Original poster: "Malcolm Watts" <m.j.watts@xxxxxxxxxxxx>
Hi Ed,
On 27 Sep 2005, at 13:37, Tesla list wrote:
> Original poster: Jim Lux <jimlux@xxxxxxxxxxxxx>
>
> At 11:17 PM 9/26/2005, you wrote:
> >Original poster: Ed Phillips <evp@xxxxxxxxxxx>
> >
> >"Excellent point, and a good example of a frequency selective system
> >that's band limited. Hmm, though, can one create an arbitrarily sharp
> >cutoff with only RC? I think not, off the top of my head."
> >
> > We used to have a guy at Hughes named Lou Weinberg who had
> > the
> >reputation of being a real expert on network synthesis. I heard him
> >say that, using only PASSIVE RC circuits, it was possible to
> >synthesize any transfer function shape if the loss were allowed to be
> >arbitrarily large. Don't know if it's true or not. Certainly with a
> >little feedback you can get arbitrarily sharp response.
> >
> >Ed
>
> I've thought a bit (and am too lazy to go find my copy of Mason and
> Zimmerman, which I'm sure would have the answer)..
>
> I think you're right. I was having trouble figuring out how to get
> sharp nulls (or more properly, sharp cutoffs). A series of RC
> networks can get you more than 180 degrees of phase shift, and if you
> sum the 180 and the 0 path appropriately, you get a null. One can
> always string up multiple nulls (at different frequencies) to get what
> you need.
>
> I don't know that you can get the classic resonance curve, though
> (except with a huge number of sections).
Twin T and Weinbridge networks spring to mind in connection with both
of those and I'm sure that's not the end of the story. I predict that
someone attempting to disprove the hypothesis about generating
transfer functions using only RC components (losses taken into
account) would have a very hard time.
Malcolm