excessive insertion loss of homebrew BCB band stop filter

Discussion in 'Homebrew and Kit Projects' started by N2LKR, Jul 4, 2021.

ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: abrind-2
ad: Left-3
ad: L-Geochron
ad: l-BCInc
ad: L-MFJ
  1. N2LKR

    N2LKR Ham Member QRZ Page


    I wasn't sure where to post this but it is homebrew and it's not working quite right, in my opinion. I just finished construction of a broadcast band stop filter so I can stop broadcast interference to my SDR when I'm listening to longwave. It is based on this design, but I tweaked it slightly using the RF Tools website to cover from 500 Khz to 1775 Khz.

    After some tweaking of the toroids as I realized that almost all of them had one or more too many turns, I managed to get the filter as close as I could.

    However, I am getting excessive insertion loss. At first I attributed it to perhaps the fact that I am using a random wire antenna with a 9:1 unun, but even using a 50 ohm dummy with an RF noise source, I am still seeing a roughly 20 dB insertion loss with the filter. Here are some images:

    filter schematic -> https://i.imgur.com/NzH1CfW.jpg

    constructed filter -> https://i.imgur.com/558C8vw.jpg


    Top image shows spectra of the filter; bottom is noise source only. Both have 50 ohm termination. Even compensating for the noise source variance, you can still see at least a 15-20 dB loss within the first 250 Khz when comparing the images and according to TinySA. And I confirm this when listening to a weaker NDB on 240 Khz, barely copyable with the filter, fairly strong without.

    I am at a loss as to what's causing my RF loss and welcome your help. This is the first time I've ever dabbled in such an adventure and just hope it won't be the last.
  2. W1VT

    W1VT Ham Member QRZ Page

    Are those ferrite toroids at the top? Many types of ferrite have too high a loss for a filter like this.

    Zak W1VT
  3. N2LKR

    N2LKR Ham Member QRZ Page

    Yes, those are ferrites. I went with them because the original plans called for them. However, out of curiosity, I'm switching them out for the iron powder tomorrow to see if anything changes.
  4. KA9JLM

    KA9JLM Ham Member QRZ Page

    It looks to me that your schematic does not match your build.

    I would start by getting the proper value caps. Using many caps to get the proper value will not work properly in your application.

    The proper type of cap is important also.

    Good Luck. Nice first try. :)
  5. N2LKR

    N2LKR Ham Member QRZ Page

    Swapped out the ferrite for the same iron toroids used in the other four. Took many more turns. However, insertion loss greatly increased! This was after pre-tuning the cores with grid dip and freq counter.

    KA9JLM, not sure what you mean unless you mean my parallel caps were not the way to go. All caps are polystyrene which I read were good for filters.
  6. N5HXR

    N5HXR Ham Member QRZ Page

    With that much loss, I'm wondering if there's a more fundamental construction error. Like, a bad solder joint, bad connector, insufficient lacquer removal, etc.

    What's the DC resistance from input to output?
  7. N2LKR

    N2LKR Ham Member QRZ Page

    Around 0.6 ohms, across center BNC in and out.
  8. N2LKR

    N2LKR Ham Member QRZ Page

    About to throw in the towel on this. I decided to couple to my signal generator and feed 300 Khz into the filter followed by a dummy load at the other end. I did some scope checking. For some reason, the very band I'm trying to monitor (below 500 Khz) is severely curtailed at the output, not as much as broadcast band, but more than it should be. Once I get above 2 Mhz, it's like there's no filter at all, very little, if any, loss. That's good on that end, but not good on longwave.

    Any final ideas would be welcome, scope checks, tuning, etc. I'm not sure how I could tune this type of filter if I needed to. Also, would a dual sided copper board be effecting it since that is what I built it upon?
  9. KP4SX

    KP4SX Premium Subscriber QRZ Page

    A less complicated filter might do the job. Maybe.
  10. KA9JLM

    KA9JLM Ham Member QRZ Page

    Yes that is it. 1 cap with the leads sort as possible would be more better.

Share This Page