sqr

[CGS2510.SEP.STUDENT.UCA at CUB.UCA.EDU]

> Ok, lets suppose we have a square router.
> O=sqr(sin(a)).
> Now , first of all, we want to have real output, not complex.
> So what are we going to do with the negative halve-wave of the sin?
> And the positive halv-wave will be distorted.
> I don't believe there is a theorem like above for muliplication or squaring,
> i.e. we can only find a numeric solution. However, you'll end up with
> a harmonic wave, lots of partials and never with half frequency or
> something like that.

This is true.  If you were implementing it though in a scenario like 
op-amps and AD633's and such the negative portion of the input to be 
rooted would quite likely end up swinging the output as far to the 
power rails as it could reach.  Since the squarer will cause all 
inputs to IT to be the same sign, a similar signed input in a closed 
loop op amp circuit would not get the feedback so that it could 
stabalize, but would rather zoom off.  Maybe if you took this into 
account at the front end by rectifying it and then flip flop an 
inverted into the final section, you could get the distortion 
(harmonic) of the rooter and still end up with a waveform similar to 
a sinewave (ie positive AND negative going).

Also, concerning modular stuff.  Has anyone gotten as interested in 
DSP as of late as I have and gotten one of the Analog Devices OR 
Motorolla OR Texas Instruments OR some other DSP evaluation kit?  I 
have Analog's ADDS 2106x coming in the mail.  Think of it... two 16 
bit ins, two 16 bit outs, quite a speedy processor, lots of text to 
read, and a C compiler!!!  And the price is so low nowadays.  You may 
not be able to write code for a full synthetic modular, but all you 
gotta do to have every effects processor known to man is download 
code from computer to evaluation board and you could change its 
entire purpose in life in just seconds!!!  I'm very excited about it, 
anyone else tries it?