I could use some help with my pneumatic automation

Hello all, I'm a bit of a newbie, so please forgive me in advanced.
I have a simple process, and need some advice.
The process and equipmentis...
Directlogic dl05, and a SMC double acting cylinder.
What I am trying to do is, apply 200# psi on a vertical stroke to a panel, 10,000 times, or until the cylinder breaks through.
I started by using a digital pressure regulator, set a predetermined pressure to cyl,( 2" cylinder, 64psi applied),
an upper(#1) and lower(#2) l.s. on cyl, that inputs to plc, input from upper l.s. sends output to sol valve, cylinder reaches lower l.s., which de-energizes sol valve.
That does not ensure full 200# of pressure, so I think I need a load cell, with module, also it does not allow my "break through" signal, which initially I intended to use a 3rd l.s(#3), which I have determined wont work, because as it hits the #2 l.s., cylinder automaticly retracts.
Is there some other sort of way of doing process without load cell....
Any advice would be greatly appreciated.
Thank you,
Motorgang

motorgang said:

Is there some other sort of way of doing process without load cell....
Any advice would be greatly appreciated.
Thank you,
Motorgang

Click to expand...

How about another cylinder partially filled with hydraulic oil. Stack the second cylinder with the air cylinder and connect the oil side of the cylinder to a hydraulic oil pressure switch. Now instead of reversing when you hit the end of the air cylinder, reverse when the pressure switch is made. You will have to figure out what the correct presuure setting on the oil cylinder pressure switch. Then you can use the end of stroke on the air cylinder to indicate when you have broke through.



.

How repeatable do you need to be?
+/-5 lbf? +/10lbf? That will determine to a large degree what kind of instrumentation you need to have and whether a pressure switch will be sufficient or whether you need a pressure transducer.

This sounds like a cyclic load testing application - you'll probably want a fast responding pressure transducer or load cell and the ability to log data if so

Another method would be, once the extended limit is made, continue to hold the output on for enough time for the pressure to complete its rise and such that, if the 'break through' were occuring, the third limit switch could be made. This will of necessity slow the process down but you wouldn't need an added pressure sensor.

OK so you are cyclic testing a piece of plastic to fail point and break thru. Break thru is defined as the piston protruding thru the broken plastic not just the plastic cracking.

ASSUMING you only want to know when break thru occurs
AND maybe stop or freeze cylinder in plase when it does.

1. Wire PLC such that one side of switch goes to air cylinder.
2. Other side of PLC switch is wired to a solid plate wire or whatever that is say 1/16" (2 mm?) beyond where the plastic normally flexes.

When cylinder breaks thru (creating a hole) it contacts the stationary plate completes circuit and PLC sets off warning bells lites freezes cylinder calls the Prez or whatever.

Another and maybe better way is a microswitch.

How fast is cycle? How many cylinder strokes until failure?

Remove extended limit switch (maybe) Extend cylinder slowly until air pressure is 200/piston area then ACCURATE pressure switch reverses cylinder?

Dan Bentler

Another option is to use the Pressure indication setup with a pressure transducer to reverse the cylinder motion. That way you can run much higher air pressure which allows the target pressure/force to be met quickly.

The transducer signal would be read very close to the cylinder port and you will probably have to read a pressure lower than required to allow for the air valve shift time that will cause pressure to override the transducer setting.

A Quick Exhaust on both cylinder ports would allow much faster cycling for shorter times for the test.

Limit switches would be placed at the Retract stroke for reversal and another at Break Through for end of test cycle.

I hope you have figured the amount of air required and have compressor output to do the test and still run the plant.

You could look at one of Peter's RMC70 motion controllers. They have special functions built in to do just what you want.

http://www.deltamotion.com


http://www.deltamotion.com/applications/appnotes/documents/prosthesis.pdf

Motorgang

I worked in a plastics test lab many, many moons ago, and we did similar tests. I am not questioning your test procedure, but merely mentioning that the speed at which you impact the test piece is an important factor. Flex tests are normally run at slow speeds where impact tests are fast. Two very different characteristics.

Having said that, if i were doing this, i would use a timer to control the cylinder stroke, and the top limit switch to end the test if failure occured, and a counter done bit to run the full test. The logic would be the same that is commonally used for a flashing light.

If you want to pursue this further email me.

The more I think about this the more I think you are going to need a load cell and the less I think a PLC will be fast enough. Unless you want to take 3 - 4 hours or so to run the test and don't really care that your test is repeatable. Maybe I am just thinking in terms of our hydraulic cyclic load testers that run 10,000 cycles in 9 minutes.

Heres the deal - the faster you try and go the more the inertia of the piston and the punch become a factor in the test. When you apply air to the cylinder its going to accelerate continuously towards the test sample. Not only that, but the acceleration rate will accelerate (refered to as jerk rate). Acceleration will continue until you hit the test sample or you reach a flow rate limitation, which will drop the force applied to the piston until it stops. If you try and do the test fast then the impact force alone could be more than 200 lbf even at pressures below 64 PSI, in which case a load cell and a fast instrument will tell you what the force was, but DL05 PLC and analog input will be way too slow to capture the impact event. Since air is highly compressible, If you go slowly then you have to wait for the pressure in the cylinder to build to 64 PSI after it stops to apply 200lbf. The nature of the stresses applied are very different from those applied by an impact force.

Without more information we can't be more specific, but I think you are going to have a very difficult time performing a consistent cyclic load test with pneumatics. In addition for a cyclic load testing you proably should be interested in displacement data as well as force data as well - but thats just a guess since all the information we have about the test sample is that it is a panel.

Of course if you are trying to do something entirley different and we are on the wrong track, let us know.

You might want to talk to these guys:
http://www.mts.com/stellent/groups/public/documents/library/dev_003990.pdf
or if you really are committed to a D.I.Y. project, at least look over their literature so you can get a better idea of what might be involved.

How about instead of hydraulic cylinder use a pneumatic diaphragm (valve actuator) This would have much less histeresis and might work faster.
Another possibility would be to use the diaphragm actuator with the housing connected to a mechanical crank on a motorized gearbox, The air in the actuator could be connected to a transmitter. Each rotation the crank would press the actuator shaft against the test piece compressing the air in diaphragm to equivalent of 200#. As soon as the air pressure spike failed to reach 200# (equivalent) the PLC would consider the test ended You would need some fine control over the pressure, perhaps a low pressure regulator that set the air pressure in diaphragm,
Roy

motorgang said:

What I am trying to do is, apply 200# psi on a vertical stroke to a panel,

Click to expand...

What are you really trying to do. Apply pressure or force?

That does not ensure full 200# of pressure

Click to expand...

Even if it did it doesn't mean you really know the true force. There is pressure on the other side of the piston. Their are only two ways to do this. One is to put a pressure transducer on each side.
NetForce=CapPressure*CapArea-RodPressure*RodArea

and....

, so I think I need a load cell,

Click to expand...

this is the other option.

with module, also it does not allow my "break through" signal, which initially I intended to use a 3rd l.s(#3), which I have determined wont work, because as it hits the #2 l.s., cylinder automaticly retracts.

Click to expand...

There are pneumatic cylinders that have slide pots built in.

Is there some other sort of way of doing process without load cell....

Click to expand...

The two pressure transducers.

What is important to know is how fast you intend to do these tests. Pneumatics will be slow compared to hydraulics or using electric cylinders.

Thanks for the input

A lot of great ideas and views, but let me be a bit more specific for those that have some questions.
The process, cycle limit of 10,000, at the rate of 14 cyles/min +/-6. The plant air is plenty sufficient to supply the demand. We are trying to ensure load of material, not impact.
I currently have meter valves set so the cycle time is 18 cycles/min, giving me a total process time of approx 9.5 hrs
The cylinder dia. is 2", and applying 63.47(64) psi, should give and applied force of 200#. On end of actuator rod, there is either a 8", or 16" disk,depending on test.
The ultimate purpose for this piece of equip, is for testing laminate tops, office related, with a thickness of 30mm. BIFMA/GSA standard, X5.5. Any desk/top must be able to withstand 200# load,(persons sitting on desk/top).
Hegewald Peschke makes the unit we are trying to reproduce. A photo can be seen at their website.
All desks that have adjustable height, must be tested, in mid adjusment range. So all desks vary in height, a micro switch that is placed under unit would have to be height adjustable, which can be made...but I was hoping for a cleaner look of machine...but that may be my option.
But i'm open to all ideas.
Thanks again,
Motorgang

Thanks for the additional information, that helps us help you. I had a look at the website and I noticed that they equipped their testers with load cells and integrated the load cells with the axis. You can't do this simply with limit switches. The force required to extend your air cylinder is mere ounces, and could in fact be negative if your test anvil that applies the force to the table top weighs enough. Since you have flow controls to govern the speed of the sir cylinder once the test anvil is in contact with the table top you then have to wait while air pressure builds in the cap end of the air cylinder and, this is important and people neglect it too often, the any remaining air pressure in the rod end of the air cylinder exhausts. As long as you have air flow into or out of the cylidner you have not achieved the desired force. The force applied by the air cylinder is

F = P_cA_p - P_r(A_p-A_r)
Where
P_c is the pressure on the cap end of the cylinder
A_p is the area of your piston
P_r is the pressure on the rod end of the cylinder
A_r is the cross section area of piston rod

So to know for sure if you have reached the force you need either
1)A load cell
or
2)Two (2) Pressure transducers (or two accurate switches) on both sides of the air cylinder.

Cylinder limit switches can then be used for retracted and over-stroke (breakthrough) position indication. If it were me I think I would be interested in measuring deflection as well so I would use a positon sensor on the cylinder instead of limit switches.

A alternate way to do this is to use a 200# weight to apply the force and a single acting air cylinder (better yet, one of Festo's pneumatic muscles) to pick it up and set it down gently. You still need a sensor to let you know that you have fully exhausted all the air from the air cylinder and that only gravity is applying the force.

Finally, what happens if your tester doesn't do what you think it is doing and something substandard gets out the door? What are the product liability issues if your test doesn't do what you think it is doing? Something to seriously think about if you want to "roll your own" tester.