By Stuart McFarlane
The name of the game in improved design of machine controls is cost and schedule. Time to market and cost are huge drivers in the development of improvements in control design. Advances in FPGA-based technology can improve performance dramatically with reduced time and development cost.
When the R&D team is looking to improve a certain area of control, providing a proof of concept (POC) rapidly and economically is extremely beneficial. This POC will provide critical information for the design of production controls to be incorporated into the target machine.
For example, the controls engineer desires to incorporate two absolute encoders with synchronous serial interface and two differential incremental encoders as inputs with analog outputs for position and velocity and an incremental digital strobe that changes increment size as a function of position. Velocity frequency response of the system must be at least 300 Hz and the strobe must be generated within 1 msec of the desired positions. Expect rapid changes and experimentation at the beginning until acceptable results are achieved or the approach abandoned.
Before the introduction of standard platforms with sufficient power to solve this type of application, the engineer would research standard products that could be combined to perform the application or try to solve it with a custom solution. Standard products typically do not have the flexibility to be combined in non-standard ways. The application described above cant be solved with off-the-shelf standard products. The requirements are just too stringent.
Custom hardware/software solutions are time-consuming and expensive, and they require specialized expertise to implement. Producing a circuit board with signal conditioning, programming a processor and getting the whole system to work correctly is reserved for volume production with larger budgets, not POC projects.
The typical outcome of this exercise was to redesign the system with a less-than-optimal approach. Innovations and advances in control design were slowed by the inability to apply high-performance hardware and software in a cost-effective and timely fashion.
Standard FPGA-based platforms and software tools have enabled applications that were previously not possible or feasible. By using digital techniques, signal-to-noise ratios have improved, analog drift has been reduced, and flexibility has been increased. In addition, high-performance, closed-loop control, tight synchronization, custom communication protocols and other benefits have been realized in many applications. A system can be assembled from these standard components in a very short timeframe.
Algorithm development can be prototyped and tested in a software test bed and deployed to hardware in a matter of days. Its possible to perform analog algorithms in the hundreds of kHz range and digital algorithms in the MHz range.
The skill set for this activity is higher than the novice level, but much lower than the expert status required for previous hardware solutions. This same platform can gather data for experimentation analysis and validation of the POC.
The information generated by the rapid development stage contributed to the requirements for a fixed-function device that could be incorporated into the machine without multiple design cycles. The same platform used for the rapid prototype can be reprogrammed to provide stimulus signals to the production controller for validation and testing. VHDL code for the production controller can be tested on the same hardware before the controller prototypes are even delivered.
Theres still room for improvement in this new technology. Higher speeds, larger programs, richer programming tools and more standard signal conditioning are all on the evolutionary path. Meeting the performance requirements of these types of systems would have been very difficult just a few years ago. Having a rapid deployment platform has enabled the use of FPGA technology for many more applications with many more to come.
Stuart McFarlane is vice president at Viewpoint Systems, a system integrator in Rochester, N.Y.
