Hydraulic Motion Control

Guest

Guest
G
This is off the topic of PLC's, but does anyone have experience using motion controllers with hydraulic servo valves? Specifically, we have a press which exerts 8000 tons of force @ a 3 degree toggle cycling 60-65 times a minute. We are flowing 45 gal/min @ 2200psi. The motion controller we are using is an (discontinued) Parker 6270 with a DY45 servo valve. There are two problems I am trying to solve:
1. On initial start-up (begining of the day) the heads tend to 'jump'
sometimes hard and sometimes not so hard, but there is no pattern
to the action.
2. When running, there is a 'fluttering' at the top of the stroke
before it comes to its position.

Like I said, I know this is off topic, but if anyone has an idea where I can find some information I would appreciate it. I can probably give some more information if it would help, this project is a couple of years old, and all of the designs/controls were done before I got here. I was just handed this project and told to 'fix it'. I have tried everything I know and could find.

Thanks in advance, the info and help on this site is outstanding, which is why I have posted this problem.
 
Opps, forgot to log in...

This is off the topic of PLC's, but does anyone have experience using motion controllers with hydraulic servo valves? Specifically, we have a press which exerts 8000 tons of force @ a 3 degree toggle cycling 60-65 times a minute. We are flowing 45 gal/min @ 2200psi. The motion controller we are using is an (discontinued) Parker 6270 with a DY45 servo valve. There are two problems I am trying to solve:
1. On initial start-up (begining of the day) the heads tend to 'jump'
sometimes hard and sometimes not so hard, but there is no pattern
to the action.
2. When running, there is a 'fluttering' at the top of the stroke
before it comes to its position.

Like I said, I know this is off topic, but if anyone has an idea where I can find some information I would appreciate it. I can probably give some more information if it would help, this project is a couple of years old, and all of the designs/controls were done before I got here. I was just handed this project and told to 'fix it'. I have tried everything I know and could find.

Thanks in advance, the info and help on this site is outstanding, which is why I have posted this problem.

Kieffer
 
Kieffer

Get hydraulic service check if you have air in the oil, and if you have nitrogen in the system may be the problem there.
Its not look like electric problem but I dont have all the info in front of me.try to avoid all hydraulic problems first.
 
From the servo side.....

Kieffer,

First of all, I do not have any experience with hydraulic servos. However, assuming your hydraulics themselves are sound (as previously mentioned) I should imagine the approach to tuning would be done in a similar fashion as that of an electric servo.

Since you have not indicated, I will layout a really general (and not wholey accurate) idea of what the three components of the servo loop.

Proportional Gain (P): Says how "hard" the system will try to get to the commanded position. The lower the setting, the more gently the servo will try.

Derivitive Gain (D): Says how to react to "upsets" in the system. If an external force were to dislodge your system from position, how will it try to get back. The larger the number the more agressively it will try.

Integral Time (I): Says how hard the system will try to "maintain" position. When the system is in steady state (rather than acceleration or deceleration) this says how hard the system will try to hold in exactly the position.

Before I suggest an approach to adjusting the PID loop, let me say it sounds to me like you Integral Time might be set just a little too high. The start-up thing could simply be a response due to a wee bit of drift after the system has been off for some time...one would certainly wonder if air is in the hydraulics if that is the case. In any event, at start up, the system trys to take control, sees an error and the I portion of the loop agressively works to get the system exactly in position.

By the same token, at the top of the stroke, when trying to hold the steady state position, the system may be working too hard to try to maintain that position...its is over- and undershooting until it finally settles out.

Now, when doing general tuning on a previously tuned system, I like to approach it first by noting my original settings (need to always be able to get back to where you started) and then take my integral Time all the way to zero. This should remove it from the system and leave only the P and I controlling the system.

Then I move the system and see how it reacts. If P and I are tuned well, the system should go to and from postion almost exactly the way you would want...but just a little bit off is ok. If it is having a hard time getting there, then either the P or the I needs adjustment.

At this point (assuming the PID setpoints utilize digital settings) I look to see if the D-gain is somewhat close to 10 times that of the P-gain. If so, it is probably in the ballpark. I Slowly increase this value until the system goes unstable. Once it does, I back off until it becomes stable again. This should be a good setting for the D-gain with respect to the P-gain.

If I cannot get the control I like with the D-gain, I set it back down to about 1/2 of it original setting. Its time to work on the P-gain. Like before, I Slowly increase the P-gain until the system goes unstable..I check this both during a move and while stationary. Once it does, I slowly back off until it becomes stable again.

Now its time to go back and do the D tuning followed by the I tuning again. Sometimes I find I have to repeat the process by setting the D-gain a little higher or lower and the readjusting the P-gain again.

When all is said and done, you will typically find the I-time is set very low and the D-gain is something like 10 times the P-gain.

I know others may do this differently that I (and I would love to hear their methods...alway nice to have other tools in your box) but that approach has worked well for me in the past.

Steve
 
One of my favorite topics

It is not clear to me what a 3 degree toggle is. I have seen toggles like in injection molding presses. A picture would help.

I will try to help with problem one while you are educating me on problem 2.

Problem 1 can be caused by the valve being out of null or not in a null position when the system is started. This means that the value should be put in a null position BEFORE APPLYING POWER. The close loop feature of the hydraulic controller should alway be disabled while hydraulic power is off. Otherwise the hydraulic controller will see a small error and try to correct the error even though it is impossible. After a while the integrator will wind up. If hydraulic power is applied then the system will jump in proportion to the integrator wind up. The motion contoller should also make the current position the set point on entering closed loop mode.

When the Parker controller starts up is in in openloop or closed loop mode?

Out-of-null valves and integrator wind up while hydraulic power is off are common problems.

Problem 2 is more difficult as a number of things can cause this. ASSUMING THAT THE SYSTEM DID WORK, THEN IT IS NOT THE TUNING PARAMETERS THAT CHANGED. It is the hydraulic or mechanical system that changed. I would first check to make sure the feed back is nothing loose mechanically. I would also output a voltage in openloop or manual mode to see how it reacts. I check for smooth motion and smooth indication from the feed back. Our controllers have a build in trend or graph feature that helps. I usually check to see how much signal it takes to make the actuator move in either direction. This checks the null and also makes sure the spool is free to move.

A valve with a dead band will appear to flutter. A dead band is a region around 0 output ( null ) where the valve does not respond or the actuator does not move. The signal builds over time and when the actuator does move, it moves too far. The actuator then repeats the process going back the other direction so it appears to flutter back and forth. It the problem is the dead band then this can be stopped by setting the integrator gain to 0. If the fluttering stops then the problem is the dead band. However, leaving the integrator off is not a solution. Fixing the valve is.

We get calls with these types of problems all the time. We we ask over and over again is what has changed? The motion controller does not change. The outputs or feedback devices may fail, but this will be easily apparent. Normally the problem is the oil temperature changes, the valve gets dirty, mechanical mounting of the feedback device is loose, something is jammed and will not move smoothly.... etc.

Steve, your tuning procedure is not very scientific. Would you like to try your hand at tuning a real hydraulic system?
 
I am using a similar system operating at 5000 psi... a far cry from 8000 tons (is that right?) but, still, a similar process except I use a "roll-yer-own" controller to regulate my pressures.

Potential problems include mechanical issues and controller issues.

After reading your description I wonder...
What is the Feed-Back method? Pressure? Or Position? Or is it Opened-Loop?

If you are using Pressure Feed-Back, then, your symptoms seem to indicate that you have air in the system.

Air is very "spongy" while oil is very firm - like a rock.

Hydraulic Seals don't react very well to air.

You say...
"On initial start-up (begining of the day) the heads tend to 'jump' sometimes hard and sometimes not so hard, but there is no pattern to the action."

That sure sounds like air... or, at least, not having a solid column of fluid. Maybe that depends on what position the device is left in over-night? As the device is cycled air is forced from the system and things then proceed smoothly... maybe.

Then you say...
"When running, there is a 'fluttering' at the top of the stroke
before it comes to its position."


When the device comes near to position (as indicated by pressure?) the controller is expecting an "oil-type" action. Air does not produce an "oil-type" action - it's too spongy. If there is still air in the system, the device will continue to move... just a bit... beyond the expected point. Maybe air is leaking past the seals and the controller is trying to maintain the pressure level?

If, on the other hand, you are using a Position Feed-Back... then the controller is suspect... or, at least, some of the operating parameters, set in the controller, are suspect.

However, the "jerking" at the start-up still sounds like it might be air. I have the same problem, occasionally, in my system and I know it is air in the system.

Are there any hydraulic flow controls on the valves?

You should fill-in a lot more about the system you have and how it is designed.
 
A tangent while waiting for more info.

Terry Woods said:
Air is very "spongy" while oil is very firm - like a rock.
I agree with the very spongy, but oil in a cylinder is similar to a very stiff spring. Each fluid has a bulk modulus that is proportional to how incompressible the fluid is. If oil wasn't compressible it would be impossible to control pressure. Controlling oil pressure is very difficult because a change in just one thousandths of an inch can change the difference in pressure across the piston by about 40 psi depending on how long the cylinder is and where the piston is in the cylinder. This means that pressure can change quite rapidly compared to position changes and the controller must be very fast or the system gain must be very low. In Terry's case the system gain is probably very low.

Some good reading about the spring effect of oil
 
Kieffer,

Does the system have a Blocking Valve? When a Blocking/Isolation Valve comes on that can cause the hydraulic system to jump from the sudden hit of pressure. I the valve is a pilot operated (2 stage) valve you might be able to add a flow control to regulate how fast the second stage comes on.

How about an accumulator on the system? What is the cylinder diameter and stroke. Does it operate in regen mode?

Help...Hydraulic systems I have work with had a check valve in the return line. Not a high set on the check valve, just to hold some oil in the tank line. WHY? (brain fade)

While checking the system how about the tank temperature at start up? Is the tank heater functioning properly?

Tank cooling? How does the heat exchanger look? How do the temperatures change through out the day?

Brian
 
Peter,

As I said, I "KNOW" I have an air problem in my system.

It only occurs when we do a shut-down and open the lines. After start-up, the air is forced out in three or four cycles and the system returns to normal - no bang for the next two-weeks.

My "Roll-yer-own" gain is pretty damned good. However, air ain't oil.

Kieffer has obviously opted to not give us the courtesy of a reply.
Oh, well. 'Sway it goes!
 
The last time I talked to you it sounded like you were still having a pressure control problem. I sounded to me like your pressure control valve was too big. Did you ever get that fixed?

If you mount the valve on top or higher than the cylinder then the air will escape on its own. Terry should know this already, but for for those that don't...

[tangent]
While we are waiting for a reply from, jokie...
Check out the July article of the month.
"The worlds largest motion platform" at Fluid Power Web

This shows how the wave motion for the new movie "Master and Commander" was done. You can see that this isn't moving around a model like in Titanic. This ship is life size. Russel Crowe was actually on this ship when it was moving around.
[/tangent]
 
Peter,

Regarding the pressure problem I described at the last Engineering Meeting...

They ran with 5000 +/-150 psi for almost 30 years. They simply didn't know any better. They thought (and still think) that the SLAM-BANG Method is the best and only method.

Then, during scheduled shutdowns, they wondered why they were spending so much time taping and gluing the damned machine back together. And why shutdowns were sometimes 2 or 3 times longer than scheduled.

Yes. I got it taken care of. I can now maintain 5000 psi +/-15psi using a common ON/OFF servo valve. I've managed to improve the control almost ten-fold.

The "hydraulic hammer" is now no more than an undulating "hydraulic hiss"... it sounds more like a varying air leak than any kind of hammer. These days, there is practically no time spent on repairing that system other than inspection and the scheduled re-builds on the valves.

Now I have only one major problem remaining. The output from the High Pressure Pumps goes directly to a large valve (we call it "the Spaceship"... it just sorta looks like an alien craft).

So, the output from the Spaceship valve can be directed to the main load, or the output can be directed to another valve which is then used to direct the HP Pump Output to either the Storage Tank or to the HP Accumulator.

The problem is the Spaceship Valve... you know, "Alien Technology" and all that...

Interestingly, the Spaceship is controlled by compressed air! How Alien can that be?

This valve is still operated using the Slam/Bang philosophy. This is necessary because of the way that the spindle is configured. Flow from the HP Pumps comes in the inlet port (common port) and is then directed to either Output Port-A or Output Port-B.

There is no problem while fully shifted to either port... the problem is during the shift. If not directed to Port-A or Port-B then the valve presents a Dead-Head to the HP Pumps! BURP!

As long as the valve is shifted very quickly (SLAM! BANG!), the pumps don't take too bad of a hit. However, if I try to eliminate the "hammer" associated with this valve action - by slowing down the valve - I then extend the Dead-Head time. BUURRRRRRP!

So... I need to eliminate the Dead-Head. I'm having the spindle re-shaped and re-chromed to eliminate the Dead-Head all together. When that is done, I'll have the flexibility I need to control the valve. Well... sorta...

I need to do one more thing. The Spaceship is controlled with air. The valve shifts from one extreme to the other. That is, as long as the Mac Valve is ON, the Spaceship shifts toward Port-A. While OFF, it shifts toward Port-B. There's no middle ground other than that occurring during the transition.

I want to be able to shift that valve to a particular position in the middle ground (using pressure feed-back) and hold it there. That is really hard to do with air - too spongy don't cha know?

So the next thing is to replace the ON/OFF Air Valve with a Double-Coil Hydraulic Servo Valve with Flow Controls.

Once that is done... the system will be "hammer-free". Hmmm... the operators might tend to fall asleep with all of the new found quiet.

Regarding the Air in my system...
All of the pumps, valves and accumulators are located about 60-Feet below the main load (a set of four SOLID rams, 30" diameter and 12-Feet tall). The rams stand upright with the driving end below the driven end.

When the plumbing is opened downstatirs, air will and does get into the system. After closing up, part of the recovery process causes air to move toward the rams. The air works its way up to the top of the cylinders near the top of the rams.

At the top of each cylinder is a set of packing. These are "hydraulically operated seals". That is, they rely on the presence of pressurized fluid to cause them to expand and create the seal.

Again, air is too spongy - and unpredictable in this case. So, when the cycle begins and pressure is applied... sometimes the packing seals quickly... and sometimes not so quickly - that produces a bang.

It only takes a couple of cycles to drive the air out of the system. From then on, it's good for the next two weeks.
 
thanks for the help thus far

st of all I want to thank everyone for the input, I had no idea I would get so much this fast. Secondly, my apologies for not responding faster, I was out this weekend...back to the topic.

I am attaching a schematic of the hydraulic system. A couple of things about it. It was designed before me and I created the drawing, my apologies if they are needed.

I am also attaching the diagram of the toggle arm in motion.

Reviewing some of the conversation thus far: When the system is off, ie. the pumps are off, I disable the axis on the motion controller. When the pumps are turned on I enable the axis. This brings me to a question about the null. I think that I may have been treating this incorrectly. While the system was on, I disabled the output to the servo valve and adjusted the null to where the head was stable. This was an arbritrary position, maybe I should investigate this a little more.
I did investigate the integral wind-up to solve the head jumping problem. I tried limiting the wind-up, limiting the output to the vavle and increasing the output over time to 'ease into' a start mode. I am going to investigate how the controller starts up --open-loop, or closed-loop.
I have not messed with the tuning too much, I have changed it for our jogging mode -- the acceleration and velocity are slow enough that I am pretty sure I am not going to 'hurt' anything.

Other information about the system which has been asked, we are using a linear trasducer for position of the cylinder(3.25" bore X 10" stroke). Also, the valve we are using is a DY90(not DY45) and we are flowing 90gpm not 45.

As I look into this more, does anyone have a good solution to removing air from a system?

Thanks again for all the help thus far, this is an R&D project which is 4 or 5 years in the making. I have only been here for a little over a year(which is about how long ago I gradutated -- so I am very green at this). The changes I have made(all program/configuration changes, none mechanical), have made some difference in the system, most for the better, it is still not perfect...and so the journey continues...

thanks

Kieffer
 

Similar Topics

I am in the metal fabricating business and have been using a pneumatically assisted hydraulic cylinder for a stamping trimming operation. This...
Replies
4
Views
2,215
This is a clip of some very well done hydraulic motion control: http://www.youtube.com/watch?v=ext_2d4C6ms&feature=youtu.be -rpoet
Replies
13
Views
10,402
Hoping our resident Delta expert will weigh in on this one.... :) Consider a somewhat hypothetical situation. I say hypothetical because the...
Replies
6
Views
3,386
Hey all, This my first year into plcs and programming. I am starting to learn how to control hydraulic cylinders. So basically a proportional...
Replies
5
Views
7,302
I am working on designing a machine that will have 4 axis of hydraulic motion control. All 4 axes are identical, but totally independent of each...
Replies
10
Views
6,620
Back
Top Bottom