[Elecraft] ultimate rejection

[Geoffrey Mackenzie-Kennedy]

Morning Merlin,

>From an old guy (hmmmpf) who has messed around with RF circuits for ages,
for what they are worth a few comments. Your idea to keep RF away from where
it should'nt be is a basic rule in good RF design, and will always be I
would think. Some digital designs get hung up because this rule is not
observed,
and there is a lot of RF about in some digital circuits.

By adding your fences, houses, barriers and getting rid of little "antennas"
is rocket science, and certainly not "dumb" !!  I have built many many HF
crystal filters for the homebrew gear, 10, 12 and 14 pole, and the isolation
of various parts from other parts is critical if a stopband attenuation
around 100db is to got, together with a good 6db / 80db shape factor and
very small passband ripple. In this playground, 6db / 80db is used more
often than 6db / 60db in design work. With VHF crystal filters using
crystals in an overtone mode (I have hombrew filters at 130Mhz and lower)
life really gets interesting - even the filter's box wants to reduce input /
output isolation. The "box" problem can be there at HF, so both the layout
and the enclosure of the filter must be thought about. But modifying an
existing filter's mechanical / electrical design can be more difficult than
starting from scratch.

I have not played with K2's filter, but apart from the input / output
isolation etc, I would suspect that the Varicaps and their bias circuitry
could also spoil "isolation" especially as pcb traces are used, but by how
much ? I have not done the sums, but I also have a hunch that the varicaps
could be one of the things that limits K2's strong signal performance. But
to be FAIR, it is my understanding that K2 started life as a QRP rig, highly
portable, and it is very nice to have the variable bandwidth filter in a
single conversion receiver. The KSB2's filter layout is compromised by lack
of space - it is much better to have the crystals in a straight line.

All said IMHO you have done the right sort of things. RF circuits that are
supposed to be separate, must not gossip.

73,

Geoff   GM4ESD


----- Original Message -----
From: <Merlarts at aol.com>
To: <elecraft at mailman.qth.net>
Sent: Tuesday, December 07, 2004 4:57 AM
Subject: [Elecraft] ultimate rejection
>
> Going first for the simple, dumb approach I began putting little RF
> "barriers" across the middle of the filter.  They consisted of a piece of
printed
> circuit material, connected to ground and soldered all along its  length.
It
> worked great!  Everything about the filter got  better:  opposite sideband
>
>
> Enough reminiscing;  I just spent about a hour with my KSB2, making  sure
> that the solder connections near the 7 pole filter were filed down, or
clipped
> as much as possible, so that they don't act as little "antennas",  thus
messing
> up the characteristics of the filter.  It worked!  The  "audio image" that
> one hears when tuning past zero beat got a lot weaker!
>
> Next, I went to local hobby shop and bought some tin.  I made little  tin
> "houses" which I placed around the matching toroids, ground them
carefully.  It
> worked too.  Next I added a little tin "fence" which  went on the bottom
the
> board, shielding the input of the filter from the  output.  It worked.
Three
> times is a charm!
>
> My point:  it ain't rocket science, but the old ideas of keeping RF  away
> from places where it shouldn't be work just as well as they ever did.
>
>