EMRFD Message Archive 1998
Message Date From Subject 1998 2008-09-03 13:15:52 timshoppa Gaussian or Equiripple filters with real crystals beyond Zverev?
for feeding into Ladpac.
Ladpac seems to be able to make its k numbers for Chebyshev and
Butterworth designs. But not for Gaussian-to-6dB, equiripple, etc.
filters. Am I understanding things correctly?
Are there programs like Ladpac that make the k numbers for
equiripple and Gaussian filters without having to fetch a copy of
Zverev, or for doing things beyond the Zverev tables? The DSP filter
gurus I know use Mathcad to generate poles for their fancy-pants
filters but I don't really grok Mathcad nor even how to turn the
poles into k numbers.
I have Carver's article (Communications Quarterly 1993) and am
slowly, slowly understanding its points about real crystals and how
the Zverev tables use q0 numbers to give k's for use with real
crystals.
AADE's filter designer has choices for Equiripple, Gaussian to 6dB
etc. but whenever I choose them I get run-time error 76, 'Path not
found'. Works fine for Chebyshev and Butterworth crystal filters
though. Am I doing something wrong, or do others have the same
problem?
Many decades ago I sat through engineering classes on filter theory
and while I could do the homework and exams, needless to say I never
built a crystal filter using what I supposedly learned! So far my
homebrew filters have been Cohn min-losses copied from other
designs, but now I want to do better especially since I've measured
my own crystals.1999 2008-09-04 09:27:27 timshoppa Re: Gaussian or Equiripple filters with real crystals beyond Zverev?
copy of Zverev a few years back. In part because I did not xerox the
whole book, just some tables!
Lots of times there are different "choices" of q1 and qn. These seem
to define (along with some math, notably EMFRFD Eq 3.12) the impedance
looking into the filter and out of the filter. There seem to be more
choices as you go to more and more filters - in Zverev for 8th order
filters on page 376 there are often 4 different choices for q1 and qn.
What's a good choice or a bad choice?
XLAD seems to like filters with same end resistance on both ends. If
q1 and qn as chosen by Zverev are very different, does this mean
the "natural" resistance (without the end shunt capacitors C-1 and C-
N) is very different on the two ends?
Tim N3QE2002 2008-09-04 19:03:00 ve7ca2 Re: Gaussian or Equiripple filters with real crystals beyond Zverev?
On my web site: http://www.shelbrook.com/~ve7ca/ there is a
"Design" page.
If you send me an email (or anyone else who is a HAM) I will provide
you with a user name and password. You will then be able to access
the various programs on my Design Page, free and no advertisements
either. Included