Help with T-Shirt Cannon

Hey CD,

My team has been working on a couple of modules for our outreach robot. Another student and I requested to be in charge of the T-shirt cannon. At first we thought this was fine but now we’re a little but stuck (He’s in charge of CADD and i do Drive Train and PR). Niether of us has taken physics yet and so we arent entirely sure how to do the calculations for pressure required. We did them on our own, and got some results that didn’t seem right, so we where wondering if CD could give us a hand. Below is the poster on which we wrote our answers:

Thanks CD!

Why don’t you think they look right? I can start explaining from there :wink:

I would not design the system around aiming for a specific distance, but rather just try to design a system that shoots as far as possible. It’s cool to try to do the physics, but kinematics does not account for many real life inefficiencies, such as air resistance that are present especially when shooting something like a t-shirt.

In order to get a good shot, you will want to make sure you have as large a flow rate as possible in your solenoid. Some other important variables that I considered when designing a t-shirt cannon are air pressure, barrel length, and volume of air needed. I am sure other, more experienced people, may have slightly different lists. If designed well, you should easily be able to shoot 50 feet and if you need to lower the range, it is fairly easy to lower the air pressure.

Well, The PSI I have on the poster has Standard Pressure added, sonit was really only about 2 PSI after the calculation, which seemed low, also, the flow rate was only marginally greater than the actual barrel volume, so it seemed to me like it wouldn’t actually push the shirt out the barrel at any real speed.

Our coach will not OK a cannon design unless it has a range of 50’ or less due to safety concerns.

All you need to limit the range is to have a pressure regulator in the system, which you should have anyways. If you use one like the frc one, you just need to adjust the nob to lower the pressure, which will limit the range. You can even use this method to change your shot on demand based on location (If you are in a small room versus a football field, you will probably want a different shot power).

Thank’s for the idea. I’ll definatly present it. I still don’t know if it will be accepted, but I’m going to push for it. That sounds a lot easier than our math homework way of doing it.

Well, The PSI I have on the poster has Standard Pressure added, sonit was really only about 2 PSI after the calculation, which seemed low, also, the flow rate was only marginally greater than the actual barrel volume, so it seemed to me like it wouldn’t actually push the shirt out the barrel at any real speed.

Our coach will not OK a cannon design unless it has a range of 50’ or less due to safety concerns.

That sounds a lot easier than our math homework way of doing it.

Right or wrong (and it doesn’t look “that wrong”), there is still a value to attempting to solve these problems using the math. Even if the error is 50%, it is still getting to within an order of magnitude of the right answer, so it has some value.

If your barrel was filled with 16 psi air, and you were able to flow air at a great enough rate to maintain 16psi throughout the full travel of the shirt down the barrel, despite it “sounding low”, it is imparting a decent force on the shirt. That 16 psi is working against ~4 square inches so you have 60pounds of force on a reasonably low mass. Additionally, you are saying your barrel has a volume of 2.1L and your fill rate is ~35 actual L/s, so your shirt is travelling the length of the barrel (2.5 ft) in 0.06sec, for an average velocity of 41ft/sec or an exit velocity of 82 ft/sec assuming constant acceleration.

I say all this not because it’s strictly right, but your math probably isn’t that far off.

So back to reality, the issue is that 35 actual L/s (which at 16psi is ~73 standard L/sec or ~150 SCFM of air. That is a LOT of air, enough that to transmit it without extreme pressure drop. Even with a 1" line, if you start at 16psi, you are losing ~0.2psi per foot of travel. At 0.5", you are losing 6.6psi per foot! Couple that with the fact that fittings, valves, regulators, and other restrictions are often much greater than your piping.

So the tl;dr on how to make a pneumatic cannon is that you need to design for higher flow, and have a nice accumulator to hold that flow really close to your barrel to minimize pressure drop. A typical design is something like a 1" solenoid right on the barrel, with an accumulator tank of sufficient volume right at the barrel. You fill it to whatever PSI you want (set with a regulator), and then fire the solenoid to shoot the T-shirt. You can test it with your teacher to determine what pressure is “safe” on the accumulator and then handle it administratively by saying “never set the regulator higher”, or adding something like a analog pressure switch and doing it in code, or adding a mechanical safety relief or even a burst disk, depending on how robust you need your safety to be.

Keep one other thing in mind though, even low pressures of air (10-20psi) can be dangerous in plastics that fail in a brittle manner (PVC) so you should aim to build as much out of more ductile metal as possible, and shield any plastic parts. I would encourage you to develop a design and come back to CD for input on ways to keep it safe.

Since a discussion on T-shirt cannon designs has been brought up, I thought I might ask this: as opposed to a pneumatic approach to the firing of the t-shirt, has anyone ever thought of using flywheels?

I had this idea an year ago, but people had concerns over shredding by the flywheels although I’m partially convinced that such a design might be better and perhaps more safer than these pneumatic t-shirt cannons that we usually see being made.

Interesting thought, always worth trying out :wink: Maybe a team with a test rig for this year could alter the spacing and give it a whirl.

My gut says for most wheel/shirt combinations that there isn’t enough compressibility/springback for it to work well and you might impact more energy shearing the layers of the T-shirt vs. each other versus propelling the mass. But… just intuition, and still would like to see someone try.

Another plus for many basic pneumatic cannon designs is being able to load by just shoving the shirt down the tube, versus having to work out a load from the back side, a way to advance the shirt into the wheels, etc.

I do think there is a potential for issues with pneumatic launchers, but having a reasonably safe design out there could probably benefit a number of teams… might add that to my offseason todo list…

You have some good ideas, I particularly like the burst disk safety idea, and it got me thinking about some upgrades to my taser-ignited propane powered potato cannon, but that’s a story for another time.

The gist of what i got from your post is that basically flow rate is equal too, if not greater than PSI in importance, because it is hard to maintain a constant PSI. We want a fast-fireing solinoid with a wide flow in between the resovior and the barrel tube to minimize the pressure drop between the two. Our current design has multiple barrels so thst presents a challange for multiple resovior connections. Thanks for the great reaponse!

A bit of experience from 2014. We had a pneumatic catapult to launch the giant ball that year. We initially used our in-house SMC Solenoids. About halfway through the season, we started having issues (some Teflon tape had gotten into the solenoids) and we started looking for alternatives. We also started to log the air pressure on both sides of the solenoid for debugging purposes and were very shocked by the pressure differential. We had up to a 30 PSI delta across the solenoids.

The key item with flow rate is that at some point increasing your PSI has zero effect on the performance of the system. There is only a finite amount of air a given solenoid can pass (as I understand it from the aerospace engineers, the air can travel no faster than the speed of sound through the solenoid, speed of sound at what pressure is TBD). For solenoids, we had very good experience with the ones from Automation Direct (low cost, HIGH flow). Related note, I also recommend using Single Solenoids (as opposed to Double Solenoids). If you do use Double Solenoids, ensure both solenoids are ‘off’ prior to firing. We found there was a noticeable performance difference with the SMCs when we went directly from Forward to Reverse on the solenoids vs. Forward-Off-Reverse (we also had a significant time we were at Off).

One final item is to watch your pneumatics connections orifice size. You’re not limited to the FRC legal 1/8" NPT or the 0.16" tubing. The larger you go the higher flow rate you can have. The tubing diameter, length, orifice sizes and quantity will have an impact on your system. You can also parallel Solenoids for greater airflow. Again, you aren’t limited by FRC rules here. Just ensure you maintain safety so you don’t hurt yourself during development.

We are using a 2-1/2" copper tube. Roughly 2-1/2" ID. About 24" long. Insertion depth past a minimum doesn’t seem to have a big impact on distance fired. 16 psi won’t push the tee-shirt out of the tube. Our shoot pressure ranges from 30 to 60 psi for normal shots. Air storage pressure is 110-120 psi with a separate shooting accumulator that is fully discharged on every shot. We adjust the shooting accumulator pressure to control the shot distance. Cannons with 3" barrels will shoot with lower pressure. How tightly the tee shirt is wrapped will greatly affect distances.

A couple of factors hard to calculate would air drag on the tee shirt and force required to move the tee shirt in the barrel. (Barrel drag)

What you really care about is PSI at the point of delivery (base of the barrel). If the T-Shirt could magically travel down the barrel without flow, you would essentially try to pull a vacuum on the tube, as an extremely small volume of starting air expands to try to fill the tube. The more air you flow into the tube, the more pressure you will have throughout the travel of the T-Shirt, and the more total force imparted.

Flow and starting pressure are linked of course, such that a higher pressure can drive a higher flow through the same size solenoid/tubing, etc. However, the pressure loss is exponential as tubing and orifice sizes decrease, and you are ultimately limited by what is called “choked flow”. Simply put, air can not travel faster than the speed of sound in the gas system, so if there is a tight enough restriction, increasing pressure further doesn’t add more flow.

Therefore, your solenoid does need to be “fast firing” but also have a large flow orifice (specified as a Cv value, which is basically the total flow across a valve for a unit pressure drop… bigger is better for this). A typical FRC solenoid will not work well for this application.

Multiple barrels and multiple connections is not as much of a design challenge as a cost challenge, if you wanted to have a solenoid valve per barrel. You could also consider (if you really wanted to push the envelope) of a revolver style design with a single barrel and a t shirt magazine that advances through it. There you still just need a single solenoid, and the main challenge is getting a good seal between the barrel and magazine.

Feel free to keep asking questions. Gas system design was my day job for a number of years.

Just checking, but is all your air storage at 120PSI or do you have any low pressure tanks? The regulator is a choke point.

We use a shooting accumulator. Essentially a low pressure tank. It is a 2" pipe 24" long. That feeds the barrel through a 1" high flow valve and a couple of inches of 1" pipe. The charge valve from the air storage tank is actually closed when shooting. It takes a couple of seconds to charge the shooting accumulator through the regulator and a FRC legal solenoid valve.

How do you open/close the air from the accumulator to the barrel?

Actually, could you post a general diagram of the existing system?

That is where you would put the solenoid.

Our findings agree with the general consensus that diameter of flow is at least as important as pressure. The key diameter is the smallest orifice between the accumulator tank and the input to the barrel. Ours is 3/4" at the solenoid valve, and with 60 psi of air in the accumulator (a cast iron tank of about 3 to 4 gallons capacity), we can toss a t-shirt from the running track over our stadium’s press box - WAY over a 50 foot lateral toss.

In addition to using pressure to adjust the range, we also regulate the “valve time”. Our current system opens for 40 or 50 ms; much longer will use a lot more air with little (if any) additional range. Shorter bursts can be used to dynamically shoot shorter range shirts or lighter objects.

It takes a lot of battery to make the amount of compressed air you’ll need to fire t-shirts. I’d suggest either having an A/C compressor to fill a storage tank to a higher pressure than your accumulator, or use a scuba tank. Stock scuba regulators bring the pressure down from about 3000 psi to where the regular FRC regulators can fine tune to the 40-60 psi we use in our accumulator (between 100 and 120 psi IIRC).

Finally, let me repeat the warning about PVC (or any other “brittle” material) being used to hold or switch pressurized air. Remember how I said how much energy is needed to compress air? Well, most of that energy is sitting in the tank. If that tank or a solenoid valve should shatter, there’s plenty of energy to accelerate shards to hazardous speeds. Work done by or to a fluid on expansion/compression is ∫PdV or (for constant pressure) PΔV. Water at 60 psi is about 200 parts per million smaller than at ambient; Air at 60 psi is about 800,000 parts per million smaller than at ambient. Because air changes volume so much more than water, compressed air has several thousand times as much energy as a similar volume of compressed water at the same pressure. NO SPRINKLER VALVES!

I worked on my team’s cannon a lot and our website has a lot of good information that you can use for ideas.

The distance the shirt will be fired can be changed dramatically by how you roll up the shirt and if you add rubber bands not. For short shots we roll them up loosely so they open up almost immediately after leaving the barrel, another is to roll it up and add a single rubber band at the back of the shirt to hold it together in flight but after firing the front of the shirt will mushroom out to slow it down and for long shots we add two rubber bands on the shirt, one on the front and one at the back. Another way to adjust distance is to change the amount of time that you leave the valve open.

Thank you, GeeTwo, for the first factual response I have seen to the “Why not a PVC T-shirt cannon?” question. At last, I am convinced.