As a university professor, I was pleased to read "Engineering Education: Who's Teaching the Teachers?" \[June 18, p. 22\]. Maintaining laboratories and the curriculum to stay on the leading edge of technology is definitely a multifaceted issue. Because I'm also a professional engineer and consult outside of the university, I can say that the best arrangements are mutually beneficial and have quantifiable deliverables.
For instance, donations of equipment can be extremely desirable if they can be used in the curriculum or in research. If no one at the university is skilled at using the equipment and there are no plans in place to develop research or curriculum around it, then there's no long-term benefit. Analog Devices provided DSP SHARC kits to Tufts University that our students used to develop senior capstone projects, which were presented at IEEE student conferences and will be used as laboratory experiments in the curriculum. Students leave the university skilled in analog products.
Software vendors, such as Altera, expect to see laboratories posted on the class Web page and made available to the general public. Altera even provides free training for professors and student researchers. Other successful collaborations include those where professors work with a company to help develop novel solutions to specific problems of interest to the company. M/A-Com Inc. has excelled in this area with its industrial/university liaison program. M/A-Com hires summer interns and professors to work on company projects. The results of the work allow the company to investigate new directions in a cost-effective manner, which in many cases lead to patents, new products, and an opportunity to influence the curriculum.
Professors gain valuable practical experience to bring back to the classroom. Plus, many participating students gain the confidence to take on more challenging courses in areas like RF and microwave design to enhance their skill set. The demand for such courses is immediately felt in enrollment at the university. Companies aware of specific hardware needs and deficiencies in the curriculum can target their donations much better than toward a "general university fund." Teradyne has sponsored the Simulation Lab at Tufts University to help develop a testing and simulation curriculum.
Finally, the university should utilize its industrial partners on curriculum review boards to assess the quality of course offerings and look for input on future research directions. Thank you for bringing attention to the importance of engineering education and industrial partnerships. Your article was extremely encouraging for any professor trying to expose students to engineering problems where the answers aren't in the back of the book.
Interest In Analog
Thank you for "Field Programmability Pervades Analog Design" \[July 9, p. 56\] on the present state of FPAA technology and related tools. But I find one element of the article that gives a sense of CAD/CAM deja vu. All of the design tools and conventions thereof are proprietary to each manufacturer's product offerings, which begs two questions: What is the current state of industry commitment to VHDL-AMS (IEEE 1076.1) tools, and will FPAA chip vendors adopt that standard?
The last Electronic Design report on VHDL-AMS tools that I saw came from Cheryl Ajluni a few years ago. The IEEE has since published numerous VHDL-AMS proceedings, but the tools technology still isn't mainstream. A more comprehensive update would be welcome, as VHDL-AMS potentially offers an easier marriage between digital and analog in mixed-signal designs without proprietary and technical migration constraints.
Electronics Design Engineer
Lattice Semiconductor of course supports VHDL and Verilog for our digital PLDs. When doing market research with potential ispPAC (programmable analog circuits) customers, we found that once they were familiar with the device concept, they preferred a simple "point-and-click" graphical technique to an HDL approach.
We concluded that the basic difference between a digital PLD and a programmable analog chip is at the heart of the matter. The PLD is an undifferentiated mass of logic that can be used to implement any digital logic construct, and can be anywhere from a few hundred gates to millions of gates. By contrast, the programmable analog chip is a small collection of specific, dedicated analog functions (like op amp, comparator, and DAC). Customers found it very easy to just click on an analog function and select its parameters from pop-up windows, and click on drag-and-drop to do on-chip routing. We provide SPICE model export capability, so the analog chip's functionality can be simulated with the circuit board's other components.
We implemented what the customers asked for. I imagine that if programmable analog chips someday reach the point where they're an undifferentiated (potentially large) mass of analog capability, which can be defined, configured, and extensively place-and-routed by users, then analog HDL support will become indispensable.
If you haven't yet checked out our software, it's free and quick to download. Or, you might just try the one-minute demo of the PAC-Designer software. Visit our Web page: www.latticesemi.com.
Vice President of New Venture Businesses
Suggestions For Low-Cost Designs
The IFD "Low-Cost, 24-V Industrial Controller Is ESD-Protected" \[July 9, p. 99\] suggests using this design as a "low-cost" controller for validating a new process in a small-scale pilot plant, instead of using an "expensive" programmable logic controller (PLC). The chips themselves might be all that's "low cost" about this whole idea. Phill Leyva must be totally discounting the cost of labor to acquire these parts, build a breadboard circuit, test it for functionality, and then write and debug an assembly language program for each process that he wishes to evaluate.
I suggest looking into some of these "large and expensive" PLC systems a little closer. At www.automationdirect.com, you can find a PLC with eight inputs and six outputs for $99, available in a variety of voltages. Software costs a whopping $99 for an excellent 32-bit Windows version of ladder logic programming. Add four channels of 4- to 20-mA analog inputs for about $50, or an LCD operator/message interface for $145. Give yourself reportablility on your PC with a $295 DDE driver package (and $15 for a cable). For about $50 more each, you can network lots of the little guys together.
If you want to spend more for less functionality, try Rockwell Automation's (Allen-Bradley's) Micro line of PLCs (20 I/O costs about $250 and the software starts at $300), or even GE's Micro version. In any event, you can quickly get operational with simple ladder logic programming. I once programmed up a network of three PLCs to operate a piece of production equipment in less than four hours.
Why not just buy an inexpensive Opto 22-type I/O board with modules and an adapter and run your pilot plant off your PC if you really like to spend a lot of time programming, but not anywhere near as much time using Basic as opposed to assembly? See www.cyberresearch.com for some low-cost products along this line. I once programmed my Osborne 1 parallel printer port as I/O and wrote a program to control it from "real English sentences" for lab testing a process. It was fun, but very labor-intensive.
Stephen H. Jones