I made this filter because of a request on BitX20 forum, to attempt to build a filter for the 40m band.
I already got some experience with the previous filter for 20m, so I proceed as follows:
1. I used the AADE Filter Design and Analysis software and tested for a few inductor values that will bring as many capacitors from the design close to standard values.
This is what I’ve chosen:
2. I decided to build three coils with an inductance slightly above the nominal 2uH (my coils indicated 2.05uH).
I found out that if I make coils from an 80cm length of 0.6mm wire and I use my Philips screwdriver with a shaft diameter of 6mm, I get 37 turns and an inductance of 2.04uH.
Later it became clear that I should have added another two turns to the coils to have the filter centered to the middle of the band (7.15MHz).
The capacitors I used were: C3, C9 = 56pF; C5. C7: 10pF; C4, C8: 180pF; C6: 232pF (150pF + 82pF).
They were all ceramic NPO and had a measured tolerance within 2% of the values marked, even if their stated tolerance may be higher.
3. I soldered all the components and run the measurements on the Spectrum Analyzer. The whole filter plot was centered to a higher frequency.
Since I didn’t want to rebuild another set of inductors, I padded the capacitors C3, C9 with another 10pF. Capacitors C5 and C7 were increased by adding a 4.7pF on each of them.
4. The measurement on SA showed the following:
The minimum loss at the center of the band is better than expected (-1.85dB at 7.15MHz, -2.76dB at 7.00MHz, -2.36dB at 7.30MHz), which is an indication the Q factor of the coils I built is higher than 100. I didn’t have to adjust the coils.
The capacitance on C3, C5, C7, C9 should have probably increased just a tiny fraction more (maybe 12pF and 5.6pF instead of 10pF and 4.7pF), but I couldn’t be bothered.
And finally, a photo of the practical circuit:
And yes, I know, the capacitors legs are a bit too long and this may spoil the rejection at the upper frequency range 🙂