Looking for a strategy to decelerate a hydraulic cylinder

[Tubularfab]

I dusted off the ancient Autocad R14 and updated the hydraulic schematic a bit. I added the missing flow control valve for the return stroke, and fixed the position of the existing forward speed control. Then I added a solenoid and another flow valve for the low speed. I am not sure if the solenoid I showed would work - would it be too much pressure on the T port? The reason I drew it that way is because there happens to be an unused valve of that configuration already mounted in the machine.

 

[milldrone]

Aren't the timer constants actually 3.0 and 1.4 seconds as opposed to 30 and 14?

Opps! I got my PLC platforms mixed up!

Thinking about your constant displacement hydraulic pump, control situation and your comment below

this machine would pretty much never be used enough to justify a serious chunk of money spent on the controls. I would definitely be better off buying a modern production machine. I was just hoping I coule improve this one a bit for now

I believe the best path for improvement would be to:

1. Leave the Gemini connected, but add a high speed counter card (H2-citro) and parallel the wires. I'm presuming this can be done, I have done this before, but it might depend on the encoder's output power.
2. Add a second motor (1.5 hp) The motor starter would be controlled by the PLC, to drive a 1 gpm pump. A pair of check valves would also have to be added to keep the two pumps from rotating the other backwards or spilling to tank.
3. Use the existing optimate for programing two additional bend angles. Assign each bend angle to two separate push buttons.

The operation would be this:

If the operator wanted to use the Gemini he could by using the original controls, nothing would change.

If the operator wanted to use the optimate, a pair of bend angles would be programmed and each line would be assigned to two separate "Optimate designated bend angle buttons". When the first angle was desired, the operator would press that button and the auto bend cycle would start. When the counter card sensed the bend was near completion, the PLC would start the second motor and pump, and unload the main pump. The smaller pump displacement would "creep" the bend arm to the final angle. At this point the small pump would be turned off, and the retraction part of the cycle would start and the main pump would retract the system. When the second angle was desired, the operator would press that button and the auto bend cycle would start. When the counter card sensed the bend was near completion, the PLC would start the second motor and pump, and unload the main pump. The smaller pump displacement would "creep" the bend arm to the final angle. At this point the small pump would be turned off, and the retraction part of the cycle would start and the main pump would retract the system.

One drawback is that you would be using the friction and resistance of the tube to do the actual "decell". I'm guessing that if you were bending small tube at high speed, the transition point from high speed to creep would need to be much earlier than when you were doing large tube at the same speed. Another drawback would be if the encoder does not have the power to drive both the Gemini and the H2-citro. If it cannot then the Gemini could not be used.

This scheme would use only two extra inputs and one output. If you desired to change the sequence valves out and add limit switches and solenoid valves you would still have some inputs and outputs available. It would also only add one card to the PLC (only 1 available slot).

 

[Tubularfab]

Vaughn - that is an interesting approach to the problem, and it makes perfect sense. I'm just not sure it would function correctly in this application. Let me explain why I'm thinking it might not work. On this machine it's important that the flow be restricted a little bit to maintain the 1800psi line pressure in the mandrel and clamp circuits. If we dropped the main pump out and were running on the little pump it seems to me the pressure would drop to whatever it needed to overcome the bending moment on the tube. Since the mandrel and clamp are tee'd into the same circuit those would drop as well, and might overcome the force needed to hold thier position??? That is my layman thought - open to debate there!

I was digging through my stock pile of hydrualic junk and took a closer look at a hydraulic power unit I picked up a while back. It's 10hp and has a Parker Variable Volume Pump on it. It's rated at 33cc/rev which I come up to being about 15.7 gal/minute at 1800rpm? Would be nice to have it and a second smaller similar pump to replace the vane pumps!

BTW - I have an H2-CTRIO on the way. I also picked up a used 6" color C-more to play with.

 

[iant]

I dusted off the ancient Autocad R14 and updated the hydraulic schematic a bit. I added the missing flow control valve for the return stroke, and fixed the position of the existing forward speed control. Then I added a solenoid and another flow valve for the low speed. I am not sure if the solenoid I showed would work - would it be too much pressure on the T port? The reason I drew it that way is because there happens to be an unused valve of that configuration already mounted in the machine.

Earlier you wanted to work with valve #16
The dwg you have here is not in that position
please confirm your intentions as this is nowhere near your original statement
The drawing modifications seem to show more knowledge in Pnuematics than Hydraulics - does this seem right

Your new valve piping drawing is incorrect -

 

[iant]

This is the Speed Control system I was reffering to
using two of your existing units.

 

[Tubularfab]

Ian - what you drew was exactly what I was describing from the first post. The new drawing does look a good bit different as I found the original did not match the machine when I started tracing out the lines. So, the position of 16 moved to the A port of the cylinder. Also, the flow control #25 was on the machine but missing from the drawing.

I did have one major mistake in updating the drawing to match the machine - I had 25 in the wrong spot. The drawing below shows the fix. I also got the text to print ot the PDF this time.

I added the slow down flow control to the feed side of the system because it makes the plumbing easier. Also, it would still function properly in the double cylinder mode, and I'm not sure the other position would? Control valve #27 happens to be mounted in the machine, and of appropriate size, so that's why I drew it in. It's not what I would consider the perfect configuration, but could it work if I plugged the b and t ports? Or, would that overpressure the valve on the t side?

Peter - I googled "hydraulics forum" yesterday, and the first post I came across was the one where you argued the fact that resistance is not in the equation for pressure. Now I understand the comment you made in this thread about that argument...

 

[Peter Nachtwey]

Peter - I googled "hydraulics forum" yesterday, and the first post I came across was the one where you argued the fact that resistance is not in the equation for pressure. Now I understand the comment you made in this thread about that argument...

The reason I get so wound up is that often the hydraulic system is not designed correctly. The end customer never calls the hydraulic 'designers' first. They call us. For some reason it is always the controller's fault. Even when it isn't we still get called because they expect the controller to fix it. That puts a strain on our tech support guys and sometimes me.
In short, poor hydraulic designers waste our time/money and I resent it.
I haven't looked at that thread in a few days. I got to see if larryman1 has given up on finding a 'pressure is resistance to flow' equation.

 

[Tubularfab]

Peter - I only have a basic understanding of simple hydraulics. Most importantly, I have little concept of what can be done with more sophisticated systems. I understand your point that resistance is not part of the pressure equation. Pressure is simply a force acting over a certain area. But can I ask you about the theory of operation the builder of this machine presented?

He mentions that the flow control to the bend cylinder must be slightly reduced (restricted) to ensure that the system pressure reaches the 1800 that the relief is set for. The mandrel and clamp cylinders actually bottom out when in position, and need to maintain full pressure to counter the forces acting on them during the bend. Since pressure is equal throughout the system it makes sense to me that the flow control is needed to get the pressure up to 1800. Otherwise, it seems to me the pressure would top out at whatever pressure is required to overcome the bending moment on the tube factored by the motion ratio and area of the piston (and friction). So is that correct? How would the system function with a variable flow pump?

Most modern benders are using servos and planetary reductions to handle the bending force. Gives easy control over speed and position...

 

[iant]

I agree with you Peter,
Many a multi - dollar Hydraulic system has been bought down by incorrect pipe sizing. frequently return lines

 

[Peter Nachtwey]

He mentions that the flow control to the bend cylinder must be slightly reduced (restricted) to ensure that the system pressure reaches the 1800 that the relief is set for.

That doesn't make sense.

The mandrel and clamp cylinders actually bottom out when in position, and need to maintain full pressure to counter the forces acting on them during the bend.

Now I think the pressure is reduced so the bend doesn't happen too quickly and hit the end of the cylinder too hard. If all your bends go to a stop the why are you worrying about decelerating?

Since pressure is equal throughout the system it makes sense to me that the flow control is needed to get the pressure up to 1800. Otherwise, it seems to me the pressure would top out at whatever pressure is required to overcome the bending moment on the tube factored by the motion ratio and area of the piston (and friction). So is that correct?

Yes, good.

How would the system function with a variable flow pump?

I don't think a variable flow pump will help much except it would use less energy. A pressure compensate pump will try to maintain a pressure within the proportional band. When the cylinders dead head against the stop or the end of the cylinder the pressure will rise and the swash plate that controls how much oil the pistons will pump will go to a neutral position. The pump will not produce any more flow. This reduces the power requirements. The problem is that a pressure compensated pump responds very slowly to pressure changes which is why I like to use accumulators.

Most modern benders are using servos and planetary reductions to handle the bending force. Gives easy control over speed and position...

It probably depends on the size of pipe you are bending. Servo motors are not the best option on bigger systems. A servo motor must be sized to provide the max power required. A lot of current is required to provide holding torque. A hydraulic system requires little power to maintain force. A hydraulic system with an accumulator requires that the pump only needs to provide the average power for a cycle.

BTW, I responded to the thread on the hydraulic forum. I don't know where these idiots keep coming from. I have been refuting these jingles for 13 years now and new idiots keep appearing. I need to find out who the master idiot is that teaches the disciples.
I am through being polite after 13 years. They cost me money and give hydraulics a bad name.

 

[Tubularfab]

Peter - the bend cylinder does not bottom, all of the other cylinders do. The bend cylinder retracts to different lengths based on the required bend angle. (the bending work is done by the cylinder retracting and pulling the chain that wraps around the sprocket on the spindle) It actually works out that the cylinder retracts 1" per 10 degrees of bend angle.

Going back to your first response on the last post "that doesn't make sense"... With all three cylinders tied to an equal pressure, and two of them needing to be counteracting forces that require close to 1800psi and the third counteracting a variable force that requires less pressure how else could the system work? What I mean is, with the one cylinder not having enough force countering it, don't we need the flow control to raise the pressure for the other 2 cylinders? I'm not arguing against you on this; I'm just trying to understand it all better.

 

[milldrone]

With all three cylinders tied to an equal pressure, and two of them needing to be counteracting forces that require close to 1800psi and the third counteracting a variable force that requires less pressure how else could the system work? What I mean is, with the one cylinder not having enough force countering it, don't we need the flow control to raise the pressure for the other 2 cylinders? I'm not arguing against you on this; I'm just trying to understand it all better.

I'm going to make a suggestion that you do not really need 1800 psi at the clamp cylinder. But what you may actually need is a pressure that is XXX psi above the bend pressure. That said in your revised schematic the sequence valves are not drawn with a drain to tank on the sensing element. This implies that the downstream pressure (bend cylinder) is additive to the spring pressure on the sequence valve. Of course what I just mentioned is rubbish if you neglected to draw the vent to tank.