[sdiy] Sallen-Key Filter & 4069

[René Schmitz]

At 08:17 07.10.01, Scott Gravenhorst wrote:
>But couldn't the forced tie point simply be where the two input
>resistors connect together?  Granted, you would 'waste' four
>of the 6 resistors.

Yes, thats it!
Umm, well, luckily I didn't say anti-series wasn't possible. ;-)

   partial schematic:

 >---D S---+---S D--+--opamp---+-->
      g    |    g   |          |
          cap      cap         |
           |        |          |
           |       gnd         |
           +-------------------+     
only the N-channels used (leave Vdd open, common is Vss)

>I see two resistors, connected at one point (common tie point
>for the 4069) and then two 'floating ends', one for the 
>filter input and one connected to the opamp input.

As Harry already pointed out, the VC-Rs like a low impedance at their
outputs (sources), because thats the reference for the control voltage. So
if that would be a virtual GND or at least fixed voltage you can simply
supply your control voltage from GND to the gates.

I made some additions to the circuit diagram above.
One sees that it would be difficult to hold both sources at a fixed
potential, as the signal is as well present. The control voltage for the
resistor is its Vgs, the signal at that point adds to the control voltage
and will give some distortion. 
OTOH if the swing is low then the signal can't change the resistance of the
VC-Rs much. (Btw. a situation as in diode and transistor ladders, but only
with a mild quadratic law!) 
One can't make large swings anyway because there are internal diodes that
may become conductive. 
Due to that anti series connection you have at least the same control
voltage for both resistors, as the sources are tied together, we only need
to tie the gates together. 

Now I think this is possible. I'm reminded of the Korg 35 circuit, where a
similar series connection (albeit with BJTs) was used for a sallen & key
filter. 

Cheers,
 René



-- 
uzs159 at uni-bonn.de
http://www.uni-bonn.de/~uzs159