Temposonic overshooting

I have been beating my head against two tempos now for awhile.

The machine is a planer of which we are controlling the side head sizes with tempos /plc-5 QB modules and hyd cylinders ( one per head)
The right head has the tempo directly in the cylinder.

The left has the magnet mounted to the head and the tempo mounted to the right head.With its own cylinder.

Most of the time the system works ok.

However frequently one head or the other will oscillate around the setpoint sometimes more violently than others.(even with nothing but pump running) It is worse/better at different setpoints. Maybe due to piston surface area??

The reason the are coupled with the tempo is to provide head biasing where the operator only has to change one offset.Ie the left head will follow any offset added to the right head.

Past practice is to null the valve to calm the head. But at differnt setpoints nuliing is again required. This is a long lasting problem through different valves and tempos.

My questions are could this simply be a setup issue? As the QB/tempo should be able to maintain position even if the cylinder has drift RIGHT?

Is the PLC-5 fast enough to do what we are trying to do with the tempos coupled together or do we need a motion controller?

As you may have guessed Im no expert. We have been nulling the valve with signal. I thought you should null it at 0 signal but that leads to a greater oscillation.

Sorry for the long first post but I am befuddled and need the advice of someone with solid positioning experience.

The Temposonics are not the problem

dlmidget said:

However frequently one head or the other will oscillate around the setpoint sometimes more violently than others.(even with nothing but pump running) It is worse/better at different setpoints. Maybe due to piston surface area??

Click to expand...

It could be:
tuning
poor mechanical linkage on the left following cylinder
a valve with dead band.

The Tempo rods should BOTH be mounted in the cylinder so there is no mechanical slop. Also, is there any hose in the system between the valve and the cylinder?

A valve with dead band ( closed center spool ) will oscillate around the set point as the integrator winds up and winds down to move the spool across the dead band.

The reason the are coupled with the tempo is to provide head biasing where the operator only has to change one offset.Ie the left head will follow any offset added to the right head.

Click to expand...

This should be easy to do if the system doesn't have any play, the system is stiff and the valve doesn't have a dead band.

Past practice is to null the valve to calm the head.

Click to expand...

That just changes the offset of the spool. The integrator in the QB module should be able to compensate with the integrator. The spool offset is not a problem.

[quite]
But at differnt setpoints nuliing is again required. This is a long lasting problem through different valves and tempos.
[/quote]
That is because you aren't really fixing the problem.

My questions are could this simply be a setup issue? As the QB/tempo should be able to maintain position even if the cylinder has drift RIGHT?

Click to expand...

Yes, it could be a tuning issue, mechanical issue ( loose linkages ) or hydraulic issue ( hose between the valve and the cylinder or valve with dead band).

Is the PLC-5 fast enough to do what we are trying to do with the tempos coupled together or do we need a motion controller?

Click to expand...

The PLC-5 AND QB should be fast enough to do this easily.

As you may have guessed Im no expert. We have been nulling the valve with signal. I thought you should null it at 0 signal but that leads to a greater oscillation.

Click to expand...

Nulling just changes the offset of the spool. You have been focusing on something that isn't the problem and letting the real problem go un fixed.

BTW, the natural frequency of a cylinder does change through out the stroke so in theory the controller gains need to change through out the stroke but in practice one tunes the gains so the response is good throughout the stroke.

My bet is there are a couple of problems involved but my suspicion is the valve has a closed center spool ( dead band ). Can you post a link to the pdf file for the valve? Then we will know for sure. So often I see people using closed center valves for servo applications. Often the hydraulic distributor is totaly ignorant of what quality of valve is required for controlling a system too.

Thanks for your response Peter. Although I don't understand much of what you write I enjoy reading your posts!

The valve is a vickers KBS DG4V-3 92L 40 PE7 H7 10 I hope the attachment worked.

Tempos are MTS=GhT024.OUD602DE4
GHS0200UD602DE1
The one that isn't in the cylinder can be caused to move the head by applying pressure to the rod by hand.I have suggested moving the tempo to the cylinder and syncing the heads with the program but didn't have confidence that it will resolve all the issues.
We have tried increasing the deadband in the configuration for the axis but it doesn't seem to affect it.
Maybe we are going the wrong direction and shoud set to 0?

I have attatched a copy of the program as well. Lad 16 rung 14 is the BTW for the QB card.

dlmidget,
I'm not going to be of any help with your problem, but the PLC program you
posted is password protected. Few will be able to open it.

password=planer

http://hydraulics.eaton.com/products/pdfs/V-VLPO-MC005-E.pdf This is the link for the pdf for the valve. I edited the other posts because that valve was one we used to get by with for awhile. It had the same problem.

Thanks Mickey forgot about that.

The Tempo rods should BOTH be mounted in the cylinder so there is no mechanical slop. Also, is there any hose in the system between the valve and the cylinder?

Click to expand...

Yes there is about 18" of hose between the valve and the cylinder.
One thing I have noticed is that the frame of the machine where the cylinder is mounted will flex as the cylinder "hunts". I have requested for the millwrights to try and "beef it up" but that may take awhile to get accomplished.

The valve should be good

The 9 in 92L indicates the valve has a zero or underlapped spool which is good.

I would fix the mechanical flexing and any slop in the linkages first. Next I would reduce the length of hose as much as possible. The valve should be mounted ON the back of the cylinder when ever possible and if it isn't possible the the hydraulic/mechanical guys need to know they are crippling the machine. The frame must not flex, but that probably is not causing your oscillation problem but it will affect how smooth and consistent the widths of your boards are. It isn't like the side heads of a planer have that much force on them. It is now where near what chipper heads have in the primary break down or chipping edgers.

Go here and get a hydraulic design guide. It will point out a lot of DOs and DON'Ts regarding hydraulic control and the reasons. It is basically a reprint of many Hydraulic and Pneumatic articles so the text doesn't get too deep. Take your time and read it. Ask questions if you don't understand something.
http://www.deltamotion.com/other/designguide/register.php

Thanks Peter, I was actually watching your application videos on the Delta site last night. I have got to get one of those backyard waterslides!

The primary line and the curve gang all are more robust applications an hit their sets with minimal gap between cants. Which is why Im so confused as to why this simple setup cannot keep position without any wood running. However, it is the only instance where one tempo/servo valve is trying to keep position with another(mechanically tied) that may be changing. That is why I suspect tuning issues. I question the overall design.

For example, if one axis moves with greater velocity wouldn't it be hard to maintain a setpoint if they are constantly adjusting? Trying to catch a target faster than it is.

I already moved all wiring which was located with some ac lines and made sure all shields were only grounded at one end. I plan on putting a scope on the power supply tommorrow. We had an instance at the sawmill where a bad power supply made the double length and log turners jump.

Does anyone suggest any power supply checks besides voltage/wave. Such as, Is it possible fo a dc power supply to keep voltage regulated but current output be reduced? I believe that was the problem with the power supply in the sawmill example because voltage read okay. When everyting tried to set up the current wasn't their to move the spools. Thats my story anyway. However, a scope was not available to indicate noise.

Other than that I'll be reading, reading, and then i'll probally read some stuff. The longer it takes to fix this the more I'll learn!