Amplifier problem.

[willpower100]

I was called in yesterday by a customer to look at a hydraulic proportional valve driven system. They have 2 non-feedback single stage proportional valves driving several cylinders. The input signal for the prop amplifiers comes from several analog adjustments on the console.

What is happening is this, when they shut off power to the system, ie, turn off the electric motor for the pump and associated input power to the 24Vdc power supply and proportional amps, the proportional valves are shifting slightly, the pump continues to produce flow for several revolutions and with the valves open slightly the hydraulic cylinders are jumping several inches.

When you remove the DIN plugs from the valves and repeat the shutdown, the valves do not shift, this then eliminates solely the hydraulics as the problem.

The valves and amps are basic older Parker models, output to the valve is PWM with a small AC dither signal, (Thats what I think anyways, as the product info on the Parker website is somewhat sketchy about exact output stage, they say something about pulsed so I assumed that is standard PWM). The amp gain is set correctly.

There does not appear to be any induced offset at the amp, the valves are A and B closed in the center envelope anyways and there is no creeping of the cylinders which would indicate that the valves are open a bit. The valves are overlapped spools, nothing very fancy.

What I am thinking is that there may be a line surge when the electric motor is shut down causing interference on the output stage of the amps or in the cables connected to the valve DIN plugs. I know that the caps in the 24Vdc power supply provide momentary power to the amps after the input is shut down until they decay. But without a input signal, nothing should move for that 1.5 sec.

A quick low tech solution would be to add solid state relays on the amp outputs to the valves, when the stop button is pressed it cuts out the contacts quickly on the relays therby isloating the proportional valve solenoids.

Any thoughts or suggestions.

[Peter Nachtwey]

I like your low tech solution. It is a must no matter how good the controller is. Normally there is an energized open valve that will close at the sign of any problems ( blocking valve ). In some cases this valve is a two positon valve with 3 ports so that the accumulator/supply oil is dumped to tank while shutting off the supply to the valve ( dump valve ). This is standard practice in the sawmill industry. You should know all about this.

Which way does the piston move? I wonder what is controlling the pulse width when the power is discharging? What if this part of the ciruit stops working and the pulse is left in the plus or minus direction while the supply is shutting down? I wouldn't blame the the power supply or line surges yet. I would look at the valve amplifier.

You should also consider posting your hydraulic qustions to the www.ifps.org forum.

[croakus]

A dump valve upstream of the prop valve should do it. If possible, the blocking valve should also dump both sides of the cylinder, otherwise pressure is trapped and if maint. needs to take a line off they're in for a rude surprise.
How is the drift started? If the cyl is currently moving, I could see some drift involved. But if the vlv is in the center position why is it moving in one particular direction? I would think both coils would be hit by any leakage. Is this a +/- 10VDC?

One thing - if the spool is shifted and power is dropped, including power to the amp, there will be no voltage/current to drive the spool to the center position, as it's not a typical spring center valve.

[willpower100]

Peter, I had thought about the blocking valve, there is no acccumulator in the system and the pump is a pcomp-load sense, the standby pressure is just a couple of hundred pounds, so I was not going to add one.

My concern is the amplifier, under no circumstances should it drive the vavle to move with no signal. This system has had this problem since it was installed years ago.

I suppose I could put my digital scope on the amp output and observe what is happening when it shuts down, wondered about what state the PWM is in on shut, either up or down, but it is always consistent, the cylinders extend an inch or better.

I can't drain both sides of the cylinders because the machine has to maintain position when shut down, the customer is forming aluminum into crude oil tankers. They may start and stop the system several times on one sheet of metal.

Thanks for the input.

[bwheat]

We had a simalar problem on an injeciton mold machine except the problem was when we turned the hydraulics off the controller kept trying to move the prop. valve. Every now and then when you turned the hydraulics back on WHAM!!! the mold would slam shut. It would shake the building.. Anyway they put a solid state relay that killed the 24 volts to the valve when the hydraulics were off.
Dont know why they just did not change the program to not command the valve to move but I think installing a relay will fix your problem.

[willpower100]

Bwheat, thanks for the input, that is what I am going to do.

[Peter Nachtwey]

Originally posted by bwheat
We had a simalar problem on an injeciton mold machine except the problem was when we turned the hydraulics off the controller kept trying to move the prop. valve. Every now and then when you turned the hydraulics back on WHAM!!! the mold would slam shut. It would shake the building.. Anyway they put a solid state relay that killed the 24 volts to the valve when the hydraulics were off.

This is not a good solution. It doesn't solve the problem with the controller output not being set to 0 while the hydraulics are disabled.

Originally posted by bwheat
Dont know why they just did not change the program to not command the valve to move but I think installing a relay will fix your problem.

I agree, 'they' should have disabled the PID when ever the controller is not in control. The prevents the PID integrator from winding up. This tell me that 'they' didn't have much experience at this.

[bwheat]

Peter Nachtwey,
I agree completely. Being new to the plant at that point and the machine was just installed and under warranty, I let the manufacturer decide what the best fix was. It was an AB processor, not a real big deal to change the code. Anyway, live and learn. I know what to do as I move on to the next one.

[Greg Gauper]

I reread the first post a few times. I could not tell if they are turning of the power supply (i.e. killing the AC power to the power supply) or are they killing the DC power to the amplifier via a relay contact?

The reason for asking is that if you switch off the AC voltage to the power supply that feeds the amplifier, you MIGHT experience unpredictable results as the power supply DC voltage decays below the correct operating voltage. A better solution would be to kill the DC voltage all at once. This would explain why it works correctly when you unplug the connector. I know of no commercially available amplifier that I would trust to hold the valve at null while the power decays!

Some proportional amplifiers (and some proportional valves w/built in electronics) have a discrete digital enable input to shut off the PWM circuit regardless of the analog command signal. If it does, you can leave the amplifier powered up and just switch the enable signal off when the pump(s) are shut down. Used in tandem with the existing scheme (shutting off power) this might also provide an acceptible solution since the amplifier MIGHT respond to the enable input dropping out a fraction of a second before the power decays and shut off the PWM circuit before the amplifier comparitor logic goes off into la la land....

[willpower100]

Ahh just what I thought was happening, without enable on the amp, I am left with isolating the outputs to the valve.

Thanks for the help.