Pistons not extending fast enough?

Hello. Today we conducted a test for a pneumatic system which we are looking to add to the robot before the next competition, but we were met with a rather unsuccessful attempt at shooting a ball. I’m posting here to hopefully get some suggestions to improve the design and/or point out any fundamental flaws with the system.

The tanks were charged at around 90 PSI (the working end pressure guage was displaying about 70 PSI, this needs to be reduced). The setup was constructed sort of arbitrarily, in that I didn’t make any exact measurements to emulate the robot in any particular way. This was set up more as a proof of concept, and unfortunately, the proof of concept seems to have failed.

Differences from the design that’ll go on the robot (see pictures):

  1. When fully retracted the pistons are attached so that the ramp is upright; this is opposite of what the actual design is at the moment, whereby the pistons will be attached so that the ramp will extend from about its current 30 degrees below horizontal to approximately 45 degrees above horizontal (or as close as possible, that being the idea angle for kinematics. This could certainly change if/when I get around to reading some of the more complex analysis that has been done on the physics of the ball). To account for this, we angled the setup (as demonstrated in the videos), though I’m not sure how closely we replicated the movement.

  2. The ramp we’re using is made of wood and lacks the metal side rails that are on the ramp currently. This would have made it easier to launch the ball as it is significantly lighter.

  3. The hinges are strap hinges instead of piano hinges as are currently supporting the ramp on the robot, though I doubt that matters much.

  4. We were shooting at a PSI that was slightly higher than the allowed 60 PSI (70); at least, that’s what our guage read.

That being said, I think that this test should have at least proved to some degree whether the concept would work. However, after doing it, I found that the ramp rotates much too slowly to be able to do anything productive; notice that even the light basketball from two years ago (rebound rumble) hardly launched off the board.

Would attaching the pistons at the very bottom of the board make that much more effective that it can launch a ball across the room? It also seems as though the pistons are extending much too slowly; is there anything we can do to fix this?

Piston setup:

Robot Ramp (in bag):

Walkthrough of testing setup:

Test 1:


Test 2:

Test 3:

Sorry, a better question is what’s the PSI on the regulator?

A single solenoid will not give you enough airflow to extend with enough force to launch the ball. What you need is a way to help the piston along. I have seen two different ways to do this.

  1. I haven’t done this, but I have read about it. You can add a spring to the arm of the piston so that it extends outward more quickly.

  2. The second way is how my team did it. You pull the piston in and have it lock in somehow(we use a gate latch). You then push the piston outward. The piston tries to extend but can’t because it is locked in. Then when the lock is released, the piston releases with more energy because you have pre-pressurized the piston. (The one problem is that our catapult system requires two huge pistons. These two pistons take a lot of air and therefore require several tanks.)

I hope this helps but it might be hard to implement on your robot because it has already been mostly built.

You will get much better results using smaller diameter cylinders. Those large ones in the picture require too large a volume of air to fill, thus the extension speed is not fast. Use much smaller cylinders (same length, smaller diameter) for a fast launch.

You may also wish to look at the flow rate on your solenoids. We are using a new solenoid we bought that has a much higher flow rate than the solenoids that were FRC legal last year and before. No flow rate restriction this year - only limit is port size.

So, I have a basic understanding of pneumatics, so I’ll attempt to advise you the best I can. Warning, this is going to be a long read, so take a deep breath, and try to understand as much as you can.

The name of the game is flow rate. How do you move air as quickly as possible into the cylinders? If you move the air in quicker, the cylinder will shoot out faster.

Before I get into this, I’m going to set up an analogy for us to follow. Think of the cylinders as buckets, and the solenoids as hoses that carry water. The goal is to fill up the buckets of water as fast as you can.

Now, based on the videos, you have one solenoid hooked up to two, fairly large, cylinders. You’re essentially trying to fill two big buckets with one hose. So, how do you make the system faster? How can you change up your variables (buckets and hoses) to make the system fill up faster?

Let’s go through a few scenarios:
Use more hoses - Of course, if you use, say, one more hose (read: solenoid), that’ll help with the flow rate, and it’ll help you fire quicker. Now, I believe the rules limit you to maximum one solenoid per cylinder, so keep that in mind.

Use smaller buckets - While this may seem counter intuitive, this is also correct. Using smaller cylinders will help you fire faster. There’s a tipping point between cylinder size that gets you the best speed vs. power. I’d recommend reaching out to a team to find out what they’re using.

Use a larger hose - Now, this is interesting. If you use a solenoid that’s rate for a faster flow rate, you can move air quicker, which will ultimately shoot the pistons out faster. FIRST doesn’t impose a solenoid flow rate limit, but given a set of constraints (price, power, etc.) you may be limited on what’s actually available to you.

Pool your water into a smaller bucket ahead of time - Now, this is a little bit more difficult to understand, but bare with me. Let’s say you pop the solenoid and let the cylinders build pressure, but you lock down the shooter. Then, at a later point in time, you manually release the shooter (via a latch or something). Here, the cylinders are essentially acting like compression springs, but you are able to get over the flow rate problem.

That’s the tip of the iceberg, and I recommend you do some more soul searching (read: researching) into your issues, your system, and which solutions would best fit your setup.

Here are some other threads that may help:
Team 842’s Amazing Machine
Creating Pneumatic Launchers
This talks more about letting the solenoids fire before you release the catapult.

Let me know if this makes sense, if there are any errors, or if you need any further assistance/clarifications.

  • Sunny G.

As mentioned, you should really be using a solenoid for each piston, we found that attaching them directly to the shooter worked better. Additionally, you may want to move your pistons farther towards the pivot point and start the catapult at a lower position. It sounds like a weird thing to do but when the catapult requires a greater amount of force to move initially it causes the air to build up in the piston first, basically doing the latch thing without an additional mechanism. Here’sa video from our first prototype, we currently have two pistons that I believe are the same size that you guys are using. You may also want to check out 842’s design, they have a great pneumatic shooter.

This might give you some ideas, and it’s the clearest picture I have of our shooter which appears to be similar in geometry to what you are attempting to accomplish. The pistons are 7" stroke, both have a small air tank on the working pressure side attached to each piston. The pistons push just short of over center (they almost line up with the hinge point on the front of the “catapult”). This puts only a slight force on the modified gate latch(not seen on the left side of the picture inside the robot, actuated by a .5" stroke piston which lets go to release the pre-pressurized shooter. The pivot is a pair of door hinges and there is a strap to prevent the pistons from smacking into the end of the cylinders whey they extend.

http://i207.photobucket.com/albums/bb115/Mcholerton/robotics/turretweight.jpg

https://www.youtube.com/watch?v=rbYTEzBO0_A&feature=em-upload_owner

I’m not quite sure I understand what you mean, but I can tell you that for each of the tests the pressure gauge on the high pressure side (before regulator) read about 90 PSI and the pressure gauge after the regulator was at a solid 70ish PSI.

Okay so I was reading that correctly. If you are using a single noid for both of the pistons, switch to 2 of them for each, like others have already said. Better flow rate means faster extension.

There isn’t actually a solenoid in the testing setup, but I do see your point with the airflow problem. Would I be able to improve this by adding two small tanks on the working side and pressurizing them to 60 PSI, plumbing each tank to its own cylinder?

As for the latch system, it seems that many teams are doing that this year. We may have to end up doing something similar, perhaps with a small stroke piston (or a rotational cylinder we have lying around?), though I would like to avoid this option as it may be at the edge of our capability technically.

In addition to what everyone else has said, be sure that you test with 60 psi working pressure as mandated in the rules. Also, I suggest that you include a hard stop before the piston reaches the end-cap of the cylinder to prevent damage to the cylinder.

P.S.
Piston is the thing that moves, while cylinder is the entire actuator.

Could you perhaps provide a link or a google search directing me to solenoids similar to those you’re using? Currently we’ve got some solenoids left over from the early 2000s, and I’m not sure what teams are using now.

As for the cylinders; if the diameter becomes too small, wouldn’t they not have enough force to push the ball + platform? I will do some calculations and figure out how much force would be needed from each piston to move the ball however, and will certainly see what the smallest piston we can use is.

FYI, R90 prohibits plumbing to solenoids two one cylinder.

I believe what is occurring is that one solenoid is currently connected to both cylinders, and he was recommending to use one solenoid per cylinder for a total of two in the entire assembly.

It’s good to be sure… Some would overlook that and see that a higher flow rate copied be a how’d by running two together. Better at this point to keep everyone well informed to save hassles at competition.

Agreed, I would not want to spend precious time between competitions tuning a system only to arrive at an event and find its illegal.

Thanks so much for the great explanation, it really helps understand the whole concept of airflow. Apart from changing from one solenoid to one per cylinder and adding a higher airflow solenoid (if possible), would there be anything else I could do to increase my airflow?

The pressure regulator also restricts flow rate somewhat. To compensate, you could add additional accumulator tanks on the low pressure side of the regulator.

What do you mean by directly attaching them to the shooter?

I will be working on modifying the catapult so that it starts in a lower position and the pistons are closer to the hinge point. I’m also going look into 842’s design a bit more, thank you for pointing that out. It seems they’re using the same size pistons as us; if they are having no issues shooting, maybe we’ll be able to do the same.

The gate latch - when your pistons retract, I’m assuming that it locks back into place until the next time you’re shooting? Also, do you have the pistons retracting with gravity or some other mechanism, or are you having them retract by sending air in the other side?