Mail Archives: geda-user/2011/12/03/10:09:06
On Dec 2, 2011, at 6:02 AM, Bob Paddock wrote:
> We've already decided our future projects will not have any Atmel
> parts of any type in them.
I've used the Atmel AT91 ARMs in a number of recent projects, and I've been reasonably satisfied.
Pluses:
1. Reasonably orthogonal design of peripherals means that, for example, you can actually just use a timer as a timer instead of having to figure out how to disentangle it from 3-phase PWM machinery.
2. Once understood, the peripherals usually turn out to be simple to use.
3. The chips work well with OpenOCD and GNU tools. I'm too busy to waste time with Windows.
4. Development boards are inexpensive, easy to acquire, and reasonably well thought out.
5. I find it quick and easy to move my stripped-down Forth (https://github.com/noqsi/LSE-ARM) to a new chip or board within this family.
Minuses:
1. Documentation fails to convey the essential simplicity of the peripherals.
2. Peripherals have implementation restrictions that are not generally tough to deal with, but they are unmentioned in documentation.
3. Example code is pure lasagna, thoroughly obfuscating the hardware/software relationships.
4. Documentation is a mashup. One would think that in the 21st century, processor manuals could include relevant details (like register addresses) where they are needed (as they did in the 1960's), but Atmel can't seem to manage real automated document generation.
John Doty Noqsi Aerospace, Ltd.
http://www.noqsi.com/
jpd AT noqsi DOT com
- Raw text -