hydraulic valve - opens on set pressure, closes no pressure

This is like an "unloader valve" (?) - which does exist - but with additional characteristics(?)

The need...

I've got a hypothetical on-paper hydraulic device.

For fatigue-testing

- while "the hydraulic cylinder is always bigger than the sample you are trying to test"

• has always meant a machine with a frame and parts distributed along a central axis, dwarfing the size of the sample it's testing

• it also means the sample will always fit *inside* the hydraulic cylinder which is testing it

I cycled up a high hill to get that inspiration, by the way, if you were wondering...

For fatigue testing samples - it has to tension and release millions of times.

If I had this valve I mention, you connect the cylinder directly to a pump - the higher its capacity the faster - more strokes per second - it will go - "strokes per second" - with "the valve" at the outlet, dumping the oil in the cylinder and flow of the pump for the time being back to the tank.

The set pressure of opening means you reach an aim maximum tension in the sample. That that valve stays fully open until the hydraulic pressure drops to (very near) zero completely unloads the sample to no load. The valve closes and the cycle repeats, etc.

Does such a valve exist?

There are computer-controlled systems with a pressure transducer and the "dump" valve opening on command. These are the "servo-hydraulic" systems which are familiar to many. That might be the option it would be necessary to use, in reality.

However - still curious if there is a stand-alone valve device which does what's wanted.

For accurate pressure control, the only thing I could think of was to use a balanced open-close valve (sliding "bobbin" ?) - but with one end pressurised by a "reference pressure system" with its own small pump, large accumulator and pressure relief valve returning to the tank. With the cylinder pressure routed to the other side of the "balanced valve". So when the cylinder pressure exceeds the reference pressure by only a small amount, the valve moves over to rapidly fully open a big dump line to tank. Then there has to be another mechanism / valve which only trips for the valve to return to closed when the cylinder pressure is about the same as atmospheric. If proven to work well, the almost constant pressure in the reference system could be taken as the peak pressure the cylinder reaches. That reference pressure is freely adjustable by turning the adjuster squeezing the spring on the relief valve of the "reference" system.

Thanks for considering.

Rich Smith

----------------------

As I understand it, you want a bistable valve with variable hysteresis between its opening and closing pressures.

I recently tricked up a relay for my solar panels that acts that way. Relays do anyway but aren't adjustable, I have it switching resistance in or out of series with the coil to set the pull-in and drop-out voltages. The reason is to protect digital meters from the voltage range just below their minimum supply requirement at dawn and dusk, where they operate strangely.

For your problem a second pilot cylinder could change the tension of the relief valve spring that opposes the pressure. You might need a small accumulator and restrictor orifice to delay the pressure change at the pilot to ensure the valve completes each operation instead of chattering between states.

The generic name for a bistable device with memory is "flip-flop".

The solution is easy with electrical control by relays and solenoid valves. You can either sense pressure or use time delays.

the example STOP/START circuit CR1 (ControlRelay1) is bistable, it remains in whichever state the last button press left it. The symbol that looks like a capacitor is a relay contact and the circles are relay coils, solenoids, motors, etc.

Look at "sequence valves."

For example:

Look at "sequence valves."

For example:

This sounds very much like the mechanism of a hydraulic shake-table driver, used for vibration testing of all kinds of equipment.

One manufacturer is Unholtz-Dickie. Look into their history, and patents assigned to them and their predecessors.

Like "Fluid-operated vibration test exciter" to John Dickie, patent US2773482A. This is basically a siren driving a shuttle piston back and forth. If the shuttle piston is prevented from moving, it will generate a cyclic stress. The addition of a dead weight to this allows the cyclic stress to ride atop a static stress.

Joe Gwinn

This sounds very much like the mechanism of a hydraulic shake-table driver, used for vibration testing of all kinds of equipment.

One manufacturer is Unholtz-Dickie. Look into their history, and patents assigned to them and their predecessors.

Like "Fluid-operated vibration test exciter" to John Dickie, patent US2773482A. This is basically a siren driving a shuttle piston back and forth. If the shuttle piston is prevented from moving, it will generate a cyclic stress. The addition of a dead weight to this allows the cyclic stress to ride atop a static stress.

Joe Gwinn

----------------------------

I considered an oscillator-based solution but didn't suggest it because they may require specialized instruments, dataloggers, digital storage oscilloscopes and spectrum analyzers, to test and debug. It's much easier to test a system that can be stopped or run slowly.

----------------------

If you want simple, a 3-port rotary valve would alternately let oil in and out of the cylinder. You could run it slowly until you determine how fast you can spin it and still attain full and minimum pressure. The valve could operate a revolution or cycle counter. If you can tolerate a little leakage it's simple enough for a model engineer to make. ( I have too many machining jobs for the sawmill and its gantry hoist in the queue to offer. )

The proper sequence valve may do it all. See Figure 2 on the page I pointed to. Connect the IN port to a tee at the cylinder's port; the OUT port resturns to tank. It'll act like a relief valve when the pressure reaches the preset, but unlike a normal relief, won't reclose until the pressure drops to a very low value.

Hi to everyone. You folk are amazing! I've been here on my own, and you come along knowing lots of solutions. I'm in information-overload with all the great leads you've given. Plus I've just done a week and a half of calculations on testing methods and really really need to give my head time-out.

You asked about this device. I'll share with you

's an idea for a metal fatigue test, particularly of welds.

So in this rig, the sample is inaccessible, inside the hydraulic cylinder immersed in the hydraulic fluid.

For the advantages you get, the disadvantages are "nothing". This test could be running for days to 10's of days, by the way. It's "high stakes".

Even where you can stop the test and approach the sample, you don't, in reality. Gets you nothing additional.

The test stops when the sample breaks...

Welding the sample to the "pistons" has more advantages than disadvantages (one would reckon). Small and stiff.

A "contraption"? Yes. However - so long as you do a "scattergun" approach, you have a significant likelihood that some try will work. You have to be very open-minded, because you can have "dead-certs" which don't work, and "shouldn't work" which delivers and then some. You can find the science was wrong, or incomplete, and by giving it a try anyway you find that out, as well as getting results.

I'll get my head around the info - the links and hints.

Like be back in a little while...

Best wishes all, Rich S

Ned - I thought this is it.

Then I realised (?) - the full flow of the pump will always be the minimum flowing through valve - which will defeat the closing action we are counting on? This device, the "kickdown valve", is for filling say a hydraulic cylinder, where the flow comes to a definitive stop at full stroke? It avoid the energy loss of pumping oil past in-effect an "intermediate-pressure" relief valve.

I suspect that constant flow from the pump would defeat it ??

I'll try to do sketches.

I have thought of a circuit I believe would act quickly at the set pressure - giving the set pressure and no more. I ran with the idea of having "separate reference pressure system" where a small pump, large accumulator and pressure in it freely set via an adjustable pressure relief valve dumping back to the reference-system tank. I'll try to sketch that too.

Rich S

That's a "vibrophore", isn't it, if you apply that conept ot fatigue testing machines?

Electro-mechanical device.

You would always use one of these if you could, for the project I'm planning (?) Test rates to 150Hz and higher. Energy consumption so low many will plug into a "domestic" wall socket.

Never met one in real life. Would desperately like to. Idea of running a sample to 20 Million cycles no problem is like a dream come true.

But the problem is when you go beyond "research test samples" to testing representations of full-sized welds.

The biggest machines are 100Tonnes-force (1000kN; 1MN).

I've indicated the discussions would get very favourable if a 250kN (25Tonne-force) "vibrophore" were available.

The rig I've sketched is for if say you needed to test a weld to hundreds of tonnes cyclic stress range.

Rich S

The proper sequence valve may do it all. See Figure 2 on the page I pointed to. Connect the IN port to a tee at the cylinder's port; the OUT port resturns to tank. It'll act like a relief valve when the pressure reaches the preset, but unlike a normal relief, won't reclose until the pressure drops to a very low value.

So in this rig, the sample is inaccessible, inside the hydraulic cylinder immersed in the hydraulic fluid.

For the advantages you get, the disadvantages are "nothing". This test could be running for days to 10's of days, by the way. It's "high stakes".

---------------------------

You might be able to significantly reduce the 80KW power demand if the pressure source is a small flywheel-driven piston built into the fixed end of the cylinder to minimize flow friction loss, that absorbs the pressure energy as it retracts. You'd add oil/bleed air until a pressure sensor signal peaks at the desired pressure.

This could require some custom machining and knowledge of using an oscilloscope. Are those acceptable for you?

When I was in the automated testing business we had to figure times for high count operations such as testing each cell on a memory chip wafer, often in conversation without a calculator. There are 86,400 seconds in a day. 1 million seconds (cycles?) is 11.57 days. A micro-year is 31.5 seconds. 1 US billion (10^9) seconds is 31.7 years.

Exactly so.

I have already done these calculations for my "beam configuration" fatigue test. Static version of it - tensile test

The benefit you have there are the highly predictable Euler-Bernoulli beam calculations for long sections. I found one fixed stroke actuator would cover every need.

The problem with the "hydraulic inner fatigue test" is it is difficult to know with much accuracy how much hydraulic oil it is going to take per stroke to reach the intended force on the sample. The end pistons seem to be the design challenge. Flexing. Cylinder analysed by "hoop stress" no problem.

Also - adjusting it so you can arrive at different forces... Because you need to explore the shape of the fatigue "S-N curve".

That's why took thought of servo-hydraulic. With "catalog" equipment you could quickly get started. Connect it up and off you go. You'd probably use a pressure transducer and servo valves, wth digital logic linking them. In reality. As the equipment would be already there and familiar.

I've drawn the servo-hydraulic drive system. Mechanical logic though. Drawn as best I can. Sorry about any offence to familiar conventions.

There's two "bobbin" valves.

The big one dumps the main power system oil to the tank. It opens when the cylinder pressure exceeds the reference pressure. That cylinder oil comes through a check-valve, so once pressurised and vlave open, it stays open. Until... The other "bobbin" valve opens when the cylinder pressure is so low it cannot overcome a spring - that opening dumping the ex cylinder pressure holding open the main bobbin valve against the reference pressure - that reference pressure reopening the main valve.

All this mechanism could be on one "pallet" - reference system and power system adjacent, with one pipe to the test-rig atking back-and-forth flows.

Sorry that's "reference pressure re-closes the main valve" - to start the next cycle.

It would be good if this system could cycle very quickly and the bigger the pump, the faster it cycles.

[snip]

In the patent reference above, the oscillation cycle is controlled by an external "variable speed motor" of unspecified kind, designated is item 25 in the figures and accompanying text.

If the motor runs slow, so does the oscillation cycle, in direct proportion.

Joe Gwinn

there's other errors. Reference pressure must hold main dump valve closed, until cylinder pressure exceeds refernce pressure. Sure there will be other errors. Got to head off now. Thanks for everything Rich S

Joe Gwinn

---------------------- If the expected fatigue life is 2 million cycles, the test time at 1 cycle per second is three weeks. I think a good solution would be a closed loop based on a pressure sensor that shows when the high and low limit pressures have been reached, so the controller can switch between the fill and dump solenoid valves as rapidly as fluid flow permits.

If I had to build a prototype of the tester the controller would be an old laptop (or desktop) with a printer port, the data bits driving a successive-approximation A/D converter to measure the pressure sensor and two control bits operating the fill and dump solenoid valves.

QBasic running in DOS gives full unhindered access to all of the printer port bits for input and output, unlike Windows. An Arduino could also work but the laptop has the advantages of a huge hard drive to store data, the keyboard for control, and the LCD on which QBasic can display the cycle count and a graph of the pressure.

This simple resistor network outputs a voltage proportional to the binary code from the port bits: The other electronics are an analog comparator (LM311) driving a printer port status bit that tells if the sensor output voltage is more or less than the DAC output, and the two high current solenoid valve drivers.

Good question, but I don't think so, as long as the valve and the return piping are sized such that the pressure at the OUT port (at full flow) is low enough that the "light spring" in Fig 2 can force the spool closed.

In other words: the pressure rises to the set point; the "control relief poppet" opens, releasing the balancing pressure on the back side of the main spool; the spool shifts open, and the pressure at the IN port drops; the control poppet closes, but; the "kickdown jet" is now open and bleeds the balancing pressure from the back of the spool, until; delta P across the spool * spool area < spring force and the valve closes.

I hope that's right. Whether this is a practical way to control your device in the real world is another matter.