FAULHABER MICROMO is the exclusive provider of FAULHABER Drive Systems for North America and brings together the highest quality motion technologies and value-added services, together with global engineering, sourcing, and manufacturing, to deliver top quality micro motion solutions. John Chandler studied Electrical Engineering at the University of Michigan and has worked in the electronic drives and motion control industry for over 34 years. John always injects a key engineering perspective into all new projects, and he enjoys working closely with OEM customers to bring exciting new products and technologies to market.
In the medical robotics industry, slotless brushless or coreless motors are frequently specified; for design engineers, what impact does this motor selection have on the corresponding drive electronics?
It’s true that these motor types are ideally suited to medical robotic applications, fundamentally due to their light weight, high power density, and zero cogging capability. However, it’s also true that these motors can be more challenging to control, with their lower inductance and rotor inertia. Most leading electronic drive suppliers have kept pace today, and they do offer the key performance features needed to effectively operate these motors. At FAULHABER, we focus on space efficient controls for slotless brushless and coreless motors. To compensate for low inductance, we provide pulse width modulation (PWM) switching frequencies up to 100 KHz. We also employ specialized PWM switching patterns to reduce current ripple. Together, both techniques serve to minimize copper loss. Likewise, our controls improve system dynamics by closing position and velocity control loops at 10 KHz. The combination of high torque and low inertia mean that these miniature motors can accelerate appreciably in only a few milliseconds. By closing control loops at 10 KHz, high gain and smooth operation are both achievable. Additional advanced control techniques like gain scheduling, feedforward, and field-oriented control (FOC) enable our controls to fully capitalize on the selection of these higher performance motors.
For medical device manufacturers, what are the benefits of partnering with a vertically integrated supplier for the design and manufacturing of motion control subsystems?
Many medical device manufacturers today seek to bring a breakthrough capability or new technology to market that leverages their unique intellectual property (IP). More often, their IP can be found within the application of some critical chemistry, imaging capability, analysis of big data, or perhaps a novel diagnostic procedure. And yet the product or device required to bring this critical IP to market often contains a complex motion control subsystem, typically consisting of motors, gearing, bearings, couplings, lead screws, lubricants, feedback sensors, and drive electronics. All of these electromechanical elements need to work together as a system, consistently and flawlessly, to deliver a critical patient or diagnostics outcome. For medical device manufacturers, success or failure depends heavily on executing a new product design with experienced engineers, validating it properly, and then being able to ramp up production, all while meeting demanding quality control standards.
Vertically integrated motion control suppliers have the component technologies, engineering, and manufacturing experience needed to optimize this key subsystem design for functionality, performance, reliability, quality, cost, and manufacturing. Vertically integrated suppliers who manufacture all of the key component technologies in-house are able to control supply and take end-to-end system responsibility, by optimizing the design integration of all critical components, and by applying best practices in each end-use case. For medical device manufacturers, suppliers certified to ISO 13485 are most qualified to assist in bringing a new product to market.
Successfully designing a handheld robotic tool can be quite challenging, particularly for an OEM engineer wanting to manage the complex control and inverse kinematic equations required on a centralized processor. In this particular case, why should an engineer consider using a distributed control architecture?
Even in physically small robotic devices, there are many compelling reasons to consider a distributed control architecture. If the device is handheld, it’s typical to have a lightweight and flexible cable to power it. With distributed control, intelligent amplifiers are mounted inside the device, close to the driven motor and feedback sensors. This makes it possible then to have a longer, lighter weight cable running into the tool, which typically carries only power and communications. Keeping servo amplifiers inside the device, close to each motor and sensor, has the added benefit of reducing design challenges relating to electromagnetic compatibility (EMC) and electromagnetic interference (EMI). To understand this better, consider that shortening the length of power connections between the motor and the servo amplifier, serves to attenuate normally one of the largest EMI transmitters in most robotic applications, directly at the source. Likewise, minimizing wire length between sensors and receiving amplifiers, also serves to shield these critical feedback signals from the adverse effects of EMI. Beyond the design challenges of reducing EMC and EMI, distributed control reduces the number of wires internal to the device, which might otherwise have to be continuously flexed during normal operation. This mechanical benefit eliminates fatigue failure, helps to extend service life, and improves the overall reliability of the tool.
The good news is that intelligent servo amplifiers with high-speed communication interfaces make it possible to garner the best of both approaches. Given the complex device geometries and challenging system dynamics found in many robotics applications, a centralized processing approach is normally preferred. However, with the selection of an appropriate communication network and intelligent servo amplifier, the problems of control processing and physical interconnection can be separated. Inverse kinematic and complex control equations common to robotics applications can be processed centrally, and also interconnected locally.
When designing a small medical device requiring an embedded motion system with 4–6 axes of control, what is the ideal or recommended network technology?
Certainly there exists more than one correct answer to this question. Two popular choices in the motion control industry today are CANopen and EtherCAT. However, before choosing any network for a motion control application, it’s good to consider not only the number of axes, but also the type of control needed for each axis, as well as the level of coordination required between axes. This helps define what information needs to be passed across the network, how often it needs to be passed, and helps to select a network with sufficient bandwidth.
Starting with the type of control required for each axis, consider where the position and velocity control loops need to be implemented. In general, if these control loops exist locally in the drive, then only target values need to pass over the network, less frequently, say perhaps from 50 to 500 times per second. In this case, the drive resident control loops can be operated at much higher frequency to dynamically stabilize the load without burdening information flow on the network.
With regard to coordination, one of the simplest forms can best be described as sequenced control. Sequenced control, for example, is where one axis needs to complete a move before another can begin a move. In this case, the trajectory generation function that produces target values for each control loop can also exist locally on the drive. This means that perhaps only “start” and “motion complete” messages need to pass over the network, at a significantly reduced rate. In this way, a lower bandwidth network can be used to effectively control and coordinate even hundreds of axes.
However, as the level of coordination between axes is increased, so too is the demand on network bandwidth. One example of higher-level coordination can be described as a vectored movement, where the required motion is described as a distance with an acceleration, velocity and deceleration rate for a point moving along a straight line, for example, in an X-Y plane. In this case, the trajectory generation for each axis is normally performed on a centralized processor. But if the vectored movement needs to be dynamic, the central processor may also need to pass both target and feedforward values to each control loop, and in return, receive status and feedback values back from each loop at a much higher cyclic rate. So, more messages per cycle, at a higher cyclic rate means more communication bandwidth is required.
CAN-based communication is very electrically robust and cost effective for embedded applications, but it has limited bandwidth when compared to real-time Ethernet networks like EtherCAT. Practically speaking, given a 4 to 6 axis embedded motion control application, with a combination of coordinated and sequenced control, CAN is a good candidate for the communications network. As many automation applications become more dynamic and robotics applications become more complex, real-time Ethernet communications, especially EtherCAT, can be the more attractive choice, even when only 4 to 6 axes of motion control are required.
With so many different motor, power transmission, sensor, and control technology options available today to OEM design engineers, what is the right process for determining the best selections?
In the beginning of a new project it normally pays to take a step back and visualize what the end deliverable actually needs to be. In a motion system, this process begins with defining the amount of power that must be delivered to the load in terms of speed and torque (or force) at a given duty cycle. With this starting point, simple thermal analysis can be performed to determine physically how large the motor is likely to be, what type of motor might work best, and whether a geared or direct drive approach might be indicated. Beyond just sizing for power, next it’s critical to consider what quality of motion needs to be delivered to facilitate the end application or process. This normally leads to a discussion of resolution, accuracy, and repeatability of the required motion. This information will help determine whether the drive solution should be open loop or closed loop.
It also helps determine what type of feedback sensor is best: optical or magnetic, linear or rotary, incremental, absolute, or multi-turn absolute. Discussion of motion quality will also lead to a discussion of bearing types, configuration options, lubricants, and some thought about where best to mount feedback sensor(s), and a holding brake, if required. Quality of motion will also lead to a discussion of required bandwidth, stiffness, and smoothness of control. As the discussion evolves, questions about environmental operating conditions like cleanliness, pressure, temperature, and humidity must be taken into consideration. When the best solution starts to take focus, consideration can then be given on how to interconnect selected elements, and how to optimize them for performance, manufacturability, quality, and cost. Finally, consideration can be given on how best to enclose the elements for the end use application. At this point, thermal modeling will typically be greatly refined, and initial prototypes can be offered for further evaluation.
When FAULHABER is contacted at the conceptualization phase of a major new project or product development effort, we respond by pulling together an interdisciplinary team of engineers representing the key stakeholders in technology, design, manufacturing, and quality, to host an open round table discussion with our customer’s design team. We know from experience that the best outcomes and best fit solutions result when we, with the customer, look at the design and manufacturing challenges holistically, and together we drive to the best final solution.