This page details how I design my pneumatic systems and the critical diameters I use for o-rings to seal. This is a subject frequently debated all across the internet and I'm tired of re-explaining it over and over, so I created this page to express my views. It should be said that EVERY developer designing such systems (paintball or otherwise) has their own methods and experience leading them to their own method for obtaining these diameters. The information found here is only my opinion however I have found it to be successful in the past so you are free to follow my example.
Factors to take into consideration:
· O-ring fit must be tight enough to provide a seal.
· O-ring fit cannot be overly tight as to increase the force needed to move the component. In other words it must be loose enough so the small amount of friction created by the o-ring is negligible. In virtually all paintball markers, the friction created by the o-ring fit is completely irreverent toward how fast the part moves. (ram/bolt speed is controlled by the valve venting speed, NOT friction).
· The more an o-ring is stretched outward, the quicker it will wear out and need replacing (theoretically).
· The more an o-ring is stretched outward, the smaller its theoretical height will become. For instance a typical size-0XX o-ring has a height of 0.070 but if stretched twice its diameter its height will be reduced down to, say, 0.045. (I don't really know off-hand, just guessing).
First thing's first, you really should use a CAD program to create your dimensions; this type of work would be pointless if done by hand.
Second. There is no set rule of diameters that can simply be "looked up". There are many online sources of design tips on how to create industrial fluid power systems using o-rings to seal (this includes both hydraulics and pneumatics) however it's my opinion that you should avoid these when working with paintball equipment. The numbers given by these industrial application notes will be far too tight a fit and your parts will never move. For instance one source suggests you crunch the o-ring by at least 10% of its height on both the inside and outside diameters...this would reduce a 0.070" o-ring to only 0.055". If you draw a picture of that you'll see why it would never work in a paintball marker (that's the type of compression you might find in a large hydraulic cylinder, where the o-ring is a quarter inch in height to begin).
Listed below is my method for designing the initial diameters to be used. After this I would have the parts prototyped, experiment with them, and make any changes at that point. This process requires time and money, but it's the only way to obtain 100% reliable numbers (at least after doing it once or twice). Experimentation can even be done on the machine during the cutting process, should you be the one doing the machining and have training to properly adjust the tool offsets.
Consider the following example. It shows a "ram" type piston component moving within a stationary bore. Shown is a size 10 o-ring used to provde a seal (magenta) along with a size 9 o-ring to act as a bumper (tan).
To find my initial diameters I follow this basic method: match the bore ID to the o-rings outside diameter, then stretch the o-ring by around 0.002"-0.003" on its inside diameter. Detailed instructions are below.
1. Draw the o-ring and make sure its dimensions are correct.
2. Edit the outside housing component. Adjust its inside diameter to match the o-ring's theoretical outside diameter. In this case it's approximately Ø0.380" (radius 0.190" in the following image).
3. Edit the piston component. Adjust the outside diameter of the o-ring groove to match the o-ring's theoretical inside diameter.
4. Draw a reference line under the groove bottom, and make it 0.002-0.003 (radius) smaller.
5. Move the groove bottom and the reference line outward until the reference line touches the o-ring's theoretical inside diameter. In this case it's approximately Ø0.244" (radius 0.122" in the following image)
These two numbers will be your critical diameters. In many cases this will leave the o-ring with a reasonably loose yet reliable fit, while at the same time maximizing the o-ring's lifespan by only stretching it as much as required. Personally I prefer to avoid "crunching" the o-ring although I know some manufacturers that do. Alternately I know some other manufacturers that prefer to stretch the o-ring by 1/32" before following a similar process. These methods work for them, although I don't prefer to follow the same trend myself. I prefer my method because it creates a seal by forcing the o-ring out into the side of the bore instead of crunching it inward. Larger diameter o-rings may need to be stretched further.
As I mentioned these should be used as starting points for the components and experimentation is required.
Tolerancing:
This is a tricky subject and open to further debate, but again I will state what I do and you can go from there. I list my critical bores as ±0.0005 to ensure the diameter is exactly where I want it. It's not so much required that the dimension not deviate (say, ±0.001) because the seal will still hold with some deviation, but I prefer critical dimensions like this to be uniform across the board and if your diameters are good then you have nothing to worry about.
Most o-rings have a listed manufacturing tolerance of ±0.0025 however in my personal experience I've found them to be a little more tightly controlled. If I had to guess I'd say ±0.001-0.0015 although I don't know for sure. Regardless, the o-ring tolerance doesn't seem to affect the diameters to the point of performance loss.
In terms of surface finish, a reamed bore will suffice just fine, or a reasonably smooth inside or outside diameter can be turned on a lathe if the finish is held to around 32µ. I suggest you DO NOT produce a mirror-shine on any surfaces coming in contact with a moving o-ring since this will work against you. It can be difficult to imagine but o-rings usually work best with a somewhat rough surface. This is because the microscopic gaps in the metal surface get filled with lubricant and help to keep the o-ring lubricated over a couple hundred/thousand cycles. If the surface is polished to a mirror shine then there will be no gaps for grease or oil, and you will be required to relubricate the o-ring MUCH more often. As mentioned, friction does not determine moving part speed, so creating a surper smooth surface is the exact opposite of what you want to do. Almost every developer in the paintball industry has learned this through their own experimentation, although there are still a few that don't understand this concept.