2 May 2018 – This blog post is intended for folks with an interest in basic practical mechanical engineering, or mechanical engineers who want a basic brush up on linear actuators. It’s mostly pretty basic stuff, which has been around for decades, but can serve as a guide to linear-motion actuators in the real world.
What triggered writing this post at this particular time is a notice crossing my desk about a video entitled “Can I Run A Linear Actuator Into A Hard Stop?” produced by global motion-control supplier Ametek. It’s an important topic that just about everyone faced with building a motorized linear-motion system needs to think about.
I first got serious about linear motion actuators in the mid-1970s as an experimental physics student. I, of course, have seen them in action since I can remember because of my father’s hobby of building powerboats. Virtually every powerboat (as opposed to sailboats) bigger than about ten feet uses manually powered linear actuators for steering linkage.
I didn’t really get into electromechanical linear actuators (linear actuators powered by electric motors) until I got involved with automated measurement systems, where steady motion or precision positioning are important. Since then, just about every system I’ve built has included a precision linear-actuator somewhere inside.
Linear Actuator Types
There are basically four main types of linear actuators: lead screw, hydraulic/pneumatic, linear-motor, and piezoelectric. I’m going to concentrate on the lead-screw type because it’s by far the most common, but I’ll drop in some info about the other types for completeness.
Piezoelectric actuators take advantage of the fact that certain anisotropic crystalline solids change their shapes when placed in an electric field. The range of motion, however, is notably microscopic, so they are best optimized to positioning things that are, well, microscopic. They’re a major enabling technology for atomic force microscopes.
Linear motors are much larger. Imagine a long, relatively narrow tray of chocolate-frosted fudge. Imagine further that the fudge is actually made of, say, barium ferrite ceramic and magnetized with magnetic north being on the frosting side and magnetic south being on the fudge side underneath.
Now, slice the fudge into strips with cuts going across the tray of fudge the short way. Finally, take every other strip out, turn it over with the frosting side down, and put it back in place.
So, you end up with the odd-numbered strips (1, 3, 5, … ) being frosting side up and the even strips (2, 4, 6, … ) frosting side down. That’s what the long stator portion of a linear motor looks like.
To make motion, however, you need an electromagnetic slide approximately as long as the fudge tray is wide, and as wide as one of the cut strips. When you energize the electromagnet, the slide will settle between two of the stips so that its north pole is as close as possible to the nearest stator south pole, while its south pole snuggles up to the nearest stator north pole.
Reversing the current through the slide’s electromagnet makes it possible to inch the slide along the stator, one strip at a time. Switching really fast makes it possible to move the slide along the stator really fast.
That’s a very rough idea of how linear motors work. They are capable of high speeds (to make, say, a rail gun), but are relatively low in the actual force department.
The pneumatic/hydraulic actuator is just a metal cylinder enclosed at one end with a moveable piston at the other. The space between the piston and the closed end is filled with some working fluid, such as air or oil. Forcing more fluid into the cylinder pushes the piston out. Pumping fluid out, pulls the piston back. Depending on details, the motion can be fast or slow, and the forces applied can be enormous. Precision of motion is, however, not so good.
A lead-screw-type linear actuator (LSLA) is a fairly complex piece of kit. Construction of the things is actually fairly simple, though, which largely accounts for their popularity.
Essentially, an LSLA consists of an ordinary reversible electric motor with a length of worm shaft fixed to its output. The worm shaft threads through a slide traveling along a track/frame that prevents the slide rotating with respect to the motor housing. The worm shaft and threaded slide form a simple screw machine to convert rotary motion of the shaft to linear motion of the slide.
A motor controller, which can be as simple as a DPST switch or as complex as an intelligent motor controller (IMC) with a microprocessor brain, supplies power and control to the motor.
Motion Control Stops
At minimum, something needs to be installed to keep the slide from either backing into the motor/shaft coupler at the proximal end of the worm shaft, or running off the distal end of the shaft. These thingd are called, not surprisingly, “stops,” and they can be mechanical, electrical, or software.
Mechanical Stops, also known as “hard” stops, are barriers attached to the frame that physically constrain the slide’s motion to a certain range. Running into a hard stop is generally considered a bad thing, and designers only put them into machines to prevent even worse outcomes that may obtain when the slide’s designed-in range is exceeded.
Electrical Stops, more often referred to as “limit switches,” are actual electrical switches mounted on the frame that are automatically actuated by the slide’s motion. Typically a designer will mount an SPST momentary switch in a bracket attached to the frame. The slide presses on the switch at the end of it’s travel, closing a set of contacts that send a logic signal to the controller alerting it to cut (or reverse) motor power. The block can also serve as a mechanical stop if the control function goes wrong.
Software Stops require adding a linear encoder to the linear actuator mechanism. There are all sorts of linear encoders, from simple lengths of resistance wire to digital optical position encoders. What they all do is send some kind of signal constantly informing the controller of where the slide is in real time. A sotware stop is then an algorithim in the controller program to say: “That’s far enough!” and trigger what happens next.
Given the choice, my preference is to rely mainly on software stops. Having a linear encoder in the mechanism gives all kinds of neat options for precision control of the system, such as positioning, speed control, and so forth, in addition to implementing the software stops.
For example, I once built an experiment to test a device to measure the attack angle of an aircraft wing. A wing’s attack angle is the angle between the relative airflow and the wing shape’s chord. It is the single most important parameter determining the wing’s lift at any given speed. There are, however, all kinds of phenomena that affect the actual attack angle, all of which change constantly in real time as the wing moves through the air. To really understand what’s going on with the wing, some means of monitoring attack angle is, shall we say, useful.
Anyway, the test protocol for the experiment called for mounting an example of the attack-angle sensor in a wind tunnel, and measuring its output at hundreds of combinations of air speed and sensor orientation. Central to the control system’s operation was a linear encoder whose output informed both the controller and the data logging computer.
The controller’s job was to hold the sensor’s orientation at a certain set point via a feedback loop just long enough to get a stable reading, then go on to the next set point. The test program set’s supervisory algorithm stepped the set point through all the orientations required, one at a time. In fact, it cycled the set point back and forth through the whole test range several times, logging data as it went.
After building and testing the whole rig, my job, as principle investigator, was reduced to setting the wind tunnel’s airspeed, then reading a novel while the system ran through the test program and logged all the data automatically.
When designing the thing, I spent a couple of days trying to figure out how to install limit switches. In the end, however, I decided it just wasn’t worth the trouble. The design I had was pretty compact to begin with. The switches available and the mounting brackets to hold them would have been bigger than the rest of the design. So, I gave up on adding limit switches and relied on software stops.
That left me in danger of running into a hard stop, though, if something went wrong with the program. There are always hard stops. Lead screws are of finite length and one of two things can happen when you come to the end: either something (a hard stop) blocks the slide motion, or it runs off the end. Both are bad.
If the electric motor rams the slide into a hard stop, it’s like the proverbial unstoppable force vs. an immovable object. Something’s gotta give and that something invariably breaks.
If, on the other hand, the slide runs off the end of the lead screw and the whole machine falls apart. That may be less destructive, but it means the entire machine has to be reassembled.
Running Into A Hard Stop
There are two rules regarding running into a hard stop:
Rule 1: DON’T DO IT!
Rule 2: ASSUME YOU CAN’T AVOID IT!
What happens if you break Rule 1 depends on the details of the mechanism’s design. Every design is different, and what happens when you go too far is different as well. The consequences are all different, but they are all more-or-less bad. There’s never a situation where running a linear actuator beyond its design limit is a good thing.
Rule 2, on the other hand, is a simple acknowledgement of Murphy’s Law: Anything that can go wrong will go wrong.
While Murphy’s Law has a statistical nature when you’re dealing with mechanical systems in use, when testing prototype systems it’s a stone-cold guarantee. And, any time you put together anything for the first time, then turn it on, you’re testing a prototype.
What Rule 2 tells you to do is think long and hard about what’s going to happen when you turn the thing on and get an unexpected surprise. You have to expect the unexpected because if you expected it, it wouldn’t be unexpected.
One of the most common surprises around linear actuators is the thing suddenly going out of control. When that happens the slide invariably runs past its design limits.
The video from Ametek is short. I hesitate to spoil it for you by telling you that the answer to the question “Can I Run A Linear Actuator Into A Hard Stop?” is “Yes.” It has to be because Rule 2 tells you it’s inevitable. Importantly, the video goes on to tell you what to do to minimize the damage when it happens.