MOTM

http://www.ear-group.net/model_23.html

Plan B has a 'traditional' ASR coming out (analog, 3 stages, no delay). The 
link has a few sound samples and graphs/charts.

I think some folk are confused about the "sample rate" versus the "shift 
rate".

In this module, it's the *same thing*.

If you feed in a pulse to the CLK IN (or, at the rate of the INT CLK):

a) the voltage at the INPUT jack is sampled by the 14-bit A/D converter
b) the ARM uP get an interrupt from the A/D ("Hey, I've finished sampling 
now!"), and then updates the 4 outputs based on the presence or absence of 
delay memory.

So, if the clock is 1Hz, then once a second, the input is sampled AND all 
the outputs get updated (this happens quite quickly, on the order of a few 
1000ths of a second).

What makes the '102 unique is that since there is no analog capacitor to 
charge up/hold in-between clocks, we can:

a) clock the ASR *really fast*, limited really to the ARM execution speed 
and the time it takes the serial SPI buss to shuffle the data from the ADC 
out to the 4 DACs. I am guessing this will wind up around 2Khz

b) there is no "droop" error (drop in output voltage over time)

c) the ARM has enough internal RAM so that we can "play around" with the 
samples, and make a sort of 'reverb for CVs'

Paul S.

However you describe it, Paul, the '102 will replace 3 or 4 modules in
my system!  Bring it on...

Chub

On 2/8/08, Paul Schreiber put forth:
>I think some folk are confused about the "sample rate" versus the "shift
>rate".
>
>In this module, it's the *same thing*.
>
>If you feed in a pulse to the CLK IN (or, at the rate of the INT CLK):
>
>a) the voltage at the INPUT jack is sampled by the 14-bit A/D converter
>b) the ARM uP get an interrupt from the A/D ("Hey, I've finished sampling
>now!"), and then updates the 4 outputs based on the presence or absence of
>delay memory.
>
>So, if the clock is 1Hz, then once a second, the input is sampled AND all
>the outputs get updated (this happens quite quickly, on the order of a few
>1000ths of a second).

Thanks for the clarification.  What confused me was the "minimum"
rate of 1KHz, which seemed like either a very slow uP clock or a very
fast sampling clock.