[sdiy] NICs, GICs, NLOs and all that

[Ian Fritz]

At 10:40 PM 1/3/2003, Magnus Danielson wrote:

>I where just about to say that I've designed both single and double ended
>gyrators with OTAs for this purpose! It took a little thinking, but I was able
>to derive the full expressions for dimensioning it. If you wish I could dig it
>up and present it here...

It would definitely be interesting to see those, if you get a chance.

>One issue with OTAs though... the offset of the OTA output may be a real 
>issue.
>I would not expect even near rail-to-rail responce without serious distorsion.
>I don't know if it has been discussed. In many cases OTAs is being used to
>drive virtual ground setups, and then you have a fixed voltage offset for the
>OTA output to operate in. Then, in other setups (especially in phasers) OTAs
>really operate on a varying output voltage offset. Most cases where this
>happends you can modify the surrounding curcuits, but this would be 
>against the
>general idea about "floating" double-ended gyrators.

Yes, you can definitely get into trouble just throwing OTAs into designs. A 
good example is the S. Franco 2-pole Butterworth lowpass (reproduced in EN) 
which may have been misdesigned because of not understanding this issue 
properly.  (The correct version is on my website.)  At one point I 
convinced myself that some of the circuits in the Curtis filter app notes 
are also wrong, for the same reason.

Of course it is always possible to consider using the 2-OTA design for a 
floating resistor.

Also the trick of replacing inductors with resistors and capacitors with 
FDNRs seems to be quite useful, as one ent of the FDNR is often grounded.

   Ian