Axis following

CharlesM

Member
Join Date
Aug 2005
Location
Arkansas
Posts
1,129
Plates.JPG


Cylinder A has an 6 inch stroke and the position sensor is inside the cylinder.

Cylinder B has an 8 inch stroke and the position sensor is attached to Plate A.

Plate B can be up to 2 inches above plate A. If cylinder B is holding position and cylinder A moves then cylinder B should move maintaining the distance from plate A to plate B.

Whats wrong with this setup?
 
The position control should be easy.

What keeps the plates A and B parallel? Isn't this more of a mechanical problem?

Is that another measuring device in the upper right hand corner tp make sure the plates remain parallel? I think you know what the problem is or is this a misery loves company sort of thing? I bet someone expects the control guy ( you ) to fix it.
 
What are you trying to do? in the correct application there may not be anything wrong with it.

With that said....why not have two position feed backs inside both cylinders then your given distance is stays the same

Can you change anything or is this what you have and you need to make it work?
 
Your drawing, if taken as shown, shows no attachment between cylinder B and plate B. Cylinder A may be taken as being attached to plate A. Which then means that cylinder A can retract plate A but cylinder B can't retract plate B.

But any real analysis would actually a LOT MORE INFORMATION than your drawing provides.
 
I didn't notice that cylinder B didn't have an internal MDT feedback.

The motion control just got a little more interesting. What you need to do is to calculate feed forward values for B and add them to B's bias. The feed forwards can be calculated by using cylinder A's velocity and acceleration and multiplying them by B's the feed forward gains.

geniusintraining said:
What are you trying to do? in the correct application there may not be anything wrong with it.
The feedback should be in both A and b cylinders for two reasons:
1. Plate A and B will vibrate causing noisy feedback and reducing the gains.
2. Axis A will not be able to use its own feed forwards. This will result in following errors unless CharlesM uses the trick above.

With that said....why not have two position feed backs inside both cylinders then your given distance is stays the same
Yes.

Can you change anything or is this what you have and you need to make it work?
Hopefully this can be changed but too often I am told the system is what it is.

CharlesM, don't you get to sanity check the designs to make sure they are controllable?
 
Is this a real situation or a thinking question for a Saturday?
Just a Friday night idea.

What keeps the plates A and B parallel? Isn't this more of a mechanical problem?

Lets assume that the plates are kept parallel.

Your drawing, if taken as shown, shows no attachment between cylinder B and plate B. Cylinder A may be taken as being attached to plate A. Which then means that cylinder A can retract plate A but cylinder B can't retract plate B.

Sorry cylinder B is attached to plate B. Its just something I did in paint this morning.

Hopefully this can be changed but too often I am told the system is what it is.

CharlesM, don't you get to sanity check the designs to make sure they are controllable?

Sanity Check? We tried that once, it didn't work.

This is just a out of the box idea. The feedback for B is attached to A and the magnet is on Plate B. The advantage to this is where ever A moves B will maintain the distance from A.

I see most of the problems you guys have come up with. The biggest is the tuning. Any error in A would be added to the error in B. I would think that B would osilate.

Putting the feedback into the cylinders will work. The trick is keeping A from pushing on B.

I think its important to think about he things you can't do and why. How many times have you not been able to do something. Then a couple of years down the road walking through a plant you see that someone else has it working.
 
Last edited:
Just because you can doesn't mean you should

CharlesM said:
I see most of the problems you guys have come up with. The biggest is the tuning. Any error in A would be added to the error in B. I would think that B would osilate.
There had better be a VERY GOOD reason for attempting to do this.
The B cylinder would just be responding to errors attempting to maintain a gap. Without feed fowards you are doomed. You would need to use the trick I suggested and that isn't as clean as putting the feedback transducers into both the A and B cylinders.

I think its important to think about he things you can't do and why. How many times have you not been able to do something. Then a couple of years down the road walking through a plant you see that someone else has it working.
But how well?

If you are just trying to keep two plattens separated by a constant distance while moving then don't do it. It is best that each cylinder has its own feedback in the cylinders. The motion controller can synchronize the axes easily this way. Both axes will be the same and setup and support will be easier.
 
I know how our mechanics would do this. Plate B needs to be distance X from plate A? Make up a block of size X, clamp it between plates B and A with a 'C' clamp and throw aaway that B cylinder. Problem solved.
 
I know how our mechanics would do this. Plate B needs to be distance X from plate A? Make up a block of size X, clamp it between plates B and A with a 'C' clamp and throw aaway that B cylinder. Problem solved.

The plates have to move. I could be 2" at the start but be 1" at the end of the cycle.

I may have missed this, but what is in the cylinders - air or oil ?

oil

If you are just trying to keep two plattens separated by a constant distance while moving then don't do it. It is best that each cylinder has its own feedback in the cylinders. The motion controller can synchronize the axes easily this way. Both axes will be the same and setup and support will be easier.

Thats how I have been doing it. But it gets more complex when you add a C & D platten. I'm just looking for other ideas.

Take a look at this press.
http://www.e-ci.com/products/pm_elec-hyd.html
 
Last edited:
CharlesM said:
Thats how I have been doing it. But it gets more complex when you add a C & D platen. I'm just looking for other ideas.
That is how they do particle board presses. Usually only one platen is controlled. The rest just float in between. The problem is that the material loaded in between the platens must be equally so the material between each platen compresses the same amount.

So what are you really trying to do? We have seen and controlled all sorts of presses.
 
Typical cycle all plattens move up (extend) to a position. Material is filled in opening. Upper ram moves down. All plattens push down at a set pressure to another position. Hold this position while upper ram pushes down to final position. Then retract plattens down to a position.

The middle step where the plattens push down and then hold a position is currently done with hard stops. These stops come in from the side while the cylinder is up then pull out before the cylinder retracts. Another option in the future will be making the cylinders large enough that they can generate enough force to hold this position. The problem with this is fitting it all in the space we have.
 

Similar Topics

Hello Experts, I'm wondering if this has been done before if possible to create an Emulate file that have motion control axis? I tried to...
Replies
0
Views
125
Hey guys, I got a carriage that can move with X and Y axe, each side got their own cylinder so left and right. It doesnt happen very often but...
Replies
0
Views
311
Hi PLCs.net! I was curious: How many axis can a B&R APC910 control? Is there a hard limit like certain AB controllers? I'm new to B&R, so would...
Replies
0
Views
349
Hi all! I am using unitronics Unistream plc, controlling two axis X and Y of polishing machine. Control is very simple as I have movement in each...
Replies
34
Views
3,633
Hello Dear Experts Is anyone knows where i can find some example to make some positionning function with S7-1200 Motion Control (v7.0) functions...
Replies
0
Views
441
Back
Top Bottom