Air Ram

Hi. I have an application where I need a ram to extend and then retract some 50mm but at a defined and repeatable acceleration and speed. Is there such a thing or do I need to go to a stepper driven system. The ram travel speed will be approx 3 metres per second.

Regards Alan Case

There are a few ways of doing this.

I have seen magnetically operated reed switches on cylinders which offer a high position accuracy. Although you need to have a magnetic cylinder plunger to operate the reed switch. Probably not the case if you are dealing with an existing installation.
Another way is to fit a linear potentiometer on to cylinder. This will give you a high accuracy on the position of your barrel.
On the subject of speed. If you are using what I call a “5 port solenoid-solenoid stay put valve” you can restrict the air of your outward movement and your inward movement by fitting regulators on the vent sides of your solenoid valve.
Of course in doing this the speed of your 50mm back stroke will be the same speed as your return to home back stroke.
If you want a two speed back stroke you need an extra on/off solenoid valve in the vent line in parallel with your regulator. So when you want a slow return the solenoid is off and the air returns to atmosphere through the regulator, and a quick return…. the solenoid is on.

This sort of problem has been mentioned in a previous posting, but I have run out of lunch break to find the link for you.

There are a lot of vendors promoting pneumatic position control these days. There was a thread on this forum not too long ago on the subject. Check the website of your favorite pneumatic components vendor to see what they have to offer. You won't match the performance of a servo or stepper or hydraulically actuated system, but if you can get adequate results with pneumatic, it will cost a lot less. Light loads and low accel/decel rates are good candidates for pneumatic actuation.

Don't overlook the fact that you may need to invest in additional equipment to condition your shop compressed air in order to use it in a positioning system.

TOL_O_MATIC offers a product called Percision Aire that boast a repeatability of +-.254mm. The system uses a magnetic partical proportional brake. You might take a look at it. Hope this helps.
RDAY

AC - ...extend and then retract some 50mm ...at... 3 metres per second.

50 / 3000 = 0.016 seconds for the whole thing to be done... with air!

Tell me I'm wrong but how about you just forget about it

Pierre. I am not after precision movement but repeatable speed. What I have is an air ram that at present extends to eject a product then it can have a dwell time and retract at any speed. The important thing is to extend at the same speed every time. At present the extend time is less than 0.012 seconds (the minimum value of time in the PLC for extend that reliably works). I use 0.015 seconds to give some lee way for low air pressure. What I am trying to do is eject a product onto a belt that is moving at 3 metres a second (max) so that the ejected product is ejected at the same speed as the belt and in the same direction so that it falls 20 mm onto the belt at a precise (within 3mm) position. I might be trying for too much with the system I have but if we never try it will never happen.
Regards Alan Case

3 Meters a second, Wow that is fast. You would need a cushion at the end of the cylinder or there is a company that makes a type of cylinder with two fow controls built into the cylinder, one for main speed control and one that controls the speed at the end of the stroke. They are used a lot in pick and place systems but I forget the name of the company, There are also inermediate stops available from a few manufacturer. I'm sorry Alan but the name of the company slips my mind right now. Another way I've seen is to use a short stroke cylinder as a stop and activate it after the main cylinder reaches it's stop, then the stronger short stroke cylinder will push it back the 50MM you need. Sorry I don't have time to help more, gotta go!

I think the only reliable way to accomplish this would be to mechanically couple the ram to the conveyor. You'll pull your hair out trying to get repeatable results with an air cylinder (Been there, done that.... Not much hair left on my head).

I'm guessing you're trying to put the product in a "pocket" on the belt? Just convert the rotary motion to linear motion with your favorite method, then enable this linear motion when you need to eject. This way, no matter what the belt speed, you match it every time. You can still retract the ram with air if you want.

I go through this battle with our engineers every time we build a cam machine, and then need to add another motion. Throwing in an air cylinder is always the cheapest, easiest way, but it never works.

Hopefully there's space available to accelerate the product before it contacts the conveyor. Accelerating from 0 to 3m/sec can't be done in 0 seconds you know! That's why we have acceleration lanes on the highway

You can do it with a servo/stepper by using the conveyor as the master speed reference, but I still think it would be best accomplished mechanically

beerchug

-Eric

I second Eric's suggestion. We accomplish many wonders with a simple technique. Here how it goes:

First, imagine an unlimited budget. Figure out all the precision devices you'll need to do this task. Then take them one by one and try and remove them from your system. In the end, being at synch will only be feasable with a mechanical gyzzmo.

But I work with Swiss engeneers. They always suprise me.

Ait is OK but does not repeat.

This is sad.

Second sanity check. I just can't see how one can drop a potato 20mm on a belt and expect it not to move, roll, or slide more than 3mm. Even if the belt was stopped it would be tough to keep a potato from rolling 3mm.

No one calculated the required acceleration rate.

Allen did not provide enough information. Not mass, speed, or distance. NO YOU DIDN'T. You didn't make it clear that 3m/s was the final or average speed and where within the 50mm the speed was to be achieved.

I can't believe the first three responders posted 'solutions' with out even knowing they were feasible. I would expect more out of Steve.

Pierre did well by doing a sanity check an pointing out that Allen's project wouldn't work pneumatically. Still Pierre had to make assumptions. He assumed that the average speed was 3m/s which was wrong.

Once Allen provided more, but still inadequate information, it was clear that Pierre's assertion was right even though his assumption about the speed was wrong.

I saw Allen's first post soon after it was posted. I quickly did a acceleration calculation assuming 3m/s was the peak speed and it would occur at 25mm after which the ram would slow down to 0 in the rest of the 50mm. The accleration rate I calculated was 180m/s^2 or 18gs. Accelerating anything at 180m/s^2, under control, would be challenging depending on the mass which Allen didn't provide. Since it is my policy not to guess or make assumptions, I decided to wait until others draw out the required information to make an intelligent decision.

I still can't make an intelligent decision and can only guess because I don't know the mass. I don't know if this event happens with every potato or just some. I STILL DON'T KNOW. I CAN ONLY GUESS.

This is why I suggested early, when the new board was created, that the ratings due to number of post was a bad idea. I wouldn't have given the first four posts any credit, in fact I would have subtracted from their post count.

This is sad. You might not like what I am saying, but you know I'm right.

Peter and every one else. I will have to continue this discussion in several weeks time as we are starting to get into areas that at present are confidential to my client.
What I will say is that I did not give full details as I was only after general ideas and did not expect you guys to do all the work for me. Regards Alan Case

Peter,

Potato?... What potato????

-Eric

Good Question.

You got me! Where did that come from? Oh well, substitute product, piece or part for potato. It doesn't change the intent of my post.

Got lost

I guess I was lost all along, the original info mentioned moving a ram 50mm ( aint that about 2 inches? Some of us US guys havent got the hang of metric) at a rate of 3 meters (thats roughly 10 feet to me) a second. The 3m per second got me. Where does that come in? Is it to match the speed of items passing by? Conveyance speed of items?
What were the size of items and the spacing/time between them?

I reckon a faux paus was made with the mention of taters, so we now wait patiently for the end results of this endeavor.

Which is more important?
Attaining a velocity of 3 meters/sec in .012 sec?
Attaining a velocity of 3 meters/sec in 50 mm?
or Moving 50 mm in .012 sec?

THEORETICALLY, you can have one and only one of those conditions!
YOU CAN NOT HAVE ALL OF THEM!

Having all three of these conditions is a physical impossibility - it's against the Laws of Physics!

"1 motion 1"


"1 motion 2"


"1 motion 3"


All of the preceding is dealing with the problem as if all theoretical conditions are perfect!

Essentially, the mass of the mover and the moved item needs to be insignificant - if so, then the inherent inertia problem becomes moot.

Then, it is also required that the moved object be in contact with the mover before the mover begins moving.

When the mover and the moved object is at the target speed, then the mover must either reduce speed immediately or maintain speed without continuing acceleration.

Basically, I see the problem as being insurmountable!! At least, with the curretly described method. There are way too many variables in the game!

My best shot at this would be to use the "Flying Cross-Cut Saw Method".

At a certain trip-point, the saw begins to track the "flying" piece of wood. At some point the speeds and position are seen to be in-sync. The saw then begins to do the cross-cut.

Coming into sync is one problem, doing the cross-cut is another.

Executing the cross-cut is nothing more than a simple trigonometry problem- speed-this-way vs. speed-that-way

Coming up to speed at the point where the cut can begin is a little tougher. The start-point has to be attained before a certain point.

It takes a certain amount of time to make the cross-cut. If the saw can't get into position before the drop-dead-point then a cut is not made.

The point being, if you know how long a cross-cut takes, and you know the physical length of the cross-cut framework, then you how much time you have to get into sync.

Different situations require the drop-dead-occurrance to be handled in different ways.

Now, back to the current issue...

Use a Servo to match speed and then drop the part, damned near exactly where it needs to be!

Use a servo to pick-up a part and then use that servo to match speed with the belt in such a manner that when the part is dropped it could easily be seen that the speed difference between the servo and the belt is zero (or damned near!).

It would be as if a pick-n-place grabber was simply picking up a part and then dropping it somewhere on a non-moving location.

The method I suggest is to try to hit a moving target from a moving and controllable platform. The target is moving at a fixed speed and the control of the "shooter" is variable by design!

This is quite different from a situation where the target is moving and the "shooter" is variable but not controllable!

I do have to say, however, this exercise has given me an idea for an accelerator device that should give reasonably consistant and reliable accelerations to items as suggested. Of course, I will stay within the Laws of Physics!