With all respect to Lee, I think he has mistaken the Cauer approximation for
something else.
The Cauer filter, also called an elliptic filter, has equiripple passband
and stopband, obtained by use of transmission zeros on the frequency axis
(notches). Far from having linear phase in the passband, the equiripple
passband implies a sharp rate of change of phase near the passband edges,
similar to the Chebyshev, which also has an equiripple passband, but with a
monotonic stopband.
I can't be certain, but I think Lee may be referring to a Gaussian filter
for that FM application. Generally, filters with smooth transitions from
passband to stopband exhibit more linear phase, and therefore more constant
delay, at the expense of stopband attenuation. There are also compromise
designs, such as the Transitional Butterworth Thompson (TBT) that trade off
the phase response against the gain response.
The usual way to have everything in an analog filter--flat passband, high
rate of attenuation in the transition band, high attenuation in the
stopband, and linear phase--would be to employ something like a Cauer filter
for the gain response, together with delay equalizers.
Garry, NI6T
> -----Original Message-----
> From: topband-bounces@contesting.com
> [mailto:topband-bounces@contesting.com]On Behalf Of R. Kline
> Sent: Friday, July 18, 2003 21:41
> To: Lee K7tjr
> Cc: topband@contesting.com
> Subject: Topband: Re: filter
>
>
> Thank you Lee. I'm sure that many people on the reflector will appreciate
> your comments.
> 73,
> Riki, 4X4NJ
> ----- Original Message -----
> From: "Lee K7tjr" <k7tjr@comcast.net>
> To: <rikik@zahav.net.il>
> Sent: Friday, July 18, 2003 10:29 PM
> Subject: filter
>
>
> > Riki... One of the filters I have used in the past was called a Cauer
> > filter.. It is designed with highly linear rate of change of
> phase through
> > the passband .. They really do not ring and they also let you decode FM
> > transmissions without distortion (frequency is merely the rate of change
> of
> > phase) in FM receivers.. They have just slightly less filter slope and
> > ultimate attenuation than most other designs so they are not
> too commonly
> > used as they require one more stage to get the same ultimate
> attenuation..
> > Regards Lee K7TJR Oregon
> >
> >
> >
>
> _______________________________________________
> Topband mailing list
> Topband@contesting.com
> http://lists.contesting.com/mailman/listinfo/topband
>
|