[sdiy] STM32 processor

[Tom Wiltshire]

On 15 Sep 2011, at 18:39, Eric Brombaugh wrote:

> On 09/15/2011 05:38 AM, Colin f wrote:
>> 
>> Even more useful, on some of the larger footprint parts, there is external
>> triggering of the DACs, with DMA update of the values.
>> For de-muxing many CV outputs, you can use a timer to generate the clock for
>> an external counter and de-mux circuit.
>> The same timer also triggers the DAC.
>> You set up an area in memory to hold the values for each de-mux output, and
>> configure the DMA controller to cycle through the DAC values on each
>> trigger.
>> The de-muxing then happens with no CPU involvement.
> 
> This is one of the most useful features of the recent MCUs we're seeing from many silicon vendors - the extreme interlinking of the peripherals, such that after setting up the timers, triggers, DMA, etc there is very little CPU involvement required. It requires a bit more imagination and a deep understanding of the SoC functions, but the upside is you can do a lot without much programming, and the timing determinism is greatly enhanced. If you're running algorithms that require more complex computation, the fact that the peripherals are running themselves leaves more bandwidth for the CPU to devote to the difficult math.

I definitely agree. I've had similar experiences with the DMA on the dsPIC. Being able to have ADC and DAC running in the background is a massive benefit. My recent designs wouldn't be possible without it.

Tom