Making window filters with cavities

Discussion in 'VHF/UHF - 50Mhz and Beyond' started by W6KCS, Dec 3, 2018.

ad: L-HROutlet
ad: l-rl
ad: Subscribe
ad: Left-3
ad: MessiPaoloni-1
ad: Left-2
ad: l-assoc
ad: L-MFJ
  1. W6KCS

    W6KCS Ham Member QRZ Page

    I appreciate all of the suggestions. Two cavities wouldn't work here, I can pass a couple of frequencies 500 Kc apart, but the insertion loss is too high between them, as shown in the plot. This typically takes several cavities, for example, here's one for high band that Telewave sells with 6 cavities, I think it's about 1 MHz wide:

    WB5WPA: Yes, the plot is two cavities in series on the same freq (the upper freq) with a critical length cable, and the type of filter I'd like to be able to build is similar, but with more cavities, as shown in the link above. What I'm trying to figure out is how they manage to do it as wide as they do with only a couple dB insertion loss. I can't seem to make it work with simple stagger-tuning, so I assume there is something else going on, like the conjugate matching I did with the critical cable, maybe in combination with some other tuning method with the cavities themselves.
    Last edited: Dec 4, 2018
  2. WB5WPA

    WB5WPA Ham Member QRZ Page

    re: "What I'm trying to figure out is how they manage to do it as wide as they do with only a couple dB insertion loss."

    The product you show as an example (depending which model you choose) renders about a 1 MHz bandpass at 146 MHz.

    Scaled down to Low band (all other factors being equal) figure roughly 1/3 that, or 300 kHz bandpass at a stated IL of roughly 2 dB.

    You could try messing with the coupling loops in your present cavities, that is the other variable you can 'play' with.

    This is straight "physics" you're playing with, and there will be 'hard' trade-offs; no magic to be had here. Been there done that with filters at various frequencies and specifically duplexer designs (including Hybrid Ring types) at ten and six meters ... that was my old webpage I ref'd earlier BTW.
    W6KCS likes this.
  3. W6KCS

    W6KCS Ham Member QRZ Page

    Good point, that makes me think that part of the problem is that I'm trying to get too much BW out of these.
  4. W6KCS

    W6KCS Ham Member QRZ Page

    I took a Telewave 450 4-cavity preselector apart today and did some measurements that gave me some clues as to how these filters are made. This one is about 3 MHz wide and the loss is only 1 dB through most of that. The first thing I noticed was that all of the cavities are tuned for the exact center of the range, they aren't stagger-tuned, which surprised me. The plot below shows one of them (superimposed over the plot of all four), but they are all nearly identical. They also have much tighter coupling than any cavity I've seen in a duplexer, with only about 0.2 dB of insertion loss per cavity. In other words, the loops were set as close to the rods as possible. Third, the cable lengths between them were all about 1/2 wavelength long and virtually the same, within 1/8" of each other. So although I still have no idea how the flat table-top filter shape is created, at least I have a place to start.

    c small.jpg
  5. W1VT

    W1VT Ham Member QRZ Page

    The flat top shape is obtained by adjusting the coupling between the resonators. Handbook of Filter Design by Zverev is the first book I requisitioned. This book describes how to calculate the coupling between filter elements.
    Last edited: Dec 7, 2018
    W6KCS likes this.

Share This Page

ad: M2Ant-1