I haven’t tried to build my own replica because of a complete absence of ability, talent and experience and a lack of access to the machinery and tools that I’d need. However, World of Replica Air Pistols reader R-Gun Pete has all these things. So, when he planned an engineering test project, guess what he decided to make? R-Gun Pete tells his story;
In 1999, I had been doing my job as a draughtsman and CAD instructor for about 9 years (1991 to 1999) and the department needed somebody bilingual to provide support on the CNC mill for the CAD/CAM course. The only other bilingual technician was already a machinist but he didn’t want to jump from conventional to computer assisted machining.
When I was asked if I would be interested, my answer was “Why not?” Since I was already involved in designing with a computer it didn’t look too bad. The only problem was that I didn’t have a clue about actual manufacturing. The goal was to get me up and running in about 18 months. This was not 18 months full time but mostly during slow periods (like summer) and my spare time.
I began to work in the shop with my colleague (the guy not wanting to make the jump) to learn the basics. He would be my mentor for the duration of the conventional portion of the training. We started doing some practice parts to learn how to operate the lathe and mill. It didn’t take me long to realize that we would spend a fair amount of time making pieces without any value other than trying some techniques. I needed something less boring, more focussed and that could be kept as a memento once the project was completed.
The best approach would be to have all the skills that I must learn, in order to achieve my supervisor’s expectations, integrated into a single project. It would need to cover conventional turning and milling as well as drilling and tapping in a variety of materials (brass, aluminum and steel) and to learn how to adjust the feed and speed accordingly. Ideally, setups, indexing and some exotic surface finish could be thrown in the pot too.
To be prepared for my modified job description (technical support of the CAD/CAM course), the project must also include CNC (Computer Numerical Control) machining…
… and mostly, it must be challenging and interesting.
Explained like that, it is clear that just finding a project is difficult enough!
I guess that the title of the article gave away the answer. I decided to make a (non-shooting) replica of a pistol or more precisely a revolver.
Now there were some legal implications and the choice had to be carefully evaluated. According to 1998 Canadian law, non-firing replicas or low velocity shooting replicas are illegal. There is an exception for replicas of antique firearms (firearms manufactured before 1898). This makes ball and cap western pistols or revolvers a nice target group. So which one to choose?
I wanted a style not readily available on the market of non-firing antique replicas (otherwise it would be a lot easier to just buy one than to make one). For me the model that fitted the bill was the Remington New Model Army 1863.
Where should I start?
I knew what I wanted but where to find the necessary information. The answer is “The Internet”. Effectively, I found what I needed. Knowing the overall length (13 ½ inches) combined with a nice profile picture, it was easy to scale it to the right size. I was able to measure directly on the printout to begin a 3D model.
For an exact reproduction, finding blueprints would be the jackpot, in this case my project would be more of an interpretation (since it would be non-firing and had to fulfill the requirements of my learning goals) so in this case blueprints were not critical. Nevertheless, exploded views (available on the net) were useful in understanding how the internal components fit together.
When I started the model, our software was SDRC I-Deas. It was a parametric 3D modeller used on Unix Workstations. The rendering was not shaded but in wireframe as shown on the next pictures.
It looks simple once completed but several important choices had to be made before reaching that point.
First I needed to decide on the extent of the mechanism that would be hidden inside the frame. I would have liked to design a working hammer and trigger that would also index the cylinder as on the actual revolver.
Unfortunately that required more parts interacting together and without any formal drawings the interaction would have to be built by trial and error. As the goal of the exercise was not to spend a lot of time designing the whole system but mostly to have an integrated project to learn machining skills, I decided to restrict myself to only a working hammer and trigger.
Now that the extent of the mechanical involvement was out of the way, I had to decide how many parts were necessary and what type of operations would be done to each of them.
Basically if I do my homework correctly the result at the end should be this:
After spending most of the summer learning the different conventional methods and appreciating the impact of feed and speed on the machined parts, I had a series of completed components.
The barrel and the sight were made of steel. The sight was turned and the tip milled to get the blade. The barrel was started from a hexagonal bar. The bore was drilled on half the length (this is a non-firing replica after all), a hole for seating the sight was bored with a flat nosed end mill and a tang with two holes were made. This tang would be anchored to the receiver.
The barrel has received a special treatment on a surface grinder which created a very smooth surface finish. I decided not to blue the steel parts and they are protected by a fine coat of oil only.
The cylinder and cylinder pin have been made from brass. Both parts required turning and milling. In the case of the cylinder, all the patterned features have been cut using the dividing head.
It has been a long time since I did the job but if I remember correctly one turn of the crank rotated the spindle by 9 degrees and 40 turns did one revolution (360 degrees) of the workpiece.
As could be seen in the previous picture some features are lined up when others are in-between, so let’s just say that there was a lot of indexing involved.
The lever is normally in one piece with a small plunger at the end that would snap in a hook attached to the barrel. Furthermore there would be a cylinder attached to that lever to compact the powder and the ball in the chamber.
All this has been simplified. The compacting cylinder was omitted and the plunger became the whole rod on which the end is cut to accommodate a small socket head cap screw threaded to the underside of the barrel.
Again both parts are brass. The plunger has been turned and milled and the main part of the lever has been shaped on the conventional mill requiring several setups to machine it.
The fake bullets were made of aluminum on a lathe and the tips were cut to shape with a file when the pieces were rotating.
Now that the basic training (conventional milling) was out of the way, I started to learn the CAM software. In fact, in I-Deas, it was fully integrated with the modeller which made it easy to try some toolpaths and modify them as necessary.
To continue my exposure to manufacturing, my supervisor started to show me how to operate the CNC mill. Also that year I was involved in the CAD/CAM course as an observer and a trainee.
In parallel, I continued to work on my own to figure the virtual machining that would end up producing the G-Code for both halves of the receiver, trigger and hammer. Those intricate parts were good candidates for CNC milling.
CNC machining is not magic, there is a lot of planning involved and the right sequence of events is essential. For a receiver half, two setups are required. The outside is drastically shaped and once machined doesn’t provide any level surfaces, on the other hand, when the side contacting the other half is observed we could see several mating surfaces sharing the same level. This would dictate the sequence. The internal details must be machined first and the outside second.
For the outside, the whole machining sequence is not shown because I wanted to concentrate on the 3D region that would be a challenge to machine by a conventional method.
In this picture where there is the notch for loading percussion caps (on a real one), I used a 3D toolpath to cut the material.
For the other side, I tried a 2D cut (X-Y) with multiple Z depths. After testing both methods, I found that the second was quickest.
After the last components were CNC machined, it was time for assembly.
Unfortunately, I don’t have pictures for all the details but essentially once the hammer and the trigger were placed on the bosses in one half of the receiver, the small detent for the cylinder was added and the barrel was pinned in position. All that was left was to put the other half on top and screw both together.
Once the main frame and barrel are assembled, the grip plates and the lever can be screwed. The cylinder with its spring is installed in the frame and the pin is pushed through until the head is in the notch then the lever can be locked by sliding the end over the barrel cap screw.
The final product looks like this:
When we switched from I-Deas to SolidWorks, I recreated my revolver in the new software to learn its features.
SolidWorks is PC based and as the new generations of CAD the rendering has been improved over the old guard.
This has been a very interesting project. I have learned a lot, I have acquired new skills and experiences and furthermore I have produced a nice souvenir that I can keep instead of unrelated practice parts that would probably have ended up in a garbage can at some point.