ODP: thru zero VCO questions

[Rene Schmitz]

At 10:32 10.09.99 +0200, Martin Czech wrote:
>I've got two objections:
>
>-I can not see how the oscillation will be stable, the toggle ff could
>be in the wrong state right from the start, or get into a wrong state
>later on due to noise or metastability. Then the direction switch will
>not work any longer, since the cap is integrated near to one power rail
>and get stuck there.

The situation is analogous to the familiar integrator-schmitt trigger
oscillator. Its not important in which direction the waveform starts. All
that is required is that at powerup the cap's voltage is within the +/-2.5V
window. (i.e. reset it via a switch at powerup)

The locking you mention can occur if the waveform hits the threshold again
immediately after a toggle has occured. If this happens within the setup
time (+ propagation delay of comparator) of the FF it can lead to problems.
And if the FM input changes polartity within this setup time it is possible
that the waveform rises beyond 2.5V until the next polarity change occurs.
It could of course happen for a pathological FM signal that the FM input
doesn't bring the signal back into the window, that would be bad. I'll work
on that problem. 

>-I can not see what advantage it should have using the nonlinear
>OTA input for more then a polarity digital signal. The linear fm signal
>will be distorted or drown in noise, and using the diff. input in "linear"
>mode will cause offset problems, too.

Ok, linearity won't be very good, but does that matter anyway ? The
sinewaves (or triangles) we'd use *are* distorted anyway, so why bother? I
think the signal can be attenuated into the linear section of the OTA. The
situation is soundwise equal to a OTA-VCA patched into a FM input.  
On the other hand this scheme would work also with a linearized OTA.
(LM13600 + diode bias) And it would also work with the SSM2164. So you
*can* get to the safe side distortion-wise, if you want.

Offset will mean that I get zero Hz (or frequency reversal) for some other
voltage than 0V, so perhaps for -42mV. This is no problem IMO, one can
easily tune this out with the linear tune pot.

Bye
  René
 



error:asciiartisnotfully  \ uzs159 at uni-bonn.de
supportedwithyourcurrent   \ http://www.uni-bonn.de/~uzs159
displaydriversettings...    \ http://members.xoom.com/Rene_Schmitz