POWER PMAC – Motion Control Innovation

Controlling the Advanced Technology Solar Telescope

The largest solar telescope in the world, with unprecedented abilities to view details of the sun, needed to have the most sophisticated control system available.

The Advanced Technology Solar Telescope (ATST) project is expected to provide the sharpest views ever taken of the solar surface. Through the use of the ATST, scientists expect to learn how cosmic magnetic fields are generated and destroyed, and learn what role the fields play in the organization of plasma structures and the impulsive releases of energy found everywhere in the universe. But more importantly, researchers are interested in the mechanisms responsible for solar variability that eventually affects the Earth.

The Telescope’s optic support structure, which includes the mirror assemblies, is expected to weigh nearly 75 tons, the mount base nearly 90 tons, and the Coude rotator 160 tons. The 84-foot diameter enclosure will be thermally controlled, highly ventilated, and will be a co-rotating hybrid that has independent rotation when the telescope is positioned at zenith. The enclosure is where the controllers and much of the instrumentation will be located.

Control and Instrumentation Needs

Overall, the instrumentation would incorporate over one hundred motion actuators at the beginning—that number will grow as instruments are added to the system. While each individual actuator will have minimal impact on the ATST Coudé environment by itself, all of the motion control hardware taken together will be a dominant factor in the generation of heat and EMI inside the telescope enclosure.

One critical area of consideration for motion control in the ATST was the expertise required to maintain multiple systems. If each area of control chose to use a different vendor for their motion controller, a tremendous burden would be placed on both the hardware and software maintenance personnel with respect to documentation, spare parts, periodic upgrades, maintenance of software code, and expertise in each of the systems.

The ATST motion control requirements had to flow down from the General Requirements & Specifications for the Design and Fabrication of ATST SPEC- 0070. A rough distillation of SPEC-0070 requirements that applied directly to the motion control hardware design and selection included a design solution that simplified components and ensured the lowest cost and highest reliability without sacrificing performance; provided an ease of removal, installation, servicing, alignment, and adjustment; and met full life expectancy of the telescope (40 years) if possible.  Finally, the controllers had to comply with NEC, minimize and offer immunity to EMI, comply with FCC reg Part 15 class B (class A for computer room), provide ESD immunity IEC 1000-4-2, and RFI immunity IEC 1000-4-3, IEC 1000-4-6, as well as provide power-line disturbance immunity per IEC 1000-4-9, IEC 1000-4-13.

Controller Selection

ATST engineering team identified vendors based on their ability to provide a product line that could run the gamut from 30+ axes of control to a simple only 2 axes of control, which would span the needs of all the instrumentation packages being used. The priority was to identify a family of motion controllers with a common Linux friendly socket-based software interface. Another interest was to find a solution that would minimize the ATST software development effort required to communicate through the socket-based interface.

All of the vendors considered used a Windows interface whereas the ATST software framework was Linux based. Each manufacturer had a different method to solve the challenge, but in each case additional development time would be required. The extra layer of software needed to interface to the controllers also required maintenance and documentation, which would require the software team to maintain another software layer.

No vendor emerged that provided everything desired for ATST instrumentation, so the vendors were judged on criteria such as company viability, price, physical size, integration capabilities, encoder options, and the elegance of the overall solution. The ATST instrument partners weighed in on the study and unanimously expressed concern over the price of the high-end options that were being evaluated, as their budgets were constrained. Since one purpose of the study was to determine a standard system that the software team would use to develop sample code (in an effort to ease instrument partner development efforts), this caused an expansion in the scope of the investigation with the addition of a few smaller-system vendors — primarily to see if the price point and development effort could be reduced.

The ATST first light instruments would all contain over 20 axes of motion each with the exception of the Visual Broadband Imager (VBI), which would contain at least 8 axis. For this reason, the 30-axes controllers were selected as the best over-all solution for comparisons. The best-value choice came down to minor differences between the vendor offerings.

After a complete analysis, Delta Tau’s new controller, Power PMAC, won ahead of the competition meeting most of the requirements defined by the ATST engineering team. Power PMAC not only supported the required price, physical size, integration capabilities, and feedback options, it also provided an elegant solution for a Linux friendly socket based interface since the controller itself is based upon a Linux OS with real time kernel. This allows use of well-defined SSH and Telnet communication protocols between the controller and other devices at ATST.

Power PMAC is equipped with a dedicated motion engine which can be used to perform a wide variety of motion profiles at all levels of complexity starting at independent and simple trajectories to coordinated and sophisticated for as many as 256 motors defined in as many as 128 independent coordinate systems. Also using a task scheduler, Power PMAC provides unrivaled access to all levels of process, which enables the user to customize the controller/computer as required by project definitions. Power PMAC can be simultaneously programmed in several programming language/environments such as script language and IEC-61131 for high level and synchronized code and ANSI C for low level programming. In Addition, Power PMAC can use the unique capabilities of MATLAB®/Simulink® and LabView® software packages which are highly desirable for scientific/academic use. Power PMAC can be used directly as a platform for executing custom designed Phase/Servo routine auto-generated code from Simulink® Realtime workshop engine and it can communicate with LabView for process control and data acquisition purposes.

Summary and Conclusion

Besides winning out over other controllers, Delta Tau’s system was also price competitive, highly capable, and offered the simplest Linux interface solution. Delta Tau also offered lower-end solutions with much lower price points putting it only slightly more expensive than the small controller that was considered.

After Delta Tau won the comparison, informal inquires were made with existing telescopes that use Delta Tau control systems. It was found that the company was generally liked and well accepted in the telescope community. It was also found that mature and well-tested Linux interfaces to Delta Tau systems were available to ATST from other users. Delta Tau customer service was reported to be good, and a forum was available with a large and very active user community. There were also consultants who specialized in developing control systems design with Delta Tau controllers. All of these reasons gave the ATST team a high degree of confidence that Delta Tau was the solution offering the best value for instrument control.

Delta Tau Controllers

Most engineers are comfortable with everything a square rail bearing is good for and use them regularly for motion support. I wanted to take a quick moment this morning to discuss the pro’s and con’s of a roller wheel bearing design.

If you are familiar with the HPLA series of products (as well as BWC, and many other competitors) you know about the fact that they utilize a wheel design instead of a square rail / recirculating bearing design. The benefits to a roller wheel design include:
•Much lighter weight systems
•Lower cost systems versus square rail designs
•More “environmentally hardened”
•Customizable
•Less maintenance required
•More forgiving for misalignment issues
Of course as with any alternate solution there are negatives as well. Some of the downsides to roller wheel designs are :

•Need a fixed preload method to eliminate carriage slop
•Shock loading can eliminate preload
•Travel straightness / flatness will not be as precise
These are just some of the pros and cons of a roller wheel bearing design. When it is time for your next motion project, consider all options before just deciding the status quo makes the most sense for the application at hand.

Often, there is skepticism or concern to use timing belt driven actuators where a high amount of axial force is required. Parker’s most popular timing belt driven actuator, the HPLA series, uses a polyurethane timing belt with reinforced wound steel cords. The steel cord gives the belt an increased tensile load rating. What does that mean? To give you a better idea of how strong these belts are, the standard belt used on an 80 mm wide product (3″ wide) has a tensile strength rating strong enough to hold an average size polar bear (did you know an average adult male polar bear weighs 900 lbs?) and a breaking strength rating strong enough to hold a Corvette.
The key point is don’t be quick to think the belt is your limiting factor when sizing an actuator. There are other mechanical components to take into account.
Things to consider before assuming the belt is the limiting factor:
•Radial loading of your motor or gearbox if it is a direct mount
•Accelerations. High accelerations on a bearings will cause the bearings to skid instead of roll which will create flat spots and failures.
•Moment loading. Be sure the carriage is properly sized to support normal or cantilevered loads.
The other important things to consider for belt failures are :

•Elongation or stretch : at it’s maximum tensile load a steel reinforced 25AT10 timing belt will elongate 4 mm / m (or roughly 0.4%) and this generally occurs very early in the machine installation/usage.
•Slippage: An improperly tensioned belt can jump teeth easily with aggressive moves resulting in loss of position.
•Tracking: A belt that doesn’t track true to the travel may move to one side and excessively load a pully/bearing creating a failure mode.
•Clamping: If a belt is not properly clamped or gripped, it will pull out of the clamping device with aggressive moves resulting in crashing units and loss of motion.
Belts are a great way to lower the cost of a general motion system versus using a ballscrew. Knowing the “gotcha’s” will help you implement a successful motion solution with the best benefit payoffs to the application at hand.

Thank you to the many customers that attended last week’s session on Delta Tau’s Power PMAC. It was clear from the feedback that this controller will play a major role in the Automation world.