How do you control the length a piston will extention/retraction?

[Racer26]

actually, i think its the opposite of what you'd said... its easy enough to flip the one variable if its wrong.

[CyberWolf_22]

For those teams that have used center off solenoids in the past, what brand and model solenoids did you use?

[ngreen]

Just to reiterate, coarse control of piston length has been around for a while. Here is a post about the setup on 1108's 2003 robot with picture and diagram. http://www.chiefdelphi.com/forums/sh...2&postcount=11

The two solenoid system work for this. I've always thought that the center stop solenoid would be a more elegant solution but never got my hands on one and the rules have changed around some as far as outside part go (I don't know what is allowed this year).

The interesting challenge would be to do fine controls, especially under load, which I haven't done much exploring in.

[R.C.]

could you use this

http://www.alibaba.com/product-gs/20..._Five_Way.html

and could anyone give more info on how to stop a ram using a three way solenoid

[Daniel_LaFleur]

Originally Posted by rc_cola1323

could you use this

http://www.alibaba.com/product-gs/20..._Five_Way.html

In previous years (including last year) valves not supplied in the KoP were not legal to use, so these would not be legal (although this could change in the future)

Originally Posted by rc_cola1323

and could anyone give more info on how to stop a ram using a three way solenoid

I'm assuming that when you say ram you mean actuator (cylinder).

You would need 2 3-way valves on the primary side of the cylinder and seal the secondary side.
The first valve is used to add pressure or seal the primary side.
The second valve is used to release pressure or seal the primary side.

When you actuate the first valve pressure will build on the primary side of the cylinder and start moving the rod. as the rod moves the gas on the secondary side will be compressed and pressure will build. When you seal the first valve the rod will continue to move until the pressure on both the primary side and secondary side of the cylinder are near equal (difference being friction).

To retract the cylinder you'll just need to activate the second valve.

*Note* the pressure on the sealed secondary side needs to be above atmosphere when the cylinder is fully retracted or the cylinder will only partially retract when the second valve is actuated unless you assist the cylinder with some external spring source like surgical tubing

*Note2* you could add 2 more valves on the secondary side and reverse the actuation from the first side to get quicker (while less efficient) response.

[Jetweb]

I have always been a little cautious about the legality of magnets on robots, but there are multiple magnetic position sensors out there that with a 2-way solenoid would give fairly accurate position. it was one of the ideas we talked about using for this year that we decided we didn't need.


http://www.bimba.com/Products/PositionSensingSolutions/

[Joe Ross]

Originally Posted by Daniel_LaFleur

In previous years (including last year) valves not supplied in the KoP were not legal to use, so these would not be legal (although this could change in the future)

<R88> allowed other pneumatic valves last year. <R87> muddies the waters a little, but it was clarified in this Q/A post. http://forums.usfirst.org/showthread.php?t=8292

[Daniel_LaFleur]

Originally Posted by Joe Ross

<R88> allowed other pneumatic valves last year. <R87> muddies the waters a little, but it was clarified in this Q/A post. http://forums.usfirst.org/showthread.php?t=8292

I wish I saw that during the build season

Originally Posted by Jetweb

I have always been a little cautious about the legality of magnets on robots, but there are multiple magnetic position sensors out there that with a 2-way solenoid would give fairly accurate position. it was one of the ideas we talked about using for this year that we decided we didn't need.


http://www.bimba.com/Products/PositionSensingSolutions/

Or you could try a String encoder

[gblake]

Originally Posted by grosh

Can we control how far the piston on a pneumatic cylinder extends and contracts? We are using the double solenoid with a dual actuated cylinder. When we turn the relay on the piston extends to its FULL length. We tried turning the relay off before the piston was in it's fully extended position, but it didn't stop. Can you control the length the piston extends with programming?

If you simply want to shorten its travel so that only goes to a single location - You use something mechanical to stop it when it has gone that far (you put something sturdy enough in front of it... ).

I'll see if I can dig up a URL to an article I recently read in Machine deisn about the math behind doing this with high precision.

Blake

[jdejoannis]

Hi,

I would love to see sketches or photos of any of the multi-valve solutions. It would be fun to try them out in the lab to learn more about pneumatics fundamentals.

Here is a thought on the KoP. It seems to me that FIRST tries to use equipment that is relatively simple and general rather than highly specialized items. The basic elements of robotics - you might say. I guess it also depends on the cost of the items in question. Does anyone have a sense of the comparative costs of regular versus specialized pneumatic components?

-Jason

[kmcclary]

Originally Posted by grosh

Can we control how far the piston on a pneumatic cylinder extends and contracts? We are using the double solenoid with a dual actuated cylinder. When we turn the relay on the piston extends to its FULL length. We tried turning the relay off before the piston was in it's fully extended position, but it didn't stop. Can you control the length the piston extends with programming?

Again, not with the kit valves UNLESS you use ONE valve to plug the exhaust ports of your cylinder's directional control valve when commanded.

However, there are two things you need to do, to make this work cleanly: Get the correct TYPE of valve, and add feedback sensing.

1) Valve change - If you go to the SMC website, you can order a Center Off version of the same series of SMC dual coil valve we use, and substitute it. In this style, turning coil 1 ON (and coil 2 OFF) moves it one way, and turning coil 2 ON (and coil 1 OFF) moves it the other way. When both coils are OFF, it plugs both ends of the cylinder, and it "stops". Buy the valve itself, and use the 12V coils and the manifold from one of the kit valves.

This however is only half of the problem, as others have mentioned air is compressable. Change the load (or the direction of gravity), and the cylinder rod may move on you because after all, you are only balancing a PAIR of forces against the load (the air charge on either side of the disc inside).

2) Next, "close the loop" by adding feedback to directly (or INDIRECTLY) detect the position of the piston, and DO something about it if it is in the wrong place.

The simplest form with is to add sensing to the thing you are operating. If the cylinder is twisting a joint, add a potentiometer to that joint. It's value is the important thing anyway, and will have some direct functional relationship to the piston's position.

If you order the MAGNETIC version of the Bimba cylinders (which we ALWAYS do!), you can sense key positions with the supplied strap on reed switches. It's also theoretically possible to whip up a sensor on the outside of the cylinder and have the magnet act like a LVDT core, but I've yet to see anyone desparate enough to directly measure the cylinder to implement one. (Again, it's easier to measure the thing you are operating!)

You can also add a String Potentiometer, or a String Encoder alongside the cylinder's back end, and attach the string to the Clevis. This gives you a direct readout of the cylinder's position, but this is the EXPENSIVE way to go, if you try to buy a commercial device vs making one. (Yet again, you're often MUCH better off watching the device the cylinder is operating!)

In all cases, if your sensor shows that the cylinder/device setup has moved too far away from where you want it to be, simply pulse the appropriate valve to move it back. This kind of closed loop control automatically compensates for load variances, such as grabbing or releasing the scoring object for this year's game from a gripper.

Does this help?

- Keith