Slowing down the action of a cylinder with flow control, mechanical or via software

[Randy Forgaard]

We are making some improvements to our robot in preparation for an off-season tournament. We have a pneumatics problem on our robot, and we're trying to find a solution that is allowable under the 2011 LogoMotion rules.

We have come up with a proposed solution for the problem, but we aren't sure if it is legal (under the rules) or practical, and are soliciting feedback.

NOTE: I have carefully read the thread, "Variable position of cylinder in compliance with <R74>," but our issue is a bit different. We don't need control over the variable position of a cylinder. That level of control is more than we require. We're just trying to reduce the speed of air flow from the exhaust port of a cylinder (see below).

Here is our issue: We have a long, pneumatically-actuated arm, with the pivot point at the bottom, that leans (rotates) down to pick up tubes. We use a double-acting cylinder to control the up-and-down movement of the arm. We are using one of the 5-port Festo valves that comes in the KOP for controlling a double-acting cylinder for this purpose.

To make the arm go down, we trigger the valve so that it sends pressurized air into the "down" port of the cylinder and exhausts from the other port of the cylinder. This sends the arm down fast (too fast, truthfully).

Our big problem comes when we reverse the valve to send the arm back up. The arm is rather long and heavy. When we send 60psi of air pressure into the "up" port of the cylinder (while exhausting from the other cylinder port), the pressure gradually builds in the cylinder for a couple of seconds until the cylinder is able to lift the arm. Then, when the cylinder lifts the arm, the arm starts going up slowly (fighting gravity), but as the arm rotates upward, there is less gravity to contend with (due to the angle of the arm), and the arm then slams very hard into the top stop bar on our robot. This makes us wince every time it happens. It really slams -- it's a wonder that the arm hasn't broken in the two tournaments we've been using the robot.

So, here's the thing. Even though we are using pneumatics to lower and raise the arm, we would like to be able to reduce the speed of the arm in both directions. We need to use the full 60psi of allowable pressure, because otherwise the cylinder cannot lift the arm from its lowest rotated down position. But using the full 60psi, the arm ends up slamming against the top stop when it is at the top and no longer fighting gravity, and we aren't happy about that. Also, in the down direction, the arm also goes too quickly.

For the down direction, the solution is fairly simple, but it doesn't appear to be allowable in the 2011 rules. We could simply use a COTS flow-control valve on the exhaust port of the Festo valve for the "down" direction of the arm, so that the air won't exhaust so quickly. A flow-control valve, for those who might not be familiar, is a simple, inexpensive, passive pneumatics part with a screw that restricts how fast the air can escape, so you can slow down the air flow. There are other, similar simple COTS pneumatics parts that achieve the same goal under different names: exhaust mufflers, speed controls, etc. They all work basically the same way -- they are passive little frobs that reduce the air flow so it doesn't move too fast. Any of these would allow the arm to go down nice and smoothly, and not overly fast. However, <R66> doesn't specifically list flow-control valves or similar devices as being legal for use this year, so it seems like we can't use a solution like that. I just want to confirm that we are reading the pneumatics rules correctly on this.

I can't quite understand why a passive flow-limiting device is not allowable. The main pressure vent plug valve will still release all stored air pressure in the system for our robot. There is no danger issue that I can see here of stored pressure, unless the flow control valve is completely closed (which would make the cylinder completely useless on the robot).

The "up" direction for the arm is also a problem, with the "slamming" at the end, and is more complicated to fix. A flow-control valve is not sufficient here. By the time the arm has rotated to the top (where it is about to "slam"), there is almost no gravity acting on the arm movement and the air exhaust is coming out of the cylinder very fast. If we used a flow-control valve for the exhaust in this direction, we would have to set the flow to be very constrained to prevent the slamming. But reducing the air flow that much would make the arm VERY slow to raise up when it is first coming up from the ground, fighting gravity -- too slow to use in competition.

A sort of obvious solution would be to use a solenoid-controlled valve on the exhaust port of the cylinder, to rapidly switch between allowing air to exhaust and blocking the exhaust. However, this would be in violation of <R74>, since we would be using multiple valves for a single commanded action of a cylinder, and this specific case is specifically clarified as illegal in Team Update #13. (This led to the <R74> discussion in this thread.)

So, here is our proposed solution for controlling the speed the air exhausts from the cylinder. If anyone sees a problem with this solution with respect to the rules, or can suggest an alternate solution, please let us know. Here's our idea:

When we are raising the arm, for the first few seconds, we will send all of the pressurized air into the "up" port on the cylinder port that raises the arm. After a few seconds, we will, under software control, slow down the arm by rapidly switching the Festo valve between the sending pressurized air into the "up" port of the cylinder and sending air into the "down" port of the cylinder. If we send air into the "up" direction 60% of the time, and send air into the "down" direction 40% of the time, the arm should still raise up, but more slowly, so it won't slam into the top stop bar of our robot.

How does that sound? It doesn't seem like it violates any of the rules, because we aren't adding any prohibited hardware, we're just using software to control the Festo valve very quickly. Also, the main pressure vent plug valve will still release all stored air pressure in the system for our robot. We haven't tried rapidly switching the Festo valve back and forth electronically before -- I wonder if that will burn out the Festo valve. Also, this solution is a kludge -- I wish we didn't have to do it this way, so if there is another way, that would be great.

Super sorry for this long message. In summary, we are looking for a way to effectively reduce the airflow out of the exhaust port on a pneumatic cylinder, in a variable manner (as the arm raises and needs to reduce the flow more), without violating the rules. And we are wondering if our software solution, above, makes any sense, or will just end up burning out our Festo valve.

I hope my problem explanation above is understandable. Any help appreciated. Thank you!

[Andrew Bates]

Have you considered just attaching some surgical tubing or springs so that when the arm is nearing it's uppermost position it has to stretch the tubing/springs. This should lessen the impact.

[kwotremb]

Sounds like you have something similar to what we did this year on our robot. We had our entire mast and arm asembly rotate at a pivot point at the base that was controlled by a pnuematic cylinder. We used the flow control valves without any issues at all (basically the simple screw ones that replaced the straight or 90s that you screw onto the festo valvues). We had them pass inspection at 3 different events so I believe there should not be any issues with you using them. I thought it worked great for us.We only had the 'up' part of the piston hooked up to the air. Then for down all we did is have a flow control valve on the output that we controlled to make sure the fee fall of the arm was at a rate we needed.

Now for you, I would think for 'up' you guys should be able to use a flow control valve on the output just as you are thinking with the down. The flow control will make sure the rate at which the air can escape is limited, or more controlled. That way when the arm is moving slowly at first, the speed is not limited at all because the air can escape as fast as air is going in the other side of the cylinder. Then when you start to get to the top the air pressure will start to build up due to the fact that it will not be able to escape fast enough and slow down the rate of the arm. Hopefully that should be enough to prevent some of the slam you are experiencing on the top. Another simple way to melp out some of the slam is to put a bumper on the arm or frame where they hit, that way the slam is less violent. We were going to do this until we designed it so our cylinder maed out before we hit a hard stop. Hopefully this helps.

[Randy Forgaard]

Originally Posted by corpralchee

Have you considered just attaching some surgical tubing or springs so that when the arm is nearing it's uppermost position it has to stretch the tubing/springs. This should lessen the impact.

I really like that idea. Don't know why we didn't think of that. I'll bring it up with the team. Thank you!

[Chris Fultz]

Flow control fittings are legal parts, and you can use them on both of the fittings of the pneumatic cylinder to control the speed in each direction.

You might want to also consider a gas shock of some type. McMaster Carr sells them in several lengths and forces.

We used one on our arm to provide some assist to the lift motors and to dampen the movement.

[Randy Forgaard]

Originally Posted by kwotremb

Sounds like you have something similar to what we did this year on our robot. We had our entire mast and arm asembly rotate at a pivot point at the base that was controlled by a pnuematic cylinder. We used the flow control valves without any issues at all (basically the simple screw ones that replaced the straight or 90s that you screw onto the festo valvues). We had them pass inspection at 3 different events so I believe there should not be any issues with you using them.

That's interesting that your simple flow-control valves screwed into the exhaust on your Festo valve passed inspection at 3 events. Sounds like we may be more worried about the flow control valves than we should be. Clearly, it seems to me that flow control valves are in the spirit of the rules, since they don't store air and are entirely passive.

Originally Posted by kwotremb

We only had the 'up' part of the piston hooked up to the air. Then for down all we did is have a flow control valve on the output that we controlled to make sure the fee fall of the arm was at a rate we needed.

That's a good idea. We have been using pressurized air to send the arm down, because "at rest," the arm is straight up in the air so it doesn't naturally fall down. But there's no reason we couldn't change the stop bar so that the top of the arm is angled out slightly, so that gravity would drop the arm with no air pressure needed, and just use a flow control valve on the exhaust to reduce the speed of the drop.

Originally Posted by kwotremb

Now for you, I would think for 'up' you guys should be able to use a flow control valve on the output just as you are thinking with the down. The flow control will make sure the rate at which the air can escape is limited, or more controlled. That way when the arm is moving slowly at first, the speed is not limited at all because the air can escape as fast as air is going in the other side of the cylinder. Then when you start to get to the top the air pressure will start to build up due to the fact that it will not be able to escape fast enough and slow down the rate of the arm. Hopefully that should be enough to prevent some of the slam you are experiencing on the top.

Your description makes sense. We experimented with a flow control valve on the exhaust port for the "up" direction, but found that the arm still slammed. But I think we must not have done enough experimenting with the valve, because it makes total sense to me that we should be able to restrict the flow enough so that it won't slam, and it also won't have any impact on the speed that the arm raises since the air can escape as fast as the air is going into the other side of the cylinder, as you note above. If indeed we can legally use a flow control valve (and, from your experience at those 3 events, it appears we can), it seems like that should solve the problem nicely for us.

Originally Posted by kwotremb

Another simple way to melp out some of the slam is to put a bumper on the arm or frame where they hit, that way the slam is less violent. We were going to do this until we designed it so our cylinder maed out before we hit a hard stop.

I'm interested in your last sentence there. You mentioned you were going to install a bumper where the arm hits the frame, but you didn't need to do that because you designed the arm so the cylinder "maed out" before it hit a hard stop. I'm not sure what the means, but it sounds promising. Could you explain a little more?

And thanks very much for your detailed recommendations. It sounds like your robot uses a very similar arm mechanism to ours.

[Randy Forgaard]

Originally Posted by Chris Fultz

Flow control fittings are legal parts, and you can use them on both of the fittings of the pneumatic cylinder to control the speed in each direction.

You might want to also consider a gas shock of some type. McMaster Carr sells them in several lengths and forces.

We used one on our arm to provide some assist to the lift motors and to dampen the movement.

That's awesome that flow control fittings are legal parts. But I can't seem to find that in the rules. <R65> says that "no pneumatic parts other than those explicitly permitted by the Pneumatic System Rules may be used on the ROBOT." <R66> lists the pneumatic parts that are specifically allowed on the robot, and flow control fittings don't seem to be on the list. However, "kwotremb," who also responded on this thread, says that his team has used flow control fittings on their robot and it has passed inspection at 3 events, so it sounds like it's not a problem. But I'm still puzzled why flow control fittings don't seem to be explicitly allowed in the rules. I may be overthinking this.

[Mr V]

Flow control valve between the solenoid and cylinder are legal. I know that a team who's lead mentor is also an inspector and has been before being associated with that team used them and I guarantee that he would not allow parts that aren't legal on his teams robot.

There are flow control valves that regulate the flow in only 1 direction. So put one that regulates the flow into the side that gets pressurized and/or one that regulates the flow out of the side that exhausts. You should be able to do what you want with one if that doesn't work add the other.

[Randy Forgaard]

Originally Posted by Mr V

Flow control valve between the solenoid and cylinder are legal. I know that a team who's lead mentor is also an inspector and has been before being associated with that team used them and I guarantee that he would not allow parts that aren't legal on his teams robot.

That's very good to know, thanks!

[jspatz1]

Everyone should take note that it helps the helpers help you when you explain your problem this thoroughly. Length of description is not a problem.

First of all, as far as I am aware there are no restrictions on using flow controls and metering valves. You can use as many COTS commercial flow controls you want as long as they are unaltered and meet the pressure requirement. You certainly want to use them on both directions of your actuator. There are clever ways to use flow controls in nested or sequential ways to achieve the flow pattern you want.

My philosophy on pneumatic actions such as this is that pneumatic acuatation is only what it is, a somewhat spongy push or pull with no live speed or force control. If it does not give you the behavior you want, you need to change what is being actuated so that it does. With the restrictions FRC puts on us (no 3-position valves, no servo-valves, no servo-regulators) you do not really have the means to control your pneumatics to match the task, so you must design the task to match the pneumatics.

I am refering to using counterweight, springs, friction, weight distribution, and shock absorbers to even your load through its travel, so your actuator yields a uniform result. It is better to design your mechanism so that it utilizes the uniform speed/force that pneumatics inherently provide, than to use complex flow control methods to try to make the pneumatics into something they are not. It is the same idea as balancing the load on motors to give a more uniform result.

We had a very similar situation this year, as did many teams, where we were pneumatically lifting our gripper from the floor to the scoring position. Obviously the same force that you need to lift the dead weight of the arm from the floor is more than you want to have at the top of the stroke. We simply found with experimentation the right length and strength of springs (or elastics) to assist the lift, so that there was prompt motion at the bottom, uniform speed throughout, and a relatively soft landing at the top. Counterbalancing your load this way may even require a negative load at the bottom (air cylinder is actually pushing the arm down to the low position) in order to get the uniform result. A spring assited lift will allow you to turn down the speed control and still get prompt motion. You can also put springs on shock cords so that they balance only a portion of the stroke, such as a spring assist at the bottom, nothing in the middle, and spring resistance at the top. Perhaps one spring or elastic can be centrally positioned to do both jobs. If you have some weight allowance left, counterweighing your arm to balance the load is a possibility. Gas springs are available and allowed to balance the load. Although I do not usually like friction methods, creating a friction source at the top of the stroke to help decelerate the load is an option. Gas shock absorbers are available to help decelerate the end of stroke. Air cylinders are available that have adjustable shock asborbtion built into the cylinder itself. The point is, some combination of balancing methods will yield the uniform motion you want, and then you can adjust the speed to suit. It takes experimentation.

Of course there are tricks you can do with flow control, such as having the arm motion physically switch a manual diverter valve from one flow control to another near the top of the stroke, or having a microswitch in the stroke switch a valve or clutch. But such tricks can be prone to adjustment, and add complexity to your system. Simply changing the nature of the load to give the desired result is the simplest and most reliable approach.

[kwotremb]

Originally Posted by Randy Forgaard

I'm interested in your last sentence there. You mentioned you were going to install a bumper where the arm hits the frame, but you didn't need to do that because you designed the arm so the cylinder "maed out" before it hit a hard stop. I'm not sure what the means, but it sounds promising. Could you explain a little more?

And thanks very much for your detailed recommendations. It sounds like your robot uses a very similar arm mechanism to ours.

Sorry, typo there.

Basically when our mast was all the way at the up we "maxed" (missed the x before) out on the stroke of the cylinder rather than hit a hard stop on the frame. We still had a little bit of hard 'slam' per say, but we were only at 35-40 psi and we felt that out mast was strong enough to take that hit, we were more worried about doing wheelies or ruining the cylinder. But with the battery and compressor in the front of the robot, it took a lot to do a wheelie. Also the cylinders should be designed to take more of an impact then we were putting it under. We were more worried about beding something if we kept hitting the mast onto the frame so we were just going to add some rubber or foam piece to add a shock absorber in that location.

What I would do with testing is start fully closed and slowly open it up as you test (that what we did for the free fall). Now I dont know if there will be enough air in the cylinder to compress to a psi high enough to counter the 60 psi you are lifting the arm up, but in theory I believe it should help.

Here is a decent pic of our robot in starting config, shows mostly what we were doing:
http://i271.photobucket.com/albums/j...1/IMG_4565.jpg

[Randy Forgaard]

Originally Posted by jspatz1

First of all, as far as I am aware there are no restrictions on using flow controls and metering valves. You can use as many COTS commercial flow controls you want as long as they are unaltered and meet the pressure requirement.

That is good to know, and seems to be the consensus on this thread. It certainly makes logical sense that flow controls should be allowed. I kind of wish I could point to a rule in the Pneumatics part of the Robot section of the 2011 rules that shows that flow controls are allowed. But there seems to be enough consensus on this point, and enough teams that have passed inspection with flow controls, that we shouldn't be concerned about it, especially at an off-season event.

Originally Posted by jspatz1

My philosophy on pneumatic actions such as this is that pneumatic acuatation is only what it is, a somewhat spongy push or pull with no live speed or force control. If it does not give you the behavior you want, you need to change what is being actuated so that it does....I am refering to using counterweight, springs, friction, weight distribution, and shock absorbers to even your load through its travel, so your actuator yields a uniform result. It is better to design your mechanism so that it utilizes the uniform speed/force that pneumatics inherently provide, than to use complex flow control methods to try to make the pneumatics into something they are not.

This also makes a lot of sense. I will discuss this design concept with our mechanical team. We already use springs in our elevator, to assist the electric motor, and there's no reason we can't use similar design ideas with the pneumatics that raise and lower our arm. Excellent point, thank you!

[Randy Forgaard]

Originally Posted by kwotremb

Basically when our mast was all the way at the up we "maxed" (missed the x before) out on the stroke of the cylinder rather than hit a hard stop on the frame.

That's a great idea as well. We could potentially replace the cylinder with one that has a somewhat shorter stroke, to prevent the slamming at the top.

Originally Posted by kwotremb

What I would do with testing is start fully closed and slowly open it up as you test (that what we did for the free fall). Now I dont know if there will be enough air in the cylinder to compress to a psi high enough to counter the 60 psi you are lifting the arm up, but in theory I believe it should help.

We'll do that. Thanks again!

[jspatz1]

Basically when our mast was all the way at the up we "maxed" (missed the x before) out on the stroke of the cylinder rather than hit a hard stop on the frame.

Originally Posted by Randy Forgaard

That's a great idea as well. We could potentially replace the cylinder with one that has a somewhat shorter stroke, to prevent the slamming at the top.

We'll do that. Thanks again!

A hard stop on the cylinder will probably not be any less problematic than a hard stop on the frame (it will still "slam"), and is very hard on the cylinder itself.

It is always easy to shorten the working length of a cylinder. Simply add a spacer between the clevis and the nose of the cylinder, so that it bottoms out externally on the spacer instead of internally on the piston. You can always use a cylinder that is a little longer than needed with this technique, but you can't get more stroke out of a cylinder that is too short.

[Tristan Lall]

There seem to be some errors in the information being given here. Let me try to clarify.

Flow control fittings are permitted (if they are "connecting fittings"). Flow control valves (and check valves, incidentally) are not permitted (because they're not any of the things permitted by <R66> and the rest of the pneumatics rules).

The fittings Chris was referring to are like an NPT-to-tubing elbow connector, with a needle-type adjustable flow control screw built in. Those are almost certainly legal, because they are usually understood to be a type of fitting. A flow control valve, that isn't also a fitting, is almost certainly illegal.

At an off-season event, do whatever the organizers require. If they're alright with teams taking reasonable liberties, go right ahead. (Usually there's no inspection.)