Solenoid design for large travel and linear force

I'm playing around in Femm, learning it. I'm trying to design a DC solenoid with a steel shell and an encapsulated plunger. The solenoid should be quite a bit longer than the plunger and I'd like the forces acting on the plunger to vary fairly linearly with plunger displacement (with 0 force when the plunger is centered), over a fairly large displacement.

I'm finding the forces increase exponentially as the plunger approaches the ends of the solenoid and I'm struggling to make the curve more linear.

I've found that increasing the # of turns and decreasing current (increasing voltage) makes things more linear but I'm at the practical maximum of turn count and still far from desired linearity. The forces I'm getting at the ends of the solenoid are sufficient for my design, it's the middle region I'm stuck on.

Any design tips appreciated!

PS I'm not married to a solenoid design, if there's some other electromagnet approach that's fairly power efficient and gives linear force vs displacement I'm all ears.

Not what you're asking for, but how about a ball screw actuator?

It would help to know what you're planning to use it for.

davegravy said:

I'm playing around in Femm, learning it. I'm trying to design a DC solenoid with a steel shell and an encapsulated plunger. The solenoid should be quite a bit longer than the plunger and I'd like the forces acting on the plunger to vary fairly linearly with plunger displacement (with 0 force when the plunger is centered), over a fairly large displacement.

I'm finding the forces increase exponentially as the plunger approaches the ends of the solenoid and I'm struggling to make the curve more linear.

I've found that increasing the # of turns and decreasing current (increasing voltage) makes things more linear but I'm at the practical maximum of turn count and still far from desired linearity. The forces I'm getting at the ends of the solenoid are sufficient for my design, it's the middle region I'm stuck on.

Any design tips appreciated!

PS I'm not married to a solenoid design, if there's some other electromagnet approach that's fairly power efficient and gives linear force vs displacement I'm all ears.

Click to expand...

You need a different approach entirely.

A single-coil solenoid is EXPECTED to lose force as its air-gap increases.

Nature of the beast to be strongest at full pull-in. Ledex utilized a ball-bearing wedge and ramp mechanism to provide for longer travel off a short stroke. See "rotary solenoid". Just one form of several using leverage or mechanical (dis)advantage.

Your application hints at a needs multiple coils and a means to transfer power?

We call those "linear motors" rather than solenoids.

See also steppers. Servos. "Voice coil" positioning. The phasing of power to multiple coils to balance force and SHIFT that balance as commanded.

OR.. use of a "torque motor", which can function as a sort of "infinite spring", and an adjustable one.

Ever catch yer bod in the doors of a lift?

Torque motors and cousins are what holds the same "non-severing" level of applied force, regardless if the closing doors caught you across the arse or just the elbow.

A(ny) "well read" engineer is constantly and curiously "inventorying" all technology as crosses the line of vision for possible future use.

B.C. Forbes had good advice:

With all thy getting, get thee understanding.

Click to expand...

Do more of that exploring of curiosity. Follow those side-trips. Expand your knowledge. Cease ye not its pursuit.

Presuming you want to be seen as knowing yer arse from yer elbow, lift doors or any other.

thermite said:

You need a different approach entirely.

A single-coil solenoid is EXPECTED to lose force as its air-gap increases.

Nature of the beast to be strongest at full pull-in. Ledex utilized a ball-bearing wedge and ramp mechanism to provide for longer travel off a short stroke. See "rotary solenoid". Just one form of several using leverage or mechanical (dis)advantage.

Your application hints at a needs multiple coils and a means to transfer power?

We call those "linear motors" rather than solenoids.

See also steppers. Servos. "Voice coil" positioning. The phasing of power to multiple coils to balance force and SHIFT that balance as commanded.

OR.. use of a "torque motor", which can function as a sort of "infinite spring", and an adjustable one.

Ever catch yer bod in the doors of a lift?

Torque motors and cousins are what holds the same "non-severing" level of applied force, regardless if the closing doors caught you across the arse or just the elbow.

A(ny) "well read" engineer is constantly and curiously "inventorying" all technology as crosses the line of vision for possible future use.

B.C. Forbes had good advice:



Do more of that exploring of curiosity. Follow those side-trips. Expand your knowledge. Cease ye not its pursuit.

Presuming you want to be seen as knowing yer arse from yer elbow, lift doors or any other.

Click to expand...

To be clear I am looking for force that decays as the air gap increases, just at a linear rather than exponential rate.

davegravy said:

To be clear I am looking for force that decays as the air gap increases, just at a linear rather than exponential rate.

Click to expand...

That's a "physics" thing, given a uniform coil and armature.

Might.. I say again MIGHT.. be do-able with some combinations of a tapered armature, progressive rate wind to the coil as to turns spacing or bobbin diameter.

The Ledex I cited:

Ledex Products - TESTCO

... may be built so as to have a shape to its ramp to "get there", but the "curve shaping" is off the back of a mechanical translation of force, linear to angular rotation, needs a linkage to deliver linear motion back again.

Ledex solenoids one fine afternoon very nearly got me machine-gunned to death by the guardians of our nation, basement of the US Capitol building.

Noisy buggers. VERY!

Demo'ing the first-ever International Autopen with built-in paper advance as happened to have 66 steps, direct pin-feed platen drive, and at the same cyclic rate as an M60 machine gun wasn't my best-ever move!

Mind.. as a 'nam vet, the sound of the pig, "outbound", was a comfort, not a stressor, so I had paid it no mind in the lab.

Thankfully, the Guvernment guy was also a vet, so laughingly waved-off the two guards sprinting at us with machine-pistols aimed!

"That will not do!".. and back to the drawing board I went!



"Revision TWO" utilized a near-as-dammit silent Bodine torque motor with mechanical tractor chain goods from Moore Business Forms Kidder Press division.



The expectation is that solenoid approaches may be "all of the above" as to inefficient use of space, input energy, and materials - hence money - for all but odd edge-cases, if-even.

The "money factor" may not be in a Physics text, but usually dominates, regardless, in the "real world".

Which may be why so many OTHER approaches dominate, moving vane meters, hard disk drive positioners, machine-tool servos, linear lift motors, maglev trains, electric warship catapults, and "rail" guns.

Not to forget air, hydraulics, and thermally driven forces, many of which are all around us, just doing their job, and largely ignored.

A common solution? Variability. Controlled. With feedback. Tailored to the need.

THEN one can do whatever is needed.

"Passive" components as have characteristics contrary to their inherent nature will fight you. And usually win the battle.

"Go active" instead.

You'll find some solenoid designs use a cone on the end of the plunger that engages a cone on the magnetic structure. Helps things a bit but IMHO you need something entirely different like a voice coil actuator similar to what BEI makes. Then you can have any force curve you like.

thermite said:

That's a "physics" thing, given a uniform coil.

Might.. I say again MIGHT.. be do-able with a progressive rate wind to the coil as to turns spacing or bobbin diameter. The expectation is that it would be an "all of the above" inefficient use of space, input energy, and materials - hence money.

The "money factor" may not be in a Physics text, but usually dominates, regardless, in the"real world".

Which may be why so many OTHER approaches dominate, and in massive numbers, hard disk drive positioners to machine-tool servos, to linear ift motors, to maglev trains, and "rail" guns.

Not to forget air, hydraulics, and thermally driven forces, many of which are all around us, just doing their job, and largely ignored.

Click to expand...

Thanks, I'll play with variable winding and bobbin diameter. I expect you're right there's a design concept that will probably be better, I just need to do my due diligence.

thermite said:

That's a "physics" thing, given a uniform coil and armature.

"Passive" components as have characteristics contrary to their inherent nature will fight you. And usually win the battle.

"Go active" instead.

Click to expand...

I hear you. Certain stakeholders in this project seem to have the perception that a control system will add unnecessary complexity (in reality I think it's just complexity they don't understand and hence fear)

I should note the application is for a variable stiffness spring. The idea being that resulting stiffness is roughly proportional to input power of the solenoid (zero stiffness when power is removed).

With this design it's actually a negative stiffness spring (force acts in the direction of deflection), but put in parallel with a positive stiffness spring it could provide variable positive stiffness.

One of the requirements is to have very low damping and high reliability/durability, which is why (virtually) frictionless design is attractive and things like screw actuators won't work.

Since you are interested in finite element analysis, I would assume you have sufficient mathematical skills to understand the principles involved. In the case of a simple solenoid, the pull curve is inherently non-linear, F = (B^2 * A)/(2 * mu). I would suggest that you find a text book that develops the equations. There ways to linearize the pull curve, such as non-uniform windings and variable permeances but these may or may not be practical. In the case of finite element analysis you would have to build a model with a stack of magnet disks, each of which has different properties to the one next to it. Then you would have to figure out how to build it.

Tom

Conrad Hoffman said:

You'll find some solenoid designs use a cone on the end of the plunger that engages a cone on the magnetic structure. Helps things a bit but IMHO you need something entirely different like a voice coil actuator similar to what BEI makes. Then you can have any force curve you like.

Click to expand...

That cone you speak of is to create a longer change in the magnetic permeance at the pole face. DC operated solenoids have a very sharp change in the pull curve as you approach the sealed condition. The taper extends the usable length of the solenoid.

Tom

would cam action allow you to shape your force/ movement curve?

surplusjohn said:

would cam action allow you to shape your force/ movement curve?

Click to expand...

??? Explain.

Tom

As usual, most people suggest all sorts of alternatives instead of attacking the actual problem. Conrad Hoffman has some good suggestions. I had a similar request for a long stroke solinoid and looked at a tapered core to spread the force out. It is a matter of controlling the amount that the magnetic path increases with the core movement. As it happened, the customer came up with a different solution so I never got to make one.

I suggest you give us a better description of the application. Is it one where the solinoid actually simulates a spring? A typical coiled spring is bidirectional, resisting motion either way from center. Is that what you want or in only one direction. Also, what are the forces required?

The nonlinearity of a regular solinoid is from the end of the armature being pulled up against the coil's core. I am envisioning a round core on a coil with a hole through it and a round rod suspended in it never touching the fixed core. It will automatically center itself longitudinally. The centering force would be proportional to the current.

Bill

9100 said:

As usual, most people suggest all sorts of alternatives instead of attacking the actual problem. Conrad Hoffman has some good suggestions. I had a similar request for a long stroke solinoid and looked at a tapered core to spread the force out. It is a matter of controlling the amount that the magnetic path increases with the core movement. As it happened, the customer came up with a different solution so I never got to make one.

I suggest you give us a better description of the application. Is it one where the solinoid actually simulates a spring? A typical coiled spring is bidirectional, resisting motion either way from center. Is that what you want or in only one direction. Also, what are the forces required?

The nonlinearity of a regular solinoid is from the end of the armature being pulled up against the coil's core. I am envisioning a round core on a coil with a hole through it and a round rod suspended in it never touching the fixed core. It will automatically center itself longitudinally. The centering force would be proportional to the current.

Bill

Click to expand...

The application is indeed for a solenoid that simulates a spring and allows for variable stiffness with very low damping. Bidirectionality is strongly preferred (0 force when the plunger is centered in the coil).

Interestingly I've found that an open-ended solenoid acts as a spring with positive stiffness (self-centering) whereas a solenoid with both ends closed is reversed and hence negative stiffness (unstable, but still with 0 force in the middle). Either behavior is acceptable for me, however I've found magnitude of stiffness I can achieve with the latter is much higher with all else (e.g. input power) being equal, I assume because there is much less leakage.

I'm looking at designs for a couple use cases:

1) 2mm displacement (+/- 1mm from 0 position) 568N/mm stiffness
2) 2m displacement (+/- 1m from 0 position) 159kN/m stiffness

I've got a decent preliminary design modeled for #1 which draws somewhere between 30-150W and has decent linearity. For #2 I have now have something quasi-linear which draws 23kW. I haven't even started thinking about cooling or air-spring or back-EMF effects or cost for either case!

I did a very rough design for #2 using an "off-the-shelf" large torque motor with active control and it came out way more power efficient.

9100 said:

As usual, most people suggest all sorts of alternatives instead of attacking the actual problem. Conrad Hoffman has some good suggestions. I had a similar request for a long stroke solinoid and looked at a tapered core to spread the force out. It is a matter of controlling the amount that the magnetic path increases with the core movement. As it happened, the customer came up with a different solution so I never got to make one.

I suggest you give us a better description of the application. Is it one where the solinoid actually simulates a spring? A typical coiled spring is bidirectional, resisting motion either way from center. Is that what you want or in only one direction. Also, what are the forces required?

The nonlinearity of a regular solinoid is from the end of the armature being pulled up against the coil's core. I am envisioning a round core on a coil with a hole through it and a round rod suspended in it never touching the fixed core. It will automatically center itself longitudinally. The centering force would be proportional to the current.

Bill

Click to expand...

You really meant the square of the current didn't you?

F=B^2*a/2mu

Tom

TDegenhart said:

You really meant the square of the current didn't you?

F=B^2*a/2mu

Tom

Click to expand...

Nope. What I have in mind would be proportional. Wattage would increase as the square of the current.

Would you consider using a permanent magnet for the armature?

Bill

9100 said:

Would you consider using a permanent magnet for the armature?

Bill

Click to expand...

Sure, please share.

OK, imagine two solinoids back to back with a common armature so that you could pull it either way and the armature would bottom on each end. Make the armature a permanent magnet and have the coil polarities to repel the movement instead of pulling. As you pushed the armature toward one end, the air gap on the opposite end would increase, reducing the repulsion and the other would decrease, increasing the opposition to the movement. You might also want to make the armature have a coil instead of permanent and control both the fixed and moving coils.

You would need a shaft of a non magnetic material and possibly taper the poles to improve linearity as Conrad suggested, and you might need a position sensor and some electronics for best linearity.

Like our idol Nikola, I have a clear vision in my head but it would take a while to sketch it..

Bill