Would someone very kindly model this BPF please?

[Chris Wilson]

It's a band pass filter designed to work on an SDR antenna input to attenuate above about 138 kHz, but when I built it and looked at it with a sig gen manually slowly swept and a spectrum analyzer it has least attenuation around 160kHz and is attenuating quite a bit down at 137khz where the reception band is. For kicks I reversed it and it seems better like that   My tracking gen won't go below 500 kHz which makes sweeping it a bit of a a pain.... Is it possible for some kind person to model it please? And to suggest a way to bench fiddle to make it as efficient as possible at 137kHz but attenuate sharply after say 140kHz.given the limitations of its simplicity? It precedes a similar BPF in the SDR (Softrock Lite II) front end itself, if you see what I mean?

[T3sl4co1l]

Did you model the antenna reactance as well?  I'm guessing it's not a flat 50 ohms.

Filters only work correctly when terminated by their rated impedances.  They can be designed for a termination at one end or the other, or both, or any mixture thereof, but never none, as there must always be at least one resistance element somewhere in the circuit.

You want the two resistances to be the antenna (radiation resistance, what supplies the signal to the circuit) and the SDR input (hopefully 50 ohms, but YMMV too?).

Tim

[amspire]

It should have a peak at 137KHz. I think your parts are not accurate enough or there are issues with your test setup.

But you can easily tweak the circuit to 137K. Just reduce the 0.16u cap until you have a peak at 137KHz. The easiest way is probably to put larger values in series.

You don't need fancy hardware for the peak detector. Just make a detector from a 1N4148 diode, a 0,1u cap and a 10K resistor across the cap. Just attach a voltmeter across the resistor.

Then do the same with the 0.005u.

Make sure you are using stable caps. Not sure what type of capacitor that blue .15uF one is, but if it is a multilayer ceramic cap, it is probably rubbish.  You would probably be better of with a polyester cap. Again with the .005u, a COG/NPO ceramic would be fine for stability, but I suspect it is not from the size. You could just go for a polyester cap there too.

Now the cores you are using for the inductors vary by over 1%/C, so a 10 degree C temperature change would change the centre frequency by over 14KHz. This is fairly low frequency  so you might have better stability using something like a gapped E core. With a gapped core, the stability of the inductance is set by the stability of the physical gap between the cores. The core material has little effect on stability.

[hagster]

This took about 2 minutes to model in QUCS. All components are ideal L's and C's.

[Neganur]

interesting software, cheers for mentioning it.

[cncjerry]

Eye opening usability improvement over LTSpice.  Some inconsistencies but overall pretty cool. 

[tec5c]

For the purpose of curiosity, I have compared hagster's results from QUCS next to NI's AWR and Keysight's ADS. (Marker frequencies are slightly off in ADS, too tired to change my snap grid sorry) Note: As with hagster's sim, all elements are ideal.