Design Challenge: How to Increase Force from Linear Motor?

Levers in the motor world is called gears

1. Simply add gearing where the with a ratio 15:1.

example a gear on motor has 10 teeth, the next gear has 150 teeth. The motor will turn 15 times for the next gear will turn once. This will effectively increase the torque on the second gear by a ratio of 15:1

2 Install a servo amp system with an amplification factor of 15:1. This will require a larger motor. A servo amp system can control very accurate movement. For extra accuracy have a course and fine cct. The course circuit is for large movement, and the fine cct involves gearing for fine control.

Thanks, I know all about gearing and servo amplifiers. However, this is a LINEAR motor. There is no rotational motion at all. The motor operates purely in a linear fashion. The best way to describe it is that it's like a normal induction motor, but the rotor and stator have been "unrolled" to a flat state. Due to this reason, there is no way that gears can help.

Thanks for the suggestion though!

Edit: This is also the largest linear motor (1100-lb peak force) we can find that fit other design constraints, so a larger motor is out of question.

You could use two gear rack and pinnion, one movable and the other fixed, and let the load attached to the pinnion shaft. Your linear motor would push one rack, and its movement actuates on the pinnion that, with a in line reduction like a planetary gear, would change this movement in translation over a fixed rack. Of course, if you need 100:1 reduction, you'll have a 1:100 movement displacement.

This arrangement can also be built using a positioning belt, just like a commercially available table scanner, but with the reduction in the gear shaft. But I'm affraid your application would be too big for commercially available parts to do that.

Just two ideas that occurred me.

Are you stuck with using a linear motor? Could you use a large rotary motor and a scissor drive to get power and linear displacement?

you can use a gear attachment whenever u wanted a constant velocity ratio and smaller gear should be at the outer side connected to roller,and the gears should be hellical type.

Hydraulic force multiplier (small cylinder in large cylinder out) a possibility?

If the roller is being moved with up to 15,000 # force, then I'd guess it's probably a largish roller. Thus, the short end of a lever might be 6-8"? If so, the long end would be 90 - 120" so the side-to-side motion at the lever end could be pretty minimal (depending on how far the roller must move). Are you sure you can't simply mount the linear motor as you would a hydraulic cylinder, and allow it to pivot? Is the movement fast?

Use the linear motor to drive a wedge?

Use a scissor linkage with the linear motor operating the long "handle" end and the roller being moved by the "cutting" end (fixed pin on one side, load on other). The linear motor would then need to translate slightly as it moved, but wouldn't need to pivot.

This application sounds like it would work well with servo-motor-driven jack screws, like those used to adjust calender roll gaps.

When you say "accurately" what do you mean? .001"? .0001"?

Thanks for all of the suggestions. A servo driven jack screw was considered, but the idea was scrapped due to the multiple connections that would be required and also the backlash.

The process needs to be repeatable to within a micron, and also accurate to within a micron.

Have also considered the scissor like linkage systems, but seems that is just getting overly complicated and would like to avoid it if possible.

The motor cannot be mounted on a pivot as it weighs 80 pounds, needs to do a 1" stoke (or more) in about 150ms, and then reverse approximately half a second later. The dynamic issues will be too much to combat to get the accuracy we require. Also, radial loading will be somewhat of a worry.

Hydraulics are also strongly opposed by the project manager due to the nature of the application.

I must say though....you hit on just about every idea we've thought up so far.

Thanks for all the suggestions!...one of them may prove useful yet. Seems it may turn into one of those things where we'll just have to choose the least of the evils and hope for the best. I was just hoping there was something we'd overlooked....

Laxman12786,

Accuracy and repeatability within 1 micron (as in one millionth of a meter) ?????????

You would likely need interferometry at the roller, among other "exotic" techniques to achieve that.

I know of no simple way to achieve that force with that accuracy, however if you overstated the accuracy:

If the only obstacle to using a (stiff enough) lever is the radial loading issue, then why not simply have the linear motor attached to a rail (Thomson or equivalent) mounted assembly to which the lever is then attached? The rail mounted assembly would be designed for the radial stresses, and the appropriately coupled linear motor would see none of them.

The underlying problem you are facing is you need both high force and accuracy. It is always better to have the appropriate linear actuator so you have a minimum of linkage to the roller (load), since added linkage introduces increased inaccuracy.

Have you looked at precision electric actuators? These are servo controlled ball screws which are very accurate positionally and in the amount of applied force.

One link that states repeatability of +/- 0.0005" (0.013mm):

http://www.danahermotion.com/products/product_detail.php?parent_id=464

GlobalSpec has many suppliers listed.

Regards, Greg

Laxman12786,

After a reread of your requirements, and taking the 1 micron repeatability with a little bit of skepticism, because I have had engineering bosses in my career, who, when I asked what tolerance to apply to something, told me with a straight face, and complete sincerity, "no tolerance, just tell them to make it dead nuts (exact)", another approach comes to mind: A "beauty and the beast" scenario.

You break up the requirements along the two lines of force (the beast) and accuracy (beauty). That is, you select an actuator with the appropriate force and speed to move the mechanism against a precision controlled stop. The actuator mechanism for the stop is selected on the basis of force and accuracy, but at a much, much slower speed and much less range of travel. For instance, if your total design range of travel is 1", +0.25", -0.05", the precision stop only has to move over a 0.30" range once the mechanism is manually fixtured for the nominal 1" movement of the force applying fast moving actuator. Backlash and mechanical deformation of the total mechanism will be taken up and accounted for in the total stroke of the force actuator and the software during operation. The underlying assumption is that the stop actuator mechanism only has to adjust at a maximum rate of say 0.001" between strokes of the force actuator, allowing for very high mechanical advantage and accuracy, while the force actuator goes about its work "dumb and happy" from a control standpoint.

I have used variations of this technique to good success in high force precision applications when doing it all with one actuator simply wasn't possible, even though it wasn't the first conceptual choice because of the added complexity. Typically, the opposing stop surfaces would be enclosed in a rubber boot or similar barrier to dirt and contamination.

Any additonal considerations such as compensations for changes in temperature would be added in as necessary.

Regards, Greg

The 1 micron repeatability is a design requirement as it has a direct result on the final product being produced. We currently have in place a process that can achieve this, but are looking to reduce time, line size, etc.

Your moving stop suggestion may be on to something. That's one we have not yet considered. As of now we are looking to run everything with a closed loop control and with pre-measured inputs (to account for incoming material variation). Because of the complexity of the measurements and adjustments that must be made in the controller, combined with the cycle time, I'm not sure a slow moving variable stop will work. It's definitely a point to consider though.

Thanks!

You are more than welcome.

The slow moving stop was just a conceptual example: it could be much faster adjusting, as in moving one of a well constrained arrangement of two wedges. The basic idea was to separate the two requirements, the rest was just to make the concept clear.

BTW: Why not become a member?

Regards, Greg

Yes, I realize the provided was just an example and it did have me thinking about other options. Thanks again!

As far as being a "member"...I believe I am, but just was not logged in at the time I posted the last comment. At any rate, I have a username and do read posts every morning.

laxman12786,

Yes, of course you are a member: I had earlier replied to you by name, but the last time, from too much jumping around on threads.... I just saw "Guest" and I forgot!

Regards, Greg

Sounds like the original constraint of using a LINEAR motor has caused the problem.

You could use a servo motor and ball screw. (More accurate version of a threaded rod driving a nut)

Correct me if I'm wrong: you need a linear actuator to do a 25- ? mm stroke with 5000 N force, low inertia (to achieve 0.17 m/s speed), 1 micron positioning accuracy and has to be electrical. Sorry for conversion, but I avoid using a mix (micron is metric).

Should be a combination of linear electric motor + positioner + actuator...

Maybe an electromagnetic artificial muscle...

Your motor looks like a bad choice. In principle, to add linear force on a given direction for a given travel , you need to put in parallel several motors. In your case, 15 motors like the one you have now! Obviously, it's not doable for several reasons. But shaking platforms are built like that!

Try to contemplate what Nature did, successfully: muscles. Several muscle fibres are connected in parallel to increase force on short strokes (parallel force vectors). To do longer strokes, fibres are connected in series (force vectors in line), building so-called long muscles. Like us, Nature cannot build any size of linear motor, so it perfected a cell and multiplies it as necessary.

I invented such a device but I have no money to build more prototypes and pay for IP.

Maybe my suggestion will help you in some way.

We need some more information such as; some details about the operating environment (dusty, sterile, zero gravity, etc.) and you mentioned "a great deal of force" as one of the design parameters. How much force is required and what is the required positioning velocity.? You mentioned the roller position accuracy requirement of 1 micron. How much does the roller weigh? Size? Suspension? What are the parameters of the position detection system? If hydraulics are a political issue does that eliminate pneumatics as well? Have you looked at hybrid servo/linear motor systems such as found in robotic assemblers?

Dear laxman12786: You will be praud solving this problem with linear electric motor! Let us know your solution when done. Think of changing basic requirements (speed vs. acuracy vs force vs, hydraulics??) Automaqtic control pf position has its limitation! Stepper motors are doing the job pretty good and let use computer programming for control. Stepping maybe used when approching a final destination / stop

How Trying something Simple??

Mount a pneumatic cylinder with the proper travel length on either side. When the motor reaches the bottom it can hit a pneumatic switch using a simple block attached to the critical point of motion. When the switch is activated it will open and send air into the cylinders thus increasing the amount of applied pressure. This can also be adjusted with the use of a regulator to increase or decrease pressure.

Hope that is not to confusing.

Talking of Einstein, Newton and Frankenstein: the former set us free to contemplate how we should bend Newton; and Newton set us free to keep Frankenstein at bay!

Are you not related to Bertrand, one of the ultimate of philosophers????

"...... and Newton set us free to keep Frankenstein at bay!"

Newton did that? I thought it was Lawrence Talbot!

Greg

Hey, this one would be a good idea. Like the power steering in automobiles, your electric servo would still control and position the actuation, but the force would be amplified by a cylinder.

But I'd still recommend hydraulics, maybe using a non-contaminating fluid if this is the case, because it's prety clear to me that a compressible working fluid would not have the required reaction speed.