General use of Quadrature thru zero VCO

[Haible Juergen]

	>BTW, thanks for posting your schematics for the frequency
shifter.  The
	>through-zero quadrature oscillator looks interesting.  Is there
anything
	>about the oscillator design that prevents it from being used as
a general
	>oscillator (i.e. can you tap into the oscillator to get a
square wave
	>output, and combine this with the triangle output to get a
sawtooth output
	>as per the normal methods)?  It would be great to have a
super-powerful
	>general VCO design that was useable for all situations (what
would be really
	>nice is a quadrature version of the Serge NTO with through-zero
linear FM);
	>does this VCO have good performance at higher frequencies?

The tri and square waves are there, of course.
Even the sine waves are, if you apply a large DC signal
at the ignal inputs of the balanced modulator chips.
(In fact, initially I wanted to use the 1496's as triangle-to-sine
converters only, and then do the modulation with RC4200's.
But then I found doing the modulation with the same chips
as the sine shaping more elegant, and SNR is no problem
with the compander anymore.)
The high frequency tracking of the oscillator is rather poor,
however. You would have to improove the expo converter,
and use faster comparators. Opamp comparators are rather
slow. You could either use a 3080-based or discrete schmitt trigger
instead. And the expo converter would use a transistor array
and a HF-track trimmer. I did not intend to go high-end in the
first place. Thought I could always upgrade this later, but
in fact I will never do this. Ok, in the kHz range the trinagle
amplitude will increase, and the sine converters will therefore
not be below 1% THD anymore, but who cares at frequency shifts in the
kHz range. For me it was important to have low THD and precise operation
at low frequencies, and still have the ability to go up
into the HF range with one turn of a knob.
If you want to make a general purpose VCO out of it, I'd suggest
you take some good old EN design for a tri/square VCO, make
direction switching with CMOS transmission gates, and build
the second comparator / integrator / feedback loop in a similar way.

JH.