RE: Bandwidth of crystal ladder filters CW and SSB


From: Jim Kortge, K8IQY (
Date: Sat Sep 28 2002 - 19:50:24 CDT

At 02:45 PM 9/28/02 -0300, you wrote:

>Hi all !
>As more and more precision and at the same
>time rather low cost per unit quartz crystals
>become available, the design of homebrew
>ladder type or "" ladder type assymetric sideband
>filters"as one of my colleagues likes to call them,becomes a very attractive
>and interesting
>plus cost saving challenge.
>I have several questions regarding those filters
>1. crystals must be properly matched to within
>a very narrow frequency range .. OK, but does
>the accuracy of the match influence the wider
>bandwidth SSB filter less than the CW much narrower filters ??

In general, the wider the filter, the less critical the matching
of the crystals becomes. That said, the better their series resonant
frequency is know, the better one can do building a predictable,
repeatable filter. Matching to within 10% is generally the accepted
standard, but one can do much better than that if a large number of
crystals are available for matching. With 100 units, 1% groups are
quite easy to do.

>2. Is it true or not, that no matter how you work
>around with these ladder filters, your response
>will ALWAYS be assymetric , in favor of the
>lower sideband ?

That is true if you don't do anything to compensate for the
holder capacitance, termed Co. However, if you cancel
Co with an inductor in parallel with the crystal, very symetrical
filters can be made. I do it all the time. Then they can be used
for lower sideband, or upper sideband, without any change to the
basic filter design. I use molded inductors, and often don't
match them for inductance values, and still wind up with decent
filters. Matching the inductors enhances the filter shape some,
but not a whole lot.

>3. Is it true that the SSB filters ( typically designed
>for between 1.8 kHz up to 3.2 kHz bandwith)
>produce much higher insertion losses than
>the narrower ( 100 hZ to 500 or 750 hZ wide )
>CW filters ???

Actually, the opposite is true. The narrower filters have the
larger insertion losses, and the wider filters have lower values.
For example, a 5 pole, 200 Hz BW CW filter might have an IL of
6-7 dB, while a corresponsing 5 pole, 2KHz filter would be on the
order of 1-2 dB. Insertion loss is also dependent on the Q of the
crystals used. Those with higher Qs make filters with lower IL.

>4. What would be the simplest and lowest cost
>design for a SSB filter... and by this I mean using
>the least number of crystals and other parts ?

I'd go with either a 4 or 5 pole Chebyshev or Cohn design, with
my preference being the Chebyshev for slightly better overall shape.
The Cohn will use one less capacitor in the 5 pole design; they
will be equal in the 4 pole design. I just designed a 5 pole,
2.1KHz Chebyshev filter and it has a shape factor of 2.7. While
that isn't spectacular, it would make a more than decent SSB rig.
With Co compenstated, it could be used for LSB or USB. It was
designed using 4.9152 MHz crystals from Fox. BTW, the last batch
of Fox crystals that I bought had Q's that averaged 165,000. Those
make good filters.

>( this question is intended to gain knowledge towards the design of the
>HURRICANE TWO, a natural replacement for our HURRICANE ONE
>double side band rig , used here in Cuba during
>the hurricane season .... )
>5. Which are the more popular and lower cost
>frequencies available from suppliers around the world ???

Almost all of the computer crystals are available for essentially
the same price until you get down below 2 MHz and above 50 MHz.
I feel that anything from 4 MHz up through 9 MHz are all good
candidates for IF filter service, with my favorite being 4.9152,
as well as others in the 4-9 MHz range. We're doing the IF filter
in the 2N2/30 at 7.23 MHz, and my SW30++ homebrew rig uses 7.68 MHz
in the IF. The 2N2/15 IF is at 4.433 MHz, since that is an easy
frequency to find in old European TV sets.

>I understand that
>3579 and 4433 are
>two of the most popular ones,

They are, but I don't like putting an IF inside a ham band. I also
don't like even frequencies, like 4.0 MHz or 6.0 MHz, etc. Too many
chances that a high powered radio service is operating there.


>HOPE this is accepted as a QRP related topic
>and that it may generate a LOOOONNNG and interesting thread...

Refreshingly so Arnie! I hope others can respond to your questions.
As you know, I'm here to help you with the Hurricane project rigs
as much as I can.

72 amigo,

Jim, K8IQY

Search QRP-L Archives

[ QRP-L Archive | ]
[ 1993 | 1994 | 1995 | 1996 | 1997 | 1998 | 1999 | 2000 | 2001 | 2002 ]


This archive was generated by hypermail 2b29 on Sun Oct 13 2002 - 10:40:05 CDT