AVR-Chat

--- In AVR-Chat@yahoogroups.com, "rtstofer" <rstofer@p...> wrote:
> My questions at this point are pretty simple:  Is the ATmega32 a 
> reasonable platform and will the WinAVR/AVR Studio 4 and STK500 work 
> together to get this thing done?  Any particular gotchas that I will 
> discover only after hours of debugging?

The STK500 is nice to have but the only thing it brings to the party
once you have your Mega32 situated comfortably in its new environs is
the ISP capability.  You can build an AVR 910-compatible ISP for
minimal cost, especially if you have a well-stocked parts bin.  On the
other hand, it takes a lot less time to program the full 32K using the
STK500 than it does with an AVR 910-compatible programmer.

If you can, when you design your hardware try to keep Port C, bits 2-5
free.  Those are the JTag I/O pins.  That will allow you to plug in a
JTag ICE without interfering with any of you application circuitry. 
I'm not sure what the JTag ICE costs - I got one free as a bonus for
attending an Atmel seminar about a year ago - but it's very nice to
have when you need it.

The only downside of using the GCC-WinAVR is that the optimizer makes
it difficult to debug (as is often the case).  This is most noticeable
when you're trying to do source-level debugging with JTagICE.  Often,
you no choice but to look at assembly code and try to figure out where
to place a breakpoint.  This is complicated by the fact that AVRStudio
displays WORD addresses but the assembly language output from GCC (the
.lss file) displays BYTE addresses.


Speaking of Atmel seminars, there may be one coming to a city near you:

http://www.atmel.com/seminar/mcu/default.asp


Don