22KHz filter

[johnsmith123]

Hello everyone,
I would like to build a very simple filter to block 22KHz.

Does anyone know where I can download a schematic?

Help much appreciated.

 

[audioguru]

What is the filter for?
Why is 22kHz present?
About the only filter that will actually block a frequency is a notch filter. A simple one set to notch 22kHz will also seriously reduce frequencies above about 5kHz and below about 80kHz. More parts are needed for a sharper notch.

A simple lowpass filter is a series resistor then a capacitor to ground. If it is set to reduce 22kHz to 1/10th its input voltage, then frequencies above about 2.2kHz will also be seriously reduced. More RC stages are needed for a sharper cutoff.

A simple highpass filter also performs poorly.

 

[Nigel Goodwin]

Your question is far too vague, you need to give exact details of what you are wanting to do, and details of what is feeding the filter, and what is fed from the filter.

 

[john1]

Hi JohnSmith,

I don't often reply to single posts,
as very often there are no more posts.
We'll see.

Here is a circuit of a 'notch filter'
They can be very effective, but i think they are intended for
sine waves.

**broken link removed**

If the 22 Kilocycle you want to filter out is not a sine,
i suspect it would not be so effective.
In which case you need to look at other ways of minimising it.

Regards, John

 

[audioguru]

Hi John,
22KHz is the same frequency no matter where it comes from, sine waves, square waves or music. A notch filter will eliminate it but not its harmonics.

Your simple "twin-T" notch filter will seriously reduce frequencies within a few octaves from its tuned frequency. Adding positive feedback to its "ground" point from an opamp buffer at its output will sharpen the notch so that it is extremely narrow. An opamp that is good at very high frequencies must be used to obtain a deep notch.
https://www.electro-tech-online.com/custompdfs/2005/09/LB-5.pdf

 

[john1]

Hi audioguru,

Thanks for that info.

Although the only problems i recall with filtering,
was with mains hum, and i didn't really filter it,
i used the filament line to inject mains hum instead.
It was on a home made oscilloscope, this was injected on to
the 'shift' plates to get rid of magnetic interference from
the mains transformer. Worked out ok.

we haven't heard back yet ...

Regards, John

 

[audioguru]

Hi John,
On my 'scope I just rotated its mains transformer to get a null. It looks odd at about 45 degrees but inside its case it can't be seen. And it works! :lol:

Where is the guy who needs a "brick wall" filter? :?:

 

[johnsmith123]

Hello guys,

First of all I would like to say thanks for all of your replies to this thread.

The filter will be used to solve a problem that is caused by a switching device that will switch between multiple LNBF's on a satellite dish. The switch works on a 22KHz tone (the receiver sends down the line to the switch) that will switch the dish's LNBF to another satellite when this tone is present.

The problem is that the 22KHz tone will bleed through to other LNBF's and cause them to switch when they are not supposed to. A simple blocking of this frequency will stop that, if installed between the switch and the LNBF.

I am not an electronics pro by any means, maybe it can be exlained better here:

This actual quote is taken from another forum discussing this same problem:
*********************************************************
"Should be easy to create an effective 22 kHz rejection filter with a properly sized inductor and a cap in parallel.
That way the high freq. IF passes through the cap with ease and the DC passes by way of the inductor, while blocking 22 kHz.

I'm guessing a cap in the range of .001uF along side of a healthy sized mH RF choke would do the trick."
*********************************************************
If someone can give me ideas, or a formula to calculate the inductor/capacitor values needed, I would certainly give it a try.

Thanks All,
JS

 

[john1]

Hi JohnSmith,

Yes an L-C will give better results than an un-assisted R-C filter.
In addition to the filter you could reduce the strength of the 22kc
signal, stray pick-up is usually very much smaller than the signal
proper. Reducing the signal may be all you need to do.

The optimum values for such a filter are those which would resonate.
So the calculation is straightforward, however you cannot rely on
calculation alone to tune a resonant circuit, allow for adjustment.
Probably by adjusting the inductor core. Variable capacitors would
not be of sufficient value to make any sensible adjustment.
So much for a passive or un-assisted arrangement.

However, an assisted R-C filter might be more to your liking, which
would need a supply source, so it may or may not suit your needs.
Something such as described above by 'audioguru' may be what you want.

First i would try reducing the signal.

Regards, John

 

[Nigel Goodwin]

However, an assisted R-C filter might be more to your liking, which
would need a supply source, so it may or may not suit your needs.
Something such as described above by 'audioguru' may be what you want.

You don't appear to have understood his question (now that he's told us what it is!), an opamp filter would be completely useless for doing what he wants.

Bear in mind it's to go in a section of coaxial cable, with no other power connections to it!. It has to perform the following requirements:

1) Pass 13V or 17V DC (switched values to select horizontal or vertical polarity), current about 250mA, from the receiver to the LNB.

2) Pass the RF signal back from the LNB to the receiver, this is about 700MHz to 2150Mhz (high UHF, low microwave), and is the first IF of the satellite receiver system.

3) To provide a further switch in the LNB (high/low band switching), a 22KHz tone can be superimposed on the 13/17V supply, it's THIS he's wanting to suppress!.

The specification of the 22KHz is:

Amplitude 0.6V +/-0.2V, frequency 22KHz +/-4KHz.

I would suggest two possibilities?.

1) A series notch filter across the cable (inner to screen) to short the 2KHz to ground.

2) A parallel notch filter in series with the inner core of the cable, to reject the 22KHz travelling up the cable.

The calculations are the same in either case, just the wiring is different.

frequency = 1/(2 x Pi x Squareroot(L x C))

With Frequency in Hz, L in Henries, and C in Farads.

However, I suspect problems may occur with layout, bearing in mind it has to work at 2GHz?.

 

[john1]

Hi Nigel,

Yes, i didn't realise exactly what this is for.
And i am not sure that filtering the 22kc to ground would help,
as this would affect the high frequency signal also.

One thing occurs to me, that is the possible use of 'nodes'
to minimise affecting the high frequency signal.

There are points along feeder lines where they can be shorted
together with almost no affect on the signal carried, maybe
a 22kc filter to ground could be inserted at such a point ... ?
This would attenuate the 22kc but hopefully leave the higher
frequency less affected.

Or maybe i still haven't grasped exactly whats going on ....
I get the impression that the line is being used to cover some
additional functions, one of which is to have a 22kc signal
superimposed upon it, to act as an instruction to something at the
far end, maybe a motor ?

After reading through again, i think the nodes idea is a non starter,
that would only work at set frequencies ....

If there are other units which might respond unwanted-ly, then maybe
they could take turns so that only one would respond at a time ... ?
That way the bleed-through wouldn't matter.

And ... if you're going to filter out the 22kc, how is it going to
do its job ?

I dunno,
I will just watch and see what comes up.

Regards, John

 

[icarusi]

Just found this thread and I'm after something similar, to block Diseqc signals, but not the continuous 22kHz band-switch tone to the LNB or the downconverted signal back to the receiver. The block could be done by a HPF, but a 'spoof' 22kHz tone would need to be generated after the block, and a detector before the block, to sense if the tone was present in the signal, to activate the spoofer generator, or deactivate it if the 22kHz continuous tone was absent. The HPF needs to pass-through 950-2150 MHz