparison between this pre-production and production Johnsons that are of great interest to all Johnson students and collectors.

An agreement was ultimately reached between Johnson Automatics and the Universal Windings Company that resulted in the establishment of the Cranston Arms Company of Cranston, Rhode Island, which would produce Johnson’s Model 1941 Rifle. Cranston Arms also produced Johnson’s Model of 1941 Light Machine Gun, which shared several design features with his semi-automatic rifle. Cranston Arms (a subsidiary of Universal Windings), Johnson Automatics, Inc. and Johnson Automatics Manufacturing Company (JAMCO) set up shot next to Universal Windings and produced its first assembly line rifle, serial number S-1 in April 1941. On April 19, 1941, the employees gathered at the factory rifle range for the first firing of serial number S-1. As Bruce Canfield states in his book, “Maynard Johnson picked up the first rifle off the line and carefully loaded it. He took aim at the target located at the other end of the 100-yard range, carefully squeezed the trigger and fired the rifle. To his astonishment, and the other witnesses’ shock and amazement, the rifle did not extract and eject the spent cartridge case and it failed to function as a semiautomatic. It is reported that Johnson nearly bit his frequently present cigar in half in irritation, frustration and rage.”

Canfield goes on to explain that the rifle could only be operated by direct manipulation of the bolt. The rifle was quickly unloaded and disassembled with all of the component parts compared against the blueprints. It was quickly discovered that Cranston Arms had failed to properly machine one of the bolt cams. The bolt was machined to specification, reinstalled into the rifle and the rifle was loaded for firing. This time Johnson’s rifle performed perfectly.

This particular rifle, Serial Number S-3, was probably manufactured the same day or within a day or two of rifle S-1 noted above. It is believed that only six of these tool room sample rifles with the “S” prefix were manufactured. The full production rifles were serial numbered in accordance with Dutch military policy, in serial numbered blocks of 9,999 rifles. The first block from serial number 0001-9999 did not have a prefix. The second block had prefix “A” and the third and final block of production had prefix “B.”

S-3 is interesting because, unlike the early R-models, it’s almost exactly like production guns. The differences are small and subtle and are detailed on the sales page,

As noted, this particular Johnson M1941 Rifle is serial number S-3 and is one of the initial tool room sample rifles manufactured by Cranston Arms. As can be clearly seen in the comparison photos, this tool room sample differed slightly from the final production model as Cranston and the Johnson team made some additional refinements prior to commencing main line production.

This rifle is in fine condition. The original barrel is in its original 22” military configuration with the sight protective ears and the bayonet lug. The front sight assembly is slightly different than production models with the machining and height of the front sight post slightly shorter on this tool room sample than the later production models. The front sight pins remain correctly staked in place and have never been removed. The barrel has 98% of its original finish. The barrel has strong rifling and a mirror bore so this will be an excellent shooting rifle. The muzzle gauges at approximately 0.5 and the muzzle crown remains perfect.

The Barrel Bushing has the correct “.30-’06” and “41” markings stamped on the face. The font is different, however, than later production rifles. Near the breech it has the “O [Gladius Sword] I” in a circle stamp found on all early Johnson 1941 Barrels. It is also correctly marked “J.A./30-06,” still crisply stamped. The Locking Bushing is in excellent condition with normal wear on the lugs. This bushing is also slightly different than production models in that the lugs here are more squared whereas the lugs on later production models were slightly rounded. The breech face remains in the white. The threads can be seen, which also differs from later production models. Both the Barrel and Locking Bushing have the matching assembly number “6664K” stamp. None of the M1941 Johnson Rifles were serial number matching so all of the Johnson 1941 Rifles will have different numbers on the bolt and barrel. The original Bayonet Lug is present. The first tool room rifle, serial number “S-1,” did not have the bayonet lug attached. It is unclear if serial number “S-2” had a bayonet lug so this could be the first of the Johnson rifles with this lug configuration.

The Receiver retains virtually all of its original parkerized finish. It has the correct “CRANSTON ARMS CO.” triangle on the right rear of the receiver. Significantly, it does not have the star stamp above the triangle. The star indicated original Dutch acceptance according to Bruce Canfield, prior to Japanese occupation of Dutch possessions in the Pacific and before these rifles were then offered to the Marine Corps. Since this was a tool room sample, it would not have received the Dutch acceptance star since it was never intended to be shipped to the Dutch. The Receiver markings are still very crisp on the top of the receiver. These markings include Johnson’s patent information, “JOHNSON AUTOMATICS” over “MODEL OF 1941,” and the manufacturer’s location, “MADE IN PROVIDENCE, R.I., U.S.” The font and size of the receiver markings are slightly different than later production models. Below that is the serial number “S-3.” All of the stampings remain crisp. The ventilated forward portion of the receiver, which becomes a ventilated top handguard, retains virtually all of its original parkerized finish.

The Rear Sight Assembly is in very fine condition and it also differs slightly from later main line production rifles. The Windage Knob is in fine condition and adjusts perfectly. The Aperture remains in its original military configuration. The Rear Sight Protective Wings retain virtually all of their original finish and differ slightly than later production models. The Rear Sight Elevator has a different font than later production models. The numbers on the right side still have the original paint that has yellowed slightly.

The original Firing Pin Stop Assembly is present and it retains all of its original blue finish. The original Firing Pin is present and it differs from later production models by slight machining differences towards the point. The Firing Pin retains virtually all of its original parkerized finish. The pin has the assembly number “J9303” on the side along with a “0” stamp towards the front. The Firing Pin Spring remains in the white.


The Bolt Catch Assembly is present and differs from later production models by the length of the machined channel.

This Johnson Model 1941 comes with a very rare and original Dutch Model leather sling. .

This Johnson Model 1941 Rifle also comes with an original and very rare Johnson Model 1941 Bayonet and leather scabbard.

This is an extraordinarily rare Johnson Model 1941 Tool Room Sample Rifle and is just the third of these pre-production rifles manufactured.

What is its provenance?

This rifle was purchased by an employee of Cranston Arms when the rifle was manufactured and was passed down through his family to his grandson. This rifle is in the exact condition it was when it was manufactured in 1941 and undoubtedly was fired by Maynard Johnson himself.

On the website, New Market is asking $16,000 for S-3; on GunBroker, bids in the $7,500 range did not meet reserve and it’s relisted. There is a buy-it-now of $

In addition to S-3, New Market has several more Johnsons for sale.

They do provide, with each listing, a very solid capsule history of the Johnson M1941, which appears to be a distillation of Bruce Canfield’s book (this is a good thing; we recently recommended it to specialists in a book-review roundup (Weaponsman Expert Book Reviews #5). (If you want early Johnson history beyond this, there’s more at the sales page for S-3, and of course the most comprehensive answer is the Canfield book).

The history of the Johnson 1941 Rifle, and its designer, is a very interesting one that began on the eve of WWII. The designer of the rifle was Melvin Maynard Johnson, Jr., who graduated from Harvard University and Harvard Law School. He was also an avid firearms enthusiast from a young age and, around the same time that he graduated from law school he was commissioned as a Second Lieutenant in the Marine Corps Reserve. Johnson took advantage of his association with the Marine Corps to pen various articles for the Marine Corps Gazette in the early 1930s. One of his articles was a general critique of the new M1 Garand Semi-Automatic Rifle. Johnson took issue in his article with the M1s “gas trap” design and its en block clip loading design, as well as several other issues with the rifle.

Melvin Johnson began an early relationship with the United Automatic Rifles Corporation in the early 1930s, initially in his role as an attorney, and began to provide the company with mechanical and engineering work on various rifle designs. The relationship did not survive but it solidified in Johnson the desire to experiment with and develop his own weapons designs. One of his first experiments, undoubtedly as a counter to the gas operated M1 Garand design, was in recoil operated automatic weapons design. Johnson eventually partnered with Marlin Firearms to build a semi-automatic rifle and light machine gun, both of which used a vertical feed design through the use of Browning Automatic Rifle 20-round magazines. The magazines developed several problems during tests at Fort Benning but Johnson was undeterred and continued developing both weapons. This led Johnson to begin work on a rotary magazine design.

In the late 1930s, Johnson founded Johnson Automatics, Incorporated, which would be the operating entity that would own the patent rights (and hopefully obtain manufacturing rights) for all of Johnson’s weapons designs. Johnson then began a determined effort to sell his designs, and his weapons, to the US Army and Marine Corps and various countries, including Great Britain and France. The rifle design that Johnson settled on, which he felt was superior to the M1 Garand rifle, was what became his Model 1941 Rifle. This rifle had a detachable barrel, a 10-round rotary magazine, and was recoil operated. The US Army, however, had settled on the M1 Garand rifle, which was then in production at Springfield Armory. Johnson continued to believe that his rifle design was just as good as John C. Garand’s design and he began to lobby members of Congress in an attempt to reopen the weapons tests that led to the adoption of the M1. Congress eventually held several rounds of hearings and, after an additional series of “head-to-head” tests, the M1 Garand was deemed superior by the Army. This left Melvin Johnson continuing to try and win a contract from the Marine Corps, which had not formally adopted a replacement for the 1903 Rifle, and from various foreign governments. Melvin Johnson had finalized his Model 1941 design by this point and now needed the assistance of an established manufacturing business to go forward with large-scale production.

An agreement was ultimately reached between Johnson Automatics and the Universal Windings Company that resulted in the establishment of the Cranston Arms Company of Cranston, Rhode Island, which would produce Johnson’s Model 1941 Rifle. Cranston Arms also produced Johnson’s Model of 1941 Light Machine Gun, which shared several design features with his semi-automatic rifle. Only a limited number of M1941 rifles had been shipped to the Dutch in the East Indies prior to the Japanese capture of Dutch possessions in the Pacific in early 1942. Some of these rifles were ultimately captured by the Japanese near the airfield at Tarikan and the port of Balikpapan in 1942. The rest were evacuated and used by the Free Dutch forces fighting in Timor through 1943. Some of the Dutch M1941 rifles were even used for a time by Australian forces fighting in Timor. The remaining M1941 rifles were then embargoed to keep them from being sent to the East Indies and possibly captured by the Japanese.

After the United States entered the war in December 1941, demand for military arms soared. By this time, the Marine Corps had followed the Army’s lead in adopting the M1 as its standard battle rifle, but M1 production was initially unable to meet demand. In addition, much to the Marine Corps’ chagrin, the Army had first priority on available supplies and on future output from Springfield Armory and Winchester, the two manufacturers of the M1 Rifle. As a result, the Johnson M1941 rifle was adopted by the Marines for issue to Marine Raiders and to newly-formed Para-Marine airborne units (because the barrel could be removed for ease of jumping the weapon), and these rifles saw action in the Solomons campaign of 1942. As M1s became available to Marine units, the Johnson rifles were withdrawn from combat use. Only a few thousand of these arms had been procured by the U.S. government before production ended in 1944, and, in addition to their limited use with the Marine Corps, some Johnson rifles were issued to clandestine O.S.S. operatives. Because the rifle was never officially adopted by the US military, and because WWII prevented any opportunity Johnson may have had for robust foreign sales, the total number of Johnson 1941 Rifles manufactured was very small, only about 30,000. Johnson Model 1941 Rifles were serial numbered in groups of 10,000, with the first 10,000 having no prefix, the second group with prefix “A,” third group with prefix “B.”

For the tool room prototype, he provides more history.

The prices seem high to us (one of our Johnsons cost us $4k, and one $700 — decades earlier, when a 1911 was $225) but New Market has sold Johnsons they were listing for $6,250 and $7,500 recently.

When the Army Resisted the M16A2, Part 2 of 3

The M16A2 was adopted by the Marines in 1983, and then by the Army in 1986. Shortly before its adoption, an Army contract analyzed the M16A2 — and found it all wrong for  the Army. The report is here:

This is the second of a three part series. In the first part, yesterday on, the Army contractors noted the specific solutions implemented on the A2 and the problems the Marines solved thereby, but complained that the problems and solutions were too USMC-specific. In this part, we’ll discuss just what they thought was wrong about the Marines’ product. In the third part, which we’ll post tomorrow, we’ll list the modifications that they suggested in lieu of or in addition to the A2 mods.

M16A1 (top) and M16A2.

As we recounted in yesterday’s post, the Army let a contract to analyze the Marines’ product-improved M16A1, originally called the M16 PIP (Product Improvement Program but in November 1983, type-classified as the M16A2. Did the A2 meet the Army’s needs for an improved rifle? The contractors recounted 17 improvements in the A2 versus the A1, and traced those improvements back to four or five fundamental goals of the Marine program: more range, accuracy and penetration at that range, more durability, and a burst-fire capability in place of the full-auto setting.

The Army contractors recognized what the USMC had done — and damned it with faint praise.

The M16A2 rifle was developed and tested by the U.S. Marine Corps. The purpose of this present analysis was to evaluate M16A2 rifle features as they relate to U.S. Army training and combat requirements. It was found that the M16A2 did not correct major shortcomings in the MI6Al and that many M16A2 features would be very problematic for the Army. Accordingly, this report provides several suggested rifle modifications which would improve training and combat performance.

The A1 shortcomings that the paper’s authors thought went unameliorated, or were worsened, by the A2 included:

  1. 25 Meter Setting: The M16A2 does not have a sight setting for firing at 25 meters, where zeroing and most practice firing occurs.
  2. Battlesight Zero: The M16A2 does not have a setting for battlesight zero, i.e., 250 meters.
  3. Aperture Size: The M16A2 probably does not have an aperture suitable for the battlesight, e.g., the single aperture used for most marksmanship training, the record fire course, the primary aperture for combat, etc. The 5mm aperture used for 0-200 meters is probably too large and the 1-3/4mm aperture used for 300-800 meters is probably too small.
  4. Sighting System: The M16A2 sighting system is too complex, i.e., elevation is changed three different ways, leaving too much room for soldier error.
  5. Sight Movement: Sight movements on the M16A2 result in changing bullet strike by different amounts; .5, 1, 1.4, and 3 minutes of angle (MOA)*. The sights intended for zeroing, .5 and 1.4 MOA, are not compatible with old Army zero targets or the new targets being fielded.
  6. Zero Recording: The M16A2 does not have a sighting system which allows for easy recording of rifle zero. Also, the zero cannot be confirmed by visual inspection.
  7. Returning to Zero: The M16A2 does not have a reliable procedure for setting an individual’s zero after changing sights for any reason, e.g., using MILES or .22 rimfire adaptors.
  8. Night Sight: The M16A2 does not have a low light level or night sight.
  9. Protective Mask Firing: The M16A2 has not been designed to aid firing while wearing a protective mask.
  10. Range Estimation: The M16A2 sight has not been designed to aid in the estimation of range

Let’s consider those, briefly. Note that every single one of those objections relates to the sights. There are no complaints about the other Marine improvements (not even the hated burst switch). Most of the sight squawks were because the sight was different from the sights of the A1, which were pretty much as Stoner, Sullivan et. al. designed them circa 1959 (the earlier AR-10 sights are different, but the later AR-15 prototypes and their descendants all used something extremely close to the M16 and M16A1 sights. (The USAF/USN M16 and the Army/Marine M16A1 differed only in the absence and presence respectively of a forward assist). Even the protective mask issue is basically a sighting problem — with the then current US M17 gas mask, the rifle had to be held canted to use carrying-handle based rear sights.

Complaints 1-5 relate only to the M16A2 sights, but 6-10 are just as applicable to the then-issued Army M16A1.

Even at the time, it was clear that optical sights were better than irons — scopes for distance and red dots for close-in work. Army special operators had already tested — on the flat range, in the tire house, and on the two-way range — such early red-dots and both-eyes-open sights such as the Single Point and the Armson Occluded Eye Gunsight (OEG). In the early 21st Century, universal optics would end the long run of the M16A2, and sweep away all these problems the 1986 Army contractors worried about. But there was no way to predict that in 1986, not with any certainty.

And that’s Part 2 of our story. Tomorrow, we’ll cover the modifications to the M16 that the authors recommended in place of the A2.

The paper is available on DTIC:


When the Army Resisted the M16A2, Part 1 of 3

The M16A2 was adopted by the Marines in 1983, and then by the Army three years later, but all of its development was done, largely on a shoestring, by the Marines.

For example, the finger bump on the A2 pistol grip? The very first prototype was built up by a Marine officer on an A1 grip, using plastic wood or body filler! Most of the modifications to the A2 were aimed at:

  1. Increased practical accuracy;
  2. Increased effective range;
  3. Increased durability; and,
  4. NATO compliance (adopting a NATO round equivalent to the FN SS109 round).

In a brief overview of the service life of the M16 series for American Rifleman in June, 2012, Martin K.A. Morgan encapsulated this history well:

In November 1983, the U.S. Marine Corps adopted a product-improved version of the M16A1 chambered for the 5.56×45 mm NATO round. The new rifle was called the M16A2 and it differed significantly from its predecessor: improved rear sights, a brass deflector, a heavier barrel and 1:7-inch rifling were among the changes. The M16A2 also replaced the M16A1’s “AUTO” selector setting with a “BURST” setting delivering three rounds with every trigger pull. The Army followed the Marine Corps’ adoption of the improved rifle in March 1986 when it ordered 100,176 M16A2 rifles from Colt. In September 1988, the U.S. government placed an initial order for 266,961 M16A2s with Fabrique Nationale’s North American subsidiary, FN Mfg., Inc. of Columbia, S.C. Late the following year, when 57,000 U.S. military personnel conducted the Operation Just Cause invasion of Panama, the M16A2 was used in combat for the first time.

For practical accuracy, the A2 had new sights, with a square front post; for range, a new round with a heavier bullet, and new rifling to match; and for durability, new stocks and handguards and significant metal reinforcement in the lower receiver’s weak areas, the pivot pin bosses and buffer tower.

The rifle was not without controversy in the Army. Indeed, contractors for the Army Research Institute for the Behavioral and Social Sciences examined the rifle and concluded that, as their paper’s abstract notes:

[U]se of the M16A2 rifle by the Army would be extremely problematic, a-fact due, in part, to the vast differences between the marksmanship training philosophies of the Army and the Marine Corps.

(The paper is here:

The Army had been researching improvements to the M16A1 for years, but hadn’t actually implemented any. In the foreword to the Army Research Institute paper, the word “problematic” crops up again and one gets the sense that the problem was this solution was Not Invented Here, and moreover, not developed the way the Army wanted to develop one. 

Referring to earlier research, they wrote:

A detailed evaluation of M16Al performance was conducted to determine adequacy, peculiarities, etc. The findings clearly indicated that the M16Al was an adequate combat rifle; however, many shortcomings were identified that should be addressed in a new rifle or any rifle Product Improvement Program (PIP).

They considered that the improvements in the A2, listed below, were suitable only for the peculiar circumstances of Marine Corps service.

The Marine Corps test results stated the following advantages for the PIP [Product Improvement Program -Ed.] rifle:

  • Ease of training (handling and ease of sight movement).
  • Improved safety (no hazard when adjusting elevation on the rear sight even with loaded weapon).
  • Increased effectiveness at long ranges (more hits, better accuracy, and greater penetration).
  • Improved handling characteristics and durability in hand-to-hand close combat.
  • Reduced barrel jump and muzzle climb during automatic and rapid fire.
  • Increased contrast and less glare with square front sight post.
  • Stronger, more durable and improved grasping characteristics of front handguard.
  • Stronger barrel with quicker twist to take advantage of increased effectiveness provided by new ammunition.
  • Improved sighting characteristics providing quick target acquisition for moving targets and better detection of targets in low level light conditions at close ranges, and more accurate long range fire by use of two modified rear sight apertures.
  • Increased ammunition conservation and more effective use of ammunition with burst control device.
  • Conformity to human factors standards by lengthening stock (alleviating bruised eyebrows, noses, and lips).
  • Stronger, more durable stock.
  • Stronger, more durable buttcap which also reduces slipping on the shoulder during firing.
  • More controllable and comfortable pistol grip contoured to the shape of the hand.
  • Improved brass deflector which protects left handed shooters from hot ejected brass casings.
  • Can use NATO type improved ammunition (XM855) which provides improved performance and penetration at long ranges.

The Army evaluators were impressed by that list of solutions, but thought they all traced back to four specific USMC objectives or requirements:

The above list of advantages is very impressive. It appears that the rifle meets the primary requirements stated by the Marines:

  • A sight adjustable to 800 meters.
  • A bullet with better accuracy at 800 meters and the capability to penetrate all known helmets and body armor at ranges of 800 meters.
  • A rifle with more durable plastic parts and barrel which will take a beating during bayonet training and extended field exercises.
  • The replacement of the full automatic capability with a burst mode which fires a maximum of three rounds with each pull of the trigger.

…but they thought that the requirements were too Marine-centric.

The list, however, represents the objective and subjective evaluation of Marine Corps personnel who are emphasizing the most positive aspects of rifle characteristics as they pertain to envisioned Marine Corps requirements.

This is the first of a three part series. In the second part, tomorrow on, the Army contractors damn the A2 with faint praise and list a litany of A1 shortcomings that they believed that the A2 did not resolve. In the third part, the modifications that they suggested in lieu of or in addition to the A2 mods are enumerated.

As it was, the contracting officer’s representative approved the paper in February, 1986. In March, and probably before any of the responsible officers read the paper, the Army went ahead and adopted the M16A2, just the way the Marines had shaken it out.

That makes this paper a time capsule.

Twists of Fate, and Rifling

What separates the winners from the losers is how a person reacts to each new twist of fate.  -Donald J. Trump.

We’re not sure about twists of Fate, but a number of you have asked us about twists of rifling. The question usually comes in the context of AR-15 rifles and their clones, with rifling twists of 1:14. 1:12, 1:9, 1:8 and 1:7 all having been used.

Can you calculate optimum twist for a given caliber and projectile? Yes, you can. There are two equations that are commonly used, Greenhill’s and Miller’s.  Let’s start with the newer one, Miller’s, which was originally proposed in Precision Shooting in March, 2005:

Miller assumes a spitzer-pointed, boat-tailed projectile. In Imperial measurements:

T is twist
30 = a constant representing: standard atmospheric conditions, and a bullet speed of approximately Mach 2 (2800 fps at sea level in standard atmospherics). If you need real precision, Miller does provide more complete equations for that, but these approximations work for rifle velocities.
m = projectile mass, decimal grains
s = gyroscopic stability factor
d = diameter, decimal inches
l =  length in calibers (i.e. length is “l” times the caliber of the projo).

Greenhill’s rule dates originally to 1879, and is frequently used by gunsmiths as it is (or was. anyway) taught as part of gunsmithing school, repeated in Hatcher’s Notebook, and included in Patrick Sweeney’s rifle gunsmithing book among many others. Sir Alfred Greenhill of the Royal Armories at Woolwich developed a number of more complex equations. (More complex than Miller’s, too). But he also provided “Greenhill’s rule of thumb.” Sweeney describes this as follows:

“The length of the bullet in calibers, multiplied by the twist rate in calibers per turn, is 150.”

The constant 150 is good for velocities to about 2800 fps. For higher velocities, as often seen with small-caliber rifles, use 180.

Some notes on twist

As a rule of thumb, the more twist, the more stable the bullet. A bullet must meet a threshold of stability to be accurate. The less twist beyond minimal stability, the less accurate the bullet, in theory, but practical accuracy doesn’t drop off until a bullet is very overstabilized. In small calibers, varmint hunters will tell you a too-fast twist will cause bullets to self-destruct from centrifugal force before overspin hurts their accuracy.

You also need enough excess stability to account for atmospheric changes. As a rule, air density decreases with increased altitude above sea level, and air density decreases with rising temperatures. Less dense air needs less spin than more dense air. This is why the original AR-15 prototypes were found to lose accuracy during Arctic testing by the Air Force — important tests for guys who might have to defend ammo igloos in Iceland, antennas in Alaska, or missiles at Minot. These prototypes had barrels made by Winchester for Armalite in 1:14 twist, then the standard .22x varmint-rifle twist (no one pops prairie dogs in -20F weather). A change to 1:12 solved the problem, at least, for 53-55 grain bullets like those in what would become M193 ball ammunition. (Lighter weight tracer rounds have always been hard to stabilize and trajectory match in 5.56mm). The change to 63 grain ammunition drove the change to a 1:7 rifling twist.

These same calculations may not scale to all types of large-caliber, high-velocity artillery pieces such as tank guns. That’s because air is not truly dimensionless; air molecules don’t scale up as projectiles do. Aerodynamicists and exterior ballisticians can compensate for this scale effect by incorporating Reynolds Numbers in their calculations. For rifle ammo, it’s not necessary or useful.

For those who just want a cheat sheet

Simplified from Sweeney, Gunsmithing Rifles, pp. 109-110

5.56 and other .22 centerfires:

Bullet weight grains Twist ratio 1:inches Velocity
> 70 8 any practical
≤ 70 9 any practical
≤ 63 12 any practical
≤ 55 14 any practical
≤ 55 15 ≥ 4100 fps
≤ 55 16 ≥ 4300 fps

Note that this is really for civilian use in “normal” climactic conditions. For military purposes where you must meet a +140ºF/-40ºF standard, you should go one twist increment slower per bullet weight increment, and understand that you will lose some ability to use weights at the extremes removed from your selected optimum round. Not much of a factor in a military application, where the fewer different DODAAC codes (ammunition stock numbers), the better, as far as the logistics elements are concerned.

7.62 NATO and other .308 centerfires:

Bullet weight grains Twist ratio 1:inches Velocity
> 220 8 any practical
≤ 220 9 any practical
≤ 170 12 any practical
≤ 168 14 any practical
≤ 150 15 any practical

Note again that this is for civilian/sporting/normal-climactic-conditions use.  And that it applies to supersonic rounds only. You must redo the calculations for the slow, heavy bullets used in suppressed applications!

For those desirous of plug-in calculators:

For those desirous of more sheet music:


3d Printed 1911 ( in .22LR Caliber)

It’s been a long time since we’ve done any kind of 3d Printed Guns update, and things are still proceeding at a staggering rate, with lots of interesting gun designs and also a real explosion of accessories.

However, here’s something everyone seems to have been anticipating: a working 3DP M1911. Except… it’s only working in .22, so far, it’s not working 100%, and as we’ll see it took several iterations to get there.  The developer goes by the name MBA Firearms. This is the current iteration:

In an Imgur album, the poster says this:

Here it is, the FIRST EVER 3D-printed 22lr semiautomatic pistol. These 1911 models were printed, assembled and test-fired by MBA. There are many firsts with this gun. This design is still in development but improvements are being made.

It’s not a very high round-count gun yet.

95rounds of .22lr through the latest model. Getting light primer strikes about 1/3rd of the time because I lightened up the mainspring too much.

Several different materials were used on developmental frames.  The upper is a .22 Kimber conversion kit. 

Changing to 910 reduced warping of the rails. The warping was partly due to the weight of the mainspring. Tests with 15# spring worked well but was obviously to light leading to light primer strikes. More testing will be done with 16# and 17# spring later which are ready to go.

The reference to 910 is to Taulman 910 nylon filament (the spool is visible). It’s strong stuff, but needs to be kept dry aggressively as possible, and may need higher temps that some printers prefer.

This image above shows the characteristic layering of 3D printing on consumer printers. The next picture shows four of the test frames. The one that’s built up into the gun is made of Taulman 910.  The white one is made of ABS. The ones that have “attached” grips actually had the grips molded in as part of the pistol.

This is the first functioning 1911 we’re aware of, to have been built on a consumer 3D printer. As such, it is a milestone of considerable significance.

Here’s the initial design and test-firing video.

And here’s a follow-up with more test firing. Better, but still not reliable.