[sdiy] Using analog for physical modeling
Eric Brombaugh
ebrombaugh1 at cox.net
Wed Sep 21 16:38:13 CEST 2011
On Sep 21, 2011, at 2:14 AM, Tom Wiltshire wrote:
>
> On 20 Sep 2011, at 23:23, Matthew Smith wrote:
>
>> Quoth Paul Perry at 21/09/11 07:45...
>>> ...What would be the 'best' (cost, size, low parts) way to
>>> go about mainng an all-digital BBD look-alike?
>>
>> A low pin-count FPGA with onboard ADC/DAC. I can think of just SO many (audio) applications!
>>
>> Dream on. (If size weren't an issue, something like Xilinx Spartan 3AN with support components might fit the bill - they've got onboard FLASH, which cuts the parts count/size.)
>
>
> Isn't that what the Princeton PT2399 is trying to be? - except that it won't go short enough to do flanging, and only just for chorus. PIty, that. So close, but no banana. They should have added a VCO modulation input too, to save the craziness people do to modulate it. Still, it's about the best modern option, and very much the successor to the original BBDs.
>
> The FPGA and a codec chip wouldn't take up much space. The various power supplies needed might be more of a pain to sort out. If it's going to be a BBD replacement, it ideally needs to run off typical analog voltages (+9V, +12V, +15V etc).
I've done a few boards for this sort of thing using low-end Xilinx Spartan3A parts. Supplies aren't really a problem since you can use LDOs for the FPGAs as the current draw is pretty low on simple audio designs with low clock rates (<50MHz). I used an outboard processor (dsPIC, ARM) to do CV digitizing and to load the FPGA bitstream, but it's perfectly possible to use flash-based FPGAs and do the processor internally. You'd still need external ADC for the CV in addition to the audio codec, so that's a tradeoff. Even low-end FPGAs have a fair amount of SRAM on-chip for short delays, but for longer delays you'd probably want an external SRAM.
The nice thing about going this route is that you can use good-quality audio codecs, larger delays and also cook up interesting effects & variations on the delay algorithm which wouldn't be possible with the canned solutions like the PT2399 or a BBD.
Eric
More information about the Synth-diy
mailing list