Hi all,
I am trying to design a simple 3 element FM broadcast band notch filter centred at around 80MHz using LC. I run simulation and calculations and got the values for a simple 3 element. I made one using SMD LC and used a spectrum analyzer with a tracking gen to test it, the notch or the cut off was set at 220MHz! So I found that odd. Both simulation and calculation shows 80MHz. see the filter circuit below:

So next I just made the parallel section, 1 element LC and tested it, and it is perfectly centred at 220MHz, with a cut off about 23dB for a single element, I dont understand why. Now I have made a 3 element in the past and it works perfectly on the FM band, that is exactly the same circuit but I didnt test it on the analyzer, I just made it, hooked it up and it works exactly as it should.

So I am just trying to find out why I am getting 230MHz indicated on the spectrum analyzer, beit a 1 element or 3 element either shunt or parallel, its just confusing me.

 

The FM broadcast band "over here" is 88MHz to 108MHz. Are you in Japan or Brazil?

 

Same here, no I am not in Brazil. I have specifically taylored the filter for the 80MHz centre because I dont want to attenuate too much of the higher frequency pass the band.

 

Then why not use a highpass filter?

 

Parasitic capacitance and inductance in your real circuit may be the problem.
Were the wires all as short as possible?

 

Parasitic capacitance and inductance in your real circuit may be the problem.
Were the wires all as short as possible?

Its a point, but no. As I mentioned, I used SMD 0603 cap and inductor, on a tiny double-sided RF PCB with a lot of stitching on the ground plane. I have been doing RF and UHF for many years, and also am an amateur radio, I only mention this so that you understand I am fully aware of some of the pitfalls of playing with RF.

The area of the board with components is about 15mm square plus an area on either side for SMA connectors via 50-ohm tracks.
From the analyzer to the board I use an N connector to SMA which are about 6 inches. So, not a lot of strays to speak of, besides, I would expect any stray would drag the centre frequency lower.

Still, something isnt quite right. As I say the perfectly working one is the same circuit on about the same size board, so the theory is right but the actual testing doesn't make any sense. 220MHz is almost double the filter cut-off freq. If I had a dip in the broadcast band plus 220MHz, I would have to assume at a push that its oscillating, but there is no dip in the broadcast band at all. It is actually a pretty good notch filter for the 220MHz, but its not meant for that!