[sdiy] uP stablized lin-exp converter

[Sowa Roman]

any cheapo DSP could do that IMHO.
But then we need better accuracy than 16 bit, so code
should work with double-resolution arithmetic (32bit)
and that's much more code, so 44kHz sample update may
be too fast. 
But wait, 33MIPS/44.1kHz gives almost 750 instructions
per sample. That should be enough even for very accurate
exp(x) aproximation. And 33MIPS DSP goes for mere 10-20$

anybody want to use floating-point DSP in expo converter?
:)

How about a DCO, which is controlled from ADC?
we have analog input and digital stability. ADCs can have
very little temerature drift. Industrial types can deliver
24bits with high linearity. Maybe not too wide bandwith,
but who wants to modulate VCO faster than 1kHz? ;)

Roman, who earlier said that VCO are going to be too precise :-P
flame attack ready


__-----Original Message-----
__From: jorgen.bergfors at idg.se [mailto:jorgen.bergfors at idg.se]
__Sent: Thursday, May 30, 2002 10:17 AM
__To: Tim Ressel; synth-diy at dropmix.xs4all.nl
__Subject: Re: [sdiy] uP stablized lin-exp converter
__
__
__
__>Is 100 updates per second
__>enough? Or is 1000 more like it?
__
__Well, let's see.. To handle audio rate FM, you would need at 
__least 44.1 kHz
__sample rate.
__Don't think this is a good idea, after all.
__
__/Jorgen
__
__
__
__---------------------------------------------------------
__This message has been scanned for viruses by MIMESweeper.
__
__Wiadomosc zostala sprawdzona pod katem wirusow przez
__oprogramowanie MIMESweeper.
__


---------------------------------------------------------
This message has been scanned for viruses by MIMESweeper.

Wiadomosc zostala sprawdzona pod katem wirusow przez
oprogramowanie MIMESweeper.