mechatronic-design

Find. Learn. Apply.

"Knowing is not enough; we must apply."

- Leonardo da Vinci

Subscribe in NewsGator Online Subscribe in Bloglines

Intelligence moves from controllers to drives

Nathan Massey

Sales Engineer
B&R Industrial Automation Corp.
Roswell, Ga.

Many designers wish to convert from central to distributed machine control. They have good reasons: Distributed (or decentral) control maximizes communication, handles additional axes concurrently, and lightens the processor's load. It accomplishes these tasks by dividing responsibilities among all drives and implementing real-time Ethernet communication. In turn, networks accommodate a larger number of synchronized axes.

Obstacles to central control

Centralized motion control with a speed interface
The controller receives position feedback from all drives on the network, calculates new setpoints, and transmits back to the drives. These constant cycles burden the motion controller and cause throughput to suffer. In this configuration, two velocity control loops run before the central controller closes the position loop.

Centralized motion control with a speed interface The controller receives position feedback from all drives on the network, calculates new setpoints, and transmits back to the drives. These constant cycles burden the motion controller and cause throughput to suffer. In this configuration, two velocity control loops run before the central controller closes the position loop.

In a centralized motion control system, each motor's position must be visible at all times to the controller. In lengthy cycles, the controller queries individual drives, calculates new parameters for all motors, and then broadcasts results to each drive in the system.

Controllers employing a speed interface receive position feedback from the drives and calculate corresponding motor speed. In turn, the controller finds a new set speed for each axis and transmits to the drives, often one at a time. Each drive must then determine current and acceleration from the new set point. On the other hand, controllers with a position interface define and transmit position set points for each axis to the drives. The drives then close both position and velocity loops.

Although centralization is the traditional control approach for motion, it overburdens processing power — especially as machines increase in complexity and require additional axes. This, naturally, creates a cost barrier: Adding depth to control loops and improving axis coordination requires more powerful and expensive processors.

Network cycle times also tend to lag in centralized systems. One delay comes from control loops inside the drive, which do not close until the motion controller calculates and broadcasts appropriate parameters. Additionally, multiple axes increase network traffic as they simultaneously attempt to transmit current position data and receive feedback.

Due to these factors, many centralized motion systems restrict axes to either eight or 16 per motion controller. Even before reaching these limits, however, system operation degrades as cycle times depend strongly on the number of axes. For many complex designs with multiple axes, centralized motion control is not a viable solution.

Going decentral

Decentralized motion control
One or more master axes broadcast position updates over the entire network and each slave axis rapidly calculates its own parameters. Axis coordination and position control are handled internally by each drive, making a central motion controller unnecessary. Consequently, cycle time improves greatly and is more economical.

Decentralized motion control One or more master axes broadcast position updates over the entire network and each slave axis rapidly calculates its own parameters. Axis coordination and position control are handled internally by each drive, making a central motion controller unnecessary. Consequently, cycle time improves greatly and is more economical.

Decentralized systems overcome many of the aforementioned obstacles by distributing motion control responsibilities among the drives in a network. Each drive manages its own axis parameters, eliminating the need for a central motion controller.-This happens when one or more axes transmit the master position, while others calculate their respective set points.

Consequently, processor and network load are reduced and system operation improves. The drives coordinate axes by distributing control functions into small, quick tasks. Now, only the master position broadcasts over the network, instead of transmitting and receiving set points for each drive. Costs decrease because motion control results do not depend on the processor or number of axes run by the controller. Decentralization also means that more axes can be configured onto a single network. The maximum number of drives depends only on the network's capabilities. Network bandwidth is maximized by multiplexing time slots for slave drives, allowing master axes to broadcast over the network every cycle.

Want to use this article? Click here for options!
© 2008 Penton Media Inc.



Videos and Webcasts


The Latest Videos from EngineeringTV.com


Email Newsletter Signup

Stay on top of the intertwined world of mechanical and electronic systems with the new Mechatronics Design newsletter from the editors of Machine Design, Electronic Design and Motion System Design. Each issue, delivered right to your desktop, covers into the latest news and technical articles on topics such as servocontrols and control theory, mechanical drives, information technology, and electronics.

>Click here to Sign Up!
Click here to view our recent eNewsletters

Brought to You By

Machine Design MSD Electronic Design

Back to Top