octave

[Martin Czech]

> Does anyone know of a site that either has schematics or block diagrams or
> something that would clue me in on how to build a device that could raise
> an analog signal a couple octaves higher?
> -mark wagy
> 

No, but it is somewhere in my brain (;->). I want to answer this
question in a bit generall way.

I noticed that some synths (Buchla, Technosaurus etc.) offer modules
with 1x, 2x, 3x, 4x .... and  1/2, 1/3, 1/4 times of the input
frequency, i think a tri or saw wave goes in and (technosaurus) all
frequencys drop out as sine waves. And I thought: How the heck do they
do it?

Now, it is obvious for me, that they convert tri/saw waves to other tri
waves with a different frequency and then use a sine shaper.

Dealing with tri waves is may be simpler, so I take this for a start:
How do we convert tri waves ?  You get the octave by simply full wave
rectifying (FWR) a tri wave (that has to be symmetrically arround 0).
And you could do this again and again, so you would end up with 8 times
the frequency of the input. You only need a FWR that is fast enough
because it will get problems at higher frequencys otherwise, the waveform
will be distorted. Thus some part of the "fundamental" will always be
there, due to the waveform distortion.

We could also do this with a saw wave, we need comparators for the
0,25,50 and 75% time of the saw and then a polarity switch, that
"folds" the saw at the respective points of time. Now the signal will
have some leaps, so we have to add some of the comparator pulse output
voltage and voila: a tri wave with double frequency.

What are we doing here ?  Ok, we have a (piecewise linear) waveform,
and the transformation is to insert a fold at a proper position, ie.  a
polarity switch and to avoid jumps by adding a pulse wave at the right
time.  By extending this scheme (especcially the one with saw waves),
we can derive (theoretically) any frequency ratio, for freqeuncy dividing
we need some kind of storage flip-flop or counter.

It is very easy to understand, but hard to explain with words.  Take a
pice of paper, draw your incomming saw wave and below your wanted
signal. Now insert "folds" into your saw wave, ie. draw the effect of
a polarity switcher. Now the signal allready goes into the same
direction as the wanted signal, but there are still some jumps, so
you'll have to find the right pulse waveform to add in order to get rid
of the jumps.  Doing this, the cicuit can be easily derived.

m.c.

------------------------------------------------------------------------
m.c. has made it finally:  3 CDs out now; 72 min. minimum; "1"
(1994-1995),"2" (95-96),"three" (96-97); experimental stuff; mostly
Eimert/Stockhausen style; but also modern popular style