Pneumatics are something that I thought I had a pretty good idea of how they worked, but as I’m reading more and more of the posts on this form I’m questioning if I know them as well I think I do. I’m considering proposing an arm design at tonight’s meeting that involves 3 pneumatic pistons. One for the base, another for the elbow, and the final one for the grabber. As I understood it, pneumatics work by controlling the releasing of pressurized air to manipulate movable pistons which when they become pressurized, expand.
But is the process controllable enough to be used on something that needs a great deal of prison and control, such as a joint in an arm? Can you acurately control the joint of the arm or would it only have one of two extremes? Please help.
In most circumstances, pneumatics can only be positioned at their extremes. There are ways to position them midway, But i would not recommend it for an arm. Motors would be a much better way to go
The easiest method is to have the joint move between mechanical stops that are just far enough apart to keep the piston of the pnuematic cylinder from bottoming out. This leaves basically two positions.
However a number of people from this forum have used different scemes for achieving control of intermediate positions. There were extensive discusions on this last year. If you search the archives for “multi-position” you should find them. I’ am not sure how precise the control is. My own experience is that the intermediate stop would be pretty imprecise. But if you use feedback from an encoder you might be able to improve that alot.
You may want to check out www.pneumaticsfirst.org and setup a little system and see what it does. It shouldn’t take any more than 20 minutes to get one cylinder and assorted parts hooked up. If you want to see even quicker than that, take the colored plugs out of the holes of the clyinder in the kit and place your mouth on one hole and blow. If nothing happens try the other end. Don’t inhale or lick your lips afterward because there is oil inside and may be a little on the outside.
I dis, and do agree with this statement. Motors do work well at controlling positioning. In fact that is all Sparky has ever used for controlling an arm. Very evident in Sparky 2, and 5. Both used in manipulation of 2x multiplier balls. There is one way that you can very effectively control the positioning of a pnuematic cylinder. The catch is that with using it as a contorler for an arm, you might have some issues with pressure. In my opinion, I find that both ways can be very productive ways of controlling an arm. And if you really want to use air, just marry 2 pistons together to get the desired mid-position stop you want, just like Moe did last year.
If you choose to use pneumatics to do this, there are a few good discussion threads from January and February 2004.
Also, some other sources -
www.cyberblue234.com / go to the multimedia and there is a video of pneumatic staging (ie starting and stopping mid-point on the travel).
also check the white papers, technical section, for a slide on the pneumatic circuit layout to do this staging. the title is “Controlling Pneumatics - Sketch”.
I am the author of the file upload if that helps the search. It was uploaded in January or February 2004.
Basically, you control the movement of the cylinder by controlling the exhaust of the solenoid valve. I believe the slide shows doing that with two switches - one to control the direction and one to start and stop the cylinder. We ended up finding a way to link the solenoids and control it with one 3 way switch (extend - hold - retract).
It is difficult to repeatedly go to the exact same stop point. Air is very compressible, so the cylinder might move a little. Most cylinders were more accurate in one direction - depending on how they were loaded.
I am not sure why more people do not know about this, but SMC offers a valve that looks just like the double solenoid but functions slightly differently.
The valve I am talking about has a center off/center pressure/center exhaust feature. Meaning you can have variable piston position using only one valve (not sure if you can get away with one relay quite yet).
The part number differs only slightly from the regular double solenoids. If you are interested, drop me a line. The only catch, currently (as of today) they are on a 4-6 week back/manufacturing order from SMC in japan as no one in the US has any in stock. This valve also fits on the manifold we used last year. (I should also note that those are not in stock either!) Argh
Pneumatic cylinders are very controlable with the use of the flow controls in the kit. They control the velocity that the cylinders move at. I would not suggest trying to use them in a situation where you need three or more positions. The first is all the way retracted and the second is all the way extended. Fred Hord
Lots of options for 3-position, closed-center valves. The hardest part is finding them in 12 VDC. Try Grainger Stock # 2G449, an ARO A312SD-012-D valve, about $114 in my older catalog. Or # 4TR82, about $100.
I didn’t see anything in McMaster-Carr, but just did a quick look.
You don’t have to use SMC - if they are backordered, go to another vendor (this is just like real engineering life; things you need are always unobtainable).
These valves are great additions to the legal parts - energize one solenoid to extend, energize the other to retract, and de-energize both to hold position.
cyberguy: i would recommend considering a lead screw for your design. the system would probably be heavier than a piston, but the functionality is very similar. the torque on the arm, or the force of the screw, would be close to if not more than that of a piston. it allows you have any position within its boundaries, and it can easily be encoded with a potentiometer. my team is planing to use one for the lower stage of our arm.
for those with experience with multi-position pistons: how firmly is the piston held at intermediate positions? the only way a normal piston stay rigid is by applying excessive force in one direction, mechanically locking it in place. if the piston is held midway between the extremes, wouldnt it act much like a pneumatic spring, or shock for that matter?
