OK, here is a headscratcher for you all. I believe that I have correctly diagnosed this problem, but I want to double check if you all would come to the same conclusion as I did.

Cylinder bore: 17”
Rod Diameter: 8”
Max operating pressure: 4,000 PSI
Max tonnage: 455 tons
Normal operating tonnage: 400 tons.

One year ago the cylinder was retrofitted with a servo valve for position and force control. The piston was replaced by a new piston modified to accept a magnetostrictive position transducer. To keep piping distances between the cylinder and the servo valve at an absolute minimum, a manifold is mounted directly on the cylinder blind end cap. The manifold ports directly into to the blind end cap through a 1” port with an o-ring between the manifold and the cap. The position sensor head mounts in the center of the cylinder cap, so the hydraulic port and manifold are to one side of the cylinder cap, centered along the radius line of the cylinder cap. The manifold also connects to the rod end of the cylinder via hard tubing.. Mounted to the manifold are a bang-bang J spool DCV valve with a P.O. check subplate and a directional control servo valve. Both control motion of the cylinder. The bang/bang valve was included for rapid coarse motion however it is rarely used and as it turns out not necessary, but it is there. The piston is always operated by the servo valve. The bang-bang valve is isolated from the servo valve by the P.O. check valve subplate between the bang-bang valve and the manifold. The manifold also has test ports, and pressure transducers for supply pressure, blind end pressure, and rod end pressure. The initial load on the rod is 80 tons. Operating tonnage is 400 tons.

The hydraulic cylinder operated with the expected precision for nearly a year. Then it began exhibiting anomalous behavior in that the piston would sometimes not move when it was first commanded to move. When the problem occurred the motion controller PID would wind up and fully open the servo valve. Valve spool position feedback confirmed that the valve is fully open and doing what it was commanded. Pump pressure was verified at 4,000 PSI. The pressure transducer for the blind end of the cylinder would show 4,000 PSI and the pressure transducer for the rod end would show 0 PSI, however the piston would not move. Pressure transducer readings were corroborated by pressure gages at the test ports. The piston would also not move if the bang-bang valve was activated. However, after about one minute the problem would apparently clear itself and the piston would begin moving and would complete its operation with its usual precision, apply full load to the process, and then retract for the next stroke. The problem would manifest itself randomly a few times a day, but over time the frequency of the problem increased, always with the problem self clearing after about a minute. The cylinder was disassembled and no mechanical damage or seal damage could be found. The seals were replaced and the newly rebuilt cylinder was put back in service. It continued to manifest the problem. An identical cylinder on a second machine began having similar problems. Then the first cylinder quit self clearing the problem; it stalled in the retracted position with full pressure applied to the piston, no pressure in the rod end, and it would remain in that position, unable to move, indefinitely until the cycle was aborted. But if the pre load was removed from the piston then it would follow its motion profile with the expected precision.

The servo valve, the magnetiostrictive positon sensor, the bang-bang valve, the hydraulic pump, and the pressure transducers are all fully functional. There is nothing wrong with the PLC logic or with the motion control card.

So, what do you all think was going on?

Can you post a schematic?
Maytag

Alaric wrote:
"The pressure transducer for the blind end of the cylinder would show 4,000 PSI and the pressure transducer for the rod end would show 0 PSI, however the piston would not move."

Without seeing the actual working circuit schematic to know exactly how the parts you mention are installed, such as, "The bang-bang valve is isolated from the servo valve by the P.O. check valve subplate between the bang-bang valve and the manifold." Also, the "J" Spool I believe is a Rexroth spool designation that has "P" Blocked, "A"" & "B" to "T" center condition but not positive that the valve is a Rexroth. The following is some possible faults that would cause the non moving cylinder.

One way this scenario would be possible is if the spool in the Servo Valve was broken and the portion blocking the Cylinder Rod End flow was moving and dropping pressure while the Cylinder Cap flow was still blocked and maintaining the 4,000 PSI pressure. I don't believe that is possible but after some of the stuff I've found on trouble shooting systems that were up to 40 years old I don't rule out anything.

The other way would be the circuit is performing as designeed but some mechanical block is taking place and stopping Cylinder motion. However, this scenario usually is evident from the press attempting to move by showing strain and some partial movement on the parts that would normally move.

Had a similar siituation on a press that had been running for 2 years and was performing in a similar manner as you describe. The cusomer had two identical presses since the pressing application was slow due to the operation and one press could not keep up with the production line. Before I was called maintenence had swapped every valve from the working press to the faulty one with absolutely no change in the symptom.
After doing some preliminary checking I advised them to check for a Hose that had a loose liner that was acting like a check valve and to replace any hoses that looked suspect. They actually found a 1/4" NPT Pipe Plug in the second line they removed that was left in the line from the manufacturer. It had rattled around in the plumbing until it was worn almost round. Removed the Plug, problem went away.
Wierd to say the least.

Here is a schematic of cylinder cap mounted manifold.



The servo valve is drawn in the power off position. When the valve OBE is powered up the spool moves to the closed center position when given 0 command.

Bud, I wondered about broken spools as well, except that the feedback from the servo valve says its moving and that the problem clears itself after a minute and then the cylinder precisely follows what it is supposed to do. The bang-bang valve stack and check subplate was replaced without results.

quote:

... I advised them to check for a Hose that had a loose liner that was acting like a check valve and to replace any hoses that looked suspect.

I would have never thought to check that, but in this case there aren't any hoses, its all direct mounted and hard plumbed.

When the press stops while operating the Servo Valve, what happens if you cycle the Bang-Bang valve?

Also, is the press cylinder bottomed out when it stops?

I've seen cylinders that are stopped by mechanical interference instead of the piston hitting bottom that put the mechanism in a bind and locked it up.

I have also seen cylinders that had bottomed out for so many cycles that the piston and end cap were mated so closely that no oil could get between them. One such situation always cycled when stopped mid-stroke but often would not cycle or would take a long time before moving when stopped on bottom.

I call it the Jo-Block effect from the fact that a new set of Johansen Blocks used to check Micrometer accuracy have to be separated by sliding them sideways if there is any oil on their faces.

One cylinder was a Miller that was the company I sold for. Miller always took a shallow spiral cut across the Head and Cap as the final operation when turning the Tube and Seal groove so there was always a path for oil to get between the Piston and End Caps.

I once had an 8 year old air cylinder that was sticking and when we took it apart the Cap End had a mirror image of the Spiral Groove on the Cylinder Head and the two parts fit very accuartely. Took a hand grinder and cut a cross from edge to edge across the piston and eliminated the problem of pausing.

Would you believe the problem got worse when the operators opened the Lubricator thinking it was a lubrication problem????

BTW if the cylinder does stop on bootom you can prove the Jo-Block sticking effect by stopping the cylinder early with a Limit Switch or some mechanical stop arrangement.

You're sneaky Alaric. Putting the same question on two forums at the same time.

Not much help from the electrical types though. Even stumped Peter.

That matches my conclusion Bud, which adds to my confidence that I have diagnosed it correctly - I have not disassembled the cylinder yet.

The original piston had a raised step about 6" in diameter and 1/4" high in the center section so that when it was fully retracted the port, which is off to the side, wouldn't be blocked. So we never had any problems before with the original piston. I scratched my head for a while over it. But then the light bulb went on and I took a fresh look at the drawings. It appears that when the manufacturer made the new piston bored for the position sensor the step was also made a little larger in diameter and it just barely overlaps the port. At first it didn't cause any problems because machining marks kept the surfaces from mating tightly, but I think that after a year of operating the piston surface and cap surface "lapped" together, making a nice metal to metal seal. For a while the metal to metal seal oozed a bit so that pressure built up enough to break the seal after about a minute and so the problem appeared to self correct. The operators actually let it run that way for about a month before they brought it to anyones attention. But eventaully the surfaces were polished enough to make a nice tight seal. If the port is blocked off then only about 1-1/2 tons of force can be generated across the cross section of the port and thats not enough to get anything to happen agaist the 80 ton preload.

To retract the piston I originally programmed the controls to just put the servo valve hard over and let the piston come all the way back. What I did for a temporary fix was program the controller to retract the piston to within about a sixteenth of an inc, and with that it functions exactly aS I would expect it. There is enough stroke to eject the part and insert the next charge but its not the long term solution I prefer. Next time it is taken down for maintenance I'm going to take a look to confirm my diagnosis first hand and then machine that step back.

This message has been edited. Last edited by: Alaric,

quote:

Not much help from the electrical types though.

I'm one of those electrical types that got stumped. It stumped me for a couple of days. I had all kinds of ideas that in retrospect might get me accused of insanity. That's why seeing others suggest the same thing at least tells me I'm not totally up in the night.

Alaric;
Have you looked at my basic book at the Ebook Link on the home page of this forum?

It is a book I used to teach Mechanical and Electrical Apprentices and Journeymen at a local ALCOA plant and other places. It's very basic but is a background for the follow-up Circuits and Trouble Shooting books.

The Electrical types got a lot out of it after they figured out NC and NO are reversed in Flud Power. I usually used Normally Passing and Normally Non-Passing for the Electricians.

Actually Elecrical types do more Fluid Power Trouble shooting than Mechanical's do.

Thanks, I'll take a look.
And thanks for the input on the stuck cylinder too.

quote:

Not much help from the electrical types though. Even stumped Peter.

I got the answer after a reading the posts. I just didn't post the answer here.

quote:

Originally posted by Peter Nachtwey:
[QUOTE]
Not much help from the electrical types though. Even stumped Peter.

I got the answer after a reading the posts. I just didn't post the answer here. It looks like we got the answer on the PLCS.net site before Alaric even posted the problem here.