CW filter/notch filter
 

I'm looking for a circuit design to a CW filter centred around 1khz that's
about 200hz wide and then rolls off very rapidly.

I'm planning to build a AFC system for SSB, which has a 1khz pilot tone that
will be used to lock the frequency correct.

I then need a circuit with quite a tight notch around 1khz to kill off the
pilot tone once it's been locked on, so you don't have an annoying whistle
on the reciever.


Can anyone possibly help with any circuit designs or where I can find
some????

I've had a look at a twin tee filter, but I'm not familiar with this design.
Would this be suitable, is it worth me reading more about it and learning,
then using this?


Thanks all,

Sam

I've found an excellent bandpass filter. It's 4 pole and is supposed to be a
CW filter. When I tune it up, it works really nicely to pass about 250hz 3dB
bandwidth, then has incredibly sharp edges so at 500hz BW, it's about 20dB
down, and then just rolls off more. This is almost perfect for what I want.

Now the only thing is finding a tight notch filter that does similar, but
the opposite, the tighter the better. Really I want to try to kill off my
1khz carrier, when it's no more than 50hz away, so it needs 100hz BW then be
rolled off at about 200hz.


After I've got these circuits sorted, I might be ready to "trial" earlier
than I originally thought.


Sam

"Samuel Hunt" wrote in message
...
I'm looking for a circuit design to a CW filter centred around 1khz that's
about 200hz wide and then rolls off very rapidly.

I'm planning to build a AFC system for SSB, which has a 1khz pilot tone
that will be used to lock the frequency correct.

I then need a circuit with quite a tight notch around 1khz to kill off the
pilot tone once it's been locked on, so you don't have an annoying whistle
on the reciever.


Can anyone possibly help with any circuit designs or where I can find
some????

I've had a look at a twin tee filter, but I'm not familiar with this
design. Would this be suitable, is it worth me reading more about it and
learning, then using this?


Thanks all,

Sam

A switched-capacitor (commutating) filter can put a potentially very
sharp notch at a frequency determined by its clock. I assume in this
AFC that you'll have access to the frequency you're trying to filter
out. If you use an 8-capacitor version of the filter, you just need to
clock it at 8* the freq you want to remove.

You CAN also design a continuous-time band-stop filter.

(Soap-box comment: seems like this would all be a lot easier if you
just digitized the signal at 8k samples/second, making detection and
removal of fs/8 pretty easy... or just transmit the SSB with a
suppressed carrier that you don't have to worry about removing.)

Cheers,
Tom

On 21 Jul 2005 16:37:00 -0700, K7ITM wrote:
A switched-capacitor (commutating) filter can put a potentially very
sharp notch at a frequency determined by its clock. I assume in this
AFC that you'll have access to the frequency you're trying to filter
out. If you use an 8-capacitor version of the filter, you just need to
clock it at 8* the freq you want to remove.

You CAN also design a continuous-time band-stop filter.

(Soap-box comment: seems like this would all be a lot easier if you
just digitized the signal at 8k samples/second, making detection and
removal of fs/8 pretty easy... or just transmit the SSB with a
suppressed carrier that you don't have to worry about removing.)

Soap-box comment: What are you nattering on about?
Oh, I see:
X-Trace: posting.google.com

In article .com,
"K7ITM" wrote:

... or just transmit the SSB with a
suppressed carrier that you don't have to worry about removing.)

Tom & Sam-

That is what I was going to suggest. It is a common practice to send a
partially reduced carrier that doesn't eat up a lot of power, but allows
you to synchronize to it using a synchronous detector.

When detected, the carrier translates to zero frequency relative to the
modulation, so it can be eliminated by a DC-blocking capacitor.

I think such a system is in use by some foreign broadcasters. (I'm
thinking one is at 21.455 MHz?) It doesn't sound quite right using an AM
detector. It can be received on an SSB receiver, but you have to
carefully tune it so music sounds right.

If you insist on the 1 KHz pilot tone approach, you could end up with both
the pilot tone and a small vestige of the suppressed carrier. Your filter
system might lock-up on either the pilot tone or the carrier!

73, Fred, K4DII

On 2005-07-21, Samuel Hunt wrote:
I'm planning to build a AFC system for SSB, which has a 1khz pilot tone that
will be used to lock the frequency correct.

If you are going to lock onto it so that that signal is exactly 1khz,
why not just cancel it with a 1khz signal 180 degrees out of phase?
This is essentially how an FIR filter would remove it digitally, but
you don't really need to generate the tone FROM the signal since you
know what it is already.

--
Ben Jackson

http://www.ben.com/

"Fred McKenzie" wrote in message
...
In article .com,
"K7ITM" wrote:

... or just transmit the SSB with a
suppressed carrier that you don't have to worry about removing.)

Tom & Sam-

That is what I was going to suggest. It is a common practice to send a
partially reduced carrier that doesn't eat up a lot of power, but allows
you to synchronize to it using a synchronous detector.

The Canadian time and frequency standard transmitters at 3330 and 7335 kHz
are USB with partially reduced carrier. I don't see the point of sending a
tone in the middle of the speech spectrum to provide a sync reference when
unsuppressing the carrier to some degree is much more economical.

Tom