processor preferences?

[eschulzeatbronxscience]

Salutations, etc.<p>I am developing an advanced high school robotics course. The course is to be heavily "hands-on" in a shop/lab environment. I would like the projects to be "scratch-built" from the MCU chip on up. Early on, I dismissed the BASIC stamp as too expensive for individual student projects. I have since done a bit of research on Microchip components as the logical alternative. The problem is that during my research, a bunch of other control chip options have presented themselves.
Atmel, TI,Motorola, etc. all show promise. I am emphasizing the use of digital control of analog devices as a theoretical framework for study, so I/O is everything.<p>Is there a "hands down" best bang for the buck programmable control chip series. Without advice, I am inclined towards the less common 16 (or more) bit types.<p>Is there a quick approach to learning C++ for programmable automation purposes?

[bwts]

Microchip PICs every time for me especially the Flash ones as they can B reprogrammed 1000's of times and need very little external support.<p>An easy way to teach C++ for control applications is to just teach C its has less going on and u can use wot ever C++ elements u need to make ur life easier, any C++ compiler will compile C.<p>B)

[L. Daniel Rosa]

The classes that I'm aware of use Motorola microcontrollers. There is also a robotics hobby group in Seattle that appears to have them as the primary choice. My preferences are with Microchip though. You will get bang-for-the-buck, and the PICs are all serially programmable (many other manufacturers are as well, but not all) which simplifies the hardware requirements.<p>With over a hundred chips in a dozen+ families in three (or so) core architectures, I'd suggest choosing the best in any group. That is, don't bother getting the 12c508 and the 12f675- only get the '675 (or the '683 if it's out by the time you start your class and the price isn't much higher). Addendum: Only get the "f" series- anything else may be a pain to impossible to erase. Also, be sure that the device in question has A/D conversion, your emphasis on digital control of analog devices necessitates this (don't worry, it's available in most families).<p>As for the programming, there are compilers for C, JAL, Basic (numerous variants, I'm not sure I'd recognize any of them), and others. Machine code will be cleaner, smaller and faster, but C will get the assignment done sooner.<p>I steer clear of uCs that have an interpreter on board. They may have useful object modules for my applications, but I've already determined that I'd rather get familiar with the core and write native code.

[Bert Russell]

Regarding programing Pic's, I've been using a Microchip "Picstart Plus" programmer. However, I have just run into an unexpected problem. The Picstart Plus seems to be limited in the size of the program it will handle. My program consists of 96 lines of code, including a look up table, and it assembles fine, but refuses to program a Pic16F84. I upgraded the Picstart to a 17C44, and it works fine for smaller programs.
So, if anyone is thinking of buying a programer, it will probably pay to go for the Promate 2, even if it is $695.

[dribach]

seeing as how this is for a course, rather than personal hobby, you might not have to worry too much about cost of the programmer vs. cost of the chip. you may only need one or two programmers that the students share during class time, meaning that the real cost savings over several years will come from using a cheaper chip. i personally have been looking into the Atmel family of 8-bit RISC chips. 8-bit means less instructions for students to have to learn, and therefore more time learning the good stuff (i also prefer writing native code). also, there's a huge selection to choose from, depending on the capabilities the chips will need, and the package you want (they have everything from DIP-8 to MLF-64). you don't hear much about these chips, but in my opinion they're definitely worth looking into.

[mediaprefect]

Remember that no matter what chip you choose, they will be able to use what they learn on other chip families. You don't need the latest & greatest. I would suggest that features such as flash memory/in-circuit programming would be of high importance in making learning more fun/less of a chore. Also since you are interested in lots of hands-on, a family such as the PIC for which you can find a wide set of published projects for could stimulate learning by studying code examples for existing projects. After you learn the basics you then build on existing work. Students can build on ideas found in code examples in articles e.g. Servo, Nuts&Volts, Circuit Cellar and the SquareOne Publishing books etc.
You can also jump to the Scenix/Ubicom family for speed etc. Thats what they are currently using at Roosevelt University.
http://www.dontronics.com/ is also a good resource for PIC and Atmel including boards & programmers.