You are in:  Components | FPGAs & Asics


Read The Magazine

Issue: 16 - 22 Dec, 2009
Get Electronics Weekly

How to design motor control on an FPGA

Wednesday 14 January 2009 12:00

With roadmaps for processor/DSP technology offering increasing performance you might wonder how FPGAs could be the next stage in the evolution of motor control.

Certainly the cost of microprocessors and DSPs is steadily decreasing and motion control system manufacturers create well-controlled and reasonably priced drives.

However, the industrial market is demanding lower costs and increasing levels of flexibility – simply making a better drive is just not enough. In the industrial space, cost reduction means creating factory-wide information systems, low-cost control networks and manufacturing systems that can be quickly and cheaply reconfigured to manufacture a different product. These requirements can be met through the use of equipment that supports Industrial Ethernet.

Industrial Ethernet (IE) is the challenge to the equipment manufacturers because there are many different competing standards and (by industrial timescales) they are evolving rapidly as each attempts to dominate the market space. Using Asics to create one-off interface solutions for each IE standard is an obvious solution but brings the burden of multiple hardware development projects and the problem of how to support these products in the field when the Asics become obsolete (due to newer versions being available with support for new/updated IE standards).

However, a board that carries a low-cost FPGA device and Ethernet PHY transceivers can support most IE standards. Protocol-specific hardware is used in programmable logic and protocol-specific software is implemented on soft processor cores like the Nios II processor.

The programmable logic and I/O of the FPGA device can also be used to implement other, motor-related digital I/O. Examples are shaft encoders, A/D converters and power circuits that require PWM outputs. Again, the flexibility of the FPGA enables you to implement as many interfaces as you need and gives you the ability to use the latest technologies such as EnDAT or BiSS encoders and sigma delta A/D converters. With an FPGA you can now also support other types of interface like CAN or PCI, as well as traditional field buses like Profibus.

With the connectivity issue solved, the next obvious step is to look at how you can save cost by integrating more functionality into the FPGA.

FPGA devices are known for DSP processing and motor drives generally require lots of processing capacity. VHDL or Verilog implementations of motor control algorithms are inherently parallel, giving high performance as well as systems that are easy to maintain and modify.

If you are developing multi-axis systems you can use logic to control multiple motors in one FPGA, but with the availability of high-speed Industrial Ethernet (for example, EtherCAT, SERCOSIII, PROFINET IRT), you have the option to use the computation of high-level control on a remote logic controller or industrial PC (IPC).

A centralised system design makes it easy to co-ordinate control of the system as a whole, as well as the benefits of increased flexibility, lower costs and a non-proprietary, ‘open’ solution where the end-user can have full algorithmic control.

This type of implementation is ultimately limited by the processing power and I/O capability of the IPC, but fortunately the high performance of real-time IE protocols and the calculation of local drive control parameters on the FPGA make it possible to easily create multi-axis IPC-based drive control systems.

The future of motion control systems has to offer reduction in system costs and implementation times. Due to their fixed features and inability to quickly adapt to newer technologies, traditional microcontroller or DSP-based solutions will struggle to keep up with the accelerating demand for flexibility in the motion control market. FPGAs deliver this flexibility as well as ways to enhance performance and reduce cost at both the drive and system level. It will be interesting to see how quickly the normally slow-moving industrial market adopts this new technology and who will profit from it most. Often it is not the strongest that survive, but those that can adapt to change the quickest.

Stefano Zammattio works for Altera Europe

Recommend this article

View the ElectronicsWeekly.com topic zones:

Electronics Weekly Zone - PowerElectronics Weekly Zone - Test & Measurement


 

Sign-up for the ElectronicsWeekly.com newsletters:

Electronics Weekly newsletters

Resources

Most Viewed

Blog roll