Atmel AVR Evaluation

AVR timeline

Timeline for AVR projects.

Atmel AVR evaluation board

The quest for the lowest possible cost for a “pick-by-light” controller lead to the Atmel AVR range which could sink enough current per high-drive output port to drive an LED. (Port A could sink 20mA). The in-system programming, plenty of I/O, claims of being a RISC with a reasonable register set looked attractive. The AT90S8515 had 8kBytes of Flash and 512 bytes of RAM, 512 bytes of in-system programmable EEPROM in a 44 pin TQFP package. The initial trials used the instruction set simulator and STK200 evaluation board (which had RS-232 rather than RS-485 on the serial port simplifying host testing).

The STK200 evaluation board could take two processors; one with analog I/O and the other purely digital. The processor’s Flash was worn out by the thousands of downloads and tests. Code had to run out of Flash as the RAM was not large enough. I don’t recall a JTAG port for single stepping. We did not have the luxury of a C compiler, (not cheap from IAR). All development was done in assembler with testing on the simulator for non-I/O, otherwise by observing recent changes. Functions were debugged one at a time. Some time later the assembler was put into tables in LaTeX to make it look like normal program listings rather than use the ugly “verbatim” output. I also wanted to put in proper typeset comments. 38 page pdf 402kB. The LaTeX formating some ten years later clashed with the multicolumn and other packages, so I still need to fix that.