Mulit-postioning with pnuematics this year?

[KenWittlief]

Since we are allowed to use more than what we have been limited to in the past, there is something Ive been wanting to try for a while, maybe this year we can.

Ive always wanted a way to be able to use a pnuematic cylinder with more control than full out or full in.

If we use extra valves on the exhaust side (the ones we dont normally put fittings on) couldnt you pressurize one side and have the cylinder go all the way out, then pressurize the other side, and have the cylinder goto the CENTER position?

if you had a position sensor, and you used the flow restrictors, couldnt you control the amount of air in each side (using more than one valve) to position it anywhere between the two extreems?

Anyone ever play around with this or try it?

any thoughts? seems like it would be perfect for controlling the position of an arm, infinitely variable.

[Rob Colatutto]

You are correct. My high school team did this in 2001 and 2003 for our arms. The system lets you stop the piston where ever you want. We never had any sensors on it but you could easily do it with some on there. The way we had it set up was, run from the regulator into a double solenoid, then the air would flow dirrectly to the extension connector (front) on the piston, but then the retraction (back) air flow would be routed through a single solenoid to with its relief valves looped together. This allowed us to stop the pneumatic at any position with some fancy programming, but you would have to ask an old 263 programmer how to do that. One thing to watch out for is the lever arm's weight on that pneumatic. We had a slight drifting problem with the arm when we had 10 pounds on the end of it 4.5 feet from the pneumatic. With the right calculations this can be solved though.

[Rickertsen2]

Since we are allowed to use more than what we have been limited to in the past, there is something Ive been wanting to try for a while, maybe this year we can.

Ive always wanted a way to be able to use a pnuematic cylinder with more control than full out or full in.

If we use extra valves on the exhaust side (the ones we dont normally put fittings on) couldnt you pressurize one side and have the cylinder go all the way out, then pressurize the other side, and have the cylinder goto the CENTER position?

if you had a position sensor, and you used the flow restrictors, couldnt you control the amount of air in each side (using more than one valve) to position it anywhere between the two extreems?

Anyone ever play around with this or try it?

any thoughts? seems like it would be perfect for controlling the position of an arm, infinitely variable.

The setup isn't ideal, but you can prolyl bootleg it to work. The cylinders are fince for this, and it is done all the time but not with this type of valve. You will also need some position feedback(I don't think the reed switches will work. They are probably not fast enough.). If you try what you describe above, i would be interested to hear the results.

[maxgebhardt]

The setup isn't ideal, but you can prolyl bootleg it to work. The cylinders are fince for this, and it is done all the time but not with this type of valve. You will also need some position feedback(I don't think the reed switches will work. They are probably not fast enough.). If you try what you describe above, i would be interested to hear the results.
well, even if the new postion switches won't work at high speed, you could try and tighten the valves to slow the speed of the extention as much as possible. my qustion is do you damage the solenoid?

[Andrew Dahl]

can we get a diagram of this



dahl

[KenWittlief]

I havent thought this through all the way, or tested anything , so Im reluctant to draw up a schematic. Someone might take it and base their design on it, only to learn much later that it wont work for some reason.

The basis idea I have is this. The pnuematic valves work buy releasing the full pressure in the line (60psi) to one end of the cylinder, and venting the other side open to the room - so its like one end of the cylinder is always disconnected.

The big valve they give us also has fittings for the vent outputs - so you could connect those to another set of valves

that would allow you to control the 'open' end - and instead of releasing all the air, contol how much is released

the second valves would not be connected to the pressureized line at all - they would be vent control valves

so when you pressurize one end of a cylinder, then switch it the other way, the first end will not loose all its air, unless you open the additional vent control valve - and depending on how long you leave it open, the air in that side will be allowed to vent out, a little, a lot, or all of it

and that should control how far the cylinder moves towards the other end, a little, a lot, or all the way.

[geo]

Does anybody has a picture or diagrams?

Last time I tried to control the position of a cylinder with a double solenoid and I attached other one in the relase end. It worked. But sometimes the cylinder retracted slowly even I close the relase end.

[Andrew Dahl]

working on it- i think i found the digrams that we used to create it on the 2001 bot



dahl

[Jedi Padawan]

Ok so if I'm getting this right you are wanting to use one cylinder in multiple positions?? is this possible even with the magnetic cylinders?? I was thinking of that but wasn't exactly sure if that was going to be possible without doing something that might "modify" it. :confused:

[Rob Colatutto]

Ok so if I'm getting this right you are wanting to use one cylinder in multiple positions?? is this possible even with the magnetic cylinders?? I was thinking of that but wasn't exactly sure if that was going to be possible without doing something that might "modify" it. :confused:
Its possible without any modifications at all.

Last time I tried to control the position of a cylinder with a double solenoid and I attached other one in the relase end. It worked. But sometimes the cylinder retracted slowly even I close the relase end.
We had the same problem at first last year untill we discovered a leak on our pneumatics system right after the storage tank. As long as there are no leaks and the system can hold over 60psi before the match starts you will not have a problem.

[ngreen]

yes, this is very possible. We used a pneumatic lift last year that could stop in multiple positions.

I drew a fast diagram in paint of how we did it last year.

Basically it involves routing all exhaust from the double solenoid through a single solenoid and putting a stop or loop on the NOT ON state of the solenoid. Make sure you stop the exhaust of the single solenoid too. Then you can program the single and double solenoids to work together. When you push one button on the joystick the single solenoid kicks on to allow the exhaust to escape to atmosphere, and when it is released the solenoid will switch to the pressured side and stop. We used two buttons, extending and retracting but made so they would trigger on side of the double solenoid and put the single solenoid into the ON state. With a little more programming you could probably be pretty exact as to when you stop.

I'll include the diagram. Any questions just ask.

I added a picture of last years robot in a halfway position. The cylinders are behind the yellow fiberglass poles.

[KenWittlief]

Thanks for the photo and diagram

and I have to say, nicely done! Your bot looks super - this would be an ideal way to actuate and arm moving up and down to grab something like

oh, I dont know, a 30" diameter ball?

[maxgebhardt]

if some of you have doubts, and the pneumatics system assembled, I used a basic demo to convince my superiors superiors that it works. if you have the plastic tube snap connectors on the exhaust, just take a plastic tubing T bracket and make a snap in, snap out stopper by leaving one end open and looping a tube in the other two. then plug it into the exhaust valve to stop the cylender mid stroke, then remove it to let it finish. it proves it can be done if anyone has any doubts, and your too lazy to rig up a solenoid.

I'm still wondering if someone can tell me if backfeeding like that causes issues with the solenoid...

[ngreen]

We used the system all last year and still use it when we do different presentations and we have never changed either solenoid so I am pretty confident that there is no problem with backfeeding. Make sure there is no leaks to the system and if will work really pretty good. Like I said, we used a buttons that the actuator length changed as long as you held the button. We used a 8" stroke and slowed the speed of the actuator by only opening the valve to them only a quarter turn. By using programming you could open for increments of time (which means a specific distance) by pressing a button once. If I remember right the actuators may have a little more give than when not using it but it was not a problem for us. A note about the single solenoid: When not powered it returns to a not on side. Any questions please ask, or set a pneumatic board up on a piece of plywood with switches from Radioshack to control you solenoid to test. Pneumatics are very easy to implement and add and their are many resources avaliable from FIRST. I'm very excited about getting to use the rotary and magnetic piston actuators this year. If they don't use the rotary actuator on our robot I might convince them to order them for me the play with. I personally think they would work great for arms for the 30" ball.

[pras870]

I've actually been looking for a way to do this for the past 2 years. I have a pretty firm grasp on pnuematics from the past 2 years, but I still fail to see how this could be done :rolleyes:

I've also found, that amazingly, pneumatics actually weigh less then a motor driven assembly. Our team has always seemed to have a weight problem, and I'm always looking for a way to incorporate pneumatics instead of motor driven for the fact that they weigh less and are actually more reliable, but I've always been under the assumption, with the materials first gives us, you can only open and close a cylinder in 2 positions (full and closed).

[ngreen]

I've actually been looking for a way to do this for the past 2 years. I have a pretty firm grasp on pnuematics from the past 2 years, but I still fail to see how this could be done :rolleyes:

I've also found, that amazingly, pneumatics actually weigh less then a motor driven assembly. Our team has always seemed to have a weight problem, and I'm always looking for a way to incorporate pneumatics instead of motor driven for the fact that they weigh less and are actually more reliable, but I've always been under the assumption, with the materials first gives us, you can only open and close a cylinder in 2 positions (full and closed).

If is very possible. I will give you the most complete explanation as to why. I'm currently working on a chemical engineering degree and one of my professor said something that applies here while speaking on diversity. Differences cause change. In pneumatics the difference in pressure cause the cylinder to extend or retract. It relies on this pressure to move. Since the pressure before the double solenoid is set and won't keep increasing past 60psi you don't have to worry about it to continously compressing. Air flowing from the 60psi side against an open exhaust at 1 atm causes the stroke to be pushes towards the one atmosphere side and the air leave the cylinder through the exhaust. If instead you route all exhaust from the cyclinder to a closed air system the cyclinder will not work or move because it is pushing against the non-atmospheric pressurized air in the tubing. Draw out a vector. When a 60psi force pushes against a 60psi force the resulting force is 0. There is no change. So we have have both choice possible. Movement and No Movement so how do we get these to work together without me walking along and doing it manually. First gave us these wonderful Festo single solenoid to use. First step is to route all exhaust through the single solenoid. Use a tee to connect the two exhaust and use a single piece of tubing to connect to the solenoid. Next plug the exhaust of the single solenoid. It is easier this way and stops it from leaking that way. Next add a loop or a stop to one of the holes on the two holed side. Note: Single solenoid work in NOt on and on and the state will revert back to not on when the power is cut. Lastly find a good programmer that will make both solenoids work together. Think that using this system, to move you must have the exhaust go through the non plugged side of the single solenoid(the one open to the atmosphere). Also, using the double solenoid you trigger just one side to switch. THere are many more things you can do with a good programmer with this system. Like pushing a button that is set to extend the cylinder for a certain time (or possibly distance).

It works. Believe me. Look at my sloppy diagram for more reference. My team wouldn't have dared built a lifting mechanism like that if we couldn't control where we stopped it. Manipulating the boxes with accuracy was a key last year. Any questions just ask. I really want to see the capabilities with a rotary actuator. Think 45 degrees.

[pras870]

I'm beginning to get an idea of what you're saying. Basically the system would be this..


Air tank -> Double Selenoid Valve

Air tanks go into a Double Solenoid Valve.
1) Double Solenoid input/output hooked up normally to cylinder
2) 2nd input/outputs connect with a T connector to the input on the single solenoid.
Single Solenoid:
1) One input/output is stopped with brass fittings.
2) Exhaust is outputted as normal
3) Stop in the other port (loop I don't really understand).

So basically when you get the two valve's to work they limit the airflor of each other, correct or incorrect?

[ngreen]

You're getting a pretty good idea. Yeah it is airflow but air flow is due to difference in pressure so you're are equalizing pressure too.

Notes: I blocked both of the exhaust port of the single solenoid so I could switch the side of the output I wishes to use. My loop simply refers to a piece of tubing that uses a quick connect to the solenoid out but is connect to a quick connect tee with a piece of tubing connecting to the other part of the tee. It is lighter than brass fittings and easier to switch sides rather than a stop (but a stop would work too.) The exhaust isn't routed normally. Both are blocked. The open output is were the exhaust from the cyclinder goes through when you want the cylinder to move.

Your basically correct. Ask question if you have any. A great way to perfect it is on a pneumatics board that way you can see how the system works. Seeing is Beleiving. My mentor hardly believed it when they saw my pneumatics board and it changed completely how our robot was built. Good luck.

[pras870]

I'm wondering if you have any close up pictures of the valves from last year. I have an idea of what they were like, but I don't fully remember the function of each port on the solenoids. That's why i'm having a problem grasping the concept actually, i so far have a firm understanding of what you're saying and it does make sense, but I'm the type of person that needs a visual to have it totally click. That, or maybe my mind just isn't thinking right because it's 3:30am. If you have aim, msg me at RFosgate15Subs, I'd really like to talk about this more.

[maxgebhardt]

I've actually been looking for a way to do this for the past 2 years. I have a pretty firm grasp on pnuematics from the past 2 years, but I still fail to see how this could be done :rolleyes:

I've also found, that amazingly, pneumatics actually weigh less then a motor driven assembly. Our team has always seemed to have a weight problem, and I'm always looking for a way to incorporate pneumatics instead of motor driven for the fact that they weigh less and are actually more reliable, but I've always been under the assumption, with the materials first gives us, you can only open and close a cylinder in 2 positions (full and closed).
well the entire point of the solenoid switch is that when it is switched not only is pressure introduced in one side of the actuator, but preasure leaves the other side through the exahust valve. so if you put a solenoid to control the exhaust valve's flow, then the pressure in the actuator equalizes and the actuator stops. that is exactly how it can be multi-positioned. like I said before, just try plugging up your ehaust valve on the solenoid while the cylender is moving and it will stop mid stroke, remove the plug to let it go.

[pras870]

after staying up all night and suffering from lack of sleep, i finally figured out a way to make it work flawlessly, oh happy day!

[Chris Fultz]

This is an excellent series of Posts.
I had never thought there was a way to do partial stops on these pneumatics. Leave it to a group of FIRST students to prove me wrong.

I think I will have our pneumatics team build some circuits and play with it to see how we can use this new knowledge.

Being able to clock the rotary pneumatic - that would be excellent! :)

[maxgebhardt]

the festo solenoid seems to have two extra contacts, is this because it inherently can close the exhaust?

[ngreen]

Are you talking about the festos from 2003 or 2004? I'm more familiar with the 2003 but I'm setting my 2004 pneumatic board up tommorow. From the 2003 I know I had to block the exhaust of the Festo to keep it from leaking through when I would stop the one side. I not sure if they can inherently close the exhaust. If you are talking the 2004 Festos, disregard this post. I get my chance to work with them tommorow after taking a week to basically design a drivetrain/chassis.

[pras870]

This is an excellent series of Posts.
I had never thought there was a way to do partial stops on these pneumatics. Leave it to a group of FIRST students to prove me wrong.

Well, that is exactly what FIRST is all about, cooperation :)

[maxgebhardt]

Are you talking about the festos from 2003 or 2004? I'm more familiar with the 2003 but I'm setting my 2004 pneumatic board up tommorow. From the 2003 I know I had to block the exhaust of the Festo to keep it from leaking through when I would stop the one side. I not sure if they can inherently close the exhaust. If you are talking the 2004 Festos, disregard this post. I get my chance to work with them tommorow after taking a week to basically design a drivetrain/chassis.
I mean the 2004 solenoid. the sleek white plastic one, with snap connectors for pneumatic hose already integrated.

[Chris Fultz]

This is still an excellent series of posts and the schematic described here will work. BUT, if the system is not properly controlled, part of the pneumatic system is operating above 60 psi if your system regulator is set at 60 psi.

Below I will try to explain some of the basics of the pneumatic system and why I say that.

WHY 1?

To hold a cylinder mid-position, the forces on either side of the actuator must be equal. This force is a function of the pneumatic pressure, but the pressures will be different.
The pressures do not ‘equalize’, the forces on either side of the moveable inner plate equalize.

WHY 2?

One end of the cylinder has a larger surface area (working area) than the other. The rod end has a smaller surface area because of the rod, and this reduces the working area of that end. If you read the pneumatics table in the book, you will notice an “extended” and a “retracted” force. The working area is why they are different.

WHY 3?

Lets use a 2” cylinder as an example.

On this cylinder, the major diameter is 2”. The rod diameter is 0.625”.

Major end: 2” dia = 1” radius, the area is pi x (radius squared). 3.14 x 1 squared
= 3.14 square inches. With 60 psi (pounds per square inch), the force of the cylinder is
3.14 x 60 = 188 pounds (as listed in the pneumatics manual).

Minor end: The major area is the same – 3.14 square inches. However, the extension rod reduces the working area. The rod diameter is 0.625”, so the radius is .3125”. Using the same formula for area, the area of the rod is 0.307 square inches. To get the working area of the minor end, you have to subtract the rod area from the major area. So, 3.14 square inches – 0.307 square inches = 2.83 square inches. With 60 psi, the force of the cylinder on this end is 2.83 x 60 = 170 pounds (as listed in the pneumatics manual).

This is why the cylinders we use will have more force in one direction than the other.
And this is why part of the pneumatic system can be higher than 60 psi (the legal limit) when you use the solenoids to do partial stops.

WHY 4?

Assume the system regulator is set at 60 psi. and the cylinder is (somehow) - half way extended. On the major end, the force applied is 188 pounds (60 psi acting on 3.14 square inches). When using the cylinders normally, with the solenoid vents opened, this force will push the air from the minor end of the cylinder and the rod will fully extend. But in this situation, the vents are closed, so the air cannot escape. The rod will continue to move until the FORCES acting on the internal plate surface are equalized.

This only happens when the air pressure on the minor end of the cylinder has increased to counteract the smaller working area. The movement of the major end of the cylinder will compress the air in the minor end – remember the minor end is a closed area because the solenoid vents are capped. The Major End will continue to extend until the pressure in the minor end reaches 66.5 psi.

WHY 5?

The minor end of the cylinder has a working area of 2.83 square inches. For this end to produce 188 pounds of force, it will require 66.5 psi (188 divided by 2.83 = 66.5 psi). This pressure is generated by the compressing action of the major end of the cylinder moving and reducing the area in the minor side of the cylinder.

By closing the vents, you have created a system that will operate above the maximum allowable pressure (60 psi) by 10%.

SOLUTION:

You can still utilize the setup described, but you must reduce the main operating pressure in the system to do so. If you want to be able to stage the extension of the 2” cylinder mid-extension, you must limit the system pressure to 54.1 psi.

WHY 6?

The pressure on the minor end becomes your limiting factor – you need to keep it at or below 60 psi. Using the 2” cylinder At 60 psi, the minor end of the cylinder can produce 170 pounds of force at a 60 psi working pressure (2.83 square inches x 60 psi). To be able to hold the cylinder mid-extension, you need to limit the force of the major end to 170 pounds – remember that the FORCES must equal, not the pressures. So, to limit to 170 pounds with a 3.14 square inch area, the pressure must be limited to 54.1 psi (170 divided by 3.14).

For the other two cylinders in the kit:

0.75” Cylinder with 0.250” dia rod – 26.5 pounds Major, 23.5 pounds Minor.
The working area of the minor (rod) end is 0.392 square inches x 60 psi = 23.5 pounds.
The working area of the major end is 0.441 square inches.
23.5 pounds divided by 0.441 square inches = 53.3 psi.
53.3 psi is the maximum allowed system pressure to allow you to hold the cylinder mid-extension and keep below the 60 psi maximum allowable in the closed end of the pneumatics.

1.5” Cylinder with 0.437” dia rod – 106 pounds Major, 97 pounds Minor.
The working area of the minor (rod) end is 1.61 square inches x 60 psi = 97 pounds.
The working area of the major end is 1.76 square inches.
97 pounds divided by 1.76 square inches = 55 psi.
55 psi is the maximum allowed system pressure to allow you to hold the cylider mid-extension and keep below the 60 psi maximum allowable in the closed end of the pneumatics.

It has been awhile since I sat in a fluids class so if I made a wrong statement here please advise and I will correct it.

[ngreen]

I'm not sure about the numbers, I'm not in fluid mechanics until next year, but I understand the concept. I still would like to re-read some of the pneumatic rules and I am still not sure on my personal thoughts as to it breaking FIRST rules, i.e. safety, or the legality of the rules. I will for sure look into this more in-depth. I know the 60 psi has been broken before i.e. robot with pneumatic arm running into wall or other robot, but I'm not sure how FIRST would take this. I see them as taking it on only if it poses a safety hazard and I know my system did not. If I was unsure of it's merits safetywise I would make changes to my current robot and not consider using it in its full capacity this coming up year. I take your comments with great seriousness because as last year rules guy I know there importance. I have on questions for you though. If the system ran in full capacity as it is now, with this possible extra compression taking place, what would be your thoughts about its safety, FIRST legality and intent, and any other you think I should consider?


The only reference I see to this says: "Working" air pressure on the robot must be no greater than 60 psi.

I do not see the exhaust air that is being compressed aka greater than 60psi as "working" but as being "worked on". By this definition I see my system as being FIRST legal and will take it that way until I get a chance to ask them the question. I would still like your thoughts though.

[Chris Fultz]

In my opinion, the system probably safely operate above 60 psi. The presssure between the tank and the first regulator valve is 120 psi, and we use similar components (especially the tubing) through-out. The 60 psi limit is probably both to add a safety factor in the system (a blown hose or other component could seriously injure a student) as well as put a limit on the power of the pneumatics. Remember the commetns from the kick-off - don't turn into a lawyer. Your goal should be to keep the system under 60 psi and it can be done just by reducing your regulated pressure.

Operating through the exhaust like you are is probably not significant, but I thought it should be mentioned. One difference in this design circuit vs. running into a wall is that running into the wall is unintentional, but building the pneumatic circuit is intentional.

I had three reasons for the post - one is that I laid out the schematic and it looks like it would work and I think it is great that studetns have found a way to do this, another was to clarify / correct a few statements about pressures and forces to be sure others walked away with the right concepts, and the third was just to be sure students were aware that part of the system had to be over 60 psi to make this work (if the system regulator was set at 60).

Please post any info you get from FIRST. And good luck in fluids.

[ngreen]

Thanks. I agree with this post completely but my only problem is I had planned to use this system with a 2" cyclinder to possibly hang. The plan was to use the retracted force, 170 pounds and the 60psi. This will cause the systems pressure to drop by 12psi and will put the force on the border of 130lbs, enough to lift the robot. Reducing this pressure to much would probably limit this ability and I would have to change to extended force to have enough power. I had reasons why I want use this system but I imagine if FIRST rules against this I can re-engineer a workable plan. Thanks for your insight into this. I had honestly never put any thought into this. I knew the compression stops the pistons (equalizing of forces not pressures) but I had not thought that it would be compression past 60 psi and possibly be ineligible. I'll try to post if I hear anything from FIRST. Thanks.

[Gary Bonner]

The pressures do not ‘equalize’, the forces on either side of the moveable inner plate equalize.


When balancing the forces, don't forget to factor in any load on the piston.

[KenWittlief]

Ok some good objections have been raised here- if anyone hasnt been following the problem, let me state it simply

if you put 60 PSI in one end of a cylinder (say to raise an arm) and then switch its valave the other way, and use an additional valve to close its exhaust vent

then the air in that cylinder has no way to get out. If someone grabs that arm and pulls it down, the air in the bottom side will be compressed past 60 psi - maybe WAY past 60 PSI

OK - this IS a serious problem, but I think I have the solution. The regulator that is use to set the pressure to 60 PSI is designed such that IT will vent air if pressure exceeds the setting on the output side

if you have setup the pnuematics as they are normally used, and someone pulled down on the raised arm, the air that is compressed in the cylinder above 60 PSI would vent out the regulator (backwards)

so the solution for the mulitposiitoning system would be to have an additional (extra) regulator between the vent connections and the blocking valve, with the output of the regulator towards the cylinder. This will release any pressure above 60 psi in that side of the system, even when the blocking valve is closed.

If you take this approach I think you will want to have your pnuematic design WELL documented so you can explain to the judges exactly what you have, and why its safe.

Ive taken the liberty of editing Ngreens diagram to show where the extra regulator would be added - note the arrow -the regulator is connected backwards, and the input side is not used.

also note the regulator will not show pressure all the time, so you will have to adjust it to 60 PSI when the vent side is charged.

[Chris Fultz]

A good thought, but I believe you will get into a cycle where you will keep increasing the pressure, then relieving it, then increasing, then relieving - eventually the cylinder will be all the way extended.

This may be a problem best solved and understood with a layout board and a few pressure gages.

[KenWittlief]

Im not sure what you mean in your last post?

I see this being used to raise and hold an arm up-so there will be a steady load on the cylinder

and the operator will hold the valve open long enough to raise of lower it where they want it.

why would it end up eventaully extended all the way? do you mean by itself?

[generalbrando]

I can see the loop thing happening so that it would be extended all the way, and I think that the way to handle this is to just reduce the overall pressure to that portion of the working system. You would have to use another regulator for Ken's idea, so why not keep the rest of the bot at 60 psi and have this part at say 50 (to be overly conservative). I'm interested in this idea and I'd like to know if anyone has any coding tips on it. As for whether or not the parts can handle more than 60 psi, I believe the tanks are rated at 120 and the pistons at 60 (correct me if wrong), so I would not suggest consistently and purposely running them higher.

[KenWittlief]

I think maybe you want it the other way, if I understand the concern.

if the vent regulator is set at 60 psi, and the main regulator is set at 60psi, then when you goto switch the direction of the cyliner, the vent regulator will spill all its air, because its being pushed backwards with 60psi from the main supply

if this is the perceived problem then you would want the vent set at 60, and the main regulator at a little less - so that it cant force the vent side to release its air. To make this system work you have to be in control of when the vent side lets air out, through the second single valve - you dont want the extra regulator venting air unless the system has been externally stressed.

I cant wait to try this!

[maxgebhardt]

I understand the idea of putting a regulator on the ehaust to prevent the pressure from jumping, however, wouldn't that mean that if the cylender was extending, then multi positioning would fail, because when the pressure would normaly jump above 60 psi, than it CANT, so the cylender would slowly extend rather than hold still. this could cause the arm to slowly extend and slowly drop the bot to the floor if it had tried to pull up to the bar...

[KenWittlief]

there would be no reason to use this approach to pull up to the bar - to do that the cylinder can extend all the way at once - no need for mulipositioning

this proposed idea is more for an arm to get the 2X ball - you could infinately adjust the position of the arm, by slowing leaking the 'vent' valve - and once you close it, it should stay where you left it.

[maxgebhardt]

well I was hoping to use it on the ball and/or the bar, since my team captain is a bit hesitant to raise the robot more than 6" to 1' in case the bot drops.

[KenWittlief]

you only need to raise it enough to be 'off' the floor - no extra points for airspace :c)

[generalbrando]

We set up this multi positioning system the way it was prescribed in that funny little drawing. It was a success. After putting a few flow controls in different places to adjust the speed and everything, it was great! It's not very hard to program it. If you set it up and look at it, you'll see that it's somewhat simple. We had an extra guage on the portion of the system that was being "looped." It does indeed sometimes go a little over 60. However, it was only a couple of PSI and you can give up a few from your main regulator to take care of that and you won't lose that much force.

A suggestion to improve the design... If you put an air tank in series with the exhausts that lead into the single solenoid, it will dampen the effect of opening it and allow you to better control the movement. You would have to precharge it in order for that to work, but I just wanted to throw that in there in case anyone was having trouble getting it to work well. The flow controls are the best way to go.

Anybody plan on using this?

[Dinush]

Our team built a system just like the one in Ken Wittlief's diagram.
Instead of stopping, when the single solenoid is closed, the piston slows down considerably, but does not stop.
- We have done all that is in our ability to stop leaks.
- is it possible, then, that we are loosing pressure through the main pressure gauge? That is, the one limiting the entering air to 60 PSI?
Thanks!
Dina.

[KenWittlief]

try increasing the setting on the secondary regulator, the one that expells excess pressure on the vent side. If you have it set to 60, and your main regulator set to 60, then I can see the second one venting on you when you switch directions

you might need to have the secondary one set to 60, and the main one set to 50 or 55

or maybe the other way around? try it and see what happens.

[Mr. Van]

This is still an excellent series of posts and the schematic described here will work. BUT, if the system is not properly controlled, part of the pneumatic system is operating above 60 psi if your system regulator is set at 60 psi.

Chris is correct here and I hope everyone read his post. It is my opinion that any team which is using this (really cool) way of multi-positioning cylinders should have the "actuation" pressure limited to 54 PSI. This ensures that the system is not DESIGNED to operate with pressures within a cylinder greater than 60 PSI.

Yes, sometimes cylinders may get "overpressurized" by some external force (another robot, wall etc.) but your pneumatics system should not be designed to take advantage of this. (In the same way you would not use a motor to compress air in a large cylinder and use it to pneumatically power a smaller cylinder at higher pressures than 60 PSI.)

Note that if you have 66 PSI in a line in your system at some point, you could USE that higher pressure by connecting another valve and another cylinder... I think we would all agree this is not legal.

So I think that we need to take precaution and design systems which never have actuation pressures (within a cylinder) greater than 60 PSI. And has already been pointed out, be sure that you include the LOAD on the cylinder as well.

-Mr. Van
Coach, 599

[ahecht]

Note that if you have 66 PSI in a line in your system at some point, you could USE that higher pressure by connecting another valve and another cylinder... I think we would all agree this is not legal.
We would all agree that using the 66PSI wouldn't be legal, but I don't think you will get a consesus that just having it there as a side-effect of use is illegal.

[ngreen]

We would all agree that using the 66PSI wouldn't be legal, but I don't think you will get a consesus that just having it there as a side-effect of use is illegal.

I agree with ahecht here. It is only controlling what air we exhaust and what air we keep. We used this system all last year and we have never had any concern of safety or legality of the system. Looking on it after the facts presents it may occasionly run a few psi above 60. All the pneumatics are rated for 125 psi for these reasons. I honestly don't have any problem with anyone who runs this system at full capacity and occasionly has air go above 60 psi, only when you are extending the cyclinder. I take the word "working" air as the air coming from the regulator through the solenoids which is definitely 60 psi and not out the exhaust. The exhaust is not working rather it is being worked upon. If you route it through another solenoid to run another cyclinder then it is working and above 60psi and illegal. But these are just my view.

Good luck to anyone trying to do this. Also, generalbrando, would putting an air tank in series with the exhaust be useless because then you lose the capacity up the line in the 120psi section and limit your ability to move large loads farther? We are only allowed 2 air tanks. I guess you could just use a lot of the tubing and practically make a storage tank.

[Gary Bonner]

It seems to me that the pressure in these setups can go much higher than 60 psi, and the proposed remedies would not necessarily work.

Take the 2” cylinder example from before to somewhat of an extreme, but with no change in the load the piston is lifting. The surface area of the extend side of the piston is 3.14 in^2, and the retract side is 2.83 in^2. Suppose we are using the cylinder to lift 180 lbs. The extend side is pressurized to 60psi and the retract side is vented to the atmosphere. This produces a lifting force of 188lbs (60 psi X 3.14 in^2). When the piston gets to the height we want, we close the retract port. The extend side remains at 60 psi and exerting a 188 lb force on the piston. This is balanced by the 180 lb load on the piston plus the equilibrium pressure reached in the exhaust side, in this case, 2.8 psi ( 8 lb / 2.83 in^2).

Now, suppose we want to lower the piston a little. We pressurize the retract side to 60 psi and vent the extend side to the atmosphere. When it gets to the desired position we close the extend port. Now the retracting force on the piston is 350 lbs [180 lb load + (60 psi * 2.83 in^2)]. To counteract this force, the extend side becomes pressurized to 111 psi (350 lbs / 3.14 in^2).

In this situation, a venting regulator on the exhaust would merely act as a flow control valve and allow the piston to slowly move. Reducing the supply pressure wouldn’t necessarily help either.

If the load on the cylinder was in the other direction (e.g. we are pulling our robot up on the bar) the pressure would be greater. Also, any change in the load after we stop the piston (e.g. a robot hanging above slips down on us) would change the pressure on the exhaust side.

[KenWittlief]

the extend side becomes pressurized to 111 psi...

that is not possible with the drawing I posted. As soon as the secondary regulator sees more than 60 psi on its output side it will vent it until the pressure drops to 60. not gradually, instantly.

its the same as if you use the pneumatics as defined in the manual - if you have a large cylinder pointing forwards on your bot, extended - and you drive into a wall at high speed, as the cylinder is pushed in the regulator vents the excess pressure - thats what they are for. they keep the pressure on the output side at the set level, no matter what happens to the cylinders.

[Leo M]

Look at page 4 of the 2004 Pneumatics Manual. The primary pressure regulator, the Norgren, is required in-line right after the tanks to limit the working pressure to 60 psig. This is a "relieving regulator" - if the pressure on the downstream side (cylinders & solenoid valves) increases above 60 psig - say, by pushing on an arm actuated by a cylinder - the regulator will relieve the pressure by venting air. There is no need to jump through hoops to limit working pressure to <= 60 psig; the main regulator takes care of that automatically.

Now, while you have the manual out, why not read the entire thing thoroughly, so you understand how all of the components work together? It will pay off in the long run, only takes about 30 minutes, and will save pages of posting.

[ngreen]

Look at page 4 of the 2004 Pneumatics Manual. The primary pressure regulator, the Norgren, is required in-line right after the tanks to limit the working pressure to 60 psig. This is a "relieving regulator" - if the pressure on the downstream side (cylinders & solenoid valves) increases above 60 psig - say, by pushing on an arm actuated by a cylinder - the regulator will relieve the pressure by venting air. There is no need to jump through hoops to limit working pressure to <= 60 psig; the main regulator takes care of that automatically.

Now, while you have the manual out, why not read the entire thing thoroughly, so you understand how all of the components work together? It will pay off in the long run, only takes about 30 minutes, and will save pages of posting.

Yes on the side between the regulator and the side of the cyclinder that is open. If you have this side open and the cyclinder is fully extended and you push it back. Air over 60 psi will be forced out of the regulator.

However on the other side of the cyclinder to where the air is stopped there is no release and there is a possibility that it will go a little over 60 psi as much as 66 psi. This would occur during a stop during an extension. During a retraction this problem wouldn't occur because of the norgen relief valve.

And don't think that reading the manual will give you answers to many of these problems. The manual is very vague and only give you very little info other some of the capabilities and pictures so that rookies will know what to look for in the kit. I am sure these people like myself have read the manual and rules several times over and have a good general idea of what to look for but the only way you can be for sure is if you hook it up and test it.

[KenWittlief]

the subject of this thread, using pneumatics in a way that will allow you to stop a cylinder between the ends of its tavel, and keep it there,

is not covered in the pneumatics manual

we are being clever with the way the cylinders are used, and blocking the vent side (that is normally open to the the room pressure)

thats why this thread is going on 4 pages - we are trying something new and want to make sure we dont:

1. violate any design rules or

2. create a system that could be dangerous

so far, I think we are on solid ground.

maybe next year the results of this thread will be IN the manual - it wouldnt be the first time that teams got so innovative that FIRST had to re-write the book :c)

[Astronouth7303]

Exuse me, but I was just poking my head in, metaphorically, and I saw this. We're doing it too. Were just taking a solenoid (the curved half-circle) and placing it before the 'memory' solenoid, which is connected to the cylinder. We haven't gotten it to work, yet, but I think that's a coding issue.

I've said my peice. I'm going back to my regular place now.

[M. Hicken]

does anyone think that you could mount a limit switch accordingly to where you want the cylinder to stop, thus it would kill power to the soleniod, would that work? I know it is not the best design, im just curious

[Biff]

does anyone think that you could mount a limit switch accordingly to where you want the cylinder to stop, thus it would kill power to the soleniod, would that work? I know it is not the best design, im just curious
Depending on how you plumb the hoses you will just send the piston the other direction. You are going to have to test things or figure them out by drawing the air flows in each state for the valve wiht power on and off. The short answer to your question is no.

[ngreen]

Depending on how you plumb the hoses you will just send the piston the other direction. You are going to have to test things or figure them out by drawing the air flows in each state for the valve wiht power on and off. The short answer to your question is no.

This is wrong. With this you could definitely do this. You can program the limit switch to change the single solenoid to its stopped state and stop the cyclinder. Yes, this would work. Any questions about set-up of the pneumatic parts refer to my diagram or ask me. The pistons will stop if you do it right, thus multi-positioning.

[ngreen]

Exuse me, but I was just poking my head in, metaphorically, and I saw this. We're doing it too. Were just taking a solenoid (the curved half-circle) and placing it before the 'memory' solenoid, which is connected to the cylinder. We haven't gotten it to work, yet, but I think that's a coding issue.

I've said my peice. I'm going back to my regular place now.

Refer to my diagram as writing on the first page and set it up exactly like it. You hook the single solenoid (curved one) after the "memory" solenoid. The curved solenoids input 1 should be the combined exhaust of the memory solenoid. Make two stops using 2 pieces of tubing and a tee on each. Put these stops on output 2 and exhaust 4. When you have this setup. Power the system and check for leaks, they do make a difference. When there are no leaks switch the memory solenoid either direction it needs to go using the two manual buttons on top. The solenoid will switch but the cyclinder won't move. Then using a finger or a screw-driver, switch the manual switch of the curved solenoid and let go. The cyclinder should go in the direction the memory solenoid selects and then stop when you let off of the manual switch of the curved solenoid. Once you get this to work and see what need to be activated or unactivated to move the cyclinder in this way and program that into a single button, multiple buttons, and single switch or a series of switches. Whatever works best. If you have questions just ask. Again the diagram is the exact one we used last year, look at the robot side and you can see the controlled cyclinder. If I find any pictures of our pneumatic board from last year I will be sure to post it. Good Luck!

[Chris Fultz]

Some follow-up /

First, this does work and we have tried it out on our 2003 robot. very controllable. One thing, you have to cycle full up and down a few times to charge up the whole system with compressed air. Then it is stable.

While experimenting, dial in the fittings to be very restrictive - it wil slow everything down and it is easier to see the efects of what you are doing.

Also, I added a pressure gage in the line. It never went above 60 psi. I don't think that is right, so I am going to change around the gages and see if it is just gage error. The math says it should be > 60, but maybe tiny leaks and volume keep it from reaching that high.

We used 2 switches. On double action to the the dual solenoid. selecting left or right will define the direction the cylinder travels. The second switch is a momentary on that is connected to the single solenoid. Pulses on this switch vent the cylinder and it moves in the direction set by switch 1.

Our programmers are trying to set up the control to do this. We are going to try some limit switches or something and see if we can stop it automatically with the control.

I have a power point slide that we used tonight in our design review. If someone can tell me how to add it to a post I will insert it. It makes it pretty clear and easy to convince the non-believers.

Like I said in an earlier post - this is awesome and thanks for sharing the idea.

[ngreen]

Chris,

Maybe you can add the powerpoint to the whitepapers and put a link from the post. That would be great. Good luck and be sure to mention anything new you find.

Thanks,

[Leo M]

Mr. Green, you are quite correct. I completely mis-interpreted the issue. Bottling up the pressure by plugging a solenoid vent port and then pushing on the cylinder will indeed raise the pressure beyond 60 psig. The regulator would never see it, being isolated from that section of the system.

Further proof - if more were needed! - that the world is a better place when FIRST and I pursue divergent pathways. I shall return to obscurity and live by the strict abstinence principle, henceforth eschewing all things robotic.

[Biff]

This is wrong. With this you could definitely do this. You can program the limit switch to change the single solenoid to its stopped state and stop the cyclinder. Yes, this would work. Any questions about set-up of the pneumatic parts refer to my diagram or ask me. The pistons will stop if you do it right, thus multi-positioning.
We are both right to a point, I took the original question as to just cutting off the power to the "memory" solenoid. If that was the case I was correct in saying that it would nor work. You are correct if it is a single action soleniod working with the "memory" soleniod. All you are using the switch for is to let the program know when to cut the air at the single solenoid. Peace.

[Chris Fultz]

"White paper" posted in the technical section.

It is two pages - one text and one picture.

[sirbleedsalot]

which page is it on

[FotoPlasma]

which page is it on
http://www.chiefdelphi.com/forums/papers.php?s=&action=single&paperid=261

[sirbleedsalot]

Thanks

[edflyerssn007]

here's a corrected version of the schematic. Our skilled engineers (students) actually got working on our pnuematics board. includes relief valves to prevent overpressures on the exhaust side.
http://ndofwrldaswenoit.tripod.com/robotics/index.html

[jacob_dilles]

wow that is genus. i am amazed. speechless. perfect thread.

[Leo M]

Against my better judgment, I am going to try to ride this horse one more time.

I have been thinking about this problem, and in experiments with the solenoids and trying out the ideas in this thread we have determined that – exactly as the manufacturer states – the valves will not change state reliably unless there is 30 psig at the ‘P’ port on the valve. In trying to make this happen with the ‘P’ port connected to the exhausting side of the cylinder, the 30 psig is not always there, and the valve will not shift reliably.

In addition, we want to avoid any rule-breaking involving going over the 60 psig limit for working air pressure, and not have to use a cumbersome solution involving regulators or relief valves in several different parts of the pneumatic circuit.

It is possible to purchase three –position solenoid valves with the center position having all ports blocked, or having the outlet ports pressurized. However, there is some question in my mind about the legality of these valves – it appears to me, at least, that the pneumatics rules and the flow chart are not in precise agreement. We would prefer to avoid the discussion with the judges if at all possible.

I have come up with a compromise solution shown in the accompanying file. Using two double solenoids, it is possible to supply air to the ‘P’ port of each valve, plug the ‘B’ outlets, direct the ‘A’ outlet of one valve to the ‘extend’ port, and the ‘A’ outlet of the other valve to the ‘retract’ port on the cylinder. To extend the piston, shift the valve on the extend port to supply air to the cylinder, and the valve on the retract port to exhaust. The 60 psig supply air at the ‘P’ ports will ensure that the valves shift when commanded. To retract, reverse the positions of both valves. And, to stop in mid-position, shift the valves so that air is supplied to both ports on the cylinder.

In the event that a force is applied to the device moved by the piston, any increase in air pressure above 60 psig will be relieved by the main regulator. Otherwise, the cylinder will be more-or-less locked in place by the 60 psig on both sides of the piston. Push hard enough and it will move, of course, but it will be possible to stop in mid-position as desired, using unquestionably legal components. The 60 psig working air pressure limit will not be exceeded, and the use of extra regulators or relief valves is avoided.

[ngreen]

Leo,

If you applied the 60 psig to each side through both the solenoids wouldn't the cyclinder just continue to slowly extend since the forces won't equalize out?

Also if you shorten the tubing from the double exhaust to the single you should not have trouble getting 30 psig for port "P".

I believe reducing the regulator to ~55 psi will take care of the 60+ psi problem. You have to give up a little strength for control.

These are just my thoughts. Have you tested that circuit yet? Just curious to how well it worked.

[Leo M]

Unfortunately, I haven't been able to experiment as much as I would like to. This is just a concept. I thought I'd throw it out to the forum and see if anyone else could make something of it. There are a lot of other considerations in designing a system like this - what exactly is attached to the piston; how do static forces such as weight, or dynamic forces of parts in motion, affect the relative motion of the mechanism; how much time and money do you have to complete the system? The complicated nature is illustrated by the length of this thread. It holds some kind of fascination for us - I just know there is a way to solve this elegantly without violating rules, spending too much money, or needlessly complicating the circuit.

Maybe if we put check valves in the air supply lines to the solenoids???
Naaah - cuts off the regulator from the cylinder.

A good, hard problem - really fights back!

[Mr. Ivey]

I have actually worked with Leo on this one. It was going to be part of our bot this year, but the original use of the piston changed, so that now the idea of multi-positioning the piston is no longer needed. But back to the point, Leo and I have spent a good amount of time working on the concept. We have also found a way to make a piston stop in mid position where you want to just about every time you want it to, and it is pretty quickly at that. I'll give you guys a picture of it full out, full in and paused in mid throw. If there are any questions, PM me. I will say first off that you are best tho have the third solenoid to back pressure the system so that you get an quick responce from the solenoid that controls exaust from the main solenoid.
ivey

[edflyerssn007]

We built and tested our pnumatics circuit and its reliable. It performs the same way every time, we used the circuit in my last post. It is one of the cooler functions of our robot. Also the 60+psi overpressure issue never arose for us, so we didn't even need to include the second regulator.

[ngreen]

We built and tested our pnumatics circuit and its reliable. It performs the same way every time, we used the circuit in my last post. It is one of the cooler functions of our robot. Also the 60+psi overpressure issue never arose for us, so we didn't even need to include the second regulator.

This way works pretty reliable but I think the members of 384 are looking for a better solution for multipositioning since the rules allow more solenoids this year. Last year when we could only use a certain number of solenoids this was one of the few and definitely the easiest way to do this. It proved to be reliable for us. We never had any troubles with it other than general lack of use. Stacking wasn't a good strategy. If they have found a better solution to multiposition it would be good. I still stick with the way I posted is reliable, easy and generally legal. But the rule change has made things different so we'll see how else you can do things. I wish I had more involvemnet with my team this year so I could have tinkered with different pneumatic ideas I had. I'd like to try out there idea. Tinker with it a little. See if you can replace double solenoids with singles and how well it works compared to how well ours works. I like to see innovation but sometimes I like something reliable and easy.

[Jonathanb]

What we used this year were off the shelf items called pilot operated check valves. You attach two single solenoids to the check valves and the cylinder will hold any position you'd like it to :D

[jacob_dilles]

what i am playing with now with spare parts (we were going to do multi postition and then they bailed out on me because it looked to "complacated") is multi postition using the 3 valve system, with limit switches and the edu-controler. so far it works with 3 (all open and all closed not included) and its pretty good. there is a needle valve on the exaust valve, and i realy have to fine tune it because if its too far in it SLOW and if its too far out it overshoots by a lot (bad edu code i guess, im not using intruputs).

by next season im gonna have a nice display board and a white paper (WITH CODE). or so i plan...

[Leo M]

OK, I took another shot at it. I may well be working under restrictions that are excessive or unnecessary, but Hey - that's part of the fun. My last sketch would not work for the reasons pointed out by Mr. Green, and also because my line-up had a shunt between the two ports on the cylinder, so it wouldn't be locked in place. Here's another attempt - it adds a 60 psig regulator ahead of each solenoid valve. Now, the shunt line is eliminated, but the system is still protected from overpressure. This design does add a component, but regulators are reliable and (usually) trouble-free. It does give you the option of setting the pressure to the extend and retract ports separately, so any force imbalance due to area can be corrected by changing the pressure. At this point in the build, it is academic, but brains need exercise as much as muscles, especially as we get older (you'll find out). Besides, this is a good thread, and I hate to see it end.....

[ngreen]

OK, I took another shot at it.

Other than being a little excessive for the set out task. This could be a neat little way of making a super accurate 3rd position. Tweak the two incoming psig to determine exactly where the the third position would be. But as for multiple position (more than three) this wouldn't work because the forces would equalize each time to find that tweaked out third position.

[Leo M]

Hmmmm - I should think it would stop anywhere in the stroke, depending on the volume of air trapped on each side of the piston when both sides are pressurized, and the pressure settings on the regulators. Now, how precise the control would be - that would take an experiment. I sure wish I had the time and equipment to play with this some more!

[ngreen]

With two incoming constant pressures the volume has little factor. The two pressure will create forces that will either equalize somewhere or the force of one will overcome the other.

The previous system I used Close and opened on the exhaust of the double solenoid such that when it was open the cyclinder would extend or retract but when it closed why extending or retracting the cyclinder would continue to compress until the forces equalized out. It didn't depend on volumes although reducing the volume will create the system to react quicker.

[Mr. Ivey]

Just a thought, if you were to use two double solenoids to control the throw of the cylinder. Use port A on one for the throw out, and on the other solenoid use port B for in. You block the other ports, both exaust and output, then route the exaust of both solenoids to the third double solenoid, and that lock when you want to stop in mid throw. I think that should allow you to position anywhere you wish, and not have the pressure equalize, but stay in the desired position. I would test it, but I'm at home and I don't have a pnuematics system to test it with here.
mark

[Chris Fultz]

We are using one double solenoid and one single and can stop the cylinder at almost any location. It does not stop instantly but there is very little creep after the switch is released.

We have a three position rocker switch wired to both solenoids.

When pushed to the right, the switch opens the double solenoid in the extend direction and opens the vent on the single solenoid to allow the cylinder to move. When released, the rocker goes to the center position which closed the single solenoid and the cylinder stops.

When the switch is pushed to the left, the same things happen only the other side of the double solenoid opens to retract the cylinder.

By simply tapping the rocker, the cylinder can be stepped up or down to an infinite number of positions.

We originally thought we needed two separate switches, one to select the direction on the double solenoid, and one to release the exhaust air through the single - but the electronics team found a way to do it with one switch.

[Biff]

Earler in a thread (I can't find it now) someone mentioned doubles with "center off"
After asking where did they get such a thing, the reply was from their local SMC dealer. This got me going in the right direction. The double valves SMC gives First, (and then us) Have a part number SY3240-5LOU on them according to SMC ordering info; this is a 3000 series 2 positon double 24 volt (but they are labled 12 volt, I suspect this is why we get them) 300 mm lead wire and I am having touble figureing out what the O is, the U is with indicator light and surge suppressor (non-polar) What our local supplier has is a SY3340-6LOU The second three in the number is what tells you it's a 3 position closed center valve, the 6 indicates 12volts. We got two and they bolt right on the body with the fittings on them, they are not pulsing and they only will "pull in" with the correct polarity, red to positive. This solved our problem, we were having trouble with our piston action not being to go from down to up, once started down you had to go all the way down. I think this is do to the nature of pilot valves. Any way, on our pretend-a-bot programming table this valve works great and we got them at about $60.00 dollars as we didn't get the wires or the fitting body. We are also putting the flow control fittings on the ea and eb holes on the fitting body and can have different speeds for up and down. Hope this helps

[Mr. Ivey]

Unfortunately the multi-positioning solenoids are illegal accoring to the flow chart. But to ngreen's posts on the pressures equalizing, or one pressure overcomming the other and the piston continuing to move, depending on how you do it, I guess, this doesn't happen. The way you described is that you are connecting the exhaust of one double solenoid to the ouputs of another, and locking the exaust depending on what point you want it. The issue causing the creep in the piston is that you are not giving a "total cutoff", nor are you regulating pressure at all positions. The creep comming from not being able to cut everything off is because you are not switching the exaust to something that is pysically cut off, the pressure in port on a solenoid can sometimes allow exaust, that's why you may have the creep. The way to stop DEAD is by completly blocking all air through the exaust of one side of a solenoid. But even then you will have some drag, like a car slamming on breaks. The key is to eliminate as much of this drag as possible. That's what Leo and I did last night. We decided we would take the task on one more time. We used all legal parts, and we have reduced the drag to less than a half inch. But there is a down to every good thing, while reducing the drag, we have to slow the throw rate of the piston, so it moves somewhat slow. It's a way of running this that we had not expected, and in doing so we found two new ways from using 3 solenoids. Thus we have cut weight in the overall function. The first method requirments 2 double action solenoids, or 1 double action and 1 single and a piloted check valve, while the second requires 2 double solenoids and 2 check valves. I won't give a long drawn out explination, but I will include diagrams, showing the two ways.

[Biff]

[QUOTE=Mr. Ivey]Unfortunately the multi-positioning solenoids are illegal accoring to the flow chart.
Rule 54 clearly states "There is no limit to the number of solenoid valves, air cylinders, and connecting fittings you may use on your robot. They must, however, be “off the shelf” pneumatic devices rated by their manufacturers for pressure of at least 125psi." The flow chart also says.
Is it an air cylinder ordered
from the Custom Cylinder
Order Form or any previous
year’s pneumatic component
except storage tanks,
pressure switches or
pressure regulators

The except and the logic structure is very confusing, So hopefully the inspectors will go with the clear wording of <R54> and not the confusing flow, of the flow chart.
Chears
[Edit] Checking the updates (10) there is an updated flow chart rev A 2/12/2004 that is easer to understand and indicates; That off the shelf and 125 psi are a YES.
Thanks.

[Pat Roche]

I've been going throught this thread and at all the skimatics i've looked at, the soliniod that feeds the piston seems to always be a double solinoid. How is it possible to have a double solinoid in that position when a double solinoid has 4 ports (two to the piston, two for exuast/pressure)? Wouldnt it have to be a single solinoid?

-Pat

[ngreen]

I've been going throught this thread and at all the skimatics i've looked at, the soliniod that feeds the piston seems to always be a double solinoid. How is it possible to have a double solinoid in that position when a double solinoid has 4 ports (two to the piston, two for exuast/pressure)? Wouldnt it have to be a single solinoid?

-Pat

Actually it is a five port manifold. You forgot the input.

You can use the double because the exhaust ports converge at a T and continue to the input of the next solenoid.

This year the solenoid combination makes no difference. You can use two single solenoids, two doubles, or one of each and this will work if you program it right. It is a control system. Last year we were limited in the number and type of solenoids we could use and this combination worked perfectly for what I wanted it to do.

BTW- Last year's and some of this year's single solenoids have five ports too. The difference between the singles and doubles is how it switches states and doesn't depend on the number of ports. (I think you were thinking of this years festo's with only one exhaust. And yes, they can be used too)

[Pat Roche]

When I am refering to single solinoids i mean the ones we have recieved in the past. The only ones i have seen have been 5 port, 1 input 2 exhuast and two feeds (like the festos from 2002 & 2003). The only double solinoid I've seen though was only four ports. So i guess my confussion stems from where your inputing with the double solinoid.

Thanks

-Pat

[Mr. Ivey]

The double solenoids I have are all SMC, the are SY3000 Serries and an NVJ5243Y. That's about all i know right on hand.
ivey

[ngreen]

When I am refering to single solinoids i mean the ones we have recieved in the past. The only ones i have seen have been 5 port, 1 input 2 exhuast and two feeds (like the festos from 2002 & 2003). The only double solinoid I've seen though was only four ports. So i guess my confussion stems from where your inputing with the double solinoid.

Thanks

-Pat

All the double solenoids I have seen from FIRST are like the drawing Mr. Ivey posted in #82. They have 5 ports (P, A, B, EA, EB).

[Pat Roche]

I figured out what you where talking about. I couldn't remember without looking on a valve that the the two exhausts were together. I was also able to make the system work too.

Thanks,
-Pat

[Liz*599]

the way our team achieved three positions from each joint, as opposed to two, was by coupling two cylinders together and then only fireing them independently, which worked very well for us.

[FadyS.]

I've actually been looking for a way to do this for the past 2 years. I have a pretty firm grasp on pnuematics from the past 2 years, but I still fail to see how this could be done :rolleyes:

I've also found, that amazingly, pneumatics actually weigh less then a motor driven assembly. Our team has always seemed to have a weight problem, and I'm always looking for a way to incorporate pneumatics instead of motor driven for the fact that they weigh less and are actually more reliable, but I've always been under the assumption, with the materials first gives us, you can only open and close a cylinder in 2 positions (full and closed).

Our team has been successfully able to get a multi-positioning system working for our arm using the pneumatics hook up talked about here.

We use the following bit of code for the positioning system:


// Position Control System (PCS) of arm
if( p3_sw_aux1 == 1 )
{
// Open arm
relay5_fwd = 1;
relay5_rev = 0;
relay6_fwd = 1;
}
else if( p3_sw_aux2 == 1 )
{
// Close arm
relay5_fwd = 0;
relay5_rev = 1;
relay6_fwd = 1;
}
else
{
relay6_fwd = 0;
}

EDIT: Relay5 = double solenoid valve
Relay 6 = Single solenoid exhaust valve.

We use this setup along with the SMC speed controllers to slow down the opening and closing of the arm and stop it when we close the exhaust valve.

[Jeffrafa]

If having >60 psi is a concern (~ 66 psi), then an easy fix would be to use a rodless cylinder, this would make both sides have the same amount of pressure when it is stopped mid-stroke.

I dont think going a little over 60 is an issue because the rule is there to make shure it is safe, and all the cylinder's have a max pressure of 150, so 66 certinly is still safe, as well as you aren't (and couldnt) use it as working force, its just a result of moving the cylinder. You still only have 60psi running in your pneumatics system.

[Astronouth7303]

There are 3 simple ways of doing this: easy, easier, and easiest. But I can't say which is which. :)

Style 1: Use 2 solenoids: Direction and Brake. Direction is a 'double throw' it has 2 ins; connect normally. Brake is a single: use on exhaust, only use A or B.

Style 2: Use 2 solenoids: Up and Down. Both are singles. Tubing could be a problem. (Exhaust through closed solenoid). the basic Idea is that when you turn on Up, it goes up. Turn on Down, it goes down.

Style 3: Use 1 solenoid: a 'double pull, double throw'. Basiclly: when off, A and B are off. When on A side, A gets pressure, B is exhaust. On B side it's opposite: B gets pressure, A is exhaust.

That should be easy enough.

[Squirrelrock]

Currently, I am in the process of trying this idea out on our 1st year robot, (umm, 2000 I think) and I will try to use the desgin in the aattached thumbnail. Besides what's already been said, has anyone had any problems doing it this way?

Thanx
Squirrel

[greencactus3]

the way our team achieved three positions from each joint, as opposed to two, was by coupling two cylinders together and then only fireing them independently, which worked very well for us.
if you used two separate stroke cylinders, you would be able to get 4 positions. 0-0, 0-1, 1-0, and 1-1 with the same stroke cylinders, 1-0 amd 0-1 are the same so only3.

and by the way.. if you are only using a single acting cylinder (ex. spring loaded or bungee loaded or just plain gravity drivin back, mulit positioning is possible with two single solenoids. i have used this method and the "normal" double and single solenoids multipositioning both with no problems ever. and for people who still dont quite get how multipositioning works, think of when you presss against the force of the air in a cylinder (only try it with a small diameter cylinder running on low pressure or you may hurt yourself). by pushing in a rod that is pushed out by the air. with the right amount of force to the rod from your pushing, the rod will stay "inbetween" the full out and full in stroke. that is multipositioning. only you dont use your hands to push in the rod but something else, gravity, elastic, or air (pressure).


*edit* oh and remember, some solenoids wont work with zero pressure buildup in the input side. (sorry, can someone word that better?)

[Squirrelrock]

oh and remember, some solenoids wont work with zero pressure buildup in the input side. (sorry, can someone word that better?)

Maybe... Some solenoids need input pressure to work?

Anyways, this morning I was able to get the cylinder to creep to midpoint by telling it to stop a little bit above the midpoint and it equalizes slowly. Also, it I tell it to stop too late, then it creeps all the way down. Could this have something to do with the lever action of gravity on the extended arm? Could this be fixed by using a spring to act against the arm?

Thanx

[Jeffrafa]

Actually, all solenoids need imput pressure in order to function. Solenoid valves themselves work just like cylinders, they use air pressure in order to shift the flow of the air, so every solenoid valve needs at least 25 - 30 psi in order to work. Some, however, have an additional port in which to supply pressure for the solenoid to function separate from working off the imput pressure - these are common in systems that carry negative pressure (vacuum).

The easiest method for multi-positioning is simply using one dual solenoid, three position valve in which the center position is all ports blocked, with this you can make the cylinder move whichever direction you want by activating each solenoid, or you can make it stop anywhere by not activating either solenoid. We had this on our hook cylinder last year and it worked flawlessly.

I can draw up a schematic of it later today if you want me to.

[ngreen]

Jeffraffa scheme seems to be the best under 2004 rules (Since three position solenoids are available and legal). I see no reason why they would limit this availability this year so the three position solenoids are the right to for the right job.

squirrelrocks schematic is a modifyed version of what was allowed in 2003 rules. The pressure regulator is the additional part to regulate the exhaust gas pressure. To explain it's need in this design read earlier post.

As far as how it is reacting there are a lot of variable to consider. The first two are important for all pneumatic systems. 1. Keep all your tubes short. Especially exhaust side. This will cause the device to react quicker. 2. Check for leaks.

The weight of the arm will affect the how quickly it reacts. The stopping force must compensate for both the pressure force (PressurexArea) and the weight (Massxgravity). When you are stopped the sum of the forces=0. I would imagine the if you were lifting the arm it might go slower up and stop quicker as opposed to lowering when it would go down quicker and take more time to stop.

There will be some play in the cyclinder when you stop it but if you make sure the pneumatics are correct and use the right cylinder at a good speed it should work.

[Rickertsen2]

Not ALL valves are air piloted. There are quite a few companies that make valves with entirely magnetically actuated spools.

BTW i would highly caution against the above setup for anything with a dynamically varying load. I will also note that changing direction wastes alot of air and is generally somewhat unstable. If the cylinder will only be required to resist in one direction, there is a much more stable and air efficient version of this.
Use a spring return cylinder. If the assembly will return by itself (IE by the weight of gravity etc, the spring is counterproductive and not necessary).
Use a 5/3 double actuated vavle with a closed center as mentioned above. Plug port 2/A/output1 (different manufacturers call it different things. Plumb port 1/input/whatever to supply. Plumb port 4/B/output2/whatever to the port on the cylinder that corresponds to the direction that you want it to apply force in. To regulate extension speed put a flow regulator in series with the supply to the valve. To regulate retraction speed, screw a flow regulator into port 5/BE/exhaust2. you could also use unidirectional flow regualtors in series with the line from the valve to the cylinder. This setup will still drift when the load changes, but it uses MUCH less air and is MUCH more stable when changing directions.

Food for thought: The REAL way to accomplish multipositioning is with a proportional valve, a position sensor, and a feedback controller of some sort.
The position sensor, porportional valve and feedback controller are normally
electronic. This wouldn't be a problem except that FIRST tells us that we can only drive solenoids from spikes. Most porportional valves take either an analog input or have some sort of digital interface. I will give people a hint and tell you that there are a number of ways that are FIRST legal an to create a entirely analog entirely pneumatic feedback position control loop. WOOT for pneumatic amplifiers differential regulators etc! In fact, the first PID controllers were entirely pneumatic. First person to come up with a workable scheme wins a prize.. well not really but i would like to see what people come up with.

[Squirrelrock]

to fix the possible leaks, we (jdiwnab and I) retubed the whole system and replaced the old valves with Festos from last year. I also modified the design so that each Festo controlled one side of the cylinder, basically A goes to the cylinder, B is a blocked loop, P is pressure, EA is blocked, and EB is open to atmosphere. This is true for both valves. Why would the piston get stuck at midway and not move up or down from there? Have I inadvertently made it so that the air cannot exhaust itself from either side?

Thanx

Squirrel

[Squirrelrock]

Solved!
Whatever problems we had we fixed. It turns out that I had blocked the wrong port on the second valve (see thumbnail on my post, about eight back from this one) :yikes: . Thanx for the help!

Squirrel

[bombadier337]

After months of reading, I finally made my own account. Anyway, I'm from team 1541, we are a rookie team, but we are doing pretty good so far. We have some video of our testing of multi-position pneumatics at http://www.sartanyac1541.com . We used the SMC to plug into the actuator, and then the exhaust is routed into the festo valve which opens whenever the extend or retract button is being held down. It took a little fooling around to get it, but now we have everything running off one spike. :-)