AVR-Chat

Zack Widup wrote:

>On Wed, 11 Jan 2006, John Samperi wrote:
>  
>
>>At 01:25 PM 11/01/2006, you wrote:
>>    
>>
>>>   Quite simple:  I have potential applicaitonsd whre a 20 Mhz processor
>>>simply isn't fast enough.
>>>      
>>>
>>Is it a hardware or software limitation? Can you rewrite part
>>(or all??) of the application in assembler perhaps? I understand
>>the case of a PWM scenario where you could be limited by the
>>PWM frequency. In other words I would NOT consider over clocking
>>a processor in a production unit, but that's just me, I like
>>things to run WELL within their specs.
>>I have been playing around with the direct digital synthesis project on 
>>Jesper's AVR page:
>>
>>http://www.myplace.nu/avr/minidds/index.htm
>>
>>If you want an output of more than a few hundred kHz, you either need to 
>>overclock the AVR, use frequency multiplication, or find something faster 
>>(if it exists). 
>>    
>>
  Oh, it exists.   One can easily find 25 and 40 MIPS capable 8051 
derivatives  out there, and there are literally DOZENS of smeiconductor 
makers with proprietary 8 bit (an.d.or 16 and/or 32 bit) 
Ucontrollers/uPorcessors out there.  But I am lazy and picky.  I want to 
do it on the AVR  architecture, even though, for reaosns alluded to in 
my reply to john, I know it cannot POSSIBLY be (unless I choose to 
implment an AVR clone running at many hundreds of Mhz in an FPGA (and 
don't laugh...www.opencoure.org (I think that's right) has the VHDL code 
in open source form to create a generic AVR processor in a Xilinx FPGA 
that will easily run at 50 Mhz.   Not counting the cot of the FPGA 
development system/programming hardware, it would cost about $25 or so 
to make of those.  I may actually do this someday, just for the hell of it))

tom