I have a couple of links to look at. I will write more on the weekend or whenever I have time

This is a .pdf print out of my Mathcad worksheet that I use to calculate natrual frequency.
ftp://ftp.deltacompsys.com/public/PDF/Mathcad%20-%20Natural%20Frequency.pdf

This is an excellent article on the bulk modulus of oil and air.
ftp://ftp.deltacompsys.com/public/PDF/SpringEffectEffBulkModl.pdf

This is an excel spread sheet that will allow you to calculate the natural frequncy
http://www.patchn.com/natfreq.xls

This will get you started and I can answer questions. The big one is "Good, now what does it mean and what do I do with it"

I will have to answer that latter. It is time to go to work.

Thanks for the information Peter. I, for one, will be looking at it and asking questions. However, I don't expect I will stop calling on you when a real circuit comes up.

I take responsibility for requesting this tutorial. Peter is well versed in this discipline and seems to enjoys instucting. Hopefully we can keep the post on a professional level and all learn from it. Maytag

Below is a link to some examples of how a system responses to changes. I show different combinations for systems with 1,2,5,10 Hz natural frequencies and how they respond to bang-bang and servo quality valves with slow ramps.

ftp://ftp.deltamotion.com/public/PDF/Natural%20Frequency%20and%20Ramps.pdf

I will try to clearn the document up a little more later.

Questions are desired because that will steer the direction of this thread.

Two things.

I noticed in the first pdf you posted that there is likely a typo. Should accumulator actually be actuator?

Second, have you ever experimented with the use of a Quincke tube to alter the natural frequncy in a circuit? How about a 1/4 wave dead end branch tube?

quote:

Originally posted by C1ay:
Two things.

I noticed in the first pdf you posted that there is likely a typo. Should accumulator actually be actuator?

Oops, you are right. I can fix this quickly. Now I know there are a couple of people paying attention.

I must confess I had to look up Quincke tube. Obviously I haven't used one and it doesn't look like the Quincke tube actually changes the natural frequency but rather supplies a longer or shorter path so that waves can cancel out. I don't see how this applies to a hydraulic actuator. The length of the Quincke tube would always have to change as a function of actuator load and position. I don't think this is practical. However, I am willing to learn if you have a link information on how a Quincke tube is used with a hydraulic actuator I would like to see it.

Back on topic. I am trying to show that valves that open or shut quickly will cause systems with low natural frequencies to oscillate whereas smooth slow moves with slower acceleration rates will not.

I am also trying to show how changes in natural frequency affect the response.

The examples shown above are not using any closed loop control or advanced control tricks except to generate the ramp. The ramp is generated using a cosine function which ramps at the frequency indicated.




Note, with a good motion controller will allow smooth motion with

Peter,
I followed your computations on that first link-the figures you gave for the oil trapped in the cylinder ends-i.e 10-this an aritrary figure I assume. You are correct "what does this mean and what do I do" Maytag

quote:

Originally posted by maytag:
Peter,
I followed your computations on that first link-the figures you gave for the oil trapped in the cylinder ends-i.e 10-this an aritrary figure I assume.

Yes, this is the oil that is trapped between the piston and the valve even if the piston is retracted or extended all the way. This volume is never 0. In this example I am assuming the valve is mounted on the cylinder so there is just a little oil that is trapped when the cylinder is retracted or extended. The trapped volume of oil is bigger on the rod side because there is a tube that runs from the valve to the rod end of the cylinder.

This is just one of those things that about the cylinder manufacturers that p!$$e$ me off. I am forced to 'guestimate' a value instead use the real value.

Now in your case it probably doesn't make much difference because the valves are mounted on a hydraulic manifold sometimes many feet from the cylinders. I have seen few if any cases where the steel or aluminum mill OEMs placed their valves in the best position. In these cases the mimumum trapped volume of oil is very high.

Some of the worst designs I have seen are in steel and alumimum plants yet this is where getting the design right is so important because of the huge masses involved.

C1ay, I fixed the error you found.

It should be noted that regen circuits lower the natural frquency of an actuator. This is because the trapped volume between the valve and the piston doesn't exist. So one must use the trapped volume of oil between the pump and the piston. If there is an accumulator in the circuit then you should use the bulk modulus of air in the rod side term in the natural freqncy equation.