A Filter Question for a Friday

[Peter Nachtwey]

Look ma Greek letters!!!

For what it's worth.

All of that and yet no specifications for the on and off times. On off times are critical for a a bang-bang positioning system. The faster Norm can turn on and off the valve the finer his control will be. Those valves are useless if it takes a 100 millisecond to open or close.

Turning the valve on and off quickly will cause pressure spikes. This is where that thread that Norm posted a link to is useful. One can determine the pressure spike as a function of flow and volume.

The rate of change in pressure is

dp/dt = BulkModulusOfOil*Flow/VolumeOfOil

I cover this in the upcoming Hydraulics and Pneumatics.

The pressure spikes can be calculated by multiplying each side by dt and making that the update period. Because we are talking in time increments dp and dt should be thought of as Δp and Δt.

Δp = β*Flow*Δt/VolumeOfOil

β is the bulk modulus of oil which is about 160,000 psi. Flow*Δt is the same as ΔVolume
You can see the units match.
All so, you can see that if in one time period the ΔVolume changes by just 0.1% the pressure will change by 160 psi. When this is multiplied by the area, which we don't know yet, the resulting change in force can cause a shudder in the system. Therefore Δt should be kept small.

I am still waiting to hear about the existence of a counter balance valve. I am not convinced that the volume change is the whole story.

 

[ndzied1]

I am still waiting to hear about the existence of a counter balance valve. I am not convinced that the volume change is the whole story.

Oops, thought I mentioned it in a previous post. No Counterbalance valve.

Didn't mean to put you on the defensive about your product. I should have said you could gold plate your RMC's but no one would pay for that even though it would look really cool.

 

[fluidpower1]

Peter;

You would love the circuit in Fluid Power Circuits Explained, REL4-7, that uses an N.C. Solenoid operated Relief Valve and a Pre-fill valve to get a fsat cycle with very low pump flow.

It is the 75 Ton Trim Press that had the 8" Bore, 7" Rod, 36' stroke cylinder and ran a 17 second cycle with a 13 GPM pump the best I remember. It had up to a 3/4" tonnage stroke.

Not my design but from a Niagra Press at a customers that I did an operations book on. The press was built in 1946 according the schematic.

No Motion Control valves but very smooth operatig at a fast speed.

 

[Peter Nachtwey]

Good no counter balance valve.

Oops, thought I mentioned it in a previous post. No Counterbalance valve.

I don't think it is the long column of oil on the top of the piston is giving you problems. I think it is the rate of change in pressure on the rod side that is giving you problems. As the actuator extends the volume on the rod side gets smaller and so the pressure rate gets bigger and bigger. I bet that reducing the Δt will help a lot. Ideally you want Δt to be infinitely small.

It would be nice to have a graph or trend of the pressures on either end of the cylinder as well as the positions.

Didn't mean to put you on the defensive about your product.

No problem, you can see I took full advantage of the opportunity.

 

[fluidpower1]

Peter wrote:
"I think it is the rate of change in pressure on the rod side that is giving you problems. As the actuator extends the volume on the rod side gets smaller and so the pressure rate gets bigger and bigger."

Is "bigger and bigger" an Engineering term when it applies to Pressure?

What makes the pressure increase on the rod side as the cylinder extends?

That is a new one to me. Mybe I can learn something.

 

[Peter Nachtwey]

Funny Bud

Peter wrote:
"I think it is the rate of change in pressure on the rod side that is giving you problems. As the actuator extends the volume on the rod side gets smaller and so the pressure rate gets bigger and bigger."

Is "bigger and bigger" an Engineering term when it applies to Pressure?

Yes as the volume get smaller and smaller.
I was trying to say that the rate of change in the rate is always increasing. See the link Norm posted. In the patchn thread you can see that the pressure increases at a increasing rate as the volume of oil is compressed.

 

[fluidpower1]

"As the actuator extends the volume on the rod side gets smaller and so the pressure rate gets bigger and bigger."


So, Do I hear you saying that a 10,000# wegiht hanging from a cylinder rod that has a 10 Sq.In. annulus area will show a higher pressure on the rod end port when it is 5" from the end of stroke than it did when it was 25" from the end of stroke?

Maybe I don't understand "Pressure Rate"??? I know the Pressure Gauge Reading on the rod port would not change if the weight was the same and the pressure on the cap end was the same.

Well maybe some added pressure from the extra weight of the added oil in the cap end. Is that what you are referring to?

 

[fluidpower1]

I think I finally got it Peter.

You mean the the time it takes to introduce the extra oil to compress the oil to a given pressure takes longer for a larger volume that for a smaller volume.

If that is what you meant, I just had a hard time getting and I agree fully.

Did a lot of Air-Oil hydraulic powered circuits in my career and always had to know the pressure and volume of the circuit to size an intensifiers high pressure ram to make sure pressure was reached before the ram bottomed out.

Another thing in those circuits that required extra volume is when they required Hose which can stretch a lot at pressure.

The 50" bore X 72" stroke press locally takes up to 65 seconds to reach 1,100 PSI at approximatelty 150 GPM when doing it's pressing cycle. However, about 60% of that is Piston Ring Bypass and product compressibility.

Took some time for it to soak in.

 

[Peter Nachtwey]

So, Do I hear you saying that a 10,000# wegiht hanging from a cylinder rod that has a 10 Sq.In. annulus area will show a higher pressure on the rod end port when it is 5" from the end of stroke than it did when it was 25" from the end of stroke?

No. The key word is pressure rate. Pressure doesn't change instantly.

Maybe I don't understand "Pressure Rate"???

The rate of change in pressure per unit of time. I usually use PSI/second or per millisecond.

I know the Pressure Gauge Reading on the rod port would not change if the weight was the same and the pressure on the cap end was the same.

Yes, but I am talking about the rate of change of pressure as a function of flow into or out of the volume of oil.

Well maybe some added pressure from the extra weight of the added oil in the cap end. Is that what you are referring to?

No, pressure doesn't change instantly. If takes a small period of time for pressures to change when a valve opens. In Norm's case when he opens the valve the pressure and force decrease but the induced load stays the same so there is a force imbalance and the load drops. Duh. When the valve shuts off the oil, the oil compresses due to the acceleration of gravity on load and the deceleration force. The two forces compress the oil to some point and then when the load has stopped decelerating there is only the load force so the oil springs back. I think this is intuitively obvious. What isn't obvious is whether this springing back would be noticeable only when exending. The only thing I can think of is that the rate of pressure change is faster there. This would cause the load to accelerate and decelerate a little faster there.

 

[fluidpower1]

"In Norm's case when he opens the valve the pressure and force decrease but the induced load stays the same so there is a force imbalance and the load drops."

Can I get a schematic of the circuit? I would also need a Material List wih Manufacturer and Part Numbers.

What is controlling the load from running away? It was stated there iis no Counter Balance valve.

Is the load the same throghout the lowering process?

 

[CharlesM]

When this is done manually people stand at each cylinder with a tape measure and radio their positions back to someone at a control panel with a toggle switch for each valve.

My bet is they just pulled the manual valves off and put on a solenoid valve. If I was to do that I might also add a sandwich flow control.

Can I get a schematic of the circuit? I would also need a Material List with Manufacturer and Part Numbers.

Good luck this is top secret stuff.

What is controlling the load from running away? It was stated there iis no Counter Balance valve.

Is the load the same throghout the lowering process?

Good question. If it had flow controls that would help with the run away.

I bet the load is coming on at the same time the position is jumping around.

 

[Tom Jenkins]

You can check if oil compressibility is the problem. The compressibility of hydraulic oil is aobut 1/2% for every 1000 psi pressure. You may have to consider both the volume of the cylinder and the volume trapped in the lines between the valve and the cylinder if that is significant.

Note that filtering may be counter productive, as it dampens the signal oscilations but also retards response.

Perhaps a servo valve that you could feather would work, or perhaps as you get near the end of stroke you could cut out the main valve and pull in a second lower flow valve that you use to feather the response. Not knowing the application you have to decide if that is feasible.

 

[fluidpower1]

"Take a look at page 62 of this companies offering of a Poppet design 4-Way valve that I believe would be much better. However the flow rate is not very high.
http://www.wandfluh.com/downloads/files/Web_Sitzv_e.pdf "

The reason I mentioned this design valve (Poppet) is it functions the same as a spool type but is more responsive.

Almost all Bang-Bang Spool valves all have Overlap and must move some distance (TIME) before any flow starts. And, almost all of them shift fully when energized even though a 10% shift would give ample flow. That means there is a pause between the Solenoid being energized and the oil flowing to start actuator movement and there will also be added delay when the solenoid is de-energized since the spool must move through some unnecessary travel before decelerating the actuator to a stop. These delays are very short and usually have no effect on a typical circuit, however, the delays can look like Days when rapid actuation is required.

The Wandfluh Poppet design is essentially Pilot Operated Check Valves so the instant they move there is flow, no lag time while coming out of overlap since all they have to do is raise off their seat. And, they only open as far as the flow pushes them so when it is time to stop the actuator the Poppet don't have to move through dead space which means they immediately start decelerating the actuator to a stop.

I have used these in applications where spool valves were too slow and got the actuator speed to the desired level.

That is why I offered this design Directional Control Valve as a possible option.

 

[ndzied1]

My bet is they just pulled the manual valves off and put on a solenoid valve. If I was to do that I might also add a sandwich flow control.

Valves have always been solenoid valves. No Manually operated valves. Same valves used in man controlled operation as automatic controlled operation

Good luck this is top secret stuff.

You just hate that don't you

I bet the load is coming on at the same time the position is jumping around.

Load is constant throughout the movement.

Note that filtering may be counter productive, as it dampens the signal oscillations but also retards response.

The process is very slow and inexact by the usual standards. I'm looking at super slow speeds and only trying to keep things pseudo-synchronized. If I can keep the cylinders within an inch of each other I'll be fine. The retarded response will more closely mimic what happens when an operator does the switching of the valves on and off.

At the beginning of the move, Things are kept to around 1/2" without problem.

I will post a schematic and some parts or at least specs when I get to a computer where I have access to the information. (less for Charles to complain about)

 

[CharlesM]

(less for Charles to complain about)

Sorry I don't mean to be that way. I am fighting some of my own top secret problems the last few weeks. So I know how hard it can be to ask for help but can't say what your doing. I have found that most people's top secret is others common knowledge.

I think it would be worth looking into putting a drive on the motor. A pump that small should have a small motor so a drive should be cheap. Then you could just open the valve and use a PID to control the speed of the motor controlling the flow of the pump. I have never done this type of thing but I know its done all the time.

http://www.plctalk.net/qanda/archive/index.php/t-25015.html

Load is constant throughout the movement.

If the load is constant then is the load keeping it from falling?