Giant Piston

Hi all,

So our plan for the robot this season was to use a huge pneumatic piston to launch the ball (~20in stroke, 2in diameter). We’ve got our system pressurized up to the max 60 psi, but we’re just not getting enough force to push the piston hard - right now it kind of just lazily slides out. We’ve considered that it could be an airflow problem as all of the storage tanks are connected linearly - we’re hoping reducing the amount of tubing, adding more storage tanks (we currently have four), and reconfiguring the setup so that all tanks could “contribute” at once will help the problem, but we haven’t tested it yet.

Does anyone have any experience firing pneumatics this large? If so, how can we get more force our of the piston? Would the above mentioned ideas work? Is there anything else we can do to make the piston hit harder?

Thanks in advance

Your cylinder is too big. Plain and simple, you’re way too big. The air just can’t flow in fast enough given the FRC pneumatics tubing size for a ram like you’re looking for.

Try 3-4 smaller-bore cylinders and/or a lever arm. OR… mechanically lock the cylinder so it can’t move when it’s initially fired, and have a tank or two after the solenoid, then trip the solenoid (preload) and unlock the cylinder a few seconds later, once it’s got some pressure in it. On smaller cylinders, it works great–we couldn’t even see the cylinder move when we tried something like that back in Week 1, it just was in one spot then in the other. (3/4" bore, probably about a 3" stroke)

BTW, point of clarification: I’m sure you’re getting enough force, once the cylinder hits 60PSI. You’re looking for speed/airflow, and getting the cylinder to 60 psi faster.

Something to think about:

if you have a 2 inch bore and 20" long piston and want it to impart just as much energy when it is almost completely deployed as when it starts you need to maintain the 60psi inlet pressure the entire time. During the operation of a piston used to shoot the ball it can be assumed that there is going to be no effective contribution from the compressor, it just isn’t fast enough to keep up.

so, the ideal gas law tells us:
P*V = nRT

For argument’s sake we will ignore the change in temperature. This tells us the pressure times the volume must be equal to a constant, so:

P_1V_1 = P_2V_2

P_1 is the pressure in the tank (120 PSI MAX)
V_1 is the total volume of all tanks
P_2 is the working pressure after actuation (60 PSI MAX)
V_2 is the volume of all air tanks + the volume of the cylinder

so, with your setup (using my assumed cylinder)
120 PSI * (574 ml 4) = ((Pi(1in^2)*20n )+(574 ml *4) * P_2
P_2 = ((120 PSI) * (574 ml 4) )/ ((Pi(1in^2)*20in ) +(574 ml *4) ) = 82.85PSI

unless you are experiencing an airflow problem.

Using several smaller diameter cylinders in tandem would solve this problem because it would allow higher flow rates into the combined cylinder-system.

Good Luck,

Our team was investigating pneumatic springs, similar to what Eric suggested, but partially filling the piston itself instead of a separate tank, then having a latch to release all of the pressure built up in the small volume of the partially extended piston. We had very good results with this technique, but we did not use as large of a piston as you, I think ours was 1-1.5". However we chose not to use this idea because we did not want to have all that force held in place by a latch, as we have had a lot of trouble with latches in the past.

spend some time looking at this robot…it shoots great, with two smaller cylinders (1.5 x 9" stroke)

I would try putting most of your tanks AFTER your regulator, instead of before your regulator so you won’t be limited by the 60 psi. However, this would mean you probably need double the tanks you would normally need without a regulator.

But honestly, I would sugest a way smaller cylinder…

As mentioned, a single cylinder that big will have a very hard time shooting the ball in any design. Here are some tips if you wish to pursue other approaches:

  1. Do not attempt to “punch” or “push” the ball straight off one cylinder. Even with tricks for getting cylinders in FRC to actuate quickly, you will have trouble getting any usable power out of this. The only direct-punchers with pneumatics that I’ve seen work are multi-stage, in that they have several pneumatics all actuating in series, so that the velocities add. Otherwise, you must use a catapult or some other device that allows you to multiply the effective velocity of your extending cylinder via. a lever arm.

  2. Make sure you remove the fitting on the far end of your pistons. This requires that you reset the piston through non-pneumatic means (it effectively turns it into a single-acting cylinder with no spring return), but maximizing the size of that orifice will have a huge effect on actuation speed.

  3. Keep tanks downstream of the regulator, one for each piston. The regulator is a massive bottleneck for airflow, and I’ve only seen a couple of designs shoot the ball well with all of the air flowing through it (most of them rely on funky geometry). You still want upstream tanks, since tanks at 60 PSI will obviously hold less air than tanks at 120PSI.

  4. If you have the money, buy high-flow solenoids. We’ve found on 4464 that they’re not really necessary to launch with pneumatics (we’re scoring from 17+ feet away with the standard ones), but they do help.

  5. If you can cope with the added complexity and hazard of a latch and large amounts of stored energy, and can’t get it working through other means, try a precharged pneumatic as described earlier in the thread. Mind that for this to work, your “loaded” position should leave your piston partly extended, else you don’t have any volume to actually precharge. I do not personally recommend these, as the forces involved are terrifying and it involves added complexity.

That said, my best advice, however discouraging, is that if you don’t have anything close to launching the ball this late in build season, you probably want to abandon ball-launching as an option and focus on getting your robot in finalized, working order and giving your drivers practice. Focus on defense, possibly with a low-goal score option or a catcher. This is going to be a game which centers around robot-robot interaction; a good defensive bot will make or break alliances.

All the time you spend on a likely whole-system redesign for your shooter is time lost from ensuring that you field a working robot with capable drivers.

Have you tried connecting your airtanks in a parallel circuit, instead of linearly?

Didn’t check your math, but I’m concerned about your units. Storage in ml and cylinder size in inches^3, did you include the conversions to add these?

Two side questions that this made me think of:

Is this air expansion adiabatic or isothermal? You’ve calculated it using the isothermal assumption. I would argue that it is adiabatic since no time is allowed for heat transfer with the outside to occur. Not sure though.

In the case that it is adiabatic the equation would become P*V^gamma = constant, where gamma for air is 1.4

Using the ideal gas law, don’t you need to calculate with absolute pressure not gage pressure? This would mean adding atmospheric (14.7 psi) pressure to the working and high pressure side. 74.7 and 134.7?

Likely doesn’t make a big difference, just curious which is correct.