Modern machine design is all about delivering higher performance at a lower cost with higher reliability. Delivering these often-conflicting performance characteristics is the mark of a great designer, and for the most common form of industrial motion, rotation, newer applications such as SCARA robots and their end effectors need torque with low mass in the smallest possible footprint. Frameless motors are an ideal choice for applications that are space-limited or must strictly control inertial mass. Kollmorgen frameless motor expert Tom Wood discusses the technology in conversation with Jim Anderton in this podcast. A transcript is provided below the recording.
Jim Anderton: Welcome to the Design World podcast. I’m Jim Anderton. You know, multiple electric motor technologies, drive and feedback systems combined with relentless cost pressures make machine design a tougher challenge than ever before. I spoke with Kollmorgen frameless motor expert Tom Wood, about those challenges and how to address them with high-performance, cost-effective technology. You know, power delivered through rotational motion has been the cornerstone of mechanical engineering since the dawn of the Industrial Revolution. Steam engines and turbines, and eventually the internal combustion engine created modern transportation industries. But the development of highly efficient electric motors in the early part of the 20th century transformed everything from industrial mass production to the household vacuum cleaner. But today it’s about high performance, specialized motors, particularly in automation applications. And for the designer, the choices are many.
Now joining me to talk about some of those choices is Tom Wood. Tom is a veteran of High-Performance Motion Systems and has integrated servo motor control solutions as a machine designer and as a motor product specialist for Kollmorgen for over 40 years. He has extensive experience designing next generation motion systems for the aerospace and defense industries, as well as for robotics applications that improve everything from workplace safety to patient mobility. He also has special expertise in frameless motor technology. Tom, welcome to the podcast.
Tom Wood: Hey, great to be with you. It’s an exciting topic.
Anderton: Yeah. Tom, there’s so much to talk about. I mean, motors, electric motors. There’s been an explosion now in new motor technology, in new ways to deliver rotational power and torque that, you know, and at the same time simplify the designer’s task, but also require more from the designer up front to extract the most from the technology.
Wood: Absolutely. And in fact, probably some of the biggest, the biggest advances that we’re seeing in motion technologies is the simplification of the mechanisms. Actually, what we’re seeing is that people are embedding the motor deeper within the machinery element itself. So for example, if you take a look at a robotic system, you don’t see motors hanging off the side of the axes. You do have some really large, robust motor applications or robotics applications from years past. But today’s robot systems, everything is closely integrated, closely coupled. You’re minimizing the losses. You’re minimizing the power, increasing efficiency and performance. And it’s several things that have helped this, primarily because as we have been able to leverage the power of microprocessor based controls and faster control loops and the electronics that drive motors, now we’re able to leverage very high resolution feedback and directly driven motor technology to get rid of the inefficiencies, get rid of the lost and wasted motion and space and heat and create smaller, faster, better kind of outcomes and machines. This is that next part. And robotics is only one element of this. You’re seeing this in virtually every application where motors have historically been used.
Anderton: Yeah. Now Tom, it’s things like variable frequency drives. I mean, they’re instrumental in that. The big shift. What are we saying? Automation. Because you can take full advantage of that. Maximum torque at zero rpm characteristic of electric motors, you know, and maybe do away with a worm gear reduction gearbox or multiple sort of torque multiplication mechanical technologies that you had to couple to that electric motor, essentially. And we’re talking now about a world in which that’s fine, but we don’t necessarily have, a three by four-foot frame and 500 pounds to play with, to, to, to mess around with these systems.
Wood: Well you know, we’ve done some we’ve done some very interesting things in the old days. And I’ll say this, and this is only 2015, 20 years ago we were so limited with servo motors. We had to very we had to very carefully analyze the load, understand that the load to inertia, a mismatch of the of the project that you’re trying to drive to the inertia of the motor itself was matched with a range and typically 5 or 10 times. Today, with advances in this same kind of microprocessor based, high speed controls and in increased resolution and feedback devices, now we’re able to put motors directly into driving the application, minimizing multi bearings support contact systems. I mean when you when you when you used to have a load, you’d have two bearings on the on the shaft of the machine. You have two bearings in your motor which meant you had to have something between them. And those very couplings are lost motion elements. You’re also introducing vibration secondary losses, more heat, all these other things that are negative to the outcomes of what machine designers are trying to get to. So when you start implementing frameless style motors into systems, specifically robotics, everything, everything from surgical, you know, robotic systems to medical systems to just improvements in general automation on the factory floor where you’re where, lean manufacturing practices have focused us into smaller and smaller spaces on the factory floor. And we have to get the machine smaller. We have to get all the work done in that. So we have to make sure that we’re not, giving problems in terms of excess heat or size on these. So, yeah, I agree with you completely trying to drive this into higher levels of technology, leveraging that technology, whether it’s electronics, whether it’s magnetics and motor design, whether it is the feedback devices, all of these are important pieces of the motion control aspect that’s benefiting the advances we’re seeing today in machine design.
Anderton: Yeah. Now, Tom, you brought up frameless motors. And intuitively, I think that you’ve got to have a frame. I mean, you’ve got you got a static element, a rotating element, basically. How can you have a frameless motor? You’ve got a hanging armature into the thin air. What do we mean by frameless motors.
Wood: Yeah, that’s a great question because a lot of times when we’re talking to folks, they say, you know what, I don’t want to have to put a motor inside my machine. I don’t want to design a motor. But fundamentally, what you’re talking about is this where you historically have a housed motor, you’ve got a frame housing, you’ve got a shaft and bearings, you’ve got a feedback device, you’ve got a connector. All those things are actually extraneous to the to what is needed from the rotor and the stator, which is the torque producing element that is found inside a motor. If you can incorporate the rotor from a given motor and find a place on a machine, look this this machine already has shafts and bearings that are that are sized appropriate to the axial and radial load characteristics. Size to the performance requirements of it someplace on that shaft. If you have the ability to mount rigidly, mount the rotor of a given motor size and then in proximity to that rotor, attach through a housing element that is simply a part or an extension of the frame of the machine itself, you’ll be able to embed the motor specifically inside the mechanism and have no lost motion in the system. No backlash, no wind up, no torsional loss, no stiffness loss characteristics because you’ve already designed that into the housing, into the mechanism that you have for the machine design and you don’t even have to worry about the machine tolerance concerns that you’ve got, because the very same machine tolerances that you have for putting together your machine that might have, let’s say you’ve got a thousandth or a thousandth and a half of total indicated run out on that shaft and bearing you’re well, well within the range where we would like to see.
Anderton: Yeah. Tom, that’s pretty provocative because from a machine design standpoint, of course, there’s always that compromise… rigidity is the goal. So obviously we’d like to mount our motor in a structure which is perfectly rigid in every, in bending moments, torsional, all of it. You can never actually achieve that. But at the same time also these are we’re talking about systems that move. There’s this motion here. And that means vibrations and other resonances that can come up in the structure as well. And if you can use things like advanced simulation, maybe you can move some mass around and you can basically tune the structure so that you can cancel out some of those resonances. Maybe you install the motor, design the motor to go into a nodal point, or you’ve got a known residence issue with a structure. How do you, basically how do frameless motors, that play into that design methodology?
Wood: What we end up doing is this, and by knowing where you’re putting this machine, this motor driving element in the system, you’re also aware of things like fundamental, first, third, the fundamental frequencies here that might excite the motion. The worst thing that you can do with typically is to have some kind of a coupling and secondary load where you no longer really have control. One of the benefits in how we can drive higher and higher performance, is that we can directly drive those higher load shaft elements. I can go to, I’ve seen applications where we exceed 1000 to 1 load inertia to so-called motor inertia mismatch on systems and maintain the integrity of very rapid positioning. I’ve seen 100 millisecond moves executed in 85 milliseconds, and then have less than 15 milliseconds of settling time to get in position without overshoot. Okay, and maintaining what could have only been done before mechanically with these kinds of high loads and but and having twice the time advantage, it went from a 200-millisecond application to a 100 millisecond application move, and it was done stably, without overshoot and without other mechanical losses. So to your point, the other thing is the whole idea that that people have about understanding and being able to do things like vibration sensing, okay, with the advances you see in control architectures for driving those servo motors, we can build in filter elements that will allow us to either run through or run past known resonance points and stay in those elements, and run at adjacent frequencies that may not give you the resonances that might excite the mechanism. Otherwise, the tuning aspect of machinery is really fundamentally improved by the embedding of the motor inside the mechanism.
Anderton: Well that’s fascinating. So it sounds like you can, get around some of those other workarounds you used to have to do. For example, like I’m shaping the RPM rise to try and blast through, a zone where, you know, you’re going to get a lot of resonance, right?
Wood: Yeah. It’s not just about setting, you know, particularly rapid accel decelerates within certain frames. This is the ability to do, on the, on the move kind of performance of, of tuning. And it can be adaptive as well. Take a look at a robotic structure, for example. And I’m, you know, a 6 or 7 degrees of freedom robotic arm. Think about the variation in the weight that is carried by that last degree of freedom in the robot. You go, let’s say if you have, in a simple collaborative style robot system, you may have 22 pounds, a ten-kilo load, okay. And you may go from zero load at that end effector to 20 kilos and still have to be able to main both the integrity of position and a fast rate of movement to come into its final position stably with motion. And one of the things that you brought up in some conversation earlier is that sometimes it isn’t just about direct drive. Sometimes it’s about the ability to leverage specialized gearing elements like harmonic drive style, strain, wave style gears with typically very high ratios. And with these high ratios, now we’re able to fundamentally drive away that load inertia mismatch. If I have a typical 100 to 1 ratio on a strain wave gearbox, which is very typical that you’ll see in a collaborative style robotic joint, okay, you reduce the reflected inertia of that load by 10,000 times. So this is one of the elements that allows us to have through 6 or 7 degrees of freedom, the ability to maintain the integrity of position and rate of movement while we pick up full, massive loads and move them through a wide range of diverse reflected inertia to the system. A combination of high bandwidth and high-performance amplifiers driving the motors’ high resolution feedback devices, and then motors that have magnetic properties that allow them to operate efficiently within a wide band of temperature requirements. So all these things come together. This is the perfect storm of technology approaching, which is why robotics has had them first in the market space.
Anderton: Yeah. I mean, I’ve always thought that those, the robot people, they’re in tough because everybody wants a more and more complex end effector. We want live tooling. We want, jigs and fixtures of enormous size, basically, and weight. But the inertial mass is a factor. So you’re swinging this big dumbbell at the end of SCARA robots, which are physically larger. So now you’re looking at something that you get 6 or 8ft of, of effectively stick out with, you know, with a couple hundred pounds of mass at the end. Oh, and by the way, we also want positioning accuracy and we want repeatability, you know, all at the same time. Well, I mean, you’re asking for everything all to say I, you know, I want a ballerina and a linebacker in the same person.
Wood: You know, exactly. And nobody ever said, yeah, I want to go. I’m willing to go slower. I’m willing to go with less precision. I’m willing to pay. They pay more for it. You know, all those things happen so yeah.
Anderton: Know, it’s these things are electromagnetic devices. Hysteresis is not your friend. How fast can you spin these things up from stop. And how fast can you stop them?
Wood: Well, actually, depending on there’s different motor architectures. There are some loads that want to have a minimal inertia so that they can leverage very, very rapid acceleration. Stop start considerations. For example, in in highly dynamic machines that you’ll have high part count but very precise rapid motions to go from point to point; you may find that you have a direct drive that has a different physical motor architecture. It might be long and skinny with a smaller diameter, smaller diameter rotor. We also find that there are applications where they have very, very high torque demands. And yet even with that, you need to make sure you can maintain, relatively rapid acceleration. But again, it’s all a matter of what is the load characteristic, what are the rates? Are you going to leverage direct drive or are you going to leverage in the case of gearing, the benefit that we’ve seen in strain wave style gearing is that it is a zero-backlash component. There is no lost motion. There’s no there’s no loss of band width relative to that mechanical system because you don’t have backlash, but you also don’t have backlash in some systems, like built in pulley mechanisms, however, you have highly variable systems that will wear and then they will change the dynamics of the system again creating maintenance headaches in the like. There is still a downside with strain wave style gearing. And that generally comes in the form of stiffness and the lack of stiffness. Well, again, here, higher resolution feedback and advanced control architectures that allow you to, to dynamically tune and make these systems artificially more stiff are really what has made collaborative style robots and large articulated robot systems much more viable in the in the market space, maintaining the integrity of higher and higher payloads while they’re going faster and while they’re maintaining the integrity position at high rates.
Anderton: Yeah, that backlash problem, that’s been a very difficult problem to solve for many of the high precision applications I’ve seen, strategies where, artificial loads are applied, you know, there’s friction devices, there’s braking or strategies where basically you basically start the machine cycle early so you can preload the. Yeah, exactly.
Wood: Right. And you know, in some cases where you’ve got gravity as your friend, you have that constant load you can kind of work with, but now you can’t work fast enough, you can’t pull G’s in terms of acceleration rates. If you’ve got gravity is helping you as part of your answer. So yeah, I think that that fundamentally it still comes down to having a team of people that when they’re building a machine, understanding what problem they’re trying to solve, understanding what rates and performance requirements you have, and then, leveraging effective mechanical, motor and control technologies to get the solutions right. And we have today more answers to these kinds of questions, because let’s face it, sometimes conventional motors are perfectly good for the application. If the problem you’re trying to solve doesn’t demand a specialized nature design, then you don’t have to do it. But we’re finding today that there are three kinds of cases that are critical for a lot of people. One is the need to drive to a more and more compact form in the mechanism, physically, the space that they have requirements, especially in cellular automation today, that you see the size of the devices, the size of the prime motion. Drivers that are here are being driven to be smaller. The second condition we often see is the fact that there is a need for extreme precision, much more so than you might have had in the past, because every single element in this, in this chain has to have a stack up of precision that will eventually lead you to the final precision of the system. So by being able to leverage direct drive and getting away from things like backlash, getting away from wind up, getting away from soft tensioning, even stiffness characteristics, we can drive to better performance, better regulated position, better accuracy, and fundamentally just being able to do the processes needed at the rate of speed. And the third condition is oftentimes there are environmental concerns or environment considerations where it’s just a physically abusive, dirty environment. Consider the fact of guys trying to do casting, sanding , abrasion or casting lines to get rid of elements of design. And that takes sanding. It takes grit, it takes blasting. Can you imagine having high precision motion and the repeated motion that has to be done by man in a suit, in the heat, in the environment today? What we’re doing is work. We’re providing force feedback devices at the end of robotic arm systems with high performance servo motors that are driving to maintain the integrity of the velocity of the material as it’s operating the surfaces. And by maintaining the integrity of that position. What we’re doing is, we’re retaining the surface condition at the highest level, even with the highly variable angle of contact force that’s applied, all those things, all these elements, the, environmental considerations, the precision considerations and driving to smaller and smaller packages, all those things feed into the potential for use of a frameless motor in a machine application.
Anderton: Tom can the motor itself give us some of that feedback sensor capability can do something with back EMF or temperature rise or are there just something things we can do? We can analyze, analyze current flow in and out of the motor to infer some things.
Wood: We absolutely can. Oftentimes what you’re trying to do, is you’re trying to take products that have standard solutions, standard elements of feedback motors and drive architectures with conventional technology. If you’re leveraging that and you’re doing, let’s say, good enough, motion control, that’s one thing you can do; sometimes the challenge, and it’s going to be it’s going to sound silly, but the challenge that people in the commercial washing machine business have, you know, they’re using frameless permanent magnet DC motors. They’re using the same kind of motors you’re seeing in high performance robots, okay. And there’s no feedback. Oftentimes they’re doing some very interesting current sensing. They’re doing some back EMF sensing. And they are doing this while they’re trying to drive. Think about your washing machine. Think of how widely varied that load is from dry clothes going in with just a little of it to wet clothes being tumbled around with an odd mass balance to it.
And yet that machine is highly tuned. What happens is somebody has to develop the control architectures, and there’s a very high-performance requirement for the drive electronics that is specifically tuned to that application, to the fundamental machinery mechanisms that are in there. And everything is smooth. There’s not a lot of wiggle room for that. Okay. Yeah. However, that we’re finding, more and more people are saying, look, I want to make my systems more and more simple. So, whether it’s maybe just the use of simple commutation techniques, we do servo motors all the time that have no feedback, what we call sensorless motor designs. I will say, though, that when you’re using conventional control architectures, it’s often easier to use some very simple, cost effective and compact feedback. Still, in these systems, even the robotic systems you see today out there, you can push the precision with which you expect the end effector to be at and the rate at which you want to do that better. When you have high resolution feedback and advanced controls, think about it. The, the washing machine example for a moment. That washing machine is not driving to a specific angular position at a really high rate at a really repeatable point. It’s in there and it’s pretty loosey-goosey in there. But the point is that it does the job that’s needed to do, and it leverages both cost quality and performance in the engagement.
Anderton: Yeah. I mean, there’s so much to talk about. We’ve just scratched the surface of this topic and every rabbit hole is fascinating going down this road. It’s incredible stuff. You’ve touched on through many, many aspects that are vital to the designer at this point. What does a designer do to find the solution they need? What’s the starting process? They pick up the phone, or do they call you or they call Kollmorgen? Do they go online? Do they do the research? What’s the starting point for finding a sensible solution to some of these motor control problems we talked about?
Wood: It’s a good question because fundamentally we start with the fact that there is a mechanism that somehow doesn’t have either the performance it needs to, or the physical construct that it needs to have. So, you’re driven by either performance or usually size, okay? Sometimes you’re driven by cost. But generally, for the machines and the mechanisms we’re looking at, you’re really driven by either a performance which is rate and precision…the two key elements there. So in order to come to somebody with motors, the first question they’re going to ask you is, well, how fast you need to go. And with what kind of accuracy and repeatability? And all you have in the system now, from understanding the mechanism, where we would ask questions after you have analyzed what machine components you might have, what rate at which you need for the end use or the end effector you’re dealing with. As motor engineering folks, we could go and we can even leverage some examples that we’ve had in the past. We can understand better what your strategy is. You might have for implementing a servo, it could easily be a conventional servo architecture. Okay. It also could very easily be something where we say,” hey, look, you really want to push the boundaries on this”. You may want to get rid of a lot of these dead components, get rid of the backlash, get rid of the problem things, and we can then we can expand on it. However, it also comes back to what power source do you have? Are you working with 48V in the system, or are you working in a in a production environment? In a factory? We’ve got 400 or 480V available. What kind of power in general you have, you know, basically is this are you looking at hundreds of watts, kilowatts, high horsepower, 20 horsepower servo motors are generally very easily seen, up to about 25 horsepower. Okay. And they can be leveraged very easily that way. I’d say that the sweet spot in servo systems is from 500W to about 7000W in size. But understanding the power requirements, understanding the power that’s available, understanding the torque, understanding the speed. Then we can start putting together the right kind of solution pieces.
Anderton: Great advice, Tom Wood , Kollmorgen, thanks for joining me today on the podcast.
Wood: Glad you’ve been here. Thank you. Take care.
Kollmorgen
kollmorgen.com
Filed Under: PODCASTS, Motors (direct-drive) + frameless motors