What's new
What's new

Control loop for hydraulic piston pump

Strostkovy

Titanium
Joined
Oct 29, 2017
I would like to make an electric log splitter for my dad for Christmas. I'm choosing electric due to it only seeing occasional use and always being in range of an outlet. However, I don't like how they tend to be either slow or weak, so I am using a variable displacement pump. All components will be surplus where possible, and I can get an affordable pump that has manual displacement control. I will be using a pressure sensor and an actuator on the pump.

All of this will be controlled by a microcontroller, with an adjustment setting to keep the amps down to what is available. My concern is oscillation in the control loop. I can make a PID control loop, but tuning sounds terrible. What I'm thinking is that I can have aggressive settings for P and D to get the displacement to decrease quickly under load (I'd like no more than 115% peak load) and limit the slew rate of the actuator when increasing the displacement.

Does this seem like a sane approach to the control loop?

Mechanically the machine is quite simple to build.
 
I would like to make an electric log splitter for my dad for Christmas. I'm choosing electric due to it only seeing occasional use and always being in range of an outlet. However, I don't like how they tend to be either slow or weak, so I am using a variable displacement pump. All components will be surplus where possible, and I can get an affordable pump that has manual displacement control. I will be using a pressure sensor and an actuator on the pump.

All of this will be controlled by a microcontroller, with an adjustment setting to keep the amps down to what is available. My concern is oscillation in the control loop. I can make a PID control loop, but tuning sounds terrible. What I'm thinking is that I can have aggressive settings for P and D to get the displacement to decrease quickly under load (I'd like no more than 115% peak load) and limit the slew rate of the actuator when increasing the displacement.

Does this seem like a sane approach to the control loop?

Mechanically the machine is quite simple to build.

What's wrong with a simple gear pump, and a manual spool valve with return kick-off ?

Will the user need an App on their phone ?
Retina scan to start it ?
Web enabled for remote usage ?
 
What's wrong with a simple gear pump, and a manual spool valve with return kick-off ?

Will the user need an App on their phone ?
Retina scan to start it ?
Web enabled for remote usage ?

I only have 2 hp available. Either I have a 20 ton ram that moves at 1/2" per second all of the time, or I have a 5 ton wedge that moves at 2" per second.

I really want it to move at 3-4" per second to the log, go to 1-2" per second while splitting,(down to a 3/8" or less for truly stubborn logs) and go through 2-3" per second as it follows through.

Two stage pumps exist but the performance is not great compared to a true constant power hydraulic ram. It's not any harder to use, other than setting a dial to the available amps. (15 or 20 for the breaker, or maybe down to 12 on a long cord or generator).
 
I split +6 cords of wood per year, slow and steady. Your variable speed machine sounds like an invitation to get dad's fingers caught in the wrong place at the wrong time. Is saving a couple minutes per cord really worth the complication and risk?
 
I really want it to move at 3-4" per second to the log, go to 1-2" per second while splitting,(down to a 3/8" or less for truly stubborn logs) and go through 2-3" per second as it follows through.

I split wood too, and I think NRDock has a good point. But sounds to me that you could just sense the pressure and choose different setpoints for the displacement: if the pressure is above X, go slow, above Y, go really slow. The pump will be run with an AC motor? What kind of actuator? The D in PID is overrated for most applications. Just P will tell you whether your idea works, and is exciting all by itself.

I think you have an interesting idea.
 
I saw one a while back that split wood going both directions, set a log on the front and start it splitting, set a log on the back and retract splitting another, slow both ways but the wood piled up fast.
 
I only have 2 hp available. Either I have a 20 ton ram that moves at 1/2" per second all of the time, or I have a 5 ton wedge that moves at 2" per second.

I really want it to move at 3-4" per second to the log, go to 1-2" per second while splitting,(down to a 3/8" or less for truly stubborn logs) and go through 2-3" per second as it follows through.

Two stage pumps exist but the performance is not great compared to a true constant power hydraulic ram. It's not any harder to use, other than setting a dial to the available amps. (15 or 20 for the breaker, or maybe down to 12 on a long cord or generator).

2-stage pump, and prince sells a valve that allows regen for faster speed.
 
I split +6 cords of wood per year, slow and steady. Your variable speed machine sounds like an invitation to get dad's fingers caught in the wrong place at the wrong time. Is saving a couple minutes per cord really worth the complication and risk?

I'm not trying to make it super fast. I'm trying to match the capability of a gas splitter that does something like an 8 second cycle time on a 16" stroke. I just want it to be able to slow as needed to get high tonnage for our regrowth oak trees we have in California. (Basically it was all clear but in the gold rush and each tree is a bunch of new sprouts from the old stumps, all combined. They have gnarly knots in them.)

3-4" per second, from what I can gather, is typical log splitter speed. I will do more research to make sure of that.
 
I split wood too, and I think NRDock has a good point. But sounds to me that you could just sense the pressure and choose different setpoints for the displacement: if the pressure is above X, go slow, above Y, go really slow. The pump will be run with an AC motor? What kind of actuator? The D in PID is overrated for most applications. Just P will tell you whether your idea works, and is exciting all by itself.

I think you have an interesting idea.

Setpoints is a neat idea. Add enough hysteresis to prevent oscillation and it would be good to go. The motor will be a regular AC TEFC motor, and the actuator will be either a large spring return rotary solenoid (if the one I have can do it) or a servo, ideally still spring return.

There is still an open center direction control valve as the surplus pumps that do decent pressure don't reverse like a zero turn mower pump does. That and a relief valve and motor overload should allow for failsafe operation.

I should clarify, the goal is not to design a faster 20 ton press, the goal is to design a 5 ton press that can slow down to get 20 tons out of 2 horsepower.
 
3-4" per second, from what I can gather, is typical log splitter speed. I will do more research to make sure of that.
Ever run one ?

I have run several for neighbors (I have a gas well for my heat) for hours and hours.
Vertical, horizontal. gas, tractor PTO, electric.
On 2 of them we rigged jib cranes to swing in 5' diameter slices the tree trimmers gave us.
Cut down trees, cut up trees, dragged them out, hauled them in, on & on.....

4" a second is too fast.

Gonna call up Rockford Safety and equip it with light curtains and "double palms" ?
 
Ever run one ?

I have run several for neighbors (I have a gas well for my heat) for hours and hours.
Vertical, horizontal. gas, tractor PTO, electric.
On 2 of them we rigged jib cranes to swing in 5' diameter slices the tree trimmers gave us.
Cut down trees, cut up trees, dragged them out, hauled them in, on & on.....

4" a second is too fast.

Gonna call up Rockford Safety and equip it with light curtains and "double palms" ?

I have run several and from memory 4" per second seems right. On a 16" ram that's probably a 7 second cycle time. If 2 or 3 is normal though, even easier for me. But 1/2" per second is too slow and I want this to be strong enough my dad never has to use a wedge ever again.
 
With the pump I can get (15gpm fully stroked at 3600 rpm, 4000 psi max) and a 4" diameter 18" stroke cylinder up to 3000 psi with a 1800 rpm motor will travel at 2.3 ips at 2.5 tons, 1.2 ips at 5 tons, 0.6 ips at 10 tons, 0.3 ips at 19 tons, and 2.8 seconds in reverse.

This seems reasonable to me.

For the control loop I am thinking position only, with a slew rate limit on increasing displacement, and a dead band before the valve actuator moves at all. I'd like to control the pump with a large solenoid and PWM, though a servo may be more practical.
 
With the pump I can get (15gpm fully stroked at 3600 rpm, 4000 psi max) and a 4" diameter 18" stroke cylinder up to 3000 psi with a 1800 rpm motor will travel at 2.3 ips at 2.5 tons, 1.2 ips at 5 tons, 0.6 ips at 10 tons, 0.3 ips at 19 tons, and 2.8 seconds in reverse.

This seems reasonable to me.

For the control loop I am thinking position only, with a slew rate limit on increasing displacement, and a dead band before the valve actuator moves at all. I'd like to control the pump with a large solenoid and PWM, though a servo may be more practical.

I think the old man this is for should sit on the porch, drink beer and watch the kids and grand kids run this.
 
Newer John Deere farm tractors use a pressure flow compensated variable displacement pump which uses a load sense hydraulic circuit to regulate the pump. This a fancy way of saying the pump will put out no more pressure than is required to do the job asked of it. The SCVs (hose outlets) are pulse width modulated to regulate flow required.

A spring inside the pump holds the swash plate in a minimum position, allowing the pump to produce around 300 psi standby pressure. When a function is activated, load sense pressure increases, shifting the swash plate into a pitched mode, which increases the pressure and moves the hydraulic cylinder. If the cylinder only requires 1800 psi to move at the flow determined, load sense determines the swash plate pitch.

I think there should be a way to use the amp draw of the electric motor as an input to control the pump. I think the pump would be best controlled hydraulically, with some sort of PWM valve governing the swash plate position.



Your application is unique in that you need to regulate the horsepower required by the pump, while maximizing the pump's output flow.
 
Newer John Deere farm tractors use a pressure flow compensated variable displacement pump which uses a load sense hydraulic circuit to regulate the pump. This a fancy way of saying the pump will put out no more pressure than is required to do the job asked of it. The SCVs (hose outlets) are pulse width modulated to regulate flow required.

A spring inside the pump holds the swash plate in a minimum position, allowing the pump to produce around 300 psi standby pressure. When a function is activated, load sense pressure increases, shifting the swash plate into a pitched mode, which increases the pressure and moves the hydraulic cylinder. If the cylinder only requires 1800 psi to move at the flow determined, load sense determines the swash plate pitch.

I think there should be a way to use the amp draw of the electric motor as an input to control the pump. I think the pump would be best controlled hydraulically, with some sort of PWM valve governing the swash plate position.



Your application is unique in that you need to regulate the horsepower required by the pump, while maximizing the pump's output flow.

I wanted to use a small cylinder and spring to actuate the swash plate, but couldn't find suitable parts affordably
 
I have run several and from memory 4" per second seems right. On a 16" ram that's probably a 7 second cycle time. If 2 or 3 is normal though, even easier for me. But 1/2" per second is too slow and I want this to be strong enough my dad never has to use a wedge ever again.

If you apply pressure to both fittings on a double acting cylinder, what happens ?
 








 
Back
Top