Time and Motion
September 6, 2010 by Steve Meyer
Motion control and mechatronics technology have grown up as a control discipline at somewhat of a disadvantage. Mechanical engineering programs tend to ignore electric motors as being, well, electrical. And electrical engineering programs ignore what’s attached to the motors as being mechanical, which it usually is. This prevents the two disciplines from correctly understanding each other and more importantly, can prevent needed breakthroughs in performance for the companies that employ mechatronics.
This has been changing in recent years with a growing number of mechatronic programs at major universities and wildly popular programs like the First Robotics Competition. Interest in mechatronics has spawned a wave of contests sponsored by manufacturers to educate young people about the technology and make future customers in the process. The greater benefit is the number of creative individuals being exposed to the technology in grade school and high school. Undoubtedly, we will all be the beneficiaries of some new inventions that will be coming in the future.
But there are still some interesting subtleties that arise in motion control. A common problem is defining coordinated motion. This is because the precise behavior has to be described BOTH as mechanical objectives and correctly modeled in the control system hardware.
There can be many axes of motion in a particular machine. But they are rarely coordinated in the absolute sense. And this is an important distinction to keep in mind during design of the control system. Most of the axes, maybe 80% of them, will require a start signal to coordinate their operation while the equipment is operating. Rarely do the axes require time synchronous control.
Want to know the secret? Simple. Think of a machine that does “Tic Tac Toe” versus one that “Draws a Circle”. Tic Tac Toe can be done with simple Cartesian linear axes with no coordination, other than a start bit and a done bit. You can have a busy signal if you want to get fancy.
Ever try to draw a circle with an Etch a Sketch? It is harder than it looks. Because every tiny point must be coordinated between the two separate sources of motion.
And when you draw a circle, what happens as the time constraint is decreased? As you go faster the acceleration and inflection points of motion become much more critical. Generally, this produces increasing error in the actual trajectory.
Which leads to the “Stump the Band” question for would-be mechatronics engineers. What is the one variable that connects all aspects of mechanical motion and electrical control together? Time
And there is no end of importance in this fact.
When you try to Draw A Circle, time is absolutely essential. The incremental change in time, delta-t, will impact how precise the circle is. And the control system programming and execution will not be of much help in regulating this. Neither will servo tuning.
For those of you planning a multi-axis system, let me share one further time-oriented thought. When you have two truly coordinated axes, and they can be anything, a servo following an ac frequency drive (don’t laugh, I did this once and it worked great) make sure that if you are using a PLC that the coordinated axes are on the same control module.
Most PLC’s use a separate processor to run up to 4 axes of motion at a time. The slave axes have to all have to be on the same module or the backplane update will limit the performance of the motion. You will see perfect performance up to some speed and then synchronism will be lost because the new position update is going through the backplane and the servo is being commanded to follow old position information.
A Phd mathematician friend once told me, after several years of work doing his dissertation, that electric motor design is a system in which there are 22 variables that are linked together in complex relationships. Since the human mind is not capable of operating at this level of complexity with all the variables simultaneously, we model these complex systems in groups of variables and perform boundary searches looking for relative optimum solutions.
That’s kind of how it is working in motion control systems. Generally there are many variables competing simultaneously within a project. The hardest thing to do is group the variables together logically and put them in order of priority. But this effort early in a given project will help you guide the project to completion even if you have to make changes as the project progresses. Staying focused on the most important objective will help to prevent becoming distracted by side issues and will help you get more done in less time.
But the really big thing is looking at operating conditions in terms of their relationship to time. And there are some interesting insights from looking at things with respect to time.
Simple things like understanding the relationship of the forces acting in the system from F=ma, in which acceleration has a time component in the denominator. In the scalar value, as the time constraint for a move decreases, the force required is increasing arithmetically. Sizing of a servo motor, for example, can quickly reach limits because the increasing force needed to accelerate the load increases the size of the motor. The motor’s own inertia now becomes a constraint working against the desired performance. And there applications where the last 10 milliseconds of move time will cause the motor sizing application to fail.
Even more interesting, and hopefully more useful, is considering the derivative of time. Not so much because of the numerical value, which isn’t as helpful, but because the first derivative of time shows up in a number of critical relationships.
The current flowing through a power mosfet has a limit of di/dt which defines the breakdown limit of the device. This is the instantaneous current rate or the rate at which the mosfet can increase the current in a specific amount of time. This limit also governs the acceleration rate of the motor. The motor acceleration can be considered the sum of the inertia of the rotor and the ability of the amplifier to put current through the stator winding.
Reversal stress in a mechanical transmission line is also limited by dT/dt, the instantaneous change in torque over a change in time. Think about the change in the drive train of a wind turbine gear reducer. The gears may be unloaded when the propellers are not turning. Then the wind comes along and pushes the propellers and sends an impulse force pushing the gears into mesh at full load in a very short amount of time. This is a bit like hammering the gears and will cause catastrophic failure.
Similarly in servo systems, reversal stress creates the same conditions every time a load is reversed. How much, if any, dwell time is allowed for the system to pause before reversing? In the metals industry, high speed stamping and forming is done with continuously rotating inertia wheels which allow the impact to occur through linking arms because the direct stress at the multi-ton force levels cannot be controlled. The same is true considering the motion of the pumping jack using in oil & gas production.
More next week.
Top 5 Electrical Considerations for Mechanical Engineers
Previously, we covered mechanical considerations for electrical engineers. Now, we give the other side a chance to speak. Here are five targeted pieces of advice for mechanical engineers responsible for electromechanical systems, from the perspective of an electrical engineer.
Mechatronics systems intelligently integrate mechanical and electrical elements to perform increasingly complex and demanding functions. When designing electromechanical systems, mechanical engineers and electrical engineers may tend to emphasize the technologies, components, and design principles from their single area of expertise—which can lead to systems with higher operating costs, increased maintenance demands, and less than optimal performance. As an electrical engineer involved in helping OEMs and manufacturers design and build mechatronic systems, I’ve seen how inefficiencies and unnecessary complexity can be unintentionally designed into machines.
Better mechatronic systems can be created when mechanical engineers consider five crucial concepts while designing manufacturing systems, to derive the greatest value and efficiency electronics systems can offer to the manufacturing process.
1: Create a clean design
Good mechatronics design starts with good mechanical design – the best electronics and electrical systems cannot compensate for poor mechanical design. The most successful designs are “clean.” They feature a strong, rigid frame, using materials and structural principles to ensure that, whatever motion the machine undergoes, its long-term stability is “engineered” in.
Make sure that rigid bearings and support are used where motors are mounted on machines; this helps prevent shafts from being sheared off due to microfractures that occur because the motor shaft is mounted out of alignment with a pillow block bearing or gearbox input planetary gear. Place motors on the machine in the best location so that operators aren’t accidentally stepping on cables and connectors and causing damage; and design machine guarding with easy access points to get to motors mounted under the wing base of the machine while still protecting them against harsh environments.
Most importantly, a clean design balances mass and motion: sturdy, durable framing that withstands years of vibration and shock, combined with lighter-weight components for the moving parts of the machine. This combination helps reduce mass, delivers more energy-efficient motion, and makes it easier to size-up smaller motor/drive components for the machine. We’ve seen a lot of very innovative mechanical machine designs over the years, and a clean design makes the largest contribution to a machine’s longevity, robustness, and lowest overall cost of ownership.
2: Directly couple the motor to the load
Effective mechatronics starts with a “clean slate” design. In the past, machines were often built around a single ac motor powering a machine line shaft, to which were attached gearboxes, pulleys, sprockets, chain drives and other mechanical devices for moving individual areas of the machine in synchronization – an approach to powering manufacturing that literally can be traced back to the dawn of the Industrial Revolution.
Consider replacing this architecture with individual servomotors coupled directly to the load you are moving. There are multiple design, machine cost, and operational advantages to this idea (which a surprising number of machine designs do not use). First, consider cost: every time you add a gearbox, you add multiple costs: it’s an additional point of failure, it has to be lubricated, and it needs spare parts. Plus, you add mechanical backlash that must be compensated for during machine commissioning every time you have a product changeover – motion and axes synchronization complexity that today’s intelligent drives and servomotors eliminate.
When you strategically locate servomotors as close as possible to the area of motion they are serving, the incremental cost of electric drive components is almost completely offset by eliminating the cost of mechanical components and labor that must be purchased, machined, assembled and configured. In particular, not having to stock multiple sets of sprockets, gears and cams, as well as the time involved in changeovers with mechanical drives, can really drive down the total cost of ownership for the machine.
Ultimately, this design approach greatly reduces windup and backlash, as well as improves machine commissioning time; and current state-of-the-art direct drives, direct motors, and linear motors let you run higher gains and improve the machine’s performance.
Consideration #3: Use electronic gearing and camming
Today’s electronic drives and motion control platforms give mechanical engineers, a powerful, flexible tool to improve the accuracy and performance of the machines you design. This technology lets you create a virtual “electronic line shaft” that can electronically synchronize all the drives and motors on the machine, eliminating the mechanical line shaft. In the process, you can dramatically improve axes synchronization and accuracy – from 1/16th or 1/32nd of an inch typical with mechanical line shafts, down to motion precision closer to hundredths or even thousandths of an inch with electronic line shafting.
And this synchronization can be accomplished with zero mechanical backlash – and fewer product jams. It also eliminates a host of mechanical adjustments to bring the machine online, as well as the operator adjustments each time the machine is stopped and restarted.
Electronic gearing and camming makes machine changeover completely programmable: For example, the use of FlexProfile technology lets operators load machine recipes with the touch of a button on the HMI screen, and the changes are made in the control and servo system to run the next product.
The FlexProfile camming technology makes it possible to build multisegmented cam profiles based on position, velocity, or time-based motion profiles. When you change a section of the electronic cam with a recipe change through the HMI, the control platform will automatically optimize the rest of the cam profile across all of the machine’s motion elements. This enables the machine to run a shorter cycle time, or provide smoother dynamics for the machine, even though a change has occurred such as a different bag seal time or flap tucking cam position on a cartoning machine.
Consideration #4: Incorporate energy-efficient technology
One of the fastest growing costs for any manufacturing operation is energy – and good mechatronic design can help control these costs through the application of electric drive and motor systems designed to save energy.
In machines that use servomotors directly coupled to critical axes of motion, and that also use electronic synchronization and camming, the proper sizing of the servo system can create a highly energy efficient machine.
Proper sizing requires an accurate assessment of several motion factors (motor by motor): How fast the axis needs to accelerate, the size of the mass you’re trying to move, and how precise the acceleration and deceleration needs to be. Undersizing will lead to strains on the drives and motors; oversizing will draw too much power to do too little work.
Some of today’s most cutting edge systems, such as the Rexroth IndraDrive Mi integrated drive/motor systems, include a highly energy efficient feature: bus sharing. Multiple drives are daisy-chained together and share power from the same bus; in many multi-axis machines, as some motors are accelerating up to speed (drawing power), others are decelerating (regeneration power). With bus sharing, rather than having to deliver maximum power to the accelerating motors and bleed off the decelerating motors into heat across a bleeder resister, power is shared, so the machine’s power consumption is significantly reduced.
A further energy-efficient technology is called regenerative power supplies. In many machines, multiple servomotors will decelerate at the same time, boosting the voltage to excess levels on the power bus. Older generation electrical drives would bleed that excess electrical energy as heat – wasting the power, and adding to the factory floor’s heat production, requiring additional cabinet cooling. With regenerative power supplies coupled to a shared bus system, what was once wasted power can now be fed back through the shared bus and sold back to the electric company.
Consideration #5: Use HMI’s for better troubleshooting
User-friendly intelligence is now available through today’s touchscreen HMIs. Machine layout drawings and schematics can be incorporated into control menus and diagnostic tools, to better manage the machine’s day-to-day operation and troubleshooting. Drawings and interactive instructional tools can not only show the precise point where a problem is – they can also step the operator through the tasks to restart production.
Advanced graphics like this can be combined with the distributed intelligence inherent in servomotor-driven machines, to prevent machine failures or faults before they happen. With such predictive maintenance, this capability lets you or machine designers set fault tolerance bands in drives and then monitor drive performance. Electric drives and motors allow a broad range of conditions to be monitored – conditions that are directly associated with mechanical performance; variations in load, temperature, vibration, torque, belt tightness, gear meshing are all mechanical events that generate changes in the torque profile of an electric drive and motor moving those machine elements. Mechanical engineers can set tolerance bands for these components, and if they exceed them, then predictive maintenance alerts can be clearly and intelligently displayed through the HMI to operators, along with specific advice about next steps to take to correct the issue before it becomes a serious production problem or something that can damage the machine.
Blending technologies for optimal value
Every electromechanical system should perform its designed function with the minimal use of energy, motion and components required to get the job done – that’s the fundamental goal of any engineer. Electrical drive and servomotor systems now offer a wealth of reliable, energy-efficient, digitally intelligent platforms to power the integrated vision of mechatronics to greater value and more innovative manufacturing and automation solutions.
Hopefully, the five considerations described here demonstrate the advantages that today’s electric drives and controls offer, helping you simplify certain mechanical design and engineering challenges and provide new resources for driving innovation and creativity in machine design.
www.boschrexroth-us.com
Thought-Controlled Computers Progressing
New instruments for imaging human brain activity, such as fMRI, offer a wonderful opportunity to study mechanisms in the brain.
Our group develops statistical machine learning algorithms to analyze fMRI data. We are specifically interested in algorithms that can learn to identify and track the cognitive processes that give rise to observed fMRI data.
Watch the video demonstration of our computer program decoding which candidate word a person is thinking about, based only on the neural activity captured in their fMRI data. The program was trained using fMRI data from other people, indicating that our different brains encode word meanings in quite similar ways.
What the data look like: In one fMRI study we trained our algorithms to decode whether the words being read by a human subject were about tools, buildings, food, or several other semantic categories. The trained classifier is 90% accurate, for example, discriminating whether the subject is reading words about tools or buildings.
The following figure shows, for each of three different subjects, the degree to which different brain locations can help predict the word's semantic category. Red and yellow voxels are most predictive. Note the most predictive regions in different subjects are in similar locations.
